Large-scale synthesis of Man₉GlcNAc₂ high mannose glycan and the effect of the glycan core on multivalent recognition by HIV antibody 2G12

Abstract

Access to homogenous high mannose glycans in high mg quantities is necessary for carbohydrate-based HIV vaccine development research. We have used directed evolution to design highly antigenic oligomannose clusters that are recognized in low-nM affinity by HIV antibodies. Herein we report an optimized large-scale synthesis of Man₉GlcNAc₂ including improved building block synthesis and a fully stereoselective 5+6 coupling, yielding 290 mg of glycan. We then use this glycan to study the effect of the GlcNAc₂ core on antigenicity of an evolved 2G12-binding glycopeptide, 10F2.

Graphical Abstract

N-linked glycosylation is a common posttranslational modification in eukaryotic proteins that are secreted or present on the cell surface.¹ The “high-mannose” structure is an unprocessed N-glycan terminating only in mannose residues, which is enriched on various viral proteins, including hepatitis C virus², Ebola virus³, coronaviruses⁴, and HIV⁵. Interaction of these glycans with the dendritic cell receptor DC-SIGN facilitates the presentation of HIV to T cells, by colocalization of these cell types^6–8. The abundance and conservation of high-mannose glycans on HIV also makes them a target of broadly-neutralizing antibodies (bnAbs).⁹ Prototypical high-mannose-binding bnAbs isolated from humans include 2G12¹⁰, which binds to multivalent clusters of Man₉GlcNAc₂ glycans, and PGT 128¹¹ and PGT122¹², which bind Man₉GlcNAc₂ 1 (Scheme 1) moieties in combination with the polypeptide backbone.

Scheme 1.

Retrosynthetic design for synthesis of Man₉GlcNAc₂ 1

There is therefore significant interest in high-mannose glycans as components of HIV vaccines. As a vaccine design approach, our lab has developed directed evolution methods that create trillions of multivalent oligomannose glycan clusters and select for optimal arrangements that are antigenic mimics of HIV.¹³ This has identified Man₉-containing glycopeptides with picomolar to nanomolar dissociation constants by antibodies such as 2G12.¹⁴ Although library generation requires sub-milligram quantities of glycans, the preparation, biophysical characterization and immunogenicity assessment of library “hits” requires milligram quantities. Low milligram quantities of Man₉GlcNAc₂ glycan can be isolated from a kilogram of soy flour; however, this process is difficult to scale up to the hundreds of mg needed for incorporation into numerous glycopeptides. Synthesis has the potential to afford larger quantities, though certain steps pose bottlenecks to material throughput. Herein we report an optimized synthesis of Man₉GlcNAc₂ 1, which yielded 290 mg of the glycan.

The most convergent and popular strategy for synthesis of Man₉GlcNAc₂ (1) and related structures has been the coupling of a branched Man₅ glycosyl donor (2) with a Man₄GlcNAc₂ hexasaccharide acceptor (3). This critical coupling, with fragments requiring many synthetic steps, is difficult to optimize, owing to the multistep synthesis of the coupling partners. Danishefsky and coworkers in 2004 reported the use of a Man₅-SPh donor 2a, protected at the C2 position of the core mannose with non-participating Bn group, and activated with Sinaÿ reagent ([(BrC₆H₄)₃N][SbF₆]).¹⁵ In an updated procedure, Danishefsky and coworkers used NIS/TMSOTf conditions to activate Man₅-SPh donor 2b with a participating (difluorobenzoate) protecting group at the core mannose C2.¹⁶ The groups of Ito¹⁷, Wong,¹⁸ and Wang¹⁹ have primarily used of Man₅-F (in some cases bearing additional saccharides at the non-reducing end), in all cases protected with non-participating Bn protecting group. Ito showed that the glycosyl fluoride, activated with Cp₂HfOTf/AgOTf was highly effective (63 % y) whereas the analogous SPh glycoside activated with NIS/AgOTf gave only trace product. Although good yields are reported for most of the above couplings, there are few apples-to-apples comparisons between substrates or conditions, and the anomeric selectivity was not reported. The isolated final quantity of deprotected high mannose glycan isolated ranges from 7–33 mg, with 115 mg as the exceptional case from Danishefsky’s second synthesis.

