Radical‐mediated brominations at ring‐positions of carbohydrates – 35 years later Free

The title reaction, namely the possibility for a direct replacement of a hydrogen atom in a carbohydrate ring by bromine, was first reported by Ferrier and Furneaux in 1977.^1,2  The transformations need to be performed under irradiaton or in the presence of radical initiators and can thus be understood by a radical mechanism (Scheme 1). They are sometimes called the “Ferrier photobromination”. The resulting products contain the bromine attached to carbons adjacent to the ring oxygen, i.e. the bonds formed are either C‐1–Br or C‐5–Br/C‐4–Br (pyranoid vs. furanoid rings); with certain compounds competitive reactions to give C‐1–Br and C‐5–Br/C‐4–Br derivatives can take place. Sporadically chlorinations have also been carried out.

The reactivity of several carbohydrate derivatives under such conditions was tested and a comprehensive survey of these studies, including suggestions for the rationalization of the observed regio‐ and stereoselectivities as well as transformations of the primary brominated products, was also published in 1991.³  Since then the reaction has been extended to new types of substrates and a broad range of subsequent transformations has led to various carbohydrate derivatives demonstrating the synthetic utility of this bromination.

The aim of the present article is to survey this type of functionalization of carbohydrate derivatives and to update the previous review more than three decades after finding the transformation. New developments in the reaction conditions and protecting groups are summarized first, followed by the brominations themselves. While in the 1991 review these reactions were grouped somewhat arbitrarily, also considering historical and chronological aspects, here the brominations are categorized according to substrate types classified by exocyclic bonds of the C‐1 centre (e.g. C‐1–O, C‐1–S, etc.). The next part deals with the transformations of the brominated sugar derivatives also outlining further synthetic uses of the obtained compounds. A brief summary of those results covered in the first review introduces each of these sections. The chapter is concluded with a tabular presentation of biological activities and utilization of the synthesized compounds.

The reactions were originally performed under the classical conditions for Wohl‐Ziegler brominations,^4,5  i.e. in refluxing CCl₄ with N‐bromosuccinimide (NBS) or bromine as the reagents in the presence of substoichiometric amounts of radical initiators like dibenzoyl peroxide (Bz₂O₂) or azobisisobutyronitrile (AIBN) or/and with irradiation. The use of ultrasound in place of the previous initiation methods was reported to give higher yield and purity for the products in slower reactions.^6–10  Advantageous addition of CBrCl₃ as a co‐solvent was mentioned in sporadic cases.³  Addition of BaCO₃ or K₂CO₃ as acid scavangers especially in reactions with Br₂ could be beneficial.

Due to its several hazardous effects (e.g. acute toxicity, specific organ toxicity to liver, kidneys, eyes, and heart, carcinogenicity, aquatic toxicity, ozone layer damages) the use of CCl₄, being otherwise an ideal solvent for these transformations, was seriously restricted, practically banned. Therefore, some research groups succesfully tried to replace CCl₄ by Cl₃CCH₃ with NBS,¹¹  and CHCl₃ or CH₂Cl₂ with Br₂.¹²  Another study¹³  showed that benzotrifluoride (PhCF₃, BTF) could be used as solvent in several cases, and the unconventional bromination reagent system¹⁴  KBrO₃–Na₂S₂O₄ in CH₂Cl₂–water biphasic solvent mixture proved also widely applicable. NBS was also shown to perform well in the latter solvent system. For chlorinations SO₂Cl₂ in CCl₄ with AIBN initiator was used.¹⁵ 

Hydroxyl groups of the sugar derivatives are generally protected by esters (benzoates preferred to acetates as the latter can undergo bromination). Recently, the use of 4‐bromobenzoate esters¹⁶  and carbonates¹⁷  was reported. From the ether type protective groups methyl and trityl¹⁸  could be applied, but benzyl ether is sensitive towards bromine radicals. The 2‐ and 4‐trifluoromethylbenzyl ethers, which are stable under some oxidative conditions,¹⁹  can, to a certain extent, survive NBS induced cleavage of benzylidene acetals,²⁰  but have, to the best of our knowledge, never been used in radical‐mediated brominations. The applicability of silyl ethers will be illustrated in Section 2.6.1. Benzylidene‐ and other aldehyde‐derived acetal protections are cleaved under the bromination conditions, however, ketone‐based derivatives have been used succesfully.

There are examples for bromination of compounds with a single free OH²¹  (Section 2.6.1) and COOH³⁰  (Section 2.2.1) groups. In other cases the presence of a free COOH group resulted in decomposition under the bromination conditions²²  (Section 2.6.1). For the protection of COOH groups besides methyl also phenacyl²³  esters were used (Section 2.7.1).

Compounds with primary carboxamide (CONH₂) substituents can be brominated without protection of the NH functionalities (Section 2.6.1). Various data are available on brominations in the presence of secondary amides (e.g. AcNH substituents) which were either masked as Ac₂N, phthalimido,²⁴  or tetrachlorophthalimido²⁵  moieties (Sections 2.2.1, 2.2.2, 2.7.1) or left unprotected^26,27  (Sections 2.5, 2.7.1).

In the 1991 review³  an attempt was made to rationalize the observed selectivities of the reactions. Since then, no focused studies have been carried out to tackle these points, nevertheless, those rules can be applied to explain new findings, as well. To give a general frame for understanding the outcome of the reactions, the factors determining selectivities are outlined here. As an illustration of these considerations Table 1 summarizes the substrates studied so far in the brominations and indicates the main products of the reactions.

Regioselectivity of the reactions is influenced by the ease of hydrogen abstraction which is determined by radical stabilities as well as stereochemical and steric effects. C‐H bonds adjacent to ring oxygens are prone to homolysis, which is reflected in the preponderant formation of α‐bromoether type compounds (cf. Scheme 1). In addition, radical stability is influenced by the substituents Y and Z; a particularly stable radical is formed and the corresponding site will be highly reactive if the so‐called capto‐dative substitution pattern is present (Y or Z is an electron withdrawing group, cf Table 1). A numerical estimate for the relevant sugar radical stabilities was given in the 1991 survey.³  An important stereochemical factor governing H‐abstraction is the axial vs. equatorial orientation of the hydrogen in pyranoid rings, the former being significantly more reactive. The steric availability of the hydrogen atom to be abstracted also contributes to the selectivity issues: bulky substituents e.g. in place of Z (cf alkyl and aryl glycosides in Table 1) were shown to direct the reaction to the C‐5 centre, while axial substituents can slow down or even totally hinder the abstraction of axial hydrogens on the same side of the ring. This may result in considerable differences in reactivities of anomers.

The stereochemistry of the products is influenced by kinetic and thermodynamic anomeric effects, both in favour of the formation of axially brominated compounds. Epimeric substrates can give common radicals which result in the same product(s). Conformation of the intermediate radicals is another important issue which can exert an effect on the formation of diastereomeric products. Glycosyl radical conformations and their consequences for stereoselectivities cannot be treated here, the reader is kindly referred to a review.²⁸ 

Antecedents:³  Simple pyranosides with O‐acyl protection having axial aglycons gave no isolable products. With equatorial aglycons C‐1‐bromination (and subsequent reactions or decomposition) occurred for methyl glucosides, while C‐5 bromides were formed from phenyl β‐D‐glucosides with increased yield for the 4‐nitrophenyl derivative.

Brominations of 4‐methoxyphenyl‐ and 2,2,2‐trichloroethyl 2‐deoxy‐2‐tetrachlorophthalimido‐β‐D‐glucopyranosides were observed to take place at both reactive centers, however, isolation in 43% yield of the 5‐bromide 1 of the latter substrate was reported only.²⁵  Phenyl β‐D‐xylopyranoside gave 2 (25%), but extended reaction times resulted in brominations of the OAc and the phenyl groups, too.²⁹  5‐Bromides of uronic acid derivatives 3 and 4 were isolated in 39%³⁰  and 65%³¹  yield, respectively.

