their 1-cyano and Cl

Carbohydrate Research. 195 (1989) C1-C2

Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

Cl

Preliminary communication

Acetylated 1-cyano and l-cyano-2-hydroxy derivatives of D-galactal and D-arabinal

LASZLO SOMSAK

Department of Organic Chemistry, Lajos Kossuth Universirr, P.O.B. 20, H-4010 Debrecen (Hungary) (Received July 25th, 1989; accepted for publication, September 7th, 1989)

Glycals and 2-hydroxyglycals¹ are useful compounds because of their versatile reactivity, and 1-nitro-²,1-sulfonyl-³,1-formyl-, and 1-hydroxymethyl-glycals⁴ have been described.

In seeking to prepare 1-cyanoglycals, 2,3,4,6-tetra-O-acetyI-l-bromo-D- galactopyranosyl cyanide (1) and its d-arabino analogue³ (5) were each treated with zinc dust in acetic acid. The glycosyl cyanides 2 amd 6 and their C-I epimers, respec- tively, were isolated subsequently by crystallization from ethanol⁶, but the yields were poor. 'H-N.m.r spectroscopy of the reactuon mixtures revealed the presence of the expected I-cyanoglycals in addition to th«e glycosyl cyanides, and the ratios of the elimination products and the a and /3 anomers were 21:59:20 from 1 and 55:25:20 from 5. Chromatography of the components of these mixtures was not efficient because of the similar RF values.

1 R = B r 3 R = H

2 R = H ₄ R OAc

5 R = B r 7 R = H

6 R = H S R OAc

0008-6215/89/$03-50 © 1989 Elsevier Science Publishers B.V.

When solutions of 1 or 5 in benzene were boiled in the presence of excess of zinc and 1 mol of pyridine, the unsaturated compounds were formed rapidly (t.l.c.

showed that 1 or 5 had disappeared after 5 min). 4,5,7-Tri-0-acetyl-2,6-anhydro-3- deoxy-D-/yxo-hept-2-enononitrile (3, 76%) from .1 had m.p. 113-115" (from ethanol), [a]g° -52° (c 1.3, chloroform). 'H-N.m.r. data (C6D6): 54.87 (dd, 1 H, 7J4 1.75, y3>5 2.5 Hz, H-3), 5.20 (ddd, 1 H, /4 5 4.5, /4 6 1.5 Hz, H-4), 5.14 (ddd, 1 H, Js6 2.5 Hz, H-5), 3.56 (m, 1 H, H-6), 3.9L-4.00 (2 s, 2 H, H-7,7'), 1.55-1.63 (3 s, 9 H, 3 Ac).

Anal. Calc. for C13H15N07: C, 52.52; H, 5.08; N, 4.71. Found: C, 52.14; H, 5.16; N, 4.53.

4,5-Di-0-acetyl-2,6-anhydro-3-deoxy-D-eo'tf^|ro-hex-2-enononitrile (7, 75%) from 5 had m.p. 94-95° (from ethanol), [a]£° +176° (c 1.2, chloroform). 'H-N.mj-.^

data (C6D6): 5 4.98 (dd, 1 H, /3>4 4, J3J5 0.7 Hz, H-3), 5.15 (ddd, 1H, 745 4,74 6.1.5 Hz, H-4), 4.88 (m, 1 H, 756 6.75 Hz, H-5), 3.26 (ddd, 1 H, J56.2.75 Hz, H-6'), 3.53 (dd, 1 H, J66.11.75 Hz, H-6), 1.59,1.64 (2 s, 6 H, 2 Ac).

Anal. Calc. for C10H„NOS: C, 53.33; H, 4.92; N, 6.21. Found: C, 52.92; H, 4.81; N, 6.15.

Reaction of 1 or 5 with 1 mol of mercury(II) cyanide in refluxing nitro- methane in the presence of catalytic amounts of silver triflate was complete within 5 min (t.l.c.). 3.4.5,7-Tetra-0-acetyl-2,6-anhydro-D-/yxo-hept-2-enononitrile (4, 81%) from 1 had m.p. 99-101° (from ethanol), [cr]g⁰ -68° (c 1.2, chloroform)..

•H-N.m.r. data (CDC13): 8 5.96 (dd, 1 H, J4 S 4.6, /4 6 1 Hz, H-4), 5.53 (dd. 1 H, JS6 1.5 Hz, H-5). 4.55 (m, 1 H, Jbl 6.3, Je'r 6.3 Hz, H-6), 4.23-^1.37 (2 s, 2 H, H-7,7'), 2.22-2.40 (4 s, 12 H, 4 Ac)'.

Anal. Calc. for C15H,7N09: C, 50.70; H, 4.82; N, 3.94. Found: C. 50.08; H, 4.77; N, 3.91.

