Synthesis of 4-O-Alkylated N-Acetylneuraminic Acid Derivatives

The synthesis of 4-O-alkyl analogues of N-acetylneuraminic acid (Neu5Ac) and the scope of the reaction are described. Activated alkyl halides and sulfonates and primary alkyl iodides give products in useful yields. The utility of the methodology is exemplified using a thiophenyl Neu5Ac building block to synthesize a 4-O-alkyl DANA analogue. These results expand the toolbox of Neu5Ac chemistry with value in drug discovery and for the design of novel tools to study the biology of Neu5Ac lectins.

* sı Supporting Information ABSTRACT: The synthesis of 4-O-alkyl analogues of N-acetylneuraminic acid (Neu5Ac) and the scope of the reaction are described. Activated alkyl halides and sulfonates and primary alkyl iodides give products in useful yields. The utility of the methodology is exemplified using a thiophenyl Neu5Ac building block to synthesize a 4-O-alkyl DANA analogue. These results expand the toolbox of Neu5Ac chemistry with value in drug discovery and for the design of novel tools to study the biology of Neu5Ac lectins.
N-Acetylneuraminic acid (Neu5Ac, 1, Figure 1) is typically found at the terminal end of glycolipids and glycoproteins that decorate the surfaces of all mammalian cell types. Neu5Ac is involved in mediating or modulating a variety of physiological and pathophysiological processes. 1 One of the most wellknown roles of Neu5Ac is in the replication cycle of the influenza virus. 2 Accordingly, substantial efforts have been placed on the development of Neu5Ac-based antivirals, 3 where modifications of the C4-position of 2-deoxy-2,3-didehydro-Nacetylneuraminic acid (DANA, 2, Figure 1) have been of central importance. 4−12 This culminated in the development of Relenza (3, Figure 1), a C4-modified analogue of 2 designed to mimic the transition state of 1 during the neuraminidase catalyzed hydrolysis reaction required for release of virus progeny from infected cells. 5 C4-modified analogues of 2 including nitrogen, 4 sulfur, 4 and deoxygenated 7 compounds are efficiently accessed via selective ring opening at position 5 of the allylic oxazoline of 2,3-didehydro-N-acetylneuraminic acid (4, Figure 1). 4 However, in the case of oxygen nucleophiles, opening occurs at position 2 of the oxazoline ring. 7 Hence, the method is not applicable to the synthesis of 4-O-modified analogues of 2 (or 1) with retained stereochemistry. C4-deoxy and C4-nitrogen analogues of 1 can, however, be accessed using the ring-opening methodology but require reinstallment of the glycosidic bond, which produces two stereoisomers of equal proportions. 13,14 The interest in Neu5Ac analogues and their roles in biological systems is constantly increasing. Therefore, efficient methods that allow site-selective modifications of the Neu5Actemplate are of great utility for studying Neu5Ac biology and for drug discovery. Methods to selectively access C4-modified analogues of 1 are scarce, with relatively few reported examples. These include carba, 15,16 keto, 16 ether, 14,17−20 nitrogen, 13 and deoxygenated 21 derivatives. A potential drawback in the development of direct methods could be the competing formation of intramolecular lactams 22−24 and lactones 25 that occur under both basic and acidic conditions. To date, examples of selectively 4-O-modified Neu5Ac analogues include 4-O-Ac, -benzyl, 26−30 -allyl, 23 -silyl, 15 -methyl, 14,17,18 -ethyl, 18 -cyanomethyl, 19 and -tert-butoxyacetate 20 groups. The electrophiles used to produce these 4-Omodified analogues all have in common that they are activated, highly reactive, and (with a few exceptions) lack the presence of β-hydrogens. Further, the commercial availability of suitable electrophiles remains limited. Herein, we set out to study the scope and the 4-O-alkylation of Neu5Ac.