Antecedents:³  O‐Peracylated aldopyranoses underwent bromination at the C‐5 centre: hexopyranoses (Y=CH₂OAcyl) gave axial 5‐bromides (α‐anomers gave lower yields than the β‐anomers, benzoates proved more stable than acetates), while the conformationally more mobile tetra‐O‐acetyl‐β‐D‐xylopyranose (Y=H) furnished C‐5 epimeric bromides with axial halogens in both compounds. Benzoylated furanosyl acetate derivatives of D‐glucose and D‐ribose gave the corresponding C‐4‐Br epimers. Anomeric esters of hexuronic acid derivatives (Y=COOR) of β‐D‐gluco and α‐L‐ido configurations gave the same axial 5‐bromide of β‐D‐glucuronic acid. Both epimers of a formally 5‐(2‐cyanoethyl)‐substituted β‐D‐xylose tetraacetate gave the same C‐5‐bromide.

Bromination of O‐per(4‐bromobenzoylated) β‐D‐glucopyranose gave the high melting C‐5 bromide 5 in 83% yield.¹⁶  O‐Peracetylated N‐acetyl‐D‐glucosamine was reported to be incompatible with the bromination conditions. The N,N‐diacetyl derivative gave an inseparable mixture of the C‐1 and C‐5 bromides, but tetrachlorophthaloyl (TCP) or phthaloyl (Pht) protection could be applied to give 6 and 7, respectively. Interestingly, the sterically more crowded 8 was obtained in higher yield than that of the anomeric 6.²⁵  α‐D‐Lyxopyranose tetraacetate gave the 5‐bromide in 35% yield, the formation of the possible other epimer was not mentioned.³²  Methyl 1,2,3,4‐tetra‐O‐acetyl‐α‐D‐glucopyranuronate was brominated to give 9 in a yield of 65–70%.^33,34 

Bromination of D‐ribofuranose tetraacetate was mentioned to give the corresponding 4‐bromide 10 as the only product, however, no experimental details were given.³⁵  D‐Fructofuranose pentabenzoate could not be brominated by NBS, however, Br₂/hν furnished 11 in 52% yield.³⁶ 

Antecedents:³  O‐Perbenzoylated phenyl 1‐thio‐β‐D‐gluco‐ and galactopyranosides brominated at the anomeric centre and ensuing reactions gave isolated enone type compounds. O‐Peracetates reacted similarly, however, these reactions suffered from overbromination in the OAc protecting groups. The α‐D‐gluco configured substrates reacted significantly slower to give the same product. Bromination of O‐peracetylated methyl (phenyl 1‐thio‐β‐D‐glucopyranoside)uronate took place both at C‐1 and C‐5 to yield the above enone type compound and the 5‐bromide, respectively. O‐Peracetylated β‐D‐glucopyranosyl phenyl sulfoxide gave acetobromoglucose on bromination, while the corresponding sulfone furnished both C‐1‐Br and C‐5‐Br derivatives in an almost equal ratio.

Bromination of O‐peracetylated β‐D‐gluco‐ and β‐D‐galactopyranosyl methoxycarbonylethyl sulfones gave moderate yields of both C‐1‐Br and C‐5‐Br derivatives 12–15, respectively (Table 2). β‐D‐Glycopyranosyl sulfonamides brominated similarly to give 16 and 17 of D‐gluco as well as 18 and 19 of D‐galacto configuration in low isolated yields together with significant amounts of the corresponding glycosyl bromides. A mechanistic rationale, based on relative radical stabilities and β‐fragmentation of sulfonamidyl radicals, was proposed to explain the regioselectivities and the formation of glycosyl bromides.³⁷ 

Antecedents:³  none.

Under treatment by NBS and irradiation, glycosyl thiohydroximates (Scheme 2, A2) undergo spirocyclization to give mixtures of epimeric oxathiazolines D2 and E2.³⁸  This spirocyclization can be understood either by the oxidative formation of biradical B2 to yield the major isomer by recombination with the known axial preference of glycosyl radicals²⁸  or by bromination of A2 to intermediate C2 and subsequent intramolecular nucleophilic substitution. The anomeric configuration of A2 had no influence on the reaction as neither the rate nor the stereoselectivity were significantly different.³⁹ 

Antecedents:³  1,N‐Dibenzoyl‐2′,3′,5′‐tri‐O‐benzoyladenosine was brominated in the 4′‐position.

O‐Peracetates of some N‐aryl‐β‐D‐glucopyranosylamines were reacted with NBS/Bz₂O₂, however, only aromatic brominations could be observed. Similarly protected N‐acetyl‐N‐aryl‐β‐D‐glucopyranosylamines remained intact under these conditions. Acetylated N‐β‐D‐cellobiosylpiperidine, as an aliphatic N‐glycosidic substrate, gave the corresponding α‐D‐cellobiosyl bromide.¹⁵ 

In an attempt to prepare glucitol spiro 1,2,4‐oxadiazolines⁴⁰  (as analogues of the spiro‐oxathiazolines shown in Scheme 2) O‐peracetylated β‐D‐glucopyranosylamidoximes (Scheme 3, A3) were treated with NBS under irradiation. Various proportions of compounds C3–E3 could be isolated whose formation can be explained by the oxidative milieu: C3 is a direct oxidation product of A3; the expected B3 can be formed by a mechanism similar to that depicted in Scheme 2, however, this compound undergoes a tautomeric ring opening followed by aromatization of the heterocycle to give D3 which is further oxidized to E3.

Bromination of a glucopyranosylpurine gave the 5′‐bromo product 20, while an unselective reaction was observed with 2′,3′,5′‐tri‐O‐benzoyluridine.¹⁵  On the other hand, benzoylated 5‐fluorouridine gave the 4′‐bromide 21, and 4′‐bromoadenosine 22 could also be isolated.⁴¹  Failure of attempts to brominate 2′‐deoxycitidine was reported.³⁵ 

Antecedents:³  none.

Reactions of variously protected glycosyl azides with NBS under irradiation or in the presence of Bz₂O₂ or AIBN (Scheme 4) resulted in the rather labile bromoiminolactones 23–31.^15,18,42  Anomeric configuration of the starting azides had a considerable bearing on the rate (but not on the yields) of the reactions:¹⁸  competitive experiments showed the relative reaction times for O‐peracetylated glycopyranosyl azides of β‐D‐manno, β‐D‐gluco, α‐D‐manno, and α‐D‐gluco configurations to be ∼2:3:6:15, respectively. In furanosyl azides the reactivities of the anomers were practically the same. A detailed mechanistic proposal suggests the formation of an anomeric radical as the initial step, which looses molecular nitrogen and rearranges to an iminyl radical whose reaction with bromine gives the final product.¹⁸  Contrary to bromination, radical‐mediated chlorination of O‐peracetylated β‐D‐glucopyranosyl azide gave the C‐5‐chloro derivative 32 (Scheme 5).¹⁵ 

Antecedents:³  none.

Bromination of O‐peracetylated β‐D‐glucopyranosyl isothiocyanate under several conditions gave acetobromoglucose and 2,3,4,6‐tetra‐O‐acetyl‐D‐glucopyranose or an unsaturated lactone in varying yields and ratios. On the contrary, chlorination resulted in the C‐5‐chloro product 33 (Scheme 5).¹⁵ 

Antecedents:³  none.