In our own efforts to produce significant quantities of Man₉GlcNAc₂ for glycopeptide synthesis, we experienced several failures in the coupling of Man₅ donor 2a with hexasaccharide acceptor 3a or 3b, despite very careful precautions to exclude moisture and purify starting materials and reagents (see SI Scheme S1). We reasoned that a participating C2 ester protecting group should at least confer exclusively α anomeric selectivity if not also better reactivity in a sluggish glycosylation (vide infra). We opted for the pivaloyl-ester 2c because of its relative stability in Zemplén deacetylation, allowing the use of acetate-protected donor 4 in bidirectional glycosylation of diol 6. Donor 4 is highly convenient to prepare in multigram quantities. As detailed below, the gold/silver-catalyzed method reported by Hotha and coworkers²⁰ was ideal for this route.

To begin the synthesis of Man₉GlcNAc₂ 1¸ we first synthesized of Man₄GlcNAc₂ hexasaccharide acceptor 3a in gram quantities (Scheme 2). Coupling of glycosyl donor 7 and acceptor 8 under NIS and TMSOTf conditions afforded protected chitobiose 9 in 78% yield; this was further treated with NaOMe in methanol at room temperature to remove the Bz group, accessing disaccharide acceptor 10 in 93% yield. Crich-Kahne β-mannosylation²¹ was performed between acceptor 10 and mannosyl sulfoxide donor 11, followed by PMB removal using DDQ, to afford desired β-anomer of trisaccharide acceptor 12 in 59% yield in over two steps. 3+3 glycosylation was then carried out between trisaccharide mannosyl donor 13 and acceptor 12 under Sinaÿ conditions, to afford a 4:1 α/β crude mixture of hexasaccharides. The desired α isomer 14 was carefully isolated by column chromatography in 58% yield. The benzylidene ring of hexasaccharide 14 was reductively opened by triethyl silane and phenylboron dichloride at −78 °C to selectively deprotect the primary -OH, affording hexasaccharide acceptor 15 in 80 % yield.

Scheme 2.

Synthesis of hexasaccharide acceptor 3b

We next proceeded with synthesis of branched pentamannose donor 2c, which itself poses some challenges. Bidirectional glycosylation of diol 6 with chloride donor 15 resulted in a mixture with some orthoester 17 and partial glycosylation products, which made chromatographic isolation of trisaccharide 16 difficult (Scheme 3). After several unsuccessful variations on this coupling (without base²², or with tetramethyl urea scavenger²³), we replaced donor 15 with glycosyl carbonate donor 4. Under Hotha’s²⁰ catalytic Au/Ag conditions (gold phosphite 18 and AgOTf), bidirectional coupling of donor 4 with diol 6 proceeded remarkably cleanly to afford grams of trisaccharide 16 in 80% yield. It is also worth noting that the Hotha coupling conditions were compatible with the presence of the phenyl thioglycoside. 30 minutes of Zemplén deacetylation resulted in selective deacetylation in the presence of Piv, affording Man₃ diol 5. Analogous bidirectional coupling of 5 with 4 under gold catalysis led to branched Man₅-SPh 2c in 72% yield.

Scheme 3.

Synthesis of branched pentamannan donor with neighboring group participants 2

Having obtained pentasaccharide donor 2c bearing C2-pivaloyl ester group, we coupled it to hexasaccharide acceptor 3 on gram scale, under Sinaÿ conditions, to afford 66% of undecasaccharide 19 as single diastereomer after 14 hours (Scheme 4). By comparison with the C2 OBn-protected donor (2a) which failed in several 5+6 coupling attempts in our hands, the C2 OPiv-protected 2c may form a more stable cationic intermediate owing to neighboring group participation, suppressing side reactions in the presence of a sluggish hexasaccharide nucleophile.^24–26 With fully protected glycan 19 in hand, global deprotection was performed in a four-step sequence. Treatment with ethylenediamine followed by acetic anhydride replaced phthalimide with acetamide groups. Zemplén methanolysis removed acetates but not the pivaloate, which was removed together with all benzyl groups in a final Birch reduction, to afford 290 mg of unprotected Man₉GlcNAc₂ 1 (63% overall yield) after two P-2 desalting column purifications.

