Syntheses of small cluster oligosaccharide mimetics

We designed multiple Small Cluster Oligosaccharide Mimetics (SCOMs) - potential glycosidase inhibitors - to be metabolically stable and small enough to enter cells or bacteria. Therefore, minimal scaffolds (urea, amide, ammonia) or simply non-glycosidic linkages of carbohydrate structures were central to our synthetic strategy, including: (a) coupling of several natural carbohydrate precursors; (b) total syntheses of aminomethyl tetrahydropyrans and their chiral amides with quinic acid; (c) glycopyranosyl cyanide reduction to prepare crowded clusters on a urea scaffold; (d) total syntheses via cycloadditions leading to amide-linked C -glycosides; (e) reduction of nitromethyl C -glycosides; and (f) a synthesis of hydroxylated 1,2-cyclohexanedicarboxylic acids.


Introduction
After the early years of carbohydrate chemistry around 1900 and Emil Fischer's structure proof of the carbohydrates, this area of chemistry received only little attention.2][3] Synthetic efforts since then have resulted in numerous structural modifications for carbohydrates.Exchange of aglycons on the anomeric carbon, exchange of the glycosidic atom (N-, C-, S-glycosides), exchange of ring substitutents, exchange of the endocyclic heteroatom (carba-, aza-, thia-derivatives), and concise streochemical transformations.Oligosaccharides are now recognized to have functions influencing the entire spectrum of cell activities.3][14] Aberrations in glycosylation have been linked to severe pathological symptoms.Monosaccharide units combine to oligosaccharides in an almost infinite number of permutations, determined by the stereochemical identity of the linked monosaccharides (e.g.glucose, galactose, mannose), their glycosidic linkage position (e.g.1→4, 1→6), their anomericity (α or β), and the overall degree of branching.This complexity, further increased by additional modifications such as sulfation or sialation, was suggested recently to serve the evolutionary purpose of "herd immunity". 15,16It has become clear over the past three decades that glycosylation is critically important to many of the signaling pathways that turn a normal cell into a cancer cell.Compounds that inhibit specific glycosylation reactions may potentially block the pathways in carcinogenesis.Carbohydrates have been recognized as novel cancer prevention agents. 17owever, a distinct disadvantage of naturally occurring carbohydrates, e. g.O-glycosidically-linked oligosaccharides, is their metabolic instability in biological systems.5][26][27][28][29] Potential targets for medical applications of these clusters have been identified. 305][26][27][28][29] Therefore, we devoted our work to the synthesis of a variety of Small Cluster Oligosaccharide Mimetics (SCOMs) 31,32 with enhanced metabolic stability, which presumably may enter the cytosol and internal cell substructures to address internal receptors.

Acylation with Quinic Lactone
Quinic acid (1, Figure 1) 33 and its derivative shikimic acid (2, Figure 1) [34][35][36] (from Illicium japonica and floridanum) are key intermediates during the biochemical synthesis of essential aromatic amino acids, such as L-tryptophan, L-phenylalanine, and L-tyrosine in plants, bacteria, and fungi.Quinic acid and shikimic acid are also required as building blocks for the assimilation of folic acid, alkaloids, and vitamins in those organisms. 37,38The quinic acid/shikimic acid pathway does not exist in mammals, which has led to extensive research activities geared towards the development of suitable antibiotic therapeutics.Especially the conversion of 3dehydroquinic acid (DHQ) to 3-dehydroshikimic acid (DHS) has been recognized as an effective target for inhibition in plants. 