Diethyl 2,3,4,6‐tetra‐O‐acetyl‐α‐ and ‐β‐D‐glucopyranosylphosphonates were brominated to give the same product 34 in 53% and 64% yields, respectively.²⁷  To get the 2‐deoxy counterpart 35 (25% isolated by chromatography as a rather unstable syrup) an anomeric mixture of the corresponding phosphonate was used.⁴³  Similarly, a mixture of both anomers was reacted to furnish the sialic acid analogue bromide 36 in 45% yield whereby 15% of the starting material was recovered.²⁷  Possible different reactivity of the anomers got no mention in these reports.

Antecedents:³  Several O‐peracetylated glycopyranosyl cyanides (2,6‐anhydro‐aldononitriles) were brominated. From the hexose‐derived compounds the β‐D‐gluco and α‐ and β‐D‐galacto configured ones gave the C‐1‐bromo products (38 and 41) in yields above 80%. The α‐D‐manno substrate furnished the analogous axial bromide 42 in 49% yield. Among pentose derivatives the α‐ and β‐D‐arabino compounds gave high yields of the same C‐1‐bromide 44, while the β‐D‐xylo and β‐D‐ribo derivatives reacted to mixtures of C‐1‐bromo epimers 43.

Since the first investigations a very large array of C‐glycosyl formic acid derivatives were studied under bromination conditions. Bromides 37–61 which were isolated in pure state are collected in Table 3 (for the sake of completeness also including some compounds actually obtained by ionic chemistry, but which could have been prepared by the radical method, too). Table 4 contains non‐isolated bromides 62–69 used immediately for further transformations. Significant difference in reaction times of the α‐ and β‐D‐pyranosyl derivatives was observed to show a 10–12 times higher reactivity for the equatorially substituted substrates leading to e.g. 37 and 41, while the reactivity of the furanoid substrates was almost the same to give e.g. product mixtures 45 and 46 or single product 53.

A specific course of the bromination of C‐(2‐deoxyglycopyranosyl)formates (Scheme 6, A6) was observed, namely the primary brominated product C6 eliminated HBr to give glycal D6 which, after bromine addition, furnished the isolated dibromide B6. Compound E6 gave F6 as the major product with some identified by‐products.⁶⁴ 

Under usual bromination conditions (Br₂/CHCl₃‐sunlight) 2,3,4,6‐tetra‐O‐acetyl‐β‐D‐galactopyranosyl thioformamide gave the corresponding 3,5‐bis(glycosyl)‐1,2,4‐thiadiazole in 80% yield. More safely reproducible results were achieved by using the non‐conventional bromination reagent system in Scheme 7 (Gly=Ac₄‐β‐D‐Glc_p, 77%; Bz₄‐β‐D‐Glc_p, 86%; Ac₄‐β‐D‐Gal_p, 80%; Ac₃‐β‐D‐Xyl_p, 62%).⁶⁵ 

Antecedents:³  Some observations on the possible formation of 2,3,4,6‐tetra‐O‐acetyl‐1‐bromo‐D‐glucopyranosylbenzene (B8e in Scheme 8) were discussed. Brominatons of O‐peracetylated C‐glycopyranosyl 1,3,4‐oxadiazoles and benzothiazoles were investigated with the β‐D‐galacto, β‐D‐xylo, and α‐D‐arabino configurations to give mostly isolable C‐1‐bromides.

Bromination of glycosylbenzenes (Scheme 8, A8 and C8) in the presence of water allowed to isolate lactols D8 as products of hydrolysis of the primarily formed bromides B8. This finding demonstrated the highly selective abstraction of hydrogen from the C‐1 position of A8 and C8. A reactivity order was also established by competitive experiments and the reaction times (C8e:C8a:A8=15:40:85 min)⁶⁶  reflected higher radical stabilization by the axial 2‐OAc substituent.²⁸ 

Antecedents:³  O‐Peracetylated glycopyranosyl chlorides of the β‐D‐gluco and β‐D‐manno configurations afforded mixtures of separable C‐1 and C‐5 bromides in very good overall yields and in a ∼5‐6 to 1 ratio in favour of the C‐1‐Br. Chlorinations, carried out with SO₂Cl₂/AIBN in CCl₄, gave similar results. The α‐D‐gluco configured chloride and bromide yielded a 1,2‐dibromide presumably in a HX‐elimination–Br₂‐addition sequence (see also ref. 67). Tetra‐O‐acetyl‐β‐D‐glucopyranosyl fluoride furnished the C‐1‐ and C‐5‐brominated compounds favouring the C‐5‐Br derivative in a ratio of ∼14 to 1. The corresponding α‐fluoride produced only the C‐5‐bromide.

It was shown later that the C‐1‐halogenated products isolated from the bromination mixtures of the β‐D‐gluco, β‐D‐galacto, and β‐D‐manno configured glycopyranosyl chlorides also contained the corresponding 1,1‐dichlorides (≤10%) besides the major 1‐bromo‐1‐chloro compounds. Formation of the minor products can be attributed to a Cl abstraction from the solvent (CCl₄). These mixtures were inseparable by chromatography, but repeated crystallizations allowed to remove the dichlorides. No C‐5‐bromide could be isolated from the reaction of the β‐D‐galacto substrate.⁶⁸  O‐Peracetylated 2‐deoxy‐2‐fluoro‐β‐D‐glucopyranosyl chloride expectedly gave a mixture of C‐1 (71) and C‐5 (72) bromo derivatives which proved inseparable.⁶⁹ 

Extensively studied brominations of several glycosyl fluorides were reported to produce C‐5‐bromides but the possible formation of the C‐1‐Br isomers was not mentioned (although such compounds might be present in the obtained mixtures). Tetra‐O‐benzoyl‐β‐D‐glucopyranosyl fluoride gave 73 in 84% yield, the α‐anomer proved to be unreactive.¹⁶  O‐Peracetylated 2‐deoxy‐2‐fluoro‐β‐D‐glycopyranosyl fluoride and 1‐fluoro‐β‐D‐glycopyranosyl fluoride produced 74 and 75, respectively, and the latter substrate was reported to be extremely unreactive.⁶⁹  β‐Fluorides of O‐peracetylated N‐acetyl‐ and N‐phthaloyl‐D‐glucosamines afforded 76²⁶  and 77,²⁴  respectively. None of 74–77 was isolated in pure state.

Tetra‐O‐acetyl‐α‐D‐galactopyranosyl fluoride gave 52% of 78,⁷⁰  and the α‐D‐manno derivative 79 was obtained in 53% yield.⁷¹  A phenacyl (β‐D‐glucopyranosyl fluoride)uronate furnished 80 (43%) selectively due to the capto‐dative nature of the C‐5 centre.²³  Interestingly, the formally also capto‐datively substituted CH₂ moiety of the phenacyl protecting group was not reported to be reactive, most probably indicating the smaller radical stabilizing capacity of the O‐acyl moiety compared to the O‐alkyl one. Bromination of tri‐O‐acetyl‐β‐D‐xylopyranosyl fluoride allowed to isolate the 5‐bromide 81 (38%) and 5,5‐dibromide 82 (19%), while the α‐anomer furnished 83 (20%).⁷² 

Antecedents:³  none.