Scheme 4.

Synthesis of Man₉GlcNAc₂-azide 20

With a substantial quantity of Man₉GlcNAc₂ in hand, we decided to test the recognition of this glycan in clustered form by HIV bnAb 2G12.²⁷ 2G12 has four glycan binding pockets that recognize up to four Man₉GlcNAc₂ glycans on the surface of HIV envelope protein gp120.^{10, 28} As a vaccine design approach, we have previously used directed evolution to design peptide ^{14, 29}and oligonucleotide^30–34 backbones that can cluster Man₉Cy (Cy = cyclohexyl linker) in tight arrangements that mimic gp120, with sub-nM to low-nM dissociation constants for 2G12. Given the possible immunogenicity of the synthetic cyclohexyl core, we were curious about whether an evolved Man₉ cluster could be recognized by 2G12 equivalently when swapping in the natural GlcNAc₂ core.

Thus, we used Fmoc solid phase peptide synthesis to prepare peptide backbone 10F2, which evolved to present Man₉Cy glycans in a low nM interaction with 2G12,¹⁴ and contains alkynes as sites for Copper-assisted Alkyne Azide “Click” (CuAAC) attachment of glycans (Figure 1A, C). A biotin tag was included at the C terminus of the peptide for immobilization in Biolayer Interferometry (BLI) assays on streptavidin biosensors. Man₉GlcNAc₂-OH 1 was converted to anomeric azide 20 under Shoda’s conditions (Scheme 4). 10F2 peptide 22 was “click”-glycosylated either with Man₉Cy-azide 21 or Man₉GlcNAc₂ 20 to afford 10F2-Man₉Cy 23³⁰ and 10F2-Man₉GlcNAc₂ 24 glycopeptides, respectively (Figure 1B, D, E).

Figure 1.

A) SPPS synthesis of 10F2 peptide 22, B) Glycopeptide formation using CuAAC click reaction, C) LC-UV chromatogram of 10F2 peptide 22, D) LC-UV chromatogram of 10F2-Man₉Cy glycopeptide 23, E) LC-UV chromatogram of 10F2-Man₉GlcNAc₂ glycopeptide 24, BioLayer Interferometry (BLI) measurements of 2G12 interacting with surface-immobilized F) 10F2-Man₉Cy glycopeptide 23 and G) 10F2-Man₉GlcNAc₂ glycopeptide 24 on streptavidin biosensors. k_on and k_off errors are standard errors of the curve fit, and K_D error is propagated from those values.

In conclusion, we have demonstrated an optimized synthetic route to access nearly 300 mg of Man₉GlcNAc₂ undecasaccharide. Synthesis of branched Man₅ donor with C2 Piv group in multi-gram scale was achieved using gold/silver glycosylation chemistry, followed by stereoselective 5+6 glycosylation. Using this synthetic glycan, we prepared a Man₉ glycopeptide containing the native GlcNAc₂ core, whose antigenicity was compared with the same peptide containing a cyclohexyl core. The ready access to this glycan will enable extensive synthesis of further evolved HIV glycopeptide antigens, their biophysical characterization and immunogenicity studies for HIV vaccine development.