39,40Several commercial herbicides selectively inhibit the DHQ-DHS conversion and prevent growth of weeds.However, disruption of the DHQ-DHS pathway by shikimic acid derivatives (3, Figure 1) has also recently been shown to cause therapeutic inhibition of Mycobacterium tuberculosis.3][44] Potent anti-influenza A+B effect has been observed for derivatives of shikimic acid inhibiting viral neuraminidase.The widely-prescribed anti-flu medication Tamiflu TM (Roche, 4, Figure 1) is not only effective in treating common flu strains, but it has also been used successfully in the treatment of the bird (avian) flu in laboratory animals.Contrary to animal experiments, Tamiflu's TM effectiveness in humans infected with the H5N1 flu strain is still controversial because only a few studies with limited statistical value have been conducted, and the understanding of the drug's action against the virus is only rudimentary.Tamiflu TM acts as influenza virus neuraminidase inhibitor with the possibility of alteration of virus particle aggregation and release.The less well-known Relenza TM (Zanamivir, GlaxoSmithKline, 5, Figure 1) 42 is also an effective inhibitor of common flu strains by the same mechanism.The problem of a sufficient supply of these two drugs in case of an influenza pandemic and the necessity for a continuous search for new structural analogs has been discussed in the literature. 45,468][49] Starting with 2-acetamido-2deoxy-D-glucopyranose (6), we prepared the corresponding benzyl glucoside (7) through acidcatalyzed acetal formation.Under these conditions, compound 7 was obtained as a mixture of anomers, which was difficult to separate or crystallize.However, the α-and β-anomer can be obtained clean after per-O-acetylation, recrystallization, and MeO -/MeOH deprotection at room temperature. 50Compound 7 was then converted to the 4,6-O-benzylidene derivative 8 with benzaldehyde and anhydrous ZnCl 2 followed by removal of the N-acetyl group in ethanolic KOH at reflux to yield 2-amino-4,6-O-benzylidene-2-deoxy-α/β-D-glucopyranose (9).In parallel, D-(-)-quininc acid was converted to the 4,5-O-isopropylidenated lactone (11), 51 which served as an electrophile in the subsequent acylation of the separated α-and β-anomers 9a/b into the potential disaccharide mimic precursors 12a/b, respectively.
We also prepared compounds 21 and 22 as regio-and diastereochemical analogs of 12, respectively.The two corresponding aminoglycosides were synthesized from D-glucose (14). 52enzylation followed by 4,6-O-benzylidenation gave compound 16 whose 2-and 3-hydroxyl groups were subsequently mesylated to afford compound 17.Upon treatment of 17 with methoxide, we were able to obtain benzyl-2,3-anhydro-4,6-O-benzylidene-α-D-allopyranoside (18).The epoxide was surprisingly stable but could be opened with NH 3 under pressure and elevated temperature.The major product (20) with D-altro-configuration was obtained through diaxial opening (Fuerst-Plattner rule) whereas the regioisomer (19) with D-gluco-configuration was only isolated in small quantities from the same reaction.When these products were combined with quinic lactone 11, SCOMs 21 and 22 were obtained, respectively.
Interestingly, the quinic amide ring in all quinamides described above had a predominant twist-boat conformation as determined by NMR in CDCl 3 (Scheme 1). 47,48This was confirmed by small gauche 3 J-values and NOEs between H3/H2a/H2b and between H5/H6a/H6b, as well as a significant 4 J between H4 and H2b (W-coupling).We observed the same conformation for the unsubstituted quinamide 13 that was obtained from 11 in a saturated methanolic solution of NH 3 at room temperature.This benzylamide derivative had been described by us earlier; however, without detailed conformational analysis (Scheme 1). 49The twist-boat conformation is pretty common for the pyranoses/tetrahydropyranes fused to five-membered acetal ring(s) (e.g.see Section 4.1), but is very rare for cyclohexane derivatives.
3][64] In the first step a per-Oacetylated sugar was treated with TiBr 4 in CH 2 Cl 2 in presence of a catalytic amount of glacial acetic acid.The resulting glycosyl bromide was in situ reduced with Zn/Cu couple.Per-Oacetylated D-glucose, D-galactose, D-glucuronic acid methyl ester, L-fucose, D-maltose, D-lactose, and D-gentiobiose gave corresponding glycals 23-30 (Figure 2) in > 90% purity prior to purification, as estimated by NMR. 62 [64] Dimerization of per-O-acetylated D-glucal was described first by Ferrier and Prasad. 653][64][66][67][68][69] 5][26][27][28][29] The non-physiological nature of the new C-C-linkages in our compounds may transfer metabolic stability also to O-glycosidic bonds still present in their structures.Dimerizations of disaccharidic glycals to tetramers required careful optimization of conditions to minimize cleavage of the disaccharidic bonds by the BF 3 -catalyst.
