A unified strategy for the synthesis of aldohexoses by boronate assisted assembly of CH2X2 derived C1-building blocks

A synthetic strategy for all aldohexoses with individually addressable protecting groups from dihalomethane C1-units is reported. The underlying synthesis of C6-sugar alcohols relies on three consecutive Matteson sequences, vinylation and bishydroxylation. Erythro and threo isomers have been realized for every glycol motif by strategic variation of the sequence.

Carbohydrates are of immense biological importance as a source of energy and as complex chiral scaffolds that participate in numerous recognition processes. 1Deciphering this "glycocode" is a task, which requires modern analytics as well as organic synthesis. 2Nowadays oligosaccharides can even be prepared in an automated manner from orthogonally protected monosaccharides. 3Syntheses of the latter oen still rely on exchiral pool strategies, each of which has to face the challenge of differentiating ve similar hydroxyl groups. 4De novo syntheses, such as the ones developed by Sharpless, 5 Danishefsky 6 or MacMillan 7 approach this problem from the bottom up.Strategies based on C 1 -building blocks, like those by Fischer, 8 Dondoni 9 and Matteson 10 could allow for maximal protecting group variability and enable isothopic labelling of individual atoms. 11owever, each of these methods has its individual limitations 12 and no unied C 1 -based strategy to aldohexoses had been reported until now.One key to our route is the Matteson homologation (MH) shown in Scheme 1A. 13 This sequence employs chiral boronic esters (1), which react with a lithiated dihalomethane and ZnCl 2 at low temperatures.In the resulting ate complex (2) electrostatic interactions 14 between the zinc bound chloride atoms and the carbenoid C-H favour an antiperiplanar arrangement of one C-X-bond relative to the boronate R-group.Upon warming 1,2-rearrangement results in the diastereoselective formation of a-halo boronates (3).Reaction with various nucleophiles yields a-chiral boronates (4) under stereochemical inversion.Thus MHs are highly useful for preparing heteroatom rich motifs. 16s shown in Scheme 1B, iterative MH and substitution with alkoxides can lead to carbohydrate like structures.This was applied by Matteson to the synthesis of L-ribose. 10While MH and substitution with LiOBn worked well for the C 1 -building block 5, and two more homologs, further homologation proved to be problematic.Attempts to react 6 with LiCHBr 2 led to intractable mixtures and the use of LiCHCl 2 only allowed for the indirect detection of product traces.The route was thus concluded by homologation with LiCH 2 Cl, which does not allow for installation of another stereocenter.For detailed discussion of this surprising limitation, an explanatory hypothesis and supporting evidence see the ESI.‡ Importantly this restricted Matteson's synthesis to ribose (an aldopentose), 15 while most biologically relevant monosaccharides are hexoses.Thus, a C 1based synthesis of aldohexoses that allows for (i) installing individual protecting groups, (ii) choosing the conguration at each stereocenter and (iii) potentially introducing isotopic labels at every individual atom, remained an open challenge. 11hemical Science