In an ongoing research project, we were interested in studying the potential of 4-O-alkyl analogues of 5 ( Figure 1) as probes targeting cell attachment during adenovirus 31−33 and coxsackievirus infections. 34,35 We hypothesized that 6 (Scheme 1) would be a suitable substrate to study O-alkylation. This previously described protective group strategy is straightforward, high yielding, and allows removal of the protective groups in a final single step. 15 Propargyl bromide was selected as the model electrophile. The lack of β-protons minimizes the competing E2 reaction, thus providing a fair measure of the effectiveness of the S N 2 reaction. In addition, the generated alkynyl product can be further modified under mild conditions. 36,37 Compound 5 38 (Figure 1) was obtained from 1 and then converted by standard methods to the known derivative 7 39 (Scheme 1). Treatment of 7 with wet trifluoroacetic acid in DCM afforded 9, which upon acetylation gave the fully protected derivative 11 in 79% yield over two steps. Treatment of 11 with TBAF in THF afforded 6 in 81% yield.
Attempts to alkylate 6 using Ba(OH) 2 /BaO in DMF, 23,30,40 K 2 CO 3 in THF, or CsCO 3 in MeCN, resulted in minimal amounts of 16. However, promising observations were made in THF using KHMDS or NaH, with NaH providing superior conversion and product formation. Standard O-alkylation conditions were screened by treating 6 with NaH on ice prior to the addition of propargyl bromide (entries 1 and 2, Table 1). In DMF, this resulted in nearly complete decomposition and only trace amounts of 16 (entry 1). However, in THF, the 4-O-propargylated derivative 17 was isolated in 50% yield over two steps (entry 2). The Odeacetylation was performed to compensate for the formation of hydrolyzed species during the reaction ( Figure S1) and, thus, facilitate isolation of the desired 4-O-alkyl product. The O-deacetylation of purified 16 gave 17 in 85% yield, the value that was used to estimate the yield for the O-alkylation (Table  1).
Alkoxide formation was studied by mixing compound 6 with NaH in DMF-d 7 and in THF-d 8 , respectively, and recording 1 H NMR spectra at two different 10 min and 1 h ( Figure S2A− F). Within 10 min, compound 6 was essentially consumed in DMF-d 7 , resulting in a complex mixture of products ( Figure  S2A,B). In contrast, only minimal signs of degradation were observed in THF-d 8 10 min postaddition of NaH ( Figure  S2D,E), and the majority of 6 was largely intact after 1 h ( Figure S2F). This prompted us to reverse the addition order, and compound 6 was mixed with propargyl bromide in the selected solvent on ice before adding NaH. Furthermore, the stoichiometry of NaH was increased from 1.1 to 2.0 equiv to ensure complete deprotonation of both the hydroxyl and acetamide of 6. These modifications drastically improved the yield of 17 in DMF (40%; entry 3, Table 1), while no significant effect was observed in THF (52% yield; entry 4). This highlights the importance of avoiding preformation of the alkoxide in DMF. The 4-O-alkylated product 17 was confirmed by 2D NMR analysis and by treatment with acetic anhydride in pyridine, which afforded 16 in 60% yield.
Common solvents for O-alkylation reactions were screened, and the yields of 17 were lower in both 1,4-dioxane (43%; entry 5, Table 1) and MeCN (35%; entry 6) compared to the reference reaction (entry 4). The reaction in toluene (entry 7) was slow, with incomplete conversion after 72 h of stirring, likely due to poor solubility, and was not processed further. Decreasing the stoichiometry of propargyl bromide to 1.1 and 2.0 equiv afforded 17 in 22% and 39% yields, respectively (entries 8 and 9), while increased stoichiometry gave 17 in  Table 1. Screening and Optimization of Reaction Conditions 44% yield (10 equiv; entry 10) and resulted in a larger concentration of side products.