Prolonged bromination of an anomeric mixture of O‐peracetylated 5‐thio‐D‐glycopyranosyl bromides gave tribromide 84 in 30% yield. 5‐Thio‐β‐D‐xylopyranosyl bromide afforded 90% of the anomeric dibromide 85 in a short reaction time, while after longer treatment tribromide 86 also appeared among the products (containing further minor polybrominated compounds in yields of 4–7%). 5‐Thio‐α‐D‐xylopyranosyl bromide reacted more sluggishly to give a mixture of 85 (45%), 86 (32%), and two other polybrominated compounds (∼10% for the two). These findings demonstrated that the presence of sulfur in the ring activated both the C‐1 and C‐5 positions towards homolysis, as occurred for the otherwise less reactive equatorial C‐H bonds.⁷³ 

Antecedents:³  Acceptor substituents, e.g. oxo (=O) or imino (=N–) groups, in positions 2 or 4 of the pyranoid ring render the respective C‐1 or C‐5 centres capto‐datively substituted which are thus prone to radical formation. O‐Perbenzoylated 1,5‐anhydro‐D‐fructose and its 2‐acyloximino derivative gave the corresponding α‐D‐arabino‐hex‐2‐ulopyranosyl bromides (C‐1‐Br compounds) in excellent yields (Scheme 9, routes F9→B9 and E9→C9, respectively); the reaction was extended also to xylose‐, cellobiose‐, maltose‐, and lactose‐derived 2‐benzoyloximino compounds. Chlorination was sluggish, although α‐D‐arabino‐hex‐2‐ulopyranosyl chloride could be isolated in a rather low yield. Both equatorial and axial anomers of an O‐perbenzoylated hexos‐4‐ulopyranose gave very high yields of the corresponding C‐5‐bromides. Monosaccharide derivatives with a 3‐en‐2‐one structural motif were brominated at the C‐5 center most probably due to a more extended delocalization of the formed radical.

For the preparation of 2‐oxoglycosyl bromides B9 by photobromination, the required substrates were obtained on route A9→D9→F9 (Scheme 9). A high yielding method was found for the direct transformation of A9 to B9 by using NBS in MeOH. This ionic reaction replaced the radical route, and bromides B9 and C9 were extensively applied in the syntheses of various oligosaccharides. As this chemistry has been authoritatively reviewed very recently, the reader is kindly referred to this excellent survey.⁷⁴ 

Antecedents:³  O‐Acetyl or ‐benzoyl protected 1,6‐anhydro‐D‐aldohexopyranoses underwent bromination with exclusive regioselectivity at the C‐6 centre to give the exo‐bromides (e.g. 87) as the sole products in the gluco, manno, galacto, ido, and talo configurations. Besides exo‐monobromides 6,6‐dibromides could also be formed from the allo, altro, and gulo isomers on prolonged reaction times, and relevant stereochemical features allowing this to happen were assessed. The reaction was investigated with maltose and lactose derivatives having 1,6‐anhydro rings, but took a different course for the latter that was not brominated in the bicyclic system. 1,5‐Anhydro‐D‐aldopentofuranoses with O‐acetal or O‐acyl protection gave isolated C‐5‐exo‐bromides in each configuration.

Later the reaction was extended to O‐peracetylated derivatives to get 88 and 89 of D‐manno and D‐galacto configuration, respectively.⁷⁵  Isotope labelled (6S)‐1,2,3,4,5,6‐²H‐1,2,3,4,5,6‐¹³C‐1,6‐anhydro‐2,3,4‐tri‐O‐benzoyl‐6‐bromo‐D‐glucopyranose was prepared in the same way.⁷⁶ 

Bromination of bicyclic L‐fucose mimics gave the less hindered bromides 90⁷⁷  and 91.⁷⁸  Compound 92 was used as an intermediate in the synthesis of herbicidin glycosides.¹⁷ 

Both pyranosides and furanosides with a phenylthio substituent attached to the exocyclic methylene unit were chlorinated with NCS in CCl₄ to give C‐6‐ (or C‐5)‐chlorides e.g. 93 by an unspecified mechanism.⁷⁹ 

Antecedents:³  Several disaccharide substrates were studied and provided nice illustrations for the governing effects of the brominations. Thus, 1,6‐anhydro‐2,3‐di‐O‐benzoyl‐4‐O‐[methyl(2,3,4‐tri‐O‐benzoyl‐β‐D‐glucopyranosyl)uronate]‐β‐D‐glucopyranose was preferentially brominated at the C‐5 of the uronate moiety, but minor amounts of the product with a second bromine in the anhydro ring could also be isolated. β‐D‐Maltose octaacetate gave the C‐5 bromide of the reducing part, and the presence of an anhydro ring in the same unit of maltose directed the bromination at the C‐6 in that moiety. Disaccharides with 2‐benzoyloximino groups, giving the corresponding glycosyl‐bromides, were mentioned in Section 2.8 and this topic has been comprehensively reviewed.⁷⁴ 

As a later example for bromination of an uronic acid containing disaccharide, compound 94 was isolated (64%) containing the bromide at the expectedly most reactive capto‐dative centre.²⁹  Reaction of cellobiosyl piperidine was notified in Section 2.4.1.

C–Br bonds are prone to reactions following both heterolytic and homolytic pathways. These possibilities have been extensively exploited with the brominated sugar derivatives. The reactions are categorized according to the above mechanistic characteristics. Reactions of structurally related anomeric halides of KDO and Neu5Ac derivatives will not be handled here.

Antecedents:³  The simplest nucleophile, the hydride ion (from LiAlH₄) was used for reductive dehalogenations of some 5‐brominated D‐glucopyranosides. The reactions gave preponderantly the configurationally inverted products thus providing an access to L‐idose derivatives. Reduction of 1‐bromo‐D‐glycopyranosyl cyanides by NaBH₄ gave anomeric mixtures of the corresponding glycosyl cyanides. Deuteration of 1,5‐anhydro‐5‐bromo‐2,3‐O‐isopropylidene‐β‐D‐ribose by LiEt₃BD took place with very highly selective inversion.

To the best of our knowledge, no further examples of H^–/D^– substitutions exist in the literature.

Antecedents:³  1‐O‐Acetyl‐2,3,5,6‐tetra‐O‐benzoyl‐4‐bromo‐β‐D‐galactofuranose gave the D‐gluco configured 4‐fluoride with AgF in CH₃CN while an epimeric mixture was obtained with AgBF₄ in Et₂O. Similar observations were made with 1‐O‐acetyl‐2,3,5‐tri‐O‐benzoyl‐4‐bromo‐β‐D‐ribofuranose to give, as the major product, the inverted 4‐fluoride with AgF and the retained one with AgBF₄. On treatment with AgF, 2,3,4,6‐tetra‐O‐acetyl‐1‐bromo‐β‐D‐glucopyranosyl chloride gave the corresponding 1‐chloro‐1‐fluoro compound with inversion of configuration and the 1,1‐difluoride was produced by using an excess of the reagent. Equilibrations of 2,3,4,6‐tetra‐O‐acetyl‐1‐bromo‐β‐D‐gluco‐ and ‐galactopyranosyl cyanides with Bu₄NBr in CCl₄ gave mixtures containing ∼10% of the α‐D‐anomers, and this allowed to estimate a rather strong anomeric effect for the CN group.

The general trend of substituting F for Br mainly with inversion by using AgF and mainly with retention by applying AgBF₄ (sometimes replaced by AgF followed by BF₃ or HF) was widely observed in syntheses of various fluorides.

