Stereoselective Polyglucuronic Acid Assembly via the Iron-Catalyzed Stereospecific Glycosylation with Glycal Epoxides

Abstract

Polyglucuronic acids (PGAs) are directly involved in essential biological processes, but catalytic methods for stereoselective PGA synthesis are underdeveloped because glucuronic acid (GlcA) donors are intrinsically difficult to activate. We report herein a convergent and stereoselective PGA assembly method via iron-catalyzed iterative and stereospecific glycosylation with glycal epoxides. This method is effective in the assembly of an array of GlcA-β-1,4-GlcA linkages in complex glycans and readily provides PGA octasaccharides in high stereoselectivity.

Polysaccharides participate in a variety of essential biological processes, and polyuronic acids play critical roles in maintaining cellular structures, enabling ion binding, and facilitating molecular recognition. , While synthetic approaches to poly-β-(1–4)-d-mannuronic acid (ManA) and α-l-guluronic acid (GulA) have been established, − methods for rapid and stereoselective synthesis of poly-β-(1–4)-d-glucuronic acid (GlcA) remain underdeveloped. This is presumably due to the intrinsically low reactivity of electron-deficient GlcA donors, which make them difficult to activate with a catalyst. − Furthermore, the nucleophilicity of the C4-OH group of a GlcA acceptor is attenuated by the strongly electron-withdrawing uronic ester moiety. As exceptions, Seeberger recently reported an automated synthetic approach to polyglucuronic acids (PGAs) that leverages iterative glycosylation with glycosyl phosphates (Figure A). This effective approach extends the backbone one sugar at a time. Zhu also developed a convergent PGA assembly method that efficiently provides PGAs up to 32-mer (Figure B). Therefore, new convergent and robust approaches for stereoselective PGA assembly are still highly desirable.

1.

(A) Polyglucuronic acid (PGA) synthesis via iterative glycosylation with glycosyl phosphates. (B) PGA synthesis via gold-catalyzed convergent glycosylation. (C) Stereoselective PGA assembly via the iron-catalyzed stereospecific glycosylation with glycal epoxides (this work).

In principle, iterative and stereospecific glycosylation with glucuronic ester epoxides could readily afford PGAs. However, most existing stereospecific methods for glucuronic ester epoxide activation are incompatible with sugar acceptors, especially hindered secondary sugar acceptors. − Recently, Jacobsen and our group − have developed metal-catalyzed, stereospecific methods to activate glucuronic ester epoxides toward a variety of nucleophiles. We discovered that iron­(III) porphyrin triflate catalysts induce highly stereospecific glycosylation of an array of hindered, secondary sugar acceptors with various glycal epoxides, which was previously challenging. −

Building upon this discovery, we envisioned that executing this iron-catalyzed stereospecific glycosylation between a glucuronic ester epoxide module and a C4-OH glucuronic ester acceptor module in an iterative and convergent manner would rapidly assemble PGAs (Figure C). Herein, we report the discovery of iron-catalyzed stereospecific and iterative glycosylation with these two challenging modules for stereoselective PGA synthesis.

We selected glucuronals 2 and 3 as the building blocks for catalyst discovery (Figure ). Epoxidation of a 4-O-silylated glucuronal 2 in a biphasic reaction medium with Oxone quantitatively afforded glucuronic ester α-epoxide S1 (Figure S1; dr >20:1; ³ J _H1–H2 = 2.4 Hz), which was azeotropically dried and used directly. Although a variety of hindered C4-OH acceptors are excellent substrates for this iron-catalyzed stereospecific glycosylation, C4-OH glucuronal acceptor 3 is a particularly difficult substrate. Extensive exploration of catalysts and other reaction conditions revealed that readily available, hemin-derived iron catalyst 1a (7 mol %) is optimal in promoting the glycosylation of 3 with the glucuronic ester epoxide derived from 2 at 0 °C, affording 4a/4b in good combined yield on a multigram scale (entry 1, 9.5 mmol scale, 74% yield, dr 4.6:1; for 4a, ³ J _H1–H2 = 7.6 Hz and ¹ J ¹³ _C1–H1 = 159.2 Hz; for 4b, ³ J _H1–H2 = 3.4 Hz and ¹ J ¹³ _C1–H1 = 171.3 Hz). Fortunately, acetylation of 4a/4b furnished 5a and 5b, which can be readily separated with a flash column. There are several noteworthy observations. First, glycosylation under cryogenic conditions (at −20 or −40 °C) does not improve the dr (entry 2), but it does induce further glycosylation of the newly generated C2-OH group in 4a/4b with the glucuronic ester epoxide, thereby decreasing the yield of 4a/4b (Figure , entry 2). Next, hemin-derived catalysts 1b and 1c with more coordinating sulfonate counteranions surprisingly led to decreased dr values (entries 3 and 4, respectively; dr from 1.6:1 to 2.5:1). We also observed that the use of TMSOTf induced significant glycal epoxide decomposition (entry 5, dr 2.4:1) and that AgOTf resulted in evidently lower reactivity and dr (entry 6, <40% conversion, dr 2.6:1). Other synthetically less accessible iron porphyrin triflate catalysts, including 1d/1e, were as effective as 1a (entry 7).