The structural assignment of glycal dimers was based upon NMR data.The original assignment for tri-O-acetyl-D-glucal dimer 31 65 was confirmed by our data, 64,66 and finally proved by X-ray crystal structure analysis (Figure 3). 66 66 The structural analysis of 31 revealed a high degree of rigidity in the molecule.Comparative analysis of the NMR and X-ray data 62,64,66 shows that there is virtually no difference between the solution and crystal conformations of both rings in compound 31.The average position of the two ring planes in 31 may be considered nearly perpendicular.The conformational equilibrium for both rings appears to be strongly biased.An absence of a solvent dependence for spin-spin coupling constants also supports these conclusions.
The results of this work combined with literature data allowed us to suggest a set of "standard" coupling constants for use in structural analysis of 2,3-dideoxy-2-enopyranose systems with α-erythro configuration in the unusual 5 H O -conformation and more common O H 5conformation (Figure 4). 62,64,66

Glycosyl cyanides from glycosyl acetates
Another way to render carbohydrate-derived structures metabolically more stable is offered by introducing a C-glycosidic cyanide functionality.In 1961 Helferich and Bettin obtained per-Oacetyl-β-D-galactopyranosyl cyanide in good yield from per-O-acetyl-α-D-galactopyranosyl bromide and mercuric cyanide in nitromethane. 70However, the analogous reaction of per-Oacetyl-α-D-glucopyranosyl bromide [71][72][73][74] gave per-O-acetyl-1,2-O-(1-cyanoethylidene)-α-Dglucopyranose 36 in 53% yield, and only 11% of the preferred per-O-acetyl-β-D-glucopyranosyl cyanide 37.The preparation of per-O-acetyl-glycopyranosyl cyanides by Myers and Lee brought little improvement. 75,768][79][80] Thus, per-O-acetyl-α/β-D-galactopyranoses could be directly converted into per-O-acetyl-β-D-galactopyranosyl cyanide.But, BF 3 -OEt 2 is hazardous, unstable, and harsh, and significant decomposition and incompatibility with glycosidic bonds were observed.However, the elimination of the unstable glycosyl bromides as intermediates was a distinct advantage.We decided to investigate the use of HgBr 2 , as a milder catalyst for this reaction (Schemes 4,5). 81 Scheme 4. Synthesis of per-O-acetyl-β-D-glucopyranosyl cyanide 37. [81][82][83] When per-O-acetyl-α-or β-D-glucopyranose was treated with excess TMS-CN and 0.1 molar equiv of HgBr 2 in nitromethane for 2 days in the absence of Hg(CN) 2 , the major product was the 1,2-cyanoethylidene derivative 36 in 59% yield, along with only 5% of the preferred cyanide 37, and about 28% of starting material (Scheme 4), [81][82][83] similar to the results of Coxon and Fletcher with Hg(CN) 2 . 71When we increased the amount of HgBr 2 to at least 0.5 equiv, the cyanoethylidene intermediate rearranged into the desired per-O-acetyl-β-Dglucopyranosyl cyanide 37 in 51% yield starting from per-O-acetyl-β-D-glucopyranose, or 65% yield from purified cyanoethylidene intermediate 36, within 1 day.Yields of rearrangement decreased to 5% when no TMS-CN was added along with the catalyst, and complete decomposition occurred on warming.The omission of Hg(CN) 2 was crucial.This suggested that rearrangement required the assistance of a mercuric isocyanide complex in conjunction with the existence of equilibrium between the exo-cyanoethylidene 36 and the acetoxonium intermediate. 82,83 84no et.al. 85 have already done extensive studies on the conformations of per-O-acetyl-1,2-O-(1-cyanoethylidene)-α-D-glycopyranoses, which were obtained from the reactions of glycosyl bromides with potassium cyanide in tetrabutylammonium bromide/acetonitrile. 86 X-Ray crystallographic 84 and NMR analyses of the enantiomerically pure cyanoethylidene intermediate 36 from our procedure showed it to exist in a twist-boat conformation of the six-membered cycle, with the cyanide exclusively in the exo-position (Figure 5).

