Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 May 31.
Published in final edited form as: J Am Chem Soc. 1982 Oct 1;104(21):5788–5789. doi: 10.1021/ja00385a041

Diastereoselection in Intermolecular Nitrile Oxide Cycloaddition (NOC) Reactions: Confirmation of the “Anti-Periplanar Effect” through a Simple Synthesis of 2-Deoxy-d-ribose†

Alan P Kozikowski 1, Arun K Ghosh 1
PMCID: PMC6543839  NIHMSID: NIHMS1024672  PMID: 31156251

We have initiated recently a program to examine the extent to which an allylic asymmetric center can control diastereoface selection in both inter- and intramolecular additions of nitrile oxides to olefins.1 While the extent of such diastereoselection appears to be relatively small when there is little to distinguish the allylic groups on a steric or electronic basis (except in intramolecular cyclizations where the allylic center is within the nonisoxazoline ring being formed),2 we now report that an allylic oxygen substituent can, on the other hand, serve as a useful control element for achieving diastereoface selectivity in [3 + 2] cycloaddition reactions. We illustrate this new concept in stereocontrol through a simple synthesis of 2-deoxy-d-ribose.

Optically active (+)-(S)-isopropylidene-3-butene-1,2-diol, prepared from isopropylidene d-glyceraldehyde by reaction with methylenetriphenylphosphorane,3 was reacted with (carboethoxy) formonitrile oxide4 to afford an 80:20 mixture of diastereomeric cycloadducts. These products were separated by gravity chromatography, and the major isomer 3 (Scheme I) was heated with sodium hydroxide to effect the following transformations: (a) ester hydrolysis; (b) decarboxylative ring opening of the isoxazoline to a β-hydroxy nitrile; (c) hydrolysis of nitrile to carboxylate. Acidification and diazomethane treatment then yielded 4 (74% overall yield from 3).5 On converting this compound to its acetate and stirring with trifluoroacetic acid, the acetate of 2-deoxy-d-ribono-1,4-lactone(5) was formed (71%). The NMR of this compound was identical with that reported previously by Mukaiyama;6 [α]24D = ‒12° (c 0.75, CH2,Cl2,); IR (thin film) 3450, 1785, 1740, 1240 cm‒1; 1H NMR (CDCl3)δ 2.07 (s, 3 H), 2.55 (dd, 1 H, J = 2.5, 18 Hz), 2.95 (dd, 1 H, J = 7, 18 Hz), 3.40−3.60 (m, 1 H), 3.85 (d, 2 H, J = 2.5 Hz), 4.40−4.60 (m, 1 H), 5.20−5.50 (m, 1 H); mass spectrum (15 eV), m/e 143, 84, 83, 53.

Scheme I.

Scheme I

Open in a new tab

Trifluoroacetic acid treatment of 4 followed by bis(3-methyl-2-buty1)borane reduction of the intermediate lactone gave 2-deoxy-d-ribose(7).6,7 The synthetic material was identical with authentic 2-deoxy-d-ribose by the standard criteria of comparison. Alternatively, the lactone 6 was silylated to give the crystalline bis(tert-butyldimethylsilyl) derivative 8 (mp 76 °C). Reduction of this product with Dibal gave the disilyl derivative 9 in 92% yield [[α]24D = +23.6° (after 8 h, c 0.096, MeOH)]. The 300-MHz 1H NMR of 9 was identical with that obtained for the product generated by silylating authentic 2-deoxy-d-ribose and chromatographically separating out the disilyl derivative.

Additionally, it was observed that acetonitrile oxide reacted with 2 to deliver after N−O bond hydrogenolysis the erythro β-hydroxy ketone as the major product (1H NMR ratio 88:12).8 By reacting 2 with the nitrile oxide drived from the tetrahydropyranyl derivative of 2-nitroethanol4 and then effecting both cleavage of the THP group and hydrogenolysis of the isoxazoline by Raney nickel/A1Cl3,/MeOH/H2O treatment, we generated nearly a single dihydroxy ketone (>94% by HPLC analysis).9 Sodium periodate cleavage of this α-hydroxy ketone followed by diazomethane treatment yielded 4 as the major isomer. These studies thus reveal that the sense of the addition of a nitrile oxide to 2 is independent of the nature of the nitrile oxide employed.

