Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2006 Sep 1.
Published in final edited form as: Carbohydr Res. 2006 Mar 13;341(10):1748–1752. doi: 10.1016/j.carres.2006.02.024

On the regioselectivity of the Hanessian–Hullar reaction in 4,6-O-benzylidene protected galactopyranosides

David Crich 1,*, Qingjia Yao 1, Albert A Bowers 1
PMCID: PMC1559505  NIHMSID: NIHMS10996  PMID: 16530738

Abstract

The N-bromosuccinimide mediated fragmentation of methyl 4,6-O-benzylidene-β-d-galactopyranoside results in the formation of methyl 4-O-benzoyl-6-bromo-6-deoxy-β-d-galactopyranoside and methyl 4-O-benzoyl-3-bromo-3-deoxy-β-d-gulopyranoside, as opposed to the methyl 6-O-benzoyl-3-bromo-3-deoxy-β-d-gulopyranoside originally reported. The kinetic methyl 4-O-benzoyl-6-bromo-6-deoxy-β-d-galactopyranoside rearranges to the thermodynamic methyl 4-O-benzoyl-3-bromo-3-deoxy-β-d-gulopyranoside under the reaction conditions, likely via a 3,6-anhydro galactopyranoside. The NBS-mediated cleavage of 4, 6-O-benzylidene acetals in the galactopyranoside series is therefore shown to conform to the regiochemistry observed in the corresponding gluco- and mannopyranoside series with preferential cleavage of the C6–O6 bond by an ionic mechanism.

Keywords: Deoxy sugars, Benzylidene acetal

1. Introduction

The N-bromosuccinimide mediated fragmentation of benzylidene acetals is an important and facile means of access to deoxy sugars.1–11 The reaction is accepted to proceed by an initial hydrogen atom abstraction to give a benzylidene radical, which then suffers fragmentation by either of the two potential pathways, as illustrated for the 4,6-O-benzylidene-type acetal in the gluco-series (Scheme 1). Both pathways were envisaged from the outset by Hanessian who, nevertheless, favored the ionic mode of fragmentation.1,3 Hullar, on the other hand, initially promoted the pure radical pathway.2 Indirect support has subsequently been provided for the ionic fragmentation,12 and this is generally accepted to be the most plausible mechanism. The waters were muddied, however, by the elegant work of Roberts who, following early work by Jeppesen,13 showed that pure radical fragmentations of 4,6-O-benzylidene acetals proceed with preferential cleavage of the primary C6–O6 bond in both the glucose and mannose series.14–17 Work from this laboratory concurs with Roberts regarding the regioselectivity of fragmentation of the 4,6-O-benzylidene acetals in both the gluco- and manno-series under free radical conditions.18

Perusal of the literature reveals that the 4,6-O-benzylidene galactopyranosides potentially afford the means of distinguishing between the two pathways for the NBS-mediated Hanessian–Hullar reaction. Thus, for the 4,6-O-benzylidene galactopyranosides the pure radical fragmentation affords mixtures of the 4- and 6-deoxy products owing to competing cleavage of the primary and secondary C–O bonds (Table 1, entry 1).14–17 On the other hand, two out of three 4,6-O-benzylidene galactopyranosides studied by Hanessian in the NBS-mediated protocol gave exclusively the 6-bromo-6-deoxy product (Table 1, entries 2 and 3), thereby strongly indicating the ionic fragmentation. Curiously, however, a third NBS-mediated example was reported to afford a mixture of the expected 6-bromo-6-deoxy product 9, and a somewhat unusual 3-bromo-3-deoxygulopyranoside 10 (Table 1, entry 4), which leaves open the possibility of alternative mechanisms. No explanation was provided by Hanessian for the formation of this unusual product 10, whose structure was based on comparison of physical data with that of an authentic sample19 of methyl 3-deoxy-β-d-xylopyranoside following hydrogenolytic removal of the bromine atom and saponification, but an ionic mechanism was later put forward by Gelas for its formation.9

Table 1.