2.

Catalyst discovery for the iron­(III) porphyrin sulfonate-catalyzed glycosylation of glucuronals with glucuronic ester epoxides. ^a Epoxidation was carried out in a biphasic reaction medium with Oxone and acetone. The glycal epoxide was dried azeotropically with toluene, assayed by ¹H NMR (dr >20:1), and then directly used. The glycosylation was carried out at 0 °C in CH₂Cl₂. Acetylation of 4a/4b was achieved using Ac₂O/pyridine to afford 5a and 5b. ^b Isolated yield. The dr was determined by ¹H NMR analysis. Iron­(III) porphyrin catalysts 1a–1e were formed in situ from the corresponding iron porphyrin chloride catalysts (10 mol %) and AgOTf (7 mol %) to prevent the formation of an excess amount of AgOTf.

To evaluate the potential of this method for GlcA-β-1,4-GlcA assembly in complex substrates, we explored an array of synthetically valuable, GlcA-based glycosyl donors and acceptors (Figure ). We observed that a 3-O-silylated glucuronal epoxide is an excellent donor for both C4-OH glucuronal and GlcA acceptors (Figure , products 6 and 7, dr from 10:1 to >20:1). To our delight, iron catalyst 1a promotes entirely stereospecific glycosylation with the epoxides generated from disaccharides 5a (GlcA-β-1,4-glucuronal) and 5b (GlcA-α-1,4-glucuronal), affording triglucuronic esters in good yields (products 8 and 9, dr >20:1). Notably, several other readily available, disaccharide-based GlcA donors, including the epoxides derived from GlcN-α-1,4-GlcA, GlcN-β-1,4-GlcA, and GalN-β-1,4-GlcA, are excellent substrates for this iron-catalyzed glycosylation, furnishing single diastereomeric trisaccharides (products 10–13, dr >20:1). The observed high β-selectivity is presumably due to the stereospecific glycosylation with α-glycal epoxides obtained in high stereoselectivity. As we previously reported, this method is effective for highly stereospecific glycosylation for GlcA epoxide activation toward nitrogen-containing glycosyl acceptors.

3.

Substrate scope for GlcA-β-1,4-GlcA assembly via iron-catalyzed stereospecific glycosylation with glycal epoxides. All yields are isolated yields. See the Supporting Information for experimental details.

With four disaccharide modules, 5a, 5b, 14a, and 14b (through desiylation using TBAF/AcOH, 90% yield) in hand, we were ready to explore the iterative glycosylation for PGA synthesis (Scheme a). Epoxidation of GlcA-β-1,4-glucuronal 5a using Oxone almost quantitatively afforded glucuronic ester α-epoxide S9 (Figure S4; dr >20:1; ³ J _H1–H2 = 2.4 Hz). We discovered that iron catalyst 1a induces highly stereospecific glycosylation of the C4-OH group from a GlcA-β-1,4-glucuronal module 14a with this epoxide, affording tetrasaccharide 15 in good yield (65% yield, dr 10:1; for 15, ³ J _H1–H2 = 7.7 Hz and ¹ J ¹³ _C1–H1 = 162.4 Hz).

1. Iron-Catalyzed Stereospecific Glycosylation with Glycal Epoxides for PGA Tetrasaccharide and Octasaccharide Synthesis .