EDGE ARTICLE
We achieved this by preparing orthogonally protected versions of prototypical sugar alcohols from CH 2 X 2 building blocks through three MHs, vinylation and bishydroxylation (Scheme 1C).By strategically combining different homologation and vinylation strategies, both erythro and threo isomers were realized.Conversion of the sugar alcohols into aldohexoses can be achieved by oxidation of either terminal hydroxyl functionality.By combining this with a short synthesis of the vinylation agent from CH 2 X 2 building blocks we paved the way for the late-stage introduction of isotopic labels.
To start the discussion with the stereochemically most basic example, the synthesis of allitol 12a is depicted in Scheme 2. It begins with the CH 2 Br 2 derived C 1 -building block 7a.MH with LiCHBr 2 and substitution with LiOPMB delivered the C 2building block 8a.Two consecutive MHs, which are followed by substitution with an alkoxide, produce a masked erythro glycol motif.Thus, a second MH and substitution with LiOBn delivered 9a, with an erythro relationship between C 2 and C 3 (as IUPAC priorities change during the route, carbon atoms are numbered according to their introduction in this article).Another homologation and substitution with LiOBn yielded 10a in 47% yield aer two steps. 18While other alkoxide based protecting groups could have been used here, a second benzyl group was chosen, to allow for conrmation of the relative conguration by direct comparison (see ESI ‡).Vinylation of 10a was achieved by Zweifel-olenation. 17Although this reaction had not been described for the sterically hindered and thermodynamically stable pinanediol boronic esters, it proceeded reasonably well aer some optimization (see ESI ‡) yielding 11a.The product contained some unidentied contaminations, which were removed aer the next step, in which Sharpless bishydroxylation delivered the desired allitol-derivative 12a (32% yield, calculated over both steps).Several silylethers at C 1 were tested but neither the use of TBS-(tert-butyldimethylsilyl) nor TBDPS-groups (tert-butyldiphenyl-sily) on the rst hydroxyl group allowed for introduction of a forth carbon atom (13/14).This was quite surprising as a third Matteson-sequence had worked well for the benzyl-derivative 6.Indeed, homologation of 13 to bromides 15 proceeded with reasonable efficiency, but subsequent substitution with LiOBn led to decomposition.This was attributed to competing nucleophilic attack of benzoxide on the silyl ether.We suspected that the latter was activated by an intramolecular O-B interaction, which simultaneously deactivates the boron-atom as an electrophile.By using the TIPS-group (triisopropylsilyl) this side reaction was avoided, through better shielding of the silicon atom (further discussion in ESI ‡). 19n order to extend the route to other diastereomers it was necessary to modify the synthesis, so that threo-glycols could be obtained.This was a particular challenge for the three stereocenters generated by Matteson homologation as the exchange of the pinanediol director is quite cumbersome. 20Thus the route was modied as shown in Scheme 3.
In order to establish a threo-relationship between C 2 and C 3 the synthesis starts with S,S-dicyclohexylethylenediol (S,S-DICHED) 21 boronic ester 7b.Homologation and substitution analogously to Scheme 2 delivered 8b-1.The greater thermodynamic stability of pinanediol boronic esters 22 allowed for transesterication to 8b-2 with (+)-pinanediol in Et 2 O, 15 as well as recovery of the precious S,S-DICHED auxiliary.However, once this card had been played and the more stable pinanediol boronate was formed, a different strategy had to be applied.In order to establish a threo relationship between C 3 and C 4 , the third Matteson homologation was followed by substitution with vinylmagnesium bromide 23  Benzyl protection and Sharpless-bishydroxylation delivered the desired glucitol 12b.Surprisingly the seemingly simple combination of a Matteson sequence and a boronate oxidation in 9b/16b proved to be quite challenging.Competing elimination reactions, such as the one from 18 to the conjugated diene 19 had to be suppressed by a strict temperature regime.Therefore, 9b was homologated as usual.Substitution with vinylmagnesium bromide to 17 required addition of the Grignard reagent at −78 °C, storage in a freezer over night at −18 °C and stirring for 4 h at 0 °C.Instead of isolating 17, H 2 O 2 was added at 0 °C and the reaction was kept at this temperature for 4 h, before the reaction was quenched by the addition of Na 2 S 2 O 3 (see ESI ‡ for further discussion).
By appropriately combining the two complementary strategies shown in Schemes 2 and 3 all but two aldohexoses should be available.However, both idose and galactose are still elusive at this point, as they have threo relationships between C 2 and C 3 as well as C 4 and C 5 and so far we have only shown how to establish an erythro relationship by Sharpless bishydroxylation.Thus Scheme 4 depicts the synthesis of mannitol 12c and its inversion to epoxide 20, which has the desired threo relationship between C 4 and C 5 .This required homologation of 9a, substitution with vinylmagnesium bromide and oxidation to 16c.Benzylation yielded 11c and Sharpless-bishydroxylation delivered mannitol 12c.Acylation of the primary hydroxide, mesylation of the remaining alcohol and epoxide closure by basic estercleavage (with concomitant substitution of the mesylate) produced 20. 27L-Mannitol 12c was chosen as a last example for several reasons: rstly to demonstrate the ease with which the routes in Schemes 2 and 3 can be combined.Secondly chemically labelled mannose-derivatives are notoriously hard to track in biological systems, which makes isotope labelled derivatives of great value. 24Finally, several mannitol derivatives are commercially available.This allowed us to conrm the relative conguration of the Sharpless bishydroxylation products by direct comparison (see ESI ‡). 25 In these reactions very strong substrate control is exhibited by the allylic stereocenter favoring the erythro-product.This is in accordance with predictions of the Houk-model 21 (Scheme 4, see ESI ‡ for complete table and discussion). 26While an oxidative method that allows for choosing the desired conguration at this last stereocenter directly would still be preferable, positioning of the diol nevertheless allows for straight forward inversion as demonstrated by conversion into 20. 27Finally two methods for converting the prepared sugar-alcohols into carbohydrates were probed on 12c (Scheme 5). 28emporary ketal protection of diol 12c and selective cleavage of the TIPS group delivered 22 in 99% yield.Parikh-Doering oxidation of the free hydroxide at C 1 proceeded under ketal cleavage leading to L-mannose derivative 23 in 86% yield.In it only the easily distinguishable primary and anomeric hydroxyl groups are unprotected.The three secondary hydroxyl groups are masked by orthogonal protecting groups, the placement of which could be easily varied.In some cases it might be advantageous to convert sugar alcohols of type 12 into an aldohexose via the other terminus (i.e., C 6 ).This was achieved by orthogonal protection of 12c, yielding 24 in 75% yield over two steps.Pivaloyl deprotection, Parikh-Doering oxidation and PMB cleavage led to 25 in 41% yield over three steps.These two options could allow for late-stage introduction of isotopic labels, as three of six carbon atoms are introduced in the nal homologation/ vinylation sequence.To enable this, a route to vinyl metal species from C 1 -building blocks (i.e., CH 2 X 2 ) was developed: bis(pinacolato)borylmethane 29 was lithiated and reacted with either dibromo-or diodomethane as described by Morken and coworkers, 30 yielding vinyl pinacol boronic ester.This product is highly volatile, but transesterication with pinanediol to 26 enabled purication by ash chromatography.Reaction of 26 with NaOMe and I 2 yielded vinyl iodide.Vinyllithium (the preferred reagent for Zweifel reactions) 17 was obtained by iodolithium exchange.Efficient reaction with a-bromoboronate 27, required transmetallation with MgBr 2 .
Scheme 4 Synthesis of protected L-mannitol 12c and inversion to epoxide 20.
Scheme 5 Synthesis of L-mannose and vinyl iodide from C 1 -units.