Fragmentation of 4,6-O-benzylidene galactopyranosides according to Hanessian and Roberts

Entry Substratea Conditionsb Products (% yield)
1 graphic file with name nihms10996t1.jpg DTBP/TIPST graphic file with name nihms10996t2.jpg
2 graphic file with name nihms10996t3.jpg NBS graphic file with name nihms10996t4.jpg
3 graphic file with name nihms10996t5.jpg NBS graphic file with name nihms10996t6.jpg
4 graphic file with name nihms10996t7.jpg NBS graphic file with name nihms10996t8.jpg
Open in a new tab
a

Galactosides 4, 6, and 8 were originally represented3 with the phenyl group axial to the dioxane ring. We have preferred to represent them with the phenyl equatorial in line with crystallographic structures20 of benzylidene protected galactose derivatives.

b

DTBP: di-tert-butylperoxide; TIPST: triisopropylsilanethiol (i-Pr3SiSH).

With a view to better understand the regioselectivity of benzylidene acetal fragmentation, we have repeated the reaction of methyl 4,6-O-benzylidene-β-d-galactopyranoside 8 with NBS, and report here our results, which lead to a minor revision of the structure of the unusual 3-bromo-3-deoxyguloside. This minor correction of structure enables a mechanism to be written bringing the regioselectivity of fragmentation in the galactose series into full agreement with that in the glucose and mannose series.

2. Results and discussion

Methyl 4,6-O-benzylidene-β-d-galactopyranoside 821 was heated to reflux with NBS and barium carbonate in 3/1 mixture of tetrachloromethane and 1,2-dichloroethane for 4 h. Chromatography on silica gel then enabled the isolation of two products, in yields of 21% and 58%, to which we assign the structures 6-bromo-6-deoxygalactoside 9 and 4-O-benzoyl-3-bromo-3-deoxyguloside 11, respectively (Scheme 2).† For the major 3-bromo-3-deoxy product 11, the location of the benzoate ester on the 4-position and not the 6-position as originally reported (Table 1) is readily apparent from the chemical shift of H-4, an apparent doublet at δ 5.49 (J = 3.0 Hz). The axial nature of the C3–Br bond is clear from the 3J coupling constant of 3.0 Hz in the apparent triplet assigned to H3 resonating at δ 4.64.‡

Our structure differs from the original report of Hanessian by the placement of the benzoate ester, something that might easily have been overlooked in 1966, and which has no consequence on the conversion through saponification and hydrogenation to methyl 3-deoxy-β-d-xylopyranoside used in the original structural proof. Nevertheless, it is this different placement of the benzoate ester that enables the NBS-mediated cleavage of 8 to be brought into line with all other known NBS cleavage reactions of 4,6-O-benzylidene acetals, be they galacto-, gluco-, or manno-, with preferential cleavage of the C6–O6 bond. Thus, the initial fragmentation reaction gives the 6-bromo-6-deoxy product 9, which under the conditions of the reaction, is in equilibrium with a ring inverted conformer 12. The population of this minor conformer is facilitated by the switch from the equatorial to the axial glycoside and the corresponding gain in anomeric stabilization. Formation of a 3,6-anhydro sugar 13 ensues and this is finally cleaved following nucleophilic ring opening by bromide at the 3-position to give 14, which relaxes to the observed 4C1 conformer 11 (Scheme 3). In strong support of this argument we note that in a reaction taken to low conversion with 44% recovered substrate, similar to the original report of Hanessian, the yield of 9 and 11 was 18% and 14%, respectively, clearly indicating the equilibrium nature of the process with 9 as the kinetic product and 11 as the thermodynamic product.§ The difference between the β- and α-galactopyranoyl series in the NBS cleavage (Table 1, entries 3 and 4) now is readily seen to be a consequence of the greater ability of the β-anomer to populate the higher energy inverted conformer. The mechanism proposed for the formation of 11 (Scheme 3) is more plausible than the mechanism written by Gelas for the formation of the alternative regioisomer 10,9 which contains a stereoelectronically improbable shift of dioxenium ion between the 4,6- and 3,4-positions, and which does not explain the difference in behavior between the α- and β-anomers.

The high degree of regioselectivity observed in the NBS-mediated fragmentation of 9, as contrasted with the lack of regioselectivity in the pure radical fragmentations of galactose-based 4,6-O-benzylidene acetals (Table 1, entry 1), provides very strong evidence that the Hanessian–Hullar fragmentation proceeds via an ionic mechanism.