^a Conditions: (a) Oxone (7.0 equiv for 5a and 5.0 equiv for 5b), 5:1 CH₂Cl₂/acetone, saturated NaHCO₃ solution, 0 °C; (b) iron catalyst 1a (7 mol %) and 14a/14b (2.0 equiv) in CH₂Cl₂ at 0 °C; (c) DMDO (1.25 equiv for 19 and 1.5 equiv for 22), CH₂Cl₂/acetone, −10 °C; (d) iron catalyst 1a (14 mol %), 20 (2.0 equiv), CH₂Cl₂, 0 °C; (e) iron catalyst 1a (7 mol %), 23 (3.0 equiv), CH₂Cl₂, −20 °C; (f) TBAF (3 equiv), AcOH (3 equiv), THF, 22 °C; (g) H₂O₂ (30% aqueous solution, 75 equiv for 24 and 150 equiv for 25), LiOH (1 M aqueous solution, 25 equiv for 24 and 50 equiv for 25), THF, 22 °C; (h) Pd­(OH)₂/C, H₂, 1 atm, MeOH/NaH₂PO₄-Na₂HPO₄ buffer (pH 7, 1:1), 22 °C. All yields are isolated yields. See the Supporting Information for details.

Surprisingly, iron-catalyzed glycosylation of GlcA-α-1,4-glucuronal module 14b with the same glycal epoxide is only moderately stereospecific, affording an inseparable diastereomeric 16a/16b mixture in excellent combined yield (71% combined yield, dr 5.3:1; for 16a, ³ J _H1–H2 = 7.8 Hz and ¹ J ¹³ _C1–H1 = 162.8 Hz; for 16b, ³ J _H1–H2 = 3.7 Hz and ¹ J ¹³ _C1–H1 = 170.6 Hz). Analogously, epoxidation of GlcA-α-1,4-glucuronal 5b quantitatively afforded glucuronic ester α-epoxide S10 (Figure S6; dr >20:1; ³ J _H1–H2 = 2.5 Hz). We observed that iron catalyst 1a is effective in inducing highly stereospecific glycosylation of both GlcA-β-1,4-glucuronal module 14a and GlcA-α-1,4-glucuronal module 14b with this epoxide, furnishing tetrasaccharides 17 and 18 in good yields (66% yield for 17, dr 10:1; for 17, ³ J _H1–H2 = 7.6 Hz and ¹ J ¹³ _C1–H1 = 162.9 Hz; 64% yield for 18, dr 10:1; for 18, ³ J _H1–H2 = 7.3 Hz and ¹ J ¹³ _C1–H1 = 162.3 Hz).

Acetylation of 15 almost quantitatively generated tetrasaccharide donor 19, which can be smoothly desilylated (TBAF/AcOH, 86% yield) to afford its corresponding acceptor, 20 (Scheme B). Epoxidation of 19 furnished a glucuronic ester α-epoxide in >90% yield (dr >20:1; ³ J _H1–H2 = 2.4 Hz), which glycosylated tetrasaccharide 20 in the presence of catalyst 1a, delivering octasaccharide 21 in good yield (Scheme B, 65% yield, dr 10:1).

As a proof of concept, we selected tetrasaccharide 19 and octasaccharide 22 (>95% yield through acetylation of 21) for derivatizaion to provide PGAs (Scheme c). Epoxidation of 19 and 22 afforded the corresponding glucuronic ester α-epoxides in high yields (>90% yield; dr >20:1; ³ J _H1–H2 = 2.4 Hz), both of which readily glycosylated primary alcohol linker 23 with a chromophore, furnishing 24 and 25, respectively (66–72% yield, dr >20:1). Desilylation of 24 and 25 with TBAF in AcOH, followed by hydrolysis (LiOH with H₂O₂) and global hydrogenolysis of benzyl groups, ultimately led to PGAs 26 and 27, respectively, in good yields (72–79%).

In conclusion, we have developed a convergent and stereoselective polyglucuronic acid assembly method via iron-catalyzed stereospecific glycosylation with glycal epoxides. This iterative and stereospecific glycosylation is effective for the synthesis of an array of GlcA-β-1,4-GlcA linkages in complex glycans, and it rapidly provides PGA tetrasaccharides and octasaccharides. Our current effort focuses on applications of this method for assembly of other synthetically challenging, uronic acid-containing complex glycans.

The data underlying this study are available in the published article and its Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.orglett.6c03166.

• Experimental procedure, characterization data for all new compounds, and selected NMR spectra (PDF)

The authors declare no competing financial interest.