Methods

Trisaccharide 16:

To a 100ml round bottom flask containing carbonate donor 4 (3.17 g, 4.93 mmol, 2.2eq.) and diol acceptor 6 (1 g, 2.27 mmol, 1 eq.) added 30 ml dry CH₂Cl₂ with syringe under nitrogen atmosphere, followed by 1g freshly flame dried of 4Å MS was added and allowed to stir for 15 min at room temperature. AgOTf (765.97 mg, 828.5 μmol) and Au-phosphite 18 (115 mg, 447.8 μmol) were added to a reaction mixture and allowed to stir for another 15 minutes at room temperature. After 15 minutes,the reaction was complete according to TLC and quenched with 500 μl of triethyl amine. Reaction mixture was then immediately filtered through a celite pad and filtrate was concentrated under reduced pressure to obtain crude, which was purified by flash column chromatography using hexane and ethyl acetate (5:1) to obtain 2.5 g of trisaccharide 16 as glassy solid in 80% yield. ¹H NMR (400 MHz, CDCl₃) δ 7.79 – 6.81 (m, 40H), 5.43 (app dd, J = 3.1, 1.8 Hz, 1H), 5.41 – 5.36 (m, 2H), 5.34 (app t, J = 3.1, 1.8 Hz, 1H), 5.19 (d, J = 1.8 Hz, 1H), 4.91 (d, J = 1.8 Hz, 1H), 4.87 (s, 1H), 4.84 (s, 1H), 4.79 – 4.60 (multiple signals, 5H), 4.58 – 4.41 (multiple signals, 7H), 4.32 – 4.25 (m, 1H), 4.16 (dd, J = 9.3, 3.1 Hz, 1H), 4.01 – 3.54 (multiple signals, 13H), 2.12 (s, 3H), 2.12 (s, 3H), 1.19 (s, 9H). ¹³C NMR (100 MHz, CDCl₃) selected signals δ 21.1, 21.2, 27.2(3C), 39.1, 66.2, 68.3, 68.5, 68.7, 69.0, 71.7, 71.9, 71.9, 72.0, 72.4, 73.4, 73.5, 73.5, 74.1, 74.2, 74.7, 74.9, 75.3, 75.5, 77.6, 78.1, 78.6, 85.8, 98.3, 100.2, 127.4(2C), 127.6, 127.7, 127.8, 127.8(2C), 127.9, 127.9(2C), 128.0(2C), 128.1(2C), 128.1, 128.1(2C), 128.2(2C), 128.2(2C), 128.3(2C), 128.4(2C), 128.4(2C), 128.4, 128.5(2C), 128.6(2C), 128.7, 129.3(2C), 131.8(2C), 133.7, 137.6, 137.8, 138.0, 138.2, 138.4, 138.5, 138.9, 170.3, 170.3, 177.4. HRMS (ESI+MS) m/z: [M+Na]⁺ : Calc. for C₈₂H₉₀O₁₈SNa: 1417.5746, found: 1417.5212

Trisaccharide diol 5c:

To a trisaccharide 16 (1.4 g, 1 mmol, 1 eq.) dissolved in Methanol (1 mL) and dichloromethane (13 mL) was added 45 μl of 25% NaOMe solution (0.2 mmol, 0.2 eq.) in 1ml MeOH dropwise. The reaction was stirred at room temperature under nitrogen for 30 minutes, the reaction was quenched with Amberlite IR-120 H-form (from basic to neutral pH~6 to 7), resin was filtered, and solvent was concentrated, and crude was purified by flash silica column chromatography by using hexane and ethyl acetate (3:2) to obtain trisaccharide diol 5 as glassy solid in 80% yield (1.05 g, 0.8 mmol). ¹H NMR (400 MHz, CDCl₃) δ 7.72 – 6.89 (multiple signals, 40H + residual CHCl₃), 5.37 (d, J = 1.7 Hz, 1H), 5.34 (dd, J = 2.4, 1.7 Hz, 1H), 5.20 (s, 1H), 4.97 (s, 1H), 4.84 – 4.77 (m, 2H), 4.73 – 4.57 (multiple signals, 7H), 4.54 – 4.41 (m, 5H), 4.29 (m, 1H), 4.14 (dd, J = 9.4, 3.0 Hz, 1H), 4.03 (d, J = 3.0 Hz, 1H), 3.97 – 3.51 (multiple signals, 14H), 2.41 (s, 1H), 2.41 (s, 1H),1.17 (s, 9H). ¹³C NMR (100 MHz, CDCl₃) δ 27.3(3C), 39.1, 66.2, 68.3, 68.6, 68.8, 69.0, 71.4, 72.0, 72.1, 72.2, 73.5, 73.5, 73.6, 74.1, 74.3, 74.7, 75.1, 75.3, 75.4, 77.4, 78.2, 79.6, 80.4, 85.8, 99.6, 102.0, 127.5, 127.5(2C), 127.6, 127.7, 127.8, 127.8, 127.8(2C), 127.9(2C), 127.9(2C), 128.0(3C), 128.0(2C), 128.1, 128.1(2C), 128.1, 128.3(2C), 128.4(2C), 128.5(2C), 128.5(2C), 128.6(2C), 128.7(2C), 128.7(2C), 129.3(2C), 131.9(2C), 133.8, 137.7, 138.0, 138.0, 138.3, 138.4, 138.5, 138.8, 177.3. HRMS (ESI+MS) m/z: [M+Na]⁺ : Calc. for C₇₈H₈₆O₁₆SNa: 1333.5534, found: 1333.5040