3. Experimental

3.1. General methods

Optical rotations were determined with an Autopol III polarimeter for solutions in CHCl3. NMR spectra were recorded for CDCl3 solutions with a Bruker Avance spectrometer. Chemical shifts are in parts per million downfield from tetramethylsilane. High resolution mass spectra were recorded with a Waters Q-TOF2 instrument.

3.2. Methyl 4-O-benzoyl-6-bromo-6-deoxy-β-d-galactopyranoside (9) and methyl 4-O-benzoyl-3-bromo-3-deoxy-β-d-gulopyranoside (11)

A solution of methyl 4,6-O-benzylidene-β-d-galactopyranoside 8 (500 mg, 1.8 mmol) in a mixture of CCl4 (28 mL) and 1,1,2,2-tetrachloroethane (9 mL) was treated with freshly recrystallized N-bromosuccinimide (360 mg, 2.0 mmol) and BaCO3 (715 mg, 3.7 mmol). The solution was deoxygenated by sparging with argon for 1 h and then was heated to reflux with stirring for 4 h. After the mixture was cooled to room temperature and filtered, it was dried (Na2SO4), concentrated, and purified by column chromatography on silica gel (2:1, hexane/EtOAc) to give first 11 (0.388 g, 1.03 mmol, 58%) as a syrup, and then 9 (0.139 g, 0.37 mmol, 21%) in the form of a viscous oil. Compound 9: [α]D +3.7 (c 1.0), lit.3 [α]D +25 (c 1.57, CHCl3); 1H NMR (500 MHz): δ 8.06 (d, 2H, J = 7.5 Hz, 2 × Hortho), 7.57 (t, 1H, J = 7.5 Hz, 1H, Hpara), 7.45 (t, 2H, J = 7.5 Hz, 2 × Hmeta), 5.68 (d, 1H, J = 1.5 Hz, H-4), 4.28 (d, 1H, J = 7.2 Hz, H-1), 3.88–3.92 (m, 2H, H-3 and H-5), 3.75 (m, 1H, H-2), 3.62 (s, 3H, CH3), 3.43 (m 2H, 2·× H-6); 13C NMR (125 MHz): δ 177.97 (C=O), 133.61 (Cpara), 130.06 (Cortho), 129.03 (Cipso), 128.53 (Cmeta), 103.96 (C-1), 73.93 (C-3 or C-5), 72.35 (C-3 or C-5), 71.59 (C-2), 70.55 (C-4), 57.45 (CH3), 28.97 (C-6); ESIMS m/z calcd for [C14H17NaO6Br]Na+: 383.0109. Found: 383.0104. Compound 11: [α]D +11.8 (c 0.56), lit.3 [α]D −6 (c 1.65, CHCl3); 1H NMR (500 MHz): δ 8.05 (d, 2H, J 7.5 Hz, 2 × Hortho), 7.61 (t, 1H, J = 7.5 Hz, Hpara), 7.47 (t, 2H, J = 7.5 Hz, 2 × Hmeta), 5.49 (d, 1H, J = 3.0 Hz, H-4), 4.70 (d, 1H, J = 7.5 Hz, H-1), 4.64 (t, 1H, J = 3.0 Hz, H-3), 4.53 (t, 1H, J = 6.5 Hz, H-5), 3.80–3.83 (m, 2H, H-2, H-6), 3.61–3.65 (m, 1H, H-6′), 3.62 (s, 3H, CH3); 13C NMR (125 MHz): δ 165.77 (C=O), 133.98 (Cpara), 130.02 (Cortho), 128.66 (Cmeta), 128.53 (Cipso), 102.15 (C-1), 72.11 (C-5), 71.62 (C-4), 67.65 (C-2), 61.23 (C-6), 57.39 (CH3), 51.82 (C-3); ESIMS m/z calcd for [C14H17NaO6Br]Na+: 383.0109. Found: 383.0117.

Supplementary data

Copies of the 1H and 13C NMR spectra for compounds 9 and 11. Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.carres.2006.02.024.

Supplementary Material

Supplementary Data

Scheme 1.

Scheme 1

Open in a new tab

Radical and ionic pathways for benzylidene fragmentation.

Scheme 2.

Scheme 2

Open in a new tab

Reaction of 8 with N-bromosuccinimide.

Scheme 3.