Glycosyl cyanides from glycals and glycal dimers
DeLas Heras et al. have shown that glycals, as well as glycosyl acetates, can be transformed with BF 3 -Me 3 SiCN in nitromethane into 2,3-unsaturated glycosyl cyanides, or glycosyl cyanides, respectively. 79The complex stereospecificity, if any, of such S N '-reactions has been discussed by March. 88More recently, alternative methods for the preparation of 2,3-unsaturated glycosyl cyanides from protected or unprotected glycals have been developed. 89,9064]92 The reagent was compatible with 1,4-and 1,6-O-glycosidic linkages, which are normally prone to cleavage in the presence of BF 3 . 7964]92 Similarly, the galactal dimer 32 gave a doubly unsaturated cyanide 46.Although the dimers 32 and 33 had only a stereochemical difference with 31 in the relative disposition of the 4-acetoxy group, the substitution of the anomeric acetoxy group was attended by elimination of AcOH, with deprotonation at C-2. [62][63][64]92 A relief of steric strain may be a driving force for this elimination. Inded, the molecular mechanics calculations (MM+) revealed a substantially higher steric strain (by 3-4 kcal/mol) in the most stable conformation of 33 when compared with 6A,6B-dideoxy analog of 31, which is a stereoisomer of 33.The remote C-4 A acetoxy group, which functions as an internal base, may facilitate the deprotonation of an intermediate oxonium ion by intramolecularly removing and transporting the proton ultimately to external CN -, like a miniature crane.Such a "crane effect" does not work in 31, which has a different orientation of the C-4 A acetoxy group.[62][63][64]92 ARKIVOC 2008 (i) 271-308 ISSN 1551-7012 Page 282 © ARKAT USA, Inc.

Reduction of glycosyl cyanides
5][96] The yield for the N-linked aminodisaccharide 50 decreased to 9% without the presence of Ac 2 O. Since the presence of Ac 2 O did not prevent migration, it appeared that O-N-acetyl migration from a 2-O-acetoxyl group was much faster than acetylation of the amines.5][96] Again, the yield of disaccharide 54 was increased in presence of the anhydride.
Cyanides are reduced with H 2 /Pd in two stages. 97The first stage is formation of aldimine, and the second is reduction to the corresponding amine.In the production of 49 and 50, O-Nacetyl migration could occur at the aldimine or amine stage.The migration at the stage of amine made it unavailable for further reaction.That no acetyl migratory product was observed in the presence of Boc 2 O suggests fast reaction of aldimines and amines with Boc 2 O at the lipophilic charcoal catalyst surface.In contrast, hydrophilic Ac 2 O, present in the bulk solution, permits acetyl migrations at the surface.The more electrophilic acylated aldimines tend to favor dimerization, which indeed occurred more readily in the presence of Ac 2 O and Boc 2 O. 81,82,94,95 This point was not appreciated by Lentz et al. in their later work with Boc 2 O, 98 probably because they did not compare hydrogenation in the presence of Ac 2 O, as we did.
From the Boc-protected aminomethyl-C-fucopyranoside synthons, a great variety of biologically interesting glycoconjugates are accessible.Transformation of the 1° amino functionality into an isocyanate allows coupling with the former 1° amine or the 2° amine dimer.Other nucleophiles would give glycoconjugates with minimal scaffolding such as unsymmetrical ureas.This could be accomplished in many ways, but most commonly by phosgenation, 99 which is economical and practical, since the advent of crystalline "triphosgene".][96] When the same deprotection and two-phase phosgenation was applied to the N-linked disaccharide 54, a novel tetra-fucopyranosyl methyl-substituted urea 57 (28%) was obtained along with the O-N-acetyl migratory side-product 50 (48%) and the stable crystalline carbamoyl chloride 56 (15%; Scheme 8) -a potentially useful reactant for preparation of more complex Cglycosidic conjugates.The yield for the tetra-substituted urea derivative was surprisingly high for a bimolecular reaction of two sterically hindered intermediates, which had to compete with the major intramolecular pathway of acetyl migration, after Boc removal.Formation of acetyl migratory side products 49 and 50 was increased if a weaker acid such as trifluoroacetic acid was employed to remove the Boc-protective group, or if Hűnig's base was used in a single non-polar organic solvent instead of a two-phase system with aqueous bicarbonate.With an excess (1/3 equivalents) of triphosgene, acetyl migratory product was minimized, and formation of the tetrasubstituted urea also ceased, but the yield of the interesting carbamoyl chloride 56 increased to 67%.[96] Scheme 8 summarizes the synthetic possibilities for the two readily obtainable Boc-aminomethyl saccharides 53 and 54.The utility of compounds 53, 54, and 56 is readily appreciated, for attaching stable β-L-fucopyranosyl residues via amide, urethane, or urea bonds to carboxyl, hydroxyl, or amino groups of proteins or other biological scaffolds, to produce e.g."neoglycoproteins."This methodology should be extendable to other sugars.