Pentasaccharide 2:

To a 100ml round bottom flask containing carbonate donor 4 (1.7 g, 2.64 mmol, 3eq.) and trisaccharide diol acceptor 5 (1.2 g, 0.914 mmol, 1 eq.) was added 30 ml dry CH₂Cl₂ with syringe under nitrogen atmosphere, followed by 1g of freshly flame dried 4Å MS, and the suspension was allowed to stir for 15 min at room temperature. AgOTf (87 mg, 338.52 μmol) and Au-phosphite (161 mg, 182.9 μmol) were added to the reaction mixture and allowed to stir for another 15 minutes at room temperature. After 15 minutes, the reaction was complete according to TLC and quenched with 500 μl of triethyl amine. Reaction mixture was then immediately filtered through a celite pad, and the filtrate was concentrated under reduced pressure, and residue was purified by flash column chromatography using hexane and ethyl acetate (4:1) to yeild pentasaccharide 2c as glassy solid in 72% yield (based on the acceptor diol 5, 1.5g; 0.6mmol). ¹H NMR (400 MHz, CDCl₃) δ 7.46 – 7.05 (multiple signals, 70H + residual CHCl₃), 5.55 – 5.48 (m, 2H), 5.38 (d, J = 1.6 Hz, 1H), 5.33 (dd, J = 3.0, 1.7 Hz, 1H), 5.23 (d, J = 1.9 Hz, 1H), 5.07 (d, J = 1.8 Hz, 1H), 5.04 (d, J = 1.8 Hz, 1H), 4.94 – 4.77 (m, 5H), 4.77 – 4.30 (m, 22H), 4.28 – 4.19 (m, 1H), 4.14 – 3.36 (multiple signals, 26H), 2.12 (s, 3H), 2.10 (s, 3H), 1.17 (s, 9H). ¹³C NMR (100 MHz, CDCl₃) δ 21.3, 21.3, 27.3(3C), 39.2, 66.7, 68.6, 68.9, 68.9, 69.0, 71.8, 72.0, 72.0, 72.1, 72.1, 72.1, 72.1, 72.2, 72.3, 72.4, 72.9, 73.4, 73.4, 73.5, 73.5, 73.5, 73.7, 74.2, 74.4, 74.6, 74.6, 74.7, 74.7, 75.1, 75.2, 75.3, 75.5, 77.4, 78.3, 78.4, 79.0, 79.7, 80.2, 85.7, 99.2, 99.6, 99.7, 101.8, 127.4(2C), 127.4, 127.5(2C), 127.6(2C), 127.6(2C), 127.6(2C), 127.7(2C), 127.7(4C), 127.7(2C), 127.7, 127.8(2C), 127.8, 127.9(2C), 128.0(2C), 128.0(2C), 128.1(2C), 128.3(2C), 128.3(2C), 128.3(6C), 128.3(2C), 128.4(2C), 128.4(2C), 128.4(6C), 128.5(5C), 128.5(2C), 128.6(2C), 128.7(2C), 129.3(2C), 131.7(2C), 134.0, 138.0, 138.2, 138.2, 138.3, 138.3, 138.4, 138.4, 138.6, 138.6, 138.7, 138.7, 138.8, 139.0, 170.3, 170.3, 177.4. HRMS (ESI+MS) m/z: [M+Na]⁺ : Calc. for C₁₃₆H₁₄₆O₂₈SNa: 2282.9653, found: 2282.9647