Phenyl 5‐fluoro‐glucoside 95 was obtained with retention in 58% yield from the corresponding 5‐bromide. Since the radical bromination of methyl glucosides took a different course (preponderant attack at C‐1, see Section 2.2.1), for the preparation of methyl 5‐fluoro‐glucosides related to 95 an alternative ionic route was also described.¹⁶  O‐Peracylated 5‐fluoro‐β‐D‐glucopyranoses 96–98 with a retained configuration at C‐5 were isolated in 61, 85, and 80% yields, respectively (97 was accompanied by a small amount (8%) of the 5‐epimer 100). 5‐Fluoride of L‐idose peracetate 99 was obtained with inversion at the C‐5 in 76% yield.¹⁶ 

For D‐glucosamine derivatives Br→F exchange at the C‐5 could be best achieved for the β‐D‐anomeric acetate with N‐phthaloyl protection: the inverted 102 was obtained in 52% yield whose epimerization gave 101 (75%). Also studied were the corresponding 2,2,2‐trichloroethyl β‐D‐glucosaminides as well as both anomeric acetates each with N‐tetrachlorophthaloyl protection, however, these gave less satisfactory results.²⁵ 

Reaction of AgF (1.25 equivs) with O‐peracetylated 1‐bromo‐β‐D‐galactopyranosyl chloride gave an inseparable mixture of 103 and 104 (57 : 43), while 3.3 equivs gave difluoride 105.^68,80 

1,5‐Difluorides 106 and 107 of β‐D‐anomeric configuration were prepared in 55 and 80% yields, respectively.¹⁶  The 2‐phthalimido compound 108 was obtained in 16% overall yield for the Br and subsequent F substitution.²⁴  O,N‐Peracetylated 5‐bromo‐β‐D‐glucosaminyl fluoride furnished the inverted 109 in a modest 8% overall yield for the bromination‐fluorination sequence.²⁶  L‐Iduronic acid derivative 111 was obtained by AgF treatment with a C‐5 inversion of the corresponding bromide which furnished a mixture of 110 and 111 (32 and 12%, respectively) on reacting with AgBF₄.²³  From 5‐bromides of α‐D‐anomeric fluorides with D‐gluco and D‐galacto configurations 112¹⁶  and 113⁷⁰  were obtained in 53 and 14% yields, respectively. To get D‐manno compounds the corresponding 5‐bromide was reacted with AgF to give 115 (55%) which was epimerized to 114 by BF₃ (55%).⁷¹  5‐Bromo‐β‐D‐xylopyranosyl fluoride gave 116 (28%) and the α‐anomer furnished a mixture of 117 (46%) and 118 (30%).⁷² 

Trifluoro derivatives 119 (5%) and 120 (3%) were prepared by AgBF₄ from the 5‐bromide of the corresponding 1,1‐difluoro glucose in the indicated overall yields for the bromination‐fluorination sequence. Each reaction was reported to be very sluggish and to give many side‐products, and this bromide did not react with AgF in CH₃CN. From a bromination mixture containing both 1‐ and 5‐bromides of 2‐deoxy‐2‐fluoro‐β‐D‐glucopyranosyl chlorides only the latter reacted with AgF to give 121 (7% for the two steps) in an overnight reaction. The 1‐bromo‐2‐deoxy‐2‐fluoro‐β‐D‐glucopyranosyl chloride was treated with an excess of AgF for 10 days to give 122 as an impure material. 1,2,5‐Trifluoride 123 of L‐ido configuration was obtained in 28% yield in an inversion reaction elicited by AgF.⁶⁹  1,5,5‐Trifluoro derivative of D‐xylose 124 was obtained from the corresponding 5,5‐dibromide in 59% yield.⁷² 

Halogen substitutions in 1‐bromo‐D‐glycopyranosyl cyanides were also studied (Scheme 10). Under kinetic conditions (18 h, rt) LiCl converted A10 (D‐galacto configuration) to a mixture of B10 and C10 (70∶30) which reached an equilibrium ratio of 15 : 85 in one week.⁸¹  Treatment of the D‐galacto configured A10 with AgF gave, expectedly, the inverted fluoride D10, while its reaction with AgBF₄ furnished the retained fluoride E10. Transformation of B10 with AgF needed forcing conditions to give E10. Reactions A10→D10 were also performed with D‐gluco, D‐xylo, and D‐arabino configured substrates, and the conformational equilibria of the pentose derivatives were examined.⁸² 

From the reaction of O‐peracetylated C‐(1‐bromo‐β‐D‐galactopyranosyl)formamide with AgF in CH₃CN the inverted fluoride 125 was isolated in ∼3% yield. The major product of this transformation was formed by solvent incorporation, and this will be discussed in Section 3.1.1.5.

4‐Fluoro‐β‐D‐ribofuranosyl derivative 126 was obtained in 30% overall yield for the bromination‐fluorination (AgF, BF₃) transformations.⁴¹ 

Antecedents:³  Hydrolysis of O‐peracylated 5‐bromo‐D‐glycopyranoses and 4‐bromo‐glycofuranoses gave the 5‐ or 4‐OH derivatives, respectively. In some cases, upon further ring opening equilibration due to the hemiacetalic nature of these compounds and subsequent loss of AcylOH, the corresponding 5‐hydroxy‐D‐glycopyranoses (with free 1‐OH) were obtained as the ultimate products. Alcoholysis of glycos‐2‐ulopyranosyl bromides resulted in aldos‐2‐ulosides, and the transformation was extended to syntheses of di‐ and trisaccharides. 5‐Bromo‐D‐xylopyranose was converted into the 5‐methoxy derivative with inversion, and similar transformations were carried out with 1,2,3,4‐tetra‐O‐acetyl‐β‐D‐glucopyranose and ethylene glycol as the nucleophiles to give unusual disaccharidic and bicyclic compounds, respectively. C‐1‐Orthoesters including spiro derivatives were prepared from 1‐bromo‐β‐D‐glucopyranosyl chloride, and a bicyclic bis‐acetal from 6‐bromo‐1,6‐anhydro‐glucopyranose. The bromine substituent in O‐peracylated 5‐bromo‐D‐glycopyranuronates, 5‐bromo‐D‐glycopyranoses, 6‐bromo‐1,6‐anhydro‐D‐glycopyranoses, and 1‐bromo‐D‐glycopyranosyl cyanides was replaced by acetoxy groups mostly with inversion of the configuration at the given reaction centre.

Acidic hydrolysis of crude bromoiminolactones furnished glyconolactones 127–131 (overall yields for bromination‐hydrolysis).⁸³ 

Methanolysis of the corresponding 1‐bromo‐glycosyl phosphonic acid gave glycoside 132 (51% overall yield for bromination‐alcoholysis).⁴³ 

Alcoholysis of the corresponding 1‐bromo‐glycosyl cyanide in the presence of AgOTf and 2,6‐lutidine gave orthoesters 133 and 138 with MeOH, and 134 and 139 with ethyleneglycol, respectively.⁴⁴  Without AgOTf in the 2‐deoxy series only the bromine was replaced to give 135, and from the 4‐epimer a mixture of methyl glycosides 136 and 137 (1∶4 ratio, 51% combined yield) was formed,⁴⁸  while with trans‐1,2‐cyclohexanediol 140 was obtained together with a HBr elimination product.⁴⁴  Reaction with AgOAc in AcOH‐Ac₂O resulted in a mixture of 141 and 142.⁴⁸  Furanoid anomers 143 were obtained by using AgOTf and 2,6‐lutidine as the promoter system.⁵⁴ 

O‐Acyl protected C‐(1‐bromo‐D‐glycopyranosyl)formamides and ‐formates (Scheme 11, C11 and D11, respectively), were reacted with several O‐nucleophiles using silver(I) salts as promoters in most cases. Hydrolysis of the respective bromides resulted in ulosonamides A11,^45,84  and methyl ulosonates B11.²²  Alkyl glycosides E11 were obtained from C11 with alcohols (R=Me, Et, nBu, tBu, Bn),⁸⁵  while for promoting transformations D11→F11 (R=Me) Hg(CN)₂‐HgCl₂ was used.⁶⁴  Phenolate salts reacted without promoters to give E11 (R=2‐ and 4‐NO₂‐C₆H₄),⁸⁵  or were generated in the reaction mixture by using K₂CO₃ in acetone to yield F11 (R=2‐NO₂‐C₆H₄ and 5‐MeO‐2‐NO₂‐C₆H₃).⁵³  When C11 was reacted in acetone, the solvent behaved as the nucleophile and incorporated in the products G11 (<10%) and H11 (>70%). A probably similar incorporation of DMSO followed by a Pummerer‐type rearrangement gave compounds I11.⁸⁴ 

Hydrolysis of 1‐bromo‐glycosyl‐benzenes was mentioned in Section 2.6.2.