Reduced stoichiometry of NaH gave 17 in 47% yield (1.0 equiv; entry 11, Table 1) with incomplete conversion, while increased stoichiometry provided 17 in 31% yield (5.0 equiv; entry 12) with larger amounts of side products, suggesting the stoichiometry of NaH should be greater than one but less than two equivalents to ensure complete conversion and minimize the formation of side products. Indeed, 1.1 equiv of NaH gave a clean reaction and complete conversion, albeit without improvement of the yield (47%; entry 13). Decreased substrate concentration gave 17 in 33% yield (0.05 M; entry 14). Pleasingly, increased concentration produced 17 in 70% yield (0.3 M; entry 15), corresponding to a 35% improvement compared to the reference reaction. Higher concentration was associated with solubility issues but provided 17 in 45% yield (1.0 M; entry 16). With the optimized conditions in hand, the stoichiometry of NaH was adjusted to 1.5 equiv as the conversion was incomplete in some reactions when using 1.1 equiv. This resulted in complete conversion of 6, providing 17 in 67% yield (entry 17). To conclude, the optimal conditions are a concentration of 0.3 M (in THF), 5.0 equiv of propargyl bromide, and 1.1−1.5 equiv of NaH.
In an attempt to further improve the yields of the Oalkylation, compounds 15 and 18 were prepared (Scheme 1). Compound 15 with its tertiary amide renders it resistant toward potential side products arising from lactamization. 22−24 Compound 15 was accessed from 11 by treatment with Boc anhydride and DMAP in THF followed by cleavage of the TBDMS group using TBAF (Scheme 1). Surprisingly, the reactivity of 15 was completely abolished toward O-alkylation (entry 1, Table 2).
Prolonged reaction times (2.5 h), heating (60°C for 16 h, with a heating mantle), and irradiation in a microwave reactor (100°C for 20 min) were inefficient in causing conversion. Compound 18 was prepared from the known derivative 19 41 including selective reduction with borane-trimethyl amine and aluminum chloride in THF affording 20 in quantitative yield that upon treatment with 2,2-dimethoxypropane and camphor sulfonic acid in MeCN gave the 9-O-benzyl-7,8-acetonide protected 18 in 96% yield. This protective group strategy is orthogonal allowing site-selective removal and functionalization of the glycerol side chain (C7, C8, and C9). Further, the protective groups have increased resistance toward hydrolysis under basic conditions. Upon O-alkylation 18 gave 22 in 57% yield (entry 2). Compound 13 was prepared in analogueous manner to 6 (Scheme 1), and upon O-alkylation afforded 23 in 74% yield (entry). Compound 23, and analogues thereof, significantly broaden the scope due to their potential for modifications at the C2-position via glycosylation, or elimination to access 4-O-alkyl DANA analogues. 19 The developed conditions were applied to synthetic intermediates 7 and 8 which provided 24 and 25 in 57% and 71% yields, respectively (entries 4 and 5). Thus, supporting access to 7-Oalkylated species. Synthetic intermediates 26 and 27 selectively afforded the 4-O-alkylated products 28 and 29 in 31% and 49% yields (entries 6 and 7), respectively. Thus, significantly decreasing the number of steps to access 4-O-alkylated analogues of 18.
Representative examples of commercial alkyl halides and sulfonates were then screened to study the scope of the 4-Oalkylation of 6 ( Figure 2). As expected, the activated alkyl bromides benzyl bromide, allyl bromide, and ethyl bromoace-tate afforded the corresponding 4-O-ethers 30 (62% yield), 31 (52% yield), and 32 (78% yield) ( Figure 2), respectively. An initial reaction with 6-iodo-1-hexyne gave 33 in a mere 10% yield. Dipolar aprotic solvents are known to increase the rate of substitution due to their ability to solvate cations, 42 and the use of DMF indeed afforded 33 in 45% yield. Using 6-chloro-1hexyne resulted in trace amounts of 33, and the addition of TBAI, or KI, did not result in the isolation of 33 in either THF or DMF. Upon O-alkylation 5-bromo-1-pentene afforded 34 in 17% yield. Ethyl tosylate gave 35 in poor yield (8%) with substantial amounts of hydrolyzed starting material. However, propargyl mesylate gave 17 in 58% yield, supporting the use of activated alkyl sulfonates. Last, we attempted to substitute 2bromopropane, which resulted in trace amounts of 36 ( Figure  2), in line with the fact that 2°halides are less reactive in Oalkylation reactions due to excess β-protons favoring an E2 pathway over the desired substitution reaction. 