Protected Man₉GlcNAc₂ 19:

To 1.5 g (0.6 mmol, 1.5 eq.) of pentasaccharide donor 2c and 1.1 g (0.4 mmol, 1 eq.) of hexasaccharide acceptor 3a in 50 mL conical flask toluene was added and mixture was azeotropically dried three times. The dry residue was dissolved in 10 mL of anhydrous acetonitrile and freshly activated 4Å molecular sieves were added, and this was allowed to stir for 20 minutes at room temperature under nitrogen atmosphere. The flask was then wrapped in aluminum foil, cooled to −10 °C and 1.0 g (1 mmol, 3 eq.) of Sinay’s reagent [(p-BrC₆H₄)₃N⁺SbCl6^-] was added. As soon as the Sinay’s reagent was added to flask a dark blue solution was observed and this reaction was allowed to warm to room temperature and stirred for 16 hours with frequent monitoring of the reaction by LCMS. After 16 hours, the reaction was quenched by addition of 5 mL of triethylamine and the blue solution turned to brown, the reaction was filtered through celite and concentrated in vacuo. The crude residue was purified by flash chromatography with 1:1:3 ethyl acetate:dichlormethane:hexane, to give 1.3 g (0.2 mmol) of undeacasaccharide 19 in 66% yield as white foam. ¹H NMR (800 MHz, CDCl₃) δ 7.74 (d, J = 7.3 Hz, 1H), 7.57 (dt, J = 29.8, 7.3 Hz, 3H), 7.48 (d, J = 7.2 Hz, 1H), 7.46 – 7.39 (m, 2H), 7.37 – 6.93 (m, 138H + residual CHCl₃), 6.89 (dd, J = 6.3, 1.8 Hz, 4H), 6.75 – 6.64 (m, 3H), 6.64 – 6.52 (m, 2H), 5.54 (dd, J = 3.3, 1.9 Hz, 1H), 5.52 – 5.48 (m, 2H), 5.16 (s, 2H), 5.13 (d, J = 1.8 Hz, 1H), 5.10 (d, J = 1.9 Hz, 1H), 5.09 – 5.02 (m, 3H), 4.99 – 4.97 (m, 1H), 4.91 (d, J = 8.5 Hz, 2H), 4.69 – 4.29 (m, 49H), 4.25 – 3.30 (m, 69H), 3.27 – 3.22 (m, 2H), 3.20 – 3.16 (m, 1H), 3.13 – 2.93 (m, 4H), 2.10 (s, 3H), 2.09 (s, 3H), 2.03 (s, 3H), 1.03 (s, 9H). ¹³C NMR (200 MHz, CDCl₃) Selected signalsδ 0.2, 21.3, 21.3, 27.2, 38.8, 70.6, 72.0, 72.0, 72.4, 72.8, 73.0, 73.1, 73.2, 73.2, 73.4, 73.5, 73.6, 74.2, 74.5, 74.9, 77.0, 77.0, 77.0, 77.2, 77.4, 77.4, 126.8, 126.9, 127.0, 127.2, 127.2, 127.2, 127.3, 127.3, 127.4, 127.4, 127.4, 127.5, 127.5, 127.5, 127.6, 127.6, 127.6, 127.6, 127.6, 127.6, 127.6, 127.7, 127.7, 127.7, 127.7, 127.7, 127.8, 127.8, 127.9, 127.9, 128.0, 128.0, 128.0, 128.1, 128.1, 128.2, 128.2, 128.2, 128.2, 128.3, 128.3, 128.3, 128.3, 128.3, 128.4, 128.4, 128.4, 128.4, 128.4, 128.4, 128.5, 128.5, 128.5, 128.6, 128.6, 128.6, 128.6, 128.7, 129.0, 131.7, 131.9, 132.0, 133.5, 133.8, 137.4, 138.1, 138.2, 138.2, 138.3, 138.3, 138.3, 138.4, 138.4, 138.6, 138.6, 138.6, 138.6, 138.7, 138.7, 138.8, 138.8, 138.9, 138.9, 138.9, 139.2, 167.5, 167.7, 168.2, 170.2, 170.3, 176.5. HRMS (ESI+MS) m/z: [M+2Na]²⁺ : Calc. for C₂₉₆H₃₀₆N₂O₆₂Na: 2464.5369, found: 2464.5453