Alcoholysis of 5‐thio‐xylopyranosylidene dibromide gave 144 (R=Me, Et, allyl) in 70–90% yields.⁷³ 

2‐Acyloximino‐glycosyl bromides were extensively used in glycosylations.^86–88  Details of this chemistry can be found in a very recent comprehensive review.⁷⁴ 

Hydrolysis of O‐peracetylated 1,6‐anhydro‐6‐bromo‐D‐glycopyranoses of gluco‐, galacto‐, and manno configurations in the presence of Ag₂CO₃ in moist acetone gave lactols 145 which were immediately oxidized to the corresponding lactones.⁷⁵  On treatment by TMSOTf in Ac₂O 1,6‐anhydro‐6‐bromo‐D‐glucopyranose furnished 146 which was further acetolyzed to 147.⁸⁹  Both hydrolysis and methanolysis of the corresponding bromides gave the substituted products 148,⁷⁷  149, and 150⁷⁸  with retention of the configuration. The lactol ring of the hydrolysed product 151 opened up to the aldehyde which was oxidized to the corresponding carboxylic acid used in the synthesis of herbicidin glycoside.¹⁷ 

Antecedents:³  O‐Peracetylated 5‐bromo‐β‐D‐xylopyranose was transformed to the 5‐acetylthio derivative with inversion at C‐5. 1,6‐Anhydro‐6‐bromo‐glucopyranose was converted to the inverted 6‐phenylthio compound. 1‐Bromo‐β‐D‐glycopyranosyl cyanides gave thioglycosides with 2‐amino‐thiophenol and glycosylidene‐spiro‐perhydro‐1,4‐thiazine derivatives with 2‐amino‐ethanethiol.

Tetraacetate of 1‐bromo‐β‐D‐galactopyranosyl cyanide was reacted with AgSCN or KSCN in refluxing CH₃NO₂ to give a 6 : 4 mixture of 152 and 153 (combined yield 75%),⁹⁰  and the transformation was extended to the β‐D‐gluco, β‐D‐xylo, and α‐D‐arabino configured substrates, too.⁹¹  Formation of isothiocyanates was not observed in these reactions. Equilibration in the presence of Bu₄NCS in CCl₄ resulted in 152 and 153 in 46 : 54 ratio. This allowed to calculate the anomeric effect of the SCN substituent, which was corroborated by X‐ray crystallographic measurements of the bond lengths and angles around the anomeric centres.⁹¹  Reaction of O‐perbenzoylated 1‐bromo‐β‐D‐ribofuranosyl cyanide and thiourea in refluxing sulfolane‐EtOH mixture followed by acidic hydrolysis furnished 37% of 154.⁵⁶  Analogous reaction of O‐perbenzoylated β‐D‐glucopyranosyl cyanide could not be elicited, however, C‐(1‐bromo‐β‐D‐glucopyranosyl)formamide reacted with thiourea to give the spiro compound 155 (82%).⁹²  The same substrate and thiophenols in acetone in the presence of K₂CO₃ gave thioglycosides 156 (R=Ph, 2‐pyridyl, 2‐benzothiazolyl) in more than 70% yields.⁸⁵  5‐Thio‐xylopyranosylidene dibromide gave 157 (91%) with ethanethiol.⁷³ 

Antecedents:³  Replacement of the bromide in tetra‐O‐acetyl‐5‐bromo‐β‐D‐xylopyranose was achieved by azide ion and a purine derivative.

Sporadic reports can be found on direct substitution by amines in the brominated compounds. Thus, by treatment with an excess of aniline, the respective O‐peracylated C‐(1‐bromo‐β‐D‐glycopyranosyl)formamides gave 158 (55%) and 159 (75%) with an inversion of the C‐1 configuration.⁸⁵  Reaction of the corresponding bridged bromo derivative with anilines, followed by removal of the O‐acetyl protecting groups, furnished amines 160 (R=H 93%, 4‐MeO 67%, 4‐NO₂ 29%) with a retention of configuration.⁷⁸  In many cases amines were obtained by reduction of azides as indicated in the next paragraphs.

Azide substitution of the bromo derivatives (generally by NaN₃ or sometimes with LiN₃ in DMF or DMSO at rt) was studied very extensively, and took place almost always with inversion of the configuration of the respective reaction centre. Phenyl 5‐bromo‐β‐D‐xylopyranoside triacetate gave 161 (45%) which was then oxidized to the corresponding nitro derivative.²⁹ 

Azide replacements in O‐acyl protected 1‐bromo‐D‐glycopyranosyl cyanides,⁸¹  (Scheme 12, A12, configurations: β‐D‐gluco, β‐D‐galacto, α‐D‐arabino), gave the inverted products C12 in very short reaction times. It was shown (with the β‐D‐galacto substrate) that in prolonged reactions cycloaddition of the azide ion to the nitrile moiety also occurred providing tetrazole D12 with the same configuration as that of C12. In case of the β‐D‐manno substrate no defined product could be isolated, but the formation of a 5‐(1‐bromoglycosyl)tetrazole was made likely. In the reaction of the β‐D‐xylo substrate, parallel formation of the epimeric C12 and E12 was observed, and mechanistic studies were carried out to explain this finding. Due to the generally inverting nature of the azide substitution, compounds E12 could be obtained from chlorides B12. Longer reaction times led to the corresponding tetrazoles D12 in these cases, too. The azido nitriles gave oxazepine derivatives in photochemical ring enlargement reactions,^93,94  and were studied together with several other glycosyl azides by CD spectroscopy for their conformational peculiarities.⁹⁵ 

From O‐peracylated C‐(1‐bromoglycopyranosyl)formamides azides 162–165 (primary amides also with the D‐arabino configuration⁵¹ ) were produced (Table 5).

A large array of azides was prepared from C‐(1‐bromo‐glycopyranosyl)formates (e.g. 166–171 in Table 5). Compound 170, on treatment by Zn/N‐methylimidazole to remove the trichloroethyl protection,⁹⁶  gave an anomeric α‐azido acid derivative, while with Zn/AcOH furnished a configurationally labile anomeric α‐amino acid.²²  Azides 172–178 were obtained from the non isolated bromides listed in Table 4, and were used for the syntheses of anomeric spirocycles and oligopeptides incorporating anomeric α‐amino acids. Anomeric α‐azido acid azide 179 was prepared by bromination and subsequent azide substitution of the corresponding anhydro‐aldonoyl chloride, and used further to get dipeptides.

O‐Acetyl protected 1‐bromo‐β‐D‐glycopyranosyl chlorides gave 1,1‐diazido derivatives 180 by using the usual reagents or in higher yields (D‐gluco: 82%, D‐galacto: 65%, D‐manno: 36%) under phase transfer catalytic conditions.⁹⁷  These compounds facilitated among others anomeric carbene generation^98,99  and synthesis of unusual 6,7‐dihydropyrano[3,4‐d]‐1,2,3‐triazoles.¹⁰⁰ 

The corresponding bridged bromo derivatives gave azides 181⁷⁷  (94%) and 182⁷⁸  (56%). 181 was reduced to the corresponding amine which was further acylated by some acid chlorides.⁷⁷ 

Reaction of a C‐(1‐bromo‐glycosyl)formamide (Scheme 13, A13) with cyanate ion was first reported for the synthesis of D‐ribofuranosylidene‐spiro‐hydantoin (type D13, 44%).⁵⁸  This kind of transformation was then extended to pyranoid compounds and thiocyanates, too. It was shown, that for the preparation of hydantoins only AgOCN could be used, and in most cases mixtures of epimers B13 and D13 were formed in which B13 of retained anomeric/spiro configuration was preponderant. On the other hand, formation of thiohydantoins C13 by thiocyanate ions was not sensitive to the counterion (Ag, K, NH₄ salts worked similarly), and the sole products of these reactions had inverted configuration at the reaction centre. Mechanistic studies were carried out to explain these findings.⁴⁵  Hydantoins (D‐gluco,^45,102  D‐galacto,⁵¹  D‐arabino,⁵¹  and D‐xylo⁴⁵ ) and thiohydantoins (D‐gluco,^12,45,46,102  D‐galacto,⁵¹  D‐arabino,⁵¹  D‐xylo,⁴⁵  and L‐rhamno⁵² ) were prepared in the indicated configurations.