Scheme 3

Open in a new tab

Mechanism for the formation of 3-bromo-3-deoxyguloside 11.

Acknowledgments

We thank the NIH (GM62160) for support of this work.

Footnotes

†

Jacobsen and Mols mention the formation of methyl 4-O-benzoyl-3-bromo-3-deoxy-β-d-gulopyranoside, that is, 11, from the Hanessian reaction of 8 with NBS, but gives neither data nor a rationale for their structural assignment.22

‡

We note that Hanessian, working at 60 MHz, characterized the purported 10 by the presence of a doublet (J = 5.0 Hz) resonating at δ 5.54 in the 1H NMR spectrum, and that, while this was obviously assigned incorrectly to the anomeric hydrogen, it is consistent with H-4 in structure 11 now proposed.

§

The mass balance in all experiments is made up of several minor unidentified products.

References

  • 1.Hanessian S. Carbohydr Res. 1966;2:86–88. [Google Scholar]
  • 2.Failla DL, Hullar TL, Siskin SB. J Chem Soc, Chem Commun. 1966:716–717. [Google Scholar]
  • 3.Hanessian S, Plessas NR. J Org Chem. 1969;34:1035–1044. [Google Scholar]
  • 4.Hanessian S, Plessas NR. J Org Chem. 1969;34:1045–1053. [Google Scholar]
  • 5.Hullar TL, Siskin SB. J Org Chem. 1970;35:225–227. doi: 10.1021/jo00826a046. [DOI] [PubMed] [Google Scholar]
  • 6.Hanessian S. Org Synth. 1987;65:243–247. [Google Scholar]
  • 7.Hanessian, S. Some Approaches to the Synthesis of Halodeoxy Sugars. In Deoxy Sugars; Hanessian, S, Ed.; Advances in Chemistry Series 74; American Chemical Society: Washington, DC, 1968; pp 159–201.
  • 8.Chana JS, Collins PM, Farnia F, Peacock DJ. J Chem Soc, Chem Commun. 1988:94–96. [Google Scholar]
  • 9.Gelas J. Adv Carbohydr Chem Biochem. 1981;39:71–156. [Google Scholar]
  • 10.Szarek WA. Adv Carbohydr Chem Biochem. 1973;28:225–306. [Google Scholar]
  • 11.Paulsen H. Adv Carbohydr Chem Biochem. 1971;26:127–195. doi: 10.1016/s0096-5332(08)60169-1. [DOI] [PubMed] [Google Scholar]
  • 12.McNulty J, Wilson J, Rochon AC. J Org Chem. 2004;69:563–565. doi: 10.1021/jo035223x. [DOI] [PubMed] [Google Scholar]
  • 13.Jeppesen LM, Lundt I, Pedersen C. Acta Chem Scand. 1973;27:3579–3585. [Google Scholar]
  • 14.Dang HS, Roberts BP, Sekhon J, Smits TM. Org Biomol Chem. 2003;1:1330–1341. doi: 10.1039/b212303g. [DOI] [PubMed] [Google Scholar]
  • 15.Cai Y, Dang HS, Roberts BP. J Chem Soc, Perkin Trans 1. 2002:2449–2458. [Google Scholar]
  • 16.Roberts BP, Smits TM. Tetrahedron Lett. 2001;42:3663–3666. [Google Scholar]
  • 17.Fielding AJ, Franchi P, Roberts BP, Smits TM. J Chem Soc, Perkin Trans 2. 2002:155–163. [Google Scholar]
  • 18.Crich D, Yao Q. J Am Chem Soc. 2004;126:8232–8236. doi: 10.1021/ja048070j. [DOI] [PubMed] [Google Scholar]
  • 19.Dahlgard M, Chastain BH, Han RJL. J Org Chem. 1962;27:929–931. [Google Scholar]
  • 20.Gururaja TL, Venugopalan P, Levine MJ. J Chem Crystallogr. 1998;28:747–759. [Google Scholar]
  • 21.Sorkin E, Reichstein T. Helv Chim Acta. 1945;28:1–17. [PubMed] [Google Scholar]
  • 22.Jacobsen S, Mols O. Acta Chem Scand. 1981;B35:163–168. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data
NIHMS10996-supplement-supp1.pdf (207.9KB, pdf)

RESOURCES