Deprotected Man₉GlcNAc₂ 1:

1.2 g (0.26 mmol) of protected Man₉GlcNAc₂ 19, 32 mL of n-butanol, 16 mL of toluene and 10 mL (1.51 mmol) of ethylenediamine were added to a 100 mL flask. The reaction was refluxed for 24 hours, then co-evaporated with dry toluene three times. The flask containing crude material was subjected to acetylation using 5.5 mL of pyridine, 1 mL of acetic anhydride and 80 mg of (0.65 mmol) of DMAP at 0 °C for 2 hours. Reaction solvent was co-evaporated with toluene three times and resulting crude was was purified on a silica column using hexane and ethyl acetate (1:1). Fractions containing product were concentrated and dissolved in 12 mL of methanol and 5 mL of THF. 13 μL of sodium methoxide (0.057 mmol) was added to reaction and after 1 hour the reaction was quenched with Amberlite IR120-H, filtered and concentrated. The crude was purified by flash chromatography with 2:3 ethyl acetate and hexane to give 770 mg crude with benzyl and pivaloyl protected product. This was crude was subjected to global deprotection using Birch reduction.

Into an oven dried 500 mL 3-necked flask, ~350 mL of ammonia was condensed under stream of nitrogen at −78 °C. 1.2 g of Na° (51 mmol, 300 eq.) was added, and the bright blue reaction was allowed to stir for 1 hour to observe persistence of blue color. Then 770 mg of intermediate crude in 5 mL of dry THF was added by syringe, and the reaction was allowed to stir for 5 hours. When rection was completed (monitored by direct infusion into ESIMS), solid NH₄Cl (4.5 g) was added portion wise until the disappearance of blue color. The ice bath was removed, allowing the mixture to warm to room temperature and ammonia was allowed to boil off under a stream of nitrogen. The crude was desalted on Biogel P-2 size exclusion gel column twice to afford 290 mg of deprotected Man₉GlcNAc₂ 1 (154.80 μmol) as a hazy glassy white solid in 63% yield over 4 steps. ¹H NMR (400 MHz, D₂O) δ 5.39 (s, 1H), 5.31 (s, 1H), 5.29 (s, 1H), 5.20 – 5.15 (m, 1H), 5.12 (s, 1H), 5.06 – 4.97 (m, 3H), 4.84 (s, 1H), 4.75 (s, 1H), 4.67 (d, J = 7.4 Hz, 1H), 4.63 – 4.54 (m, 1H), 4.21 (d, J = 3.1 Hz, 1H), 4.13 (s, 1H), 4.11 – 3.44 (multiple signals, 78H), 2.05 (s, 3H), 2.02 (s, 3H). ¹³C NMR (101 MHz, D₂O) selected signals δ 21.9, 22.2, 53.6, 55.0, 56.0, 60.0, 60.1, 60.8, 60.9, 61.0, 61.0, 61.1, 61.1, 65.0, 65.0, 65.4, 65.6, 66.7, 66.8, 66.8, 66.9, 66.9, 67.0, 69.2, 69.4, 69.9, 69.9, 69.9, 69.9, 69.9, 69.9, 69.9, 69.9, 70.0, 70.0, 70.0, 70.0, 70.2, 70.2, 70.3, 70.3, 70.3, 70.3, 71.1, 71.9, 72.4, 72.6, 73.1, 73.1, 73.1, 73.2, 73.2, 73.2, 73.3, 74.1, 74.5, 74.5, 78.4, 78.4, 78.5, 78.5, 78.5, 78.6, 78.8, 78.8, 78.8, 79.2, 79.7, 81.0, 90.4, 94.7, 97.9, 99.5, 100.1, 100.5, 100.5, 100.7, 101.3, 102.1, 102.1, 102.1, 102.2, 174.3, 174.4. HRMS (ESI+MS) m/z: [M+2H]²⁺: Calc. for C₇₀H₁₁₈N₂O₅₆: 942.3296, found: 942.3302