We have shown that O-acetyl protective groups are removable under very mild conditions, with NEt 3 /MeOH/H 2 O. 100 The base and the byproduct AcOH are smoothly removed by azeotropic water vapor distillation in vacuo, or presumably also by freeze drying, without damage to delicate biochemical substrates.

Henry Condensation
One of the most common methods for the chain extension of carbohydrates is the Henry condensation of pentoses and hexoses with nitromethane in the presence of a strong base.In 1946 Sowden and Fisher condensed 4,6-O-benzylidene-D-glucopyranose 59 (Scheme 9) with nitromethane in the presence of methoxide to give D-heptitol 60 and nitromethyl β-D-glucopyranoside 63 in 21% and 5% yield, respectively. 101Later, Petrus et.al. improved the total yields of unprotected nitromethyl D-hexopyranosides, but their approach is lengthy and complicated, requires careful workup with adjustment of pH, and produces isomeric mixtures. 102rong bases (HO -, MeO -) in these procedures in protic solvents lead to open-chain products (i.e.60, Scheme 9). 101Their hot dehydration gives cyclic forms, reverting only very slowly to acyclic at low temperatures. 103Presumably, acyclic forms, having more hydroxy groups, are stabilized by solvation in protic solvents.We had synthesized 82,104-107 directly the cyclic compounds in aprotic solvents with a highly active bifunctional 108 catalyst 2-hydroxypyridine 109 and DBU/molecular sieves 104 with an improved yield of 77%. 105We avoided protic solvents because of the strong basicity of DBU, which would provide just another way of creating RO -in ROH.The cyclic product 63 was produced in a single step along with two minor byproducts 60 and 64. 104e reinvestigated later 82,106,107 our original condensation 104,105 of nitromethane with 4,6-O-benzylidene-D-glucopyranose 59 to nitromethyl 4,6-O-benzylidene β-D-glucopyranoside 63 and discovered that 2-HP was unnecessary.A time study, with careful TLC analysis and flash chromatography of the reaction mixture with DBU and molecular sieves revealed sequential products, which are conveniently explained as outlined in the mechanistically simplified Scheme 9.
The condensation progressed similarly in THF and dioxane, with 4,6-O-benzylidene-D-glucopyranose via 5,7-O-benzylidene-1-deoxy-1-nitro-D-glycero-D-gulo-heptitol 60, which subsequently changed into nitromethyl β-D-glucopyranoside 63.The former was observed accumulating before being further converted during at least 50 hours at room temperature to the cyclic product 63, along with some 5,7-O-benzylidene-1,2-dideoxy-1-nitro-2-nitromethyl-D-gluco-heptitol 64, in yields similar to those already reported. 104,105In THF or dioxane, accumulation of the final acyclic product 64 was dependent on temperature, time, and solvent: less decomposition was observed in dioxane.In these condensations, we could not determine if the presence of 2-HP had any effect at all, at any stage of the reaction.In any case, the first heptitol 60 and the desired 63 were kinetically controlled intermediates, which, given enough time, were transformed into the "thermodynamic" product 64, by a still slower overall rate.
When similar conditions were employed for unprotected D-glucose 58, there was no reaction for prolonged periods of time.Heating only resulted in almost complete decomposition.This was expected since unprotected sugars are typically not very soluble in THF or dioxane.But in acetonitrile, pyridine or a mixture of these at 50 °C, the condensations were fairly clean. 82,106,107he condensation progressed to completion within one day, and only nitromethyl β-Dglucopyranoside 62 (Scheme 9) was detected and isolated when 5:1 acetonitrile/pyridine (38% yield) or pyridine was used as the solvent (57% yield).For completion and maximal yield, a stoichiometric amount of DBU was needed.