Conclusions
All in all we have developed a highly modular route that opens up a vast eld of opportunities for the synthesis of differentially protected sugar alcohols and carbohydrates.By combining the different approaches depicted in Schemes 2-4 a wide variety of C 6 -sugar alcohols and by extension all natural and unnatural aldohexoses become available.Two consecutive MHs, followed by substitution with an alkoxide lead to erythro-C 3 building block 9a.To introduce a C 2 -C 3 threo-relationship the chiral director can be changed from DICHED to pinanediol.The transesterication proceeds readily and allows for recovery of the valuable DICHED auxiliary.Both enantiomers of DICHED and pinanediol are available, so that all stereoisomers of C 3building blocks of type 9 are accessible.To gain control over the relationship between C 3 and C 4 a strategic crossroad was incorporated in the next homologation.A vinyl group was introduced either by Zweifel olenation (erythro) or Matteson substitution (threo).Thereby C 3 -building blocks of type 9 can be converted into vinyl tetrol of type 11, again with the potential for making all stereoisomers.For installing the nal glycol moiety Sharpless bishydroxylation was employed.Unfortunately overwhelming substrate control only allowed for the direct synthesis of sugar alcohols of type 12 with a C 4 -C 5 -erythro conguration.In order to obtain a C 4 -C 5 -threo conguration at this position conversion into epoxide 20 was necessary.Some rst attempts at epoxide opening to a diol of type 12 (with a C 4 -C 5 -threo conguration) were plagued by side reactions (see ESI ‡).Fortunately this only affects the synthesis of monosaccharides with both C 2 -C 3 -threo and C 4 -C 5 -threo congurations (i.e., galactose and idose).For these cases the corresponding epoxides might be better converted into hexoses along the lines of Shapless's carbohydrate synthesis. 31In all other cases conversion into the desired aldohexoses can be achieved by appropriate cyclisation via C 1 or C 6 due to the orthogonal silyloxy group at C 1 (Scheme 5).Another advantage of these two cyclisation options arises as half of the carbon scaffold is introduced in the last homologation/vinylation sequence.The required vinyl metal species can be prepared from two (CH 2 X 2 derived) C 1 building blocks (Scheme 5).By choosing the appropriate cyclisation route an isotopic label could thus be placed at every position in the aldohexose scaffold.Thus this C 1 based de novo approach to aldohexoses is uniquely suited for the synthesis of labelled aldohexoses, which we plan to pursue in the near future.

Scheme 2
Scheme 2 Synthesis of protected allitol and associated challenges.