Man₉GlcNAc₂-azide 20:

Fully deprotected Man₉GlcNAc₂ 1 (115 mg, 0.06 mmol, 1 eq.) was dissolved in D₂O (5 mL) and cooled at 0 °C in 20 mL reaction vial. 2-chloro-1,3-dimethylimidazolinium chloride (DMC) (204 mg, 1.5 mmol, 25 eq.), 2,6-lutidine (353 μL, 3 mmol, 50 eq. and sodium azide (793 mg, 12 mmol, 200eq) was added to reaction vial and allowed to warm to room temperature (ice was allowed to melt) and incubated for 18 hours. The reaction was monitored by taking 0.5 mL crude in NMR tube and recorded after each 3 hours until reaction completed upto 18 hours. Reaction mixture was loaded on desalting Biogel P-2 size exclusion gel column and repeated two more times which afforded 81 mg of Man₉GlcNAc₂-azide 20 in 70% yield. ¹H NMR (400 MHz, D₂O) δ 5.43 (s, 1H), 5.36 (s, 1H), 5.34 (s, 1H), 5.17 (s, 1H), 5.12 – 4.99 (overlap signals, 3H), 4.89 (s, 1H), 4.78 (s, 1H), 4.63 (d, J = 7.0 Hz, 1H), 4.25 (d, J = 3.1 Hz, 1H), 4.18 (s, 1H), 4.17 – 3.59 (multiple signals, 78H), 2.10 (s, 3H), 2.08 (s, 3H). ¹³C NMR (100 MHz, D₂O) δ 22.1, 22.2, 54.5, 55.0, 59.9, 60.0, 61.0, 61.1, 65.0, 65.1, 65.4, 65.6, 66.8, 66.8, 66.8, 66.9, 66.9, 67.0, 67.0, 69.4, 69.9, 70.0, 70.2, 70.3, 71.2, 71.9, 72.2, 72.7, 73.1, 73.2, 73.2, 73.4, 74.1, 74.5, 76.4, 78.4, 78.6, 78.7, 78.8, 78.9, 81.0, 88.5, 98.0, 99.6, 100.2, 100.6, 100.8, 101.3, 102.2, 102.2, 174.5, 174.7. HRMS (ESI+MS) m/z: [M+Na]⁺: Calc. for C₇₀H₁₁₇N₅O₅₅Na: 1930.6410, found: 1930.6078

Binding analysis of Glycopeptides and 2G12 by BLI (Biolayer Interferometry).

Glycopeptide bearing biotin was loaded (120s) onto a streptavidin biosensor as 250nM solution in buffer 1 (20mM Tris pH 7.5, 150 mM NaCl, 0.2mg/ml BSA, 0.02% v/vTween-20). The sensor was washed with buffer 1 for 120s. The sensor was washed with buffer 3 (10 nM glycine HCl, pH 2.5) for 120s and followed by again buffer 1 for 120s.

The sensor was then equilibrated with buffer 2 (20mM Tris pH 7.5, 150 mM NaCl, 2mg/ml BSA, 0.1% v/v Tween-20) for 120s. 2G12 (prepared in buffer 2) was associated at several concentrations (1, 2, 4, 8, 16, 32 and 64nM) for 300s, followed by dissociation into blank buffer 2 for 300s. After each dissociation sensor was regenerated to remove remaining 2G12 by treatment with buffer 3 for 120s, followed by 120s of wash with buffer 1. Throughout the experiment shaker rate was set at 2200 rpm. Data were fit to 1:1 binding model.