One can rationalize the production of 2-deoxy-d-riboseas the major product of the above scheme through the following two factors: (a) cycloaddition occurs preferentially through a transition state resembling conformer A; 10 (b) addition of the nitrile oxide occurs anti to the C−O bond (the anti-periplanar effect).11 This latter factor is due presumably to the minimization of secondary antibonding orbital interactions as predicted on a theoretical basis by the work of Houk et al. The explanation for such stereoselection is, of course, related closely to that offered by Anh in support of the Felkin type transition state, i.e., the addition of a nuclephile to a carbonyl compound (bearing an α-asymmetric center) anti to the large group (the one having the lowest energy σ∗C2−X orbital).12

For comparison with current aldol technology, we list in Table I the erythro/threo ratios observed for the reactions of 1 with various carbon nucleophiles.

Table I.

Aldol Stereoselection Observed in the Addition of Various Carbon Nucleophiles to 1

graphic file with name nihms-1024672-t0003.jpg
Open in a new tab

graphic file with name nihms-1024672-f0002.jpg

In conclusion, we suggest that the anti-directing effect of an allylic oxygen should be very useful in cycloadding chiral or achiral nitrile oxides to chiral olefinic units so as to produce β-hydroxy carbonyl compounds (aldol fragments) in a stereopredictable fashion. A variety of molecular systems that can be constructed through applications of this stereochemical concept are now under study. 15

Acknowledgment.

We are indebted to the National Institutes of Health (HL-20579) and the Camille and Henry Dreyfus Foundation for support of this work. We are grateful to Professor Houk for helpful discussions.

References

  • (1).Kozikowski AP; Chen YY J. Org. Chem 1981, 46, 5248. [Google Scholar]; Kozikowski AP; Stein PD J. Am. Chem. soc 1982, 104, 4023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (2).Kozikowski AP; Chen YY Tetrahedron Left 1982, 23, 2081. [Google Scholar]
  • (3).Crawford RJ; Lutener SB; Cockcroft RD Can. J. Chem 1976, 54,3364. [Google Scholar]
  • (4).Kozikowski AP; Adamczyk MJ Org. Chem, in press. [Google Scholar]
  • (5). The 1H NMR of 4 has also been compared with the 1H NMR of the product formed from the reaction of the anion of ethyl acetate with 1, which is known to be an 85:15 mixture of diastereomers with 4 (ethyl ester) predominating. The ABX patterns of the α-methylene protons were identical.
  • (6).Murakami M; Mukaiyama T Chem. Left 1982, 241. [Google Scholar]
  • (7).Nakaminami G; Shioi S; Sugiyama Y; Isemura S; Shibuya M; Nakagawa M Bull. Chem. SOC. Jpn 1972, 45, 2624. [Google Scholar]
  • (8). The structure of the major isomer was verified by hydrolyzing 4 to the corresponding β-hydroxy acid and reacting this compound with excess methyllithium. The 1H NMR of the resulting 8-hydroxy ketone matched that displayed by the major isomer present in the hydrogenolysis mixture of the isoxazolines prepared from 2 and acetonitrile oxide.
  • (9).Kozikowski AP; Adamczyk M Tetrahedron Letf 1982, 23, 3123. [Google Scholar]
  • (10).Karabatsos GJ; Fenoglio DJ Top. Stereochem 1970, 5, 167. [Google Scholar]
  • (11).Caramella P; Rondan NG; Paddon-Row MN; Houk KNJ Am. Chem. SOC 1981, 103,2438 For a related observation in a Diels-Alder reaction, see: Franck, R. W.; John, T. V.; Olejniczak, K.; Blount, J. F. Ibid. 1982, 104, 1106. The dominance of the anti-bonding effect of an allylic oxygen substituent was first suggested by Franck in this communication. [Google Scholar]
  • (12).Anh NT Top. Curr. Chem 1980, 88, 145. [Google Scholar]
  • (13).Suzuki K; Yuki Y; Mukaiyama T Chem. Letr 1981, 1529. [Google Scholar]
  • (14).Heathcock CH; Young SD; Hagen JP; Pirrung MC; White CT; VanDerveer DJ Org. Chem 1980, 45, 3846. [Google Scholar]
  • (15). The reactions of nitrile oxides with 3-buten-2-ol and its derivatives are less selective than for 2, except in the case where the hydroxyl group is protected as its tert-butyldimethylsilyl ether derivative. These results constitute unpublished observations by several groups: Kozikowski, A. P.; Ghosh, A. K.; Houk, K. N.; Moses, S., at the University of Pittsburgh, and Jager, V. and co-workers at the University of Würzburg.

RESOURCES