Silylated thymine was reacted, as an N‐nucleophile, with the corresponding bromo compounds to give derivatives 183⁵⁷  (56%) and 184⁵⁵  (69% for the mixture, the epimers could be separated), which were converted further to functionally modified oligonucleotides.⁵⁵ 

Nitriles reacted with C‐(1‐bromo‐D‐glycopyranosyl)formamides in the presence of silver salts (Ag₂CO₃ or AgOTf) in a Ritter‐type reaction exemplified by the D‐galacto configured substrate in Scheme 14. Primary amides A14 (R¹=H) were transformed, probably via intermediate B14 undergoing a tautomeric ring opening and a further tautomerization, into C14. Compounds of type C14 were obtained in D‐gluco and D‐arabino configurations, as well (R²=CH₃, CH₃CH₂, CH₂=CH, CH₂=CHCH₂, (CH₃)₃C, CH₃OCH₂).¹⁰³  Reactions of substituted amides A14 (R¹≠H) stopped at B14 which could be isolated and opened by a mild acidic hydrolysis to give peptides D14 incorporating anomeric α‐amino acids (e.g. R¹=CH₂CO₂Me, R²=CH₂NHCO₂Bn).¹⁰⁴ 

Antecedents:³  none.

6‐Bromo‐1,6‐anhydro‐D‐mannose triacetate was reacted with trimethylsilylated carbon nucleophiles in the presence of AgOTf in CH₂Cl₂ to give 185 (R=allyl, phenylethynyl, hept‐1‐ynyl, allenyl) in 47–63% yields. In toluene 185 (R=4‐Me‐C₆H₄) was obtained while with other aromatic reaction partners iPrCN proved the best solvent to yield 66–76% of 185 (R=2‐MeO‐5‐Me‐C₆H₃, 2,5‐di‐MeO‐C₆H₃, 2‐furyl).¹⁰⁵ 

Antecedents:³  The O‐peracetylated bromosugar derivatives were prone to HBr eliminations elicited by bases (generally DBU) or to reductive eliminations induced by a metal (almost exclusively Zn in AcOH). Thus, 5‐bromo‐uronates and 5‐bromo‐β‐D‐xylose gave the corresponding 4‐acetoxy‐ or 4‐deoxy‐hex‐4‐enopyranuronate and 4‐acetoxy‐ or 4‐deoxy‐pent‐4‐enopyranose derivatives, respectively. In the case of 5‐bromo‐hexopyranoses the elimination may form endo‐ and exocyclic double bonds. Thus, from 5‐bromo‐β‐D‐glucopyranose esters, HBr elimination gave the endo‐alkene, while Zn/AcOH led to the exo‐methylene derivative as the main products. Interestingly, with the same substrate, endo elimination of HBr took place instead of substitution with NaCN, NaOBz, or CsF, however, NaSAc or NaI furnished the unsaturated products in exo sense. Following these lines a 5‐exo‐methylene derivative could be obtained from the reducing‐end 5‐bromide of octa‐O‐acetyl‐maltose by Zn/AcOH. In β‐D‐gluco‐ or ‐galactofuranose peresters 3,4‐endo double bonds were formed by DBU, and reductive elimination furnished the 4‐exo‐alkenes. Acetylated 1‐bromoglycosyl cyanides gave 1‐cyano‐2‐acetoxy‐glycals with DBU, but cleaner transformations were induced by Hg(CN)₂/AgOTf, and aldonolactones were isolated on treatment by Hg(OAc)₂ in DMSO. Reductive elimination was perfomed with Zn in refluxing benzene in the presence of Et₃N or pyridine to yield 1‐cyano‐glycals. Reaction of O‐peracetylated 1‐bromo‐β‐D‐glucopyranosyl chloride with DBU or DABCO gave the 1‐chloro‐2‐acetoxy‐glucal.

Unsaturated uronate 186 was obtained by DBU from the corresponding 5‐bromide.³¹  Formation of unsaturated phosphonic esters 187 (76%) and 188 (72%) was brought about by Zn/Cu in EtOH.²⁷  The Zn/N‐base method for reductive elimination was extended to prepare C‐(hex‐1‐enopyranosyl)formamides (e.g. 189⁵⁰  70% by Zn/N‐methylimidazole in refluxing EtOAc) and ‐heterocycles (e.g. 190¹⁰⁶  38% by Zn/Py in refluxing benzene). Based on these experiences, a general method¹⁰⁷  was elaborated for the preparation of glycals¹⁰⁸  from the corresponding O‐peracylated glycosyl bromides that can be regarded as the aprotic variant of the classical Fischer‐Zach synthesis of glycals. Mechanistic studies were also disclosed.¹⁰⁹ 

DBU induced elimination of HCl from 3‐(1‐chloro‐β‐D‐glycopyranosyl)propenes produced glycosylidene‐butadienes 191 (D‐gluco 42%, D‐galacto 46%, D‐manno 45%).¹¹⁰  Elimination of HBr from 1‐bromo‐β‐D‐galactopyranosyl chloride by DABCO gave 192 in 33% yield. Attempted reductive elimination to get a 1‐chloro‐glycal failed under several conditions: the only isolable product was tri‐O‐acetyl‐D‐galactal.⁶⁸  From the corresponding 1,1‐dibromide 1,5‐anhydro‐1‐bromo‐5‐thio‐D‐threo‐pent‐1‐enitol triacetate (193, 96%) was obtained by DBU.⁷³  Unsaturated disaccharide 194 was obtained by DBU in DMF in 75% yield.²⁹  Exomethylene derivative 195 (72%) was prepared from 4‐bromo‐D‐fructofuranose pentabenzoate by the Zn/N‐methylimidazole method in refluxing EtOAc, and used further for polymerization studies.³⁶ 

Antecedents:³  Tributyltin‐hydride/deuteride in the presence of a radical initiator (mostly AIBN) was applied to replace bromine by H/D. Methyl glycosides and 1‐O‐acetates of 5‐bromo‐β‐D‐glucopyranuronates gave mixtures of D‐gluco and L‐ido isomers of the corresponding uronic acid derivatives with isolated yields in the range of 44–64% and 28‐38%, respectively. With a 4‐deoxy derivative, the D‐gluco compound was formed almost quantitatively, and retention of the C‐5 configuration was also very high starting with a D‐galacto substrate. Under similar conditions 1‐bromo‐β‐D‐glycopyranosyl cyanides gave mostly α/β mixtures (D‐galacto 4∶6, D‐arabino 2∶8, D‐manno β only) of the corresponding glycosyl cyanides. These findings indicated the outstanding importance of the neighbouring substituent in governing the reductions. 1,5‐Anhydro‐5‐bromo‐pentofuranose and 1,6‐anhydro‐6‐bromo‐hexopyranose derivatives including disaccharides were deuterated in this way at C‐5 and C‐6, respectively, with very high (often exclusive) stereoselectivities. This allowed to establish conformational preferences around the C‐5–C‐6 bond of pyranoid compounds by NMR spectroscopy. Reduction of 5‐bromo‐5‐cyanoethyl‐β‐D‐xylopyranose gave the D‐gluco configured product.