Condensations with L-fucose 65 were monitored by TLC and flash chromatography, and gave sequential products as outlined in Scheme 10. 82,106,107 Within 12 h in THF more than half of the starting L-fucose was converted into corresponding 1,6-dideoxy-1-nitro-L-glycero-D-mannoheptitol 66.By the end of day 3, most of L-fucose had been transformed into the acyclic heptitol 66, and the subsequent conversion into cyclic nitromethyl β-L-fucopyranoside 67 was followed for another 5 days, during which time a significant portion of the desired cyclic product 67 had also been slowly converted into 1,1,6-trideoxy-1-nitro-2-nitromethyl-L-galacto-heptitol 68.Thus, the maximal yield for product 67 required minimizing decomposition and the yields of 66 and 68.In dioxane, only 67 was isolated as the product (62%).On the contrary, 50% 68 vs. 5% 67 were obtained in DMF.The results suggest that solubility might be the most important factor to affect Henry condensations of unprotected sugars.A possible further transformation of the synthesized nitro-compounds is illustrated by the reduction of nitromethyl β-D-glucopyranoside 63 with a specially prepared elemental iron in water/THF under CO 2 atmosphere to aminomethyl-C-glucopyranoside 69 (Scheme 11). 104

Condensation with Malononitrile
As described above, Henry condensation with nitromethane of partially protected and nonprotected pyranoses with a free anomeric hemiacetal function in presence of DBU and molecular sieves produces C-glycosides, which molecules incorporated one or two molecules of nitromethane (an example of L-fucose reaction is shown on Scheme 10).][106][107] Inspired by this result, we explored a possibility of a similar synthesis of C-glycosides using malononitrile (MN) instead of nitromethane as an active methylene component (Scheme 12). 110,111Malononitrile is a versatile compound of exceptional reactivity. 112,1135][116] MN has not been used for preparation of C-glycosides, though their formation using dialkyl malonates was studied extensively.L-Fucose 65 was selected as a starting carbohydrate.We expected products similar to those obtained in reaction with nitromethane (Scheme 10).However, the reaction was much faster (1 h instead of several days), and neither mono-adduct 70 nor bis-adduct 71 (Scheme 12), which are analogs of 67 and 68 (Scheme 10), was obtained.Instead, two cyclic compounds 73 (oil) and 74 (white crystals) in a ratio ~5:1 (by NMR) were isolated by a column chromatography of the reaction mixture (Scheme 13). 110,111The bicyclic structure 74 was proven by X-ray crystallography (Figure 6). 110,118nterestingly, the solution conformation (in CD 3 OD) of the side chain in 74 estimated from the vicinal spin-spin couplings in 1 H NMR correlated well with the crystal structure data. 111e results of the synthesis may be rationalized as a sequence of (a) Knoevenagel condensation producing intermediate 72, then (b) Michael addition of another MN resulting in formation of bis-adduct 71, and then (c) a base-induced double cyclization.The latter can occur via either the dihydrofuran derivative 73, or the dihydropyran derivative 73a.However, we isolated only one monocyclic compound, which structure could not be assigned based on NMR data.In an attempt to prepare a crystalline derivative appropriate for X-ray analysis, we acetylated product 73 (Scheme 14) and obtained the furan derivative 75, which NMR parameters perfectly matched the parameters of similar furan derivatives described in literature. 119This result suggests the structure 73 as the intermediate monocyclic compound isolated in the reaction of L-fucose with MN. 110,111
Acid catalyzed reactions between homoallylic alcohols and aldehydes or acetals have been used before for the preparation of 2,6-disubstituted di-and tetrahydropyran derivatives (see Ref. 92 and references therein).We chose an expandable two-component scheme of cyclization with a possibility of various functional groups in readily available starting compounds.Use of phthalimido derivatives allowed for syntheses of (protected) aminomethyl C-glycopyranosides (Scheme 15).
Thus, we obtained a series of four racemic phthalimido-protected aminomethyl C-dideoxyglycopyranosides 84-87, which produced with aq.Me 2 NH 144 the dimethylammonium salts of phthalamic acids.They were hydrolyzed after evaporation of excess Me 2 NH by heating with aq.H 2 SO 4 (70 o C), followed by extraction of phthalic acid with CH 2 Cl 2 and Et 2 O. Neutralization by NH 3 aq.and subsequent evaporation followed by column chromatography gave compounds 88-91 (80-86%), characterized and analyzed as their N-acetyland/or N,Otriacetyl derivatives, and used as building blocks for pseudodisaccharides.Acylation of benzylamine with the isopropylidenated γ-lactone of quinic acid 11 145,146 provided us with H 1 NMR chemical shifts and coupling constants for the quinoyl part of compound 96.Similarly we converted racemic C-glycosides 88-91 into their quinamides 97-103 (i.e. into diastereomeric Cpseudodisaccharides; Scheme 15).