43 To exemplify the utility of the developed methodology, we purified intermediate 37 and treated it with TfOH and NIS in DCM, 44 affording the 4-O-propargyl DANA analogue 38 in 87% yield (Scheme 2), thus confirming access to 4-O-alkyl Table 2. O-alkylation of Diversely Protected Neu5Ac Building Blocks a a All reactions were conducted in THF (0.3 M substrate) and performed by treating a stirred solution of 5.0 equiv of propargyl bromide and substrate with 1.1−1.5 equiv of NaH (specific details in Supporting Information). n.r. = no reaction. Yield over two steps. In summary, we have systematically studied O-alkylation of Neu5Ac derivatives and provided insights into the scope of the reaction for preparation of tool compounds and starting points for drug discovery. 51 ■ EXPERIMENTAL SECTION General Chemical Procedures. 1 H NMR and 13 C NMR spectra were recorded with a Bruker DRX-400 spectrometer at 400 and 100 MHz, respectively, or with a Bruker DRX-600 spectrometer at 600 and 150 MHz, respectively. NMR experiments were conducted at 298 K in CD 3 OD (residual solvent peak = 3.31 ppm, δH and 49.00 ppm, δC) or CDCl 3 (residual solvent peak = 7.26 ppm, δH and 77.16 ppm, δC). Liquid chromatography−mass spectrometry (LC−MS) data were recorded by detecting positive/negative ions (electrospray ionization, ESI) on an Agilent 1290 Infinity II-6130 Quadrupole using H 2 O/CH 3 CN (0.1% formic acid) as the eluent system or on an Agilent 1290 Infinity-6150 Quadrupole using YMC Triart C18 (1.9 μm, 20 mm × 50 mm column) and H 2 O/CH 3 CN (0.1% formic acid) as the eluent system. High-resolution mass spectrometry (HRMS) data were recorded on an Agilent 1290 binary LC System connected to an Agilent 6230 Accurate-Mass Time-of-Flight (TOF) LC−MS (ESI+), which was calibrated with Agilent G1969−85001 ES-TOF Reference Mix containing ammonium trifluoroacetate, purine, and hexakis(1H,1H,3H-tetrafluoropropoxy)phosphazine in 90:10 CH 3 CN/H 2 O. Semipreparative high-performance liquid chromatography (HPLC) was performed on a Gilson system using a YMC-Actus Triart C18, 12 nm, S-5 μm, 250 mm × 20.0 mm with a flow rate of 20 mL min −1 , detection at 214 nm, and eluent system A with aqueous 0.005% formic acid, and B with CH 3 CN 0.005% formic acid. Thinlayer chromatography (TLC) was performed on silica gel 60 F254 (Merck) with detection under ultraviolet (UV) light and/or development with 5% H 2 SO 4 in EtOH and heat. Flash chromatography was performed using a Biotage Isolera One system and purchased prepacked silica gel cartridges (Biotage Sfar Silica). Lyophilization was performed by freezing the diluted CH 3 CN/water solutions in a dry ice−acetone bath or liquid nitrogen and then employing an Alpha 3−4 LSCbasic freeze-dryer. Organic solvents were dried using a Glass Contour Solvent System (SG Water USA). All commercial reagents were used as received. All target compounds were ≥95% pure according to HPLC UV traces, unless otherwise noted.
General Procedure for O-Alkylation (GP1): Exemplified with the Synthesis of 16. An oven-dried vial was charged with a magnetic stirring bar and compound 6 (40 mg, 0.086 mmol). The vial was placed under nitrogen, and THF (0.3 mL) followed by propargyl bromide (48 μL, 0.43 mmol, 5 equiv) were added. The mixture was cooled to an ice-bath temperature, and NaH (3.8 mg, 0.095 mmol, 1.1 equiv) was added in portions. After 10 min, the reaction was allowed to perform at room temperature for an additional 2 h (monitored by TLC/EtOAc; R f = 0.43). After completion, the reaction was quenched by the addition of a few drops of sat. aq NH 4 Cl, and the solvents were removed under reduced pressure. The crude product was directly used in the deacetylation step, unless otherwise noted.