Phenyl 2,3‐di‐O‐acetyl‐5‐bromo‐4‐deoxy‐4‐fluoro‐β‐D‐glucopyranosid‐uronic acid was deuterated by Bu₃SnD to give a ∼1 : 1 mixture of the D‐gluco and L‐ido epimers from which 196 could be isolated in only 5% yield.³⁰  Contrary to that of the β‐anomer, reduction of methyl 1,2,3,4‐tetra‐O‐acetyl‐α‐D‐glucopyranuronate with Bu₃SnH gave the D‐gluco and L‐ido isomers in 1 : 3 ratio and 197 was prepared in 67% yield.^33,34  Deuteration of 4‐bromo‐D‐ribofuranose tetraacetates gave 1∶4 mixtures of 198.³⁵  Reduction of non‐isolated 1‐(2‐cyano‐ or 2‐phosphonoethyl)‐1‐chloro‐D‐glucopyranosyl derivatives gave 199 and 200, respectively, thereby providing access to β‐D‐configured C‐glucopyranosyl alkane type compounds.¹¹¹  Reaction of 3‐(1‐chloro‐β‐D‐glycopyranosyl)propenes with Bu₃SnH resulted in the formation of C‐allyl glycosides 201 (D‐gluco 50%, D‐galacto 51%, D‐manno 57%).¹¹⁰  1,5‐Anhydro‐L‐rhamnulose 202 was prepared by reduction of the corresponding ulosyl bromide in 71% yield.¹¹²  Isotope labelled (6R)‐1,2,3,4,5,6‐²H‐1,2,3,4,5,6‐¹³C‐1,6‐anhydro‐2,3,4‐tri‐O‐benzoyl‐D‐glucopyranose was prepared by Bu₃SnH treatment of the corresponding 6‐bromide.⁷⁶ 

Radicals can undergo substitutions according to mechanisms more complex than atom abstraction. Those reactions applied to brominated sugars are illustrated in Scheme 15. Thus, as shown in route a, radicals R˙ may add to alkenes (C‐1 or C‐4/5 sugar radicals are rendered nucleophilic by the ring oxygen, therefore, ideal partners are electron deficient olefines with X=electron withdrawing group) and subsequent hydrogen abstraction gives the alkylated product. Route b shows reactions with alkenes having a radical leaving group (L_R, in the forthcoming examples Bu₃Sn, RSO₂) that makes the intermediate prone to β‐fragmentation to furnish the product of allyl substitution.

Antecedents:³  Treatment of a C‐5 epimeric mixture of O‐peracetylated 5‐bromo‐β‐D‐xylopyranoses with acrylonitrile in the presence of Bu₃SnH and a radical initiator resulted in the formation of a mixture of D‐gluco and L‐ido configured 5‐bromo‐5‐cyanoethyl‐β‐D‐xylopyranoses.

Among transformations according to route a in Scheme 15 reaction of O‐peracetylated 1‐bromo‐β‐D‐glucopyranosyl chloride with acrylonitrile in the presence of Bu₃SnH gave complex mixtures under a variety of conditions. In THF and Et₂O products of mono‐ and di‐hydrodehalogenation preponderated, but in benzene the mixture contained mainly 203 (16%) and the products of elimination of HCl in both exo (22%) and endo (25%) fashion.^113,114  Combination of this substitution and Bu₃SnH reduction of 203 and also that of non‐isolated 204¹¹¹  gave C‐β‐D‐glucopyranosyl derivatives 199 and 200, respectively, discussed in the previous section.¹¹⁵ 

Allylation of 5‐bromo‐β‐D‐glucopyranose pentaacetate following route b gave an inseparable mixture of 205 and 206 (10 : 3, combined yield 76%).¹¹⁶  Starting with 5‐bromo‐α‐D‐lyxopyranose tetraacetate, 207 and 208 were prepared (68% for a 3 : 1 mixture), and the former could be isolated by HPLC.³²  Reactions of O‐peracetylated 1‐bromo‐β‐D‐glycopyranosyl chlorides with Bu₃SnCH₂CH=CH₂ under irradiation produced 209 (D‐gluco 86%, D‐galacto 51%, D‐manno 31%) together with minor amounts of hydrolysis products.^110,117  To get bis‐allyl compounds 210 the 1,1‐Cl,Br derivatives or 209 were reacted with 6 equiv of Bu₃SnCH₂CH=CH₂ under various conditions, however, complex mixtures were obtained (containing among others hydrolysis, reduction, and HCl endo‐elimination products of 209) from which 210 could be isolated in moderate yields (D‐gluco 24%, D‐galacto 24%, D‐manno 34%).¹¹⁸  From an anomeric mixture of O‐perbenzoylated 1‐bromo‐D‐ribofuranosyl cyanides the C‐1 epimeric allylated compounds were obtained in a 5∶1 β/α ratio, and the β‐isomer 211 was isolated in 42% yield.⁵⁶  4‐Allyl tetrofuranose derivative 212 was prepared from the corresponding 4‐chloro‐tetrofuranose.^119,120 

Tri‐O‐benzoyl‐1,6‐anhydro‐6‐bromo‐D‐glucopyranose was reacted with a series of stannanes to 6‐exo‐alkylated bicycles 213 (R=allyl, 2‐methylallyl, (2‐benzyloxymethyl)allyl, methyl β‐acrylyl, ethyl β‐acrylyl) in 30–50% yields.^121,122  The use of allyl sulfones as reagents significantly improved the yields for 213 (R=allyl 67%, 2‐chloroallyl 71%).¹²²  To the best of our knowledge, there is only one example for the introduction of a nitrile group in the bromo sugar derivatives, namely the formation of 214 with tBuNC by a similar β‐fragmentation to that of route b.¹²¹ 

Antecedents:³  none.

Reaction of dibromide 215 with Bu₃SnH resulted in a high yielding elimination of bromine to give the glycal derivative 216.⁶⁴ 

A unique transformation of bromoiminolactones was brought about by highly reactive metal‐graphites such as Zn/Ag‐graphite or more favourably C₈K in THF (Scheme 16). Thus, N‐metalation resulted in a ring opening to a cyano‐alkoxide intermediate which, in the presence of electrophiles, gave aldononitriles 217–223.¹²³ 

A tabular presentation of the biological/biochemical investigations with the synthesized compounds may kindly orient the reader on the usefulness of derivatives 224–257. Some hints on the synthetic pathway are also given (Table 6).

Radical reactions are in the forefront of synthetic and biological chemistry. Radical bromination, and to a much lesser extent chlorination, of carbon atoms adjacent to ring oxygen in carbohydrate derivatives, pioneered by Robin Ferrier and coworkers and followed by many others, have become extremely valuable procedures for specific functionalization of sugars. This methodology has become part of the common chemical property, and seems so evident and predictable that citations of the first papers are continuously diminishing. The diversity of substrates investigated so far has been enormously increased and several of them have only recently been studied. The propensity of the carbon‐halogen bond to both heterolytic and homolytic cleavage facilitates a great variety of transformations, among others highly selective attachment of further heteroatomic or carbon substituents to the ring, regio‐ and stereoselective isotope labelling, access to synthetically useful unsaturated derivatives, synthesis of compounds with uncommon structures like spirocycles, sugar‐peptide hybrides, polymers, etc. On the other hand, the versatility of the reactions has allowed to obtain excellent molecular tools for studies in glycobiochemistry and chemical glycobiology, e.g. enzyme inhibitors and inactivators, as well as antibacterial agents. Thus, for the fourth decade of its “life”, radical‐mediated halogenation of carbohydrates has grown up to a solid and reliable synthetic method and provider of sophisticated substances for glycobiological investigations.

Support by the Hungarian Scientific Research Fund (OTKA CK77712), TÁMOP 4.2.1/B‐09/1/KONV‐2010‐0007 and TÁMOP‐4.2.2./B‐10/1‐2010‐0024 projects co‐financed by the European Union and the European Social Fund is acknowledged.