C-Glycosides via Cycloadditions
The aqueous hetero-Diels-Alder reaction has been investigated by A. Lubineau et.al., [147][148][149][150][151] for the synthesis of dihydropyran cyclo-adducts.We wanted to explore their use, through dihydroxylation of the double bond for syntheses of C-glycosides, 152,153 as mimetics of naturally occuring 2,6-dideoxy sugars. 154Technical 1,3-pentadiene (piperylene) and aq.glyoxylic acid are inexpensive commercial products, which were used as starting materials.The reaction was regioselective in the intended way, most likely because of the electron donating property of the methyl group of pentadiene.The cisand transacids were obtained in comparable amounts (Scheme 17). 152,153Their direct separation by chromatography was unsuccessful.However, methylation of the crude acid product 108, 109 rendered the mixture 110, 111 separable on TLC, and by column chromatography.In this way we obtained suitable starting materials for the following dihydroxylations.The cis-isomer 111 gave only one major racemic product of cis-dihydroxylation (112).It is not clear why the molecule 111 does not allow OsO 4 approach from both sides.Similarly, the epoxidation of 111 gave a crude epoxide 113 (Scheme 17) in 84% yield, the NMR and TLC of which showed no evidence of the epoxide with opposite configuration.Its purification by column chromatography proved to be difficult, due to its instability.As expected, acid hydrolysis (113→114) gave a single major product, which could be isolated pure after esterification to 115 (81% overall; Scheme 17).It is advantageous, that any "opposite epoxide" in the crude mixture also would give preferentially the same diaxial hydrolysis product 114. 143is-Dihydroxylation of trans-isomer 110 gave a mixture of stereoisomeric 116 and 117 (ratio 3:1 by NMR; overall yield 73%). Compound 116 was crystallized from the mixture (57%).Compound 117 was not successfully isolated from the mother liquor, but its 1 H-NMR parameters were determined from the mixture.
The epoxidation of trans-ester 110 gave two epoxides 118 and 119 as detected by TLC.Their mixture was subjected to hydrolysis and methylation to give a mixture of trans-dihydroxy esters 120 and 121, which was separated by column chromatography (yields 38% and 43% respectively).
We also tried to separate the carboxylic acids 108 and 109 by the reaction of NBS with their mixture.Only the trans-isomer 108 reacted with NBS to form a bromolactone, and the lactone 122 was easily separated from the reaction mixture, but the yield was too low (15-25%).The nature of the side reaction of the trans-isomer with NBS could not be determined.The bromolactone 122 was cleaved by NaOMe to give the single epoxide 118 which was hydrolyzed to give a single carboxylic acid 120, providing another structure proof for it (Scheme 17).
][157][158] The inhibition was found to be specific with regard to the configuration of inhibitor, the type of enzyme, and even the particular source of enzyme.Most noticeably, these compounds inhibited α-and β-D-glucosidases.The isomers 130 and 131 appeared to be more selective inhibitors.It is worth mentioning that the compound 129e produced the same inhibition effect upon β-D-glucosidase from A. oryzae, as a ~10-fold higher concentration of methyldeoxynojirimycin.
The results of our studies proved that simple and readily available racemic derivatives of 1,2-cyclohexanedicarboxylic acid can be potent inhibitors for α-and β-D-glucosidases The presence of at least two carboxylic groups and one hydroxy group was found to be essential for efficient inhibition.Its magnitude depended on configuration of substituents and increased for βglucosidase with lengthening of the inhibitor's alkoxy group.Further studies towards the enhancement of inhibitory activity may be focused on separation of enantiomers, on variation of the substituent configuration, of the length and structure of the lipophilic groups, and of the nature of heteroatoms.These compounds may be considered as promising precursors for combination with natural fragments (e. g. shikimic acid; see Sections 2, 7) for the synthesis of diastereomeric SCOMs.

Scheme 12 .
Scheme 12. Expected products from the reaction of L-fucose with malononitrile.

Scheme 13 .
Scheme 13.Condensation of L-fucose and malononitrile in presence of DBU.