Column chromatography (benzene-ether-hexane, 6:3:1) on silica gel gave 3,4,5-tri-0-acetyl-2.6-anhydro-D-eryihro-hex-2-enononitrile (8, 82%), R? 0.39, [a]g° +162° (c 1.5, chloroform), from 5. 'H-N.m.r. data (CDC13): 8 5.82 (dd, 1 H, 7«4.5,/4.61.25 Hz, H-4), 5.32 (m, 1 H , /S 63 . 5 Hz, H-5), 4.21 (ddd, 1 H,766.11.25 Hz, H-6), 4.12 (dd, 1 H, /_S6. 8 Hz, H-6'), 2.12-2.22 (3 s, 9 H, 3 Ac).

Reactions of 1 on a 5-10-g scale were performed without difficulty.

The utility of the 1-cyanoglycals in synthesis is being investigated.

R E F E R E N C E S

1 R. J. FERRIER. in W. PIGMAN AND D . HORTON ( E d s . ) , The Carbohydrates: Chemistry and Bio- chemistry, 2nd edn.. Vol. IB, Academic Press, New York, 1980, pp. 843-879.

2 F . BAUMBERGER. D . BEER, M . CHRISTEN, R . PREWO. AND A . VASELLA, Helv. Chim. Acta. 69 (1986) 1191-1204.

3 R . J . FERRIER. R . H . FURNEAUX, AND P . C . T Y L E R , Carbohydr. Res., 58 (1977) 397-404; J . F . CASSIDY AND J . M . WILLIAMS, Tetrahedron Lett., 2 7 ( 1 9 8 6 ) 4 3 5 5 - 4 3 3 8 .

4 H . M . DEITINGER. G . KURZ. AND J . LEHMANN, Carbohydr. Res., 7 4 (1979) 301-307.

5 L . SOMSAK. G V . BATTA. AND I . FARKAS, Carbohydr. Res., 124 ( 1 9 8 3 ) 4 3 - 5 1 . 6 L. SOMSAK. Gv. BATTA. AND I. FARKAS. Tetrahedron Lett., 2 7 (1986) 5877-5880.

L. Somsak, I. Bajza, G. Batta 1265

Preparation of 2,6-Anhydro-3-deoxyhept-(or hex-)2-enononitri!es (1-Cyanoglycals) from 1-Bromo-D-glycosyI Cyanides with Zinc under Aprotic Conditions

László Somsak*, István Bajza, and Gyula Batta Department of Organic Chemistry, Lajos Kossuth University, P.O.B. 20, H-4010, Debrecen, Hungary

Received July 4, 1990

Key Words: 1-Bromo-D-glycosyl cyanides, reactions with zinc / 1-Cyanoglycals, aprotic synthesis of / Carbohydrates Acetylated 1-bromo-D-glycosyl cyanides 1 - 5 react with zinc

dust in acetic acid, acetic acid/water, or 2-propanol to give mixtures of acetylated 1-cyanoglycals 6—9 and anomeric pairs of glycosyl cyanides 10-14. Reaction of 1 - 5 in refluxing

benzene in the presence of one equivalent of triethylamine or especially pyridine predominantly leads to the formation of 1-cyanoglycals.

Glycals " [l,5-anhydro-2-deoxyhex-(or pent-)l-enitols] are among the most useful and variously transformable monosaccharide de- rivatives. Some of their 1-substituted analogues such as 1-nitro-", 1-sulfonyl-³', and l-formylglycals⁴¹ are also described. Since our efforts have been focused on the investigation of the reactivity of 1-bromo-D-glycosyl cyanides⁵¹ 1—5, which are closely related to acetobromo sugars (representing the classical starting material for glycal synthesis), and because 1-cyanoglycals with the exception of a 1-cyano-L-rhamnal" derivative have generally been unknown, we have tried to prepare this class of substituted glycals. We expected that the well-known procedure, i.e. the reaction of an acetylated glycosyl bromide with zinc dust in acetic acid ", which has also been applied to some structurally related 5-bromo-D-glucopyranuronate derivatives⁷¹, would be suitable for this purpose. However, from the reaction of 2,3,4,6-tetra-O-acetyl-l-bromo-D-galactopyranosyl cyanide (2) and its D-arabino analogue 5 under the above conditions the anomeric pairs of the corresponding acetylated glycosyl cya- nides llaß and 14aß, respectively, have been isolated by crystal- lization in poor yields". Trials to improve the yields especially for the 1,2-cis anomers 11a and 14 ß have not been successful, and a careful analysis of the 'H-NMR spectra of the reaction mixtures shows the presence of the expected 1-cyanoglycals and significant proportions of the anomeric glycosyl cyanides. This paper describes how the way to a satisfactory synthesis of 1-cyanoglycals" has been found.

Following the first observations", a systematic investigation of the reactions of 1-bromo-D-glycosyl cyanides 1—5 with zinc dust in protic media has been carried out (Table 1). Reactions in acetic acid at room temperature (entries 1, 5, 14, 17, and 22) give the elimination and substitution products in various ratios. In the case of the D-maiino compound 3 (entry 14), the trans-diaxial arrange- ment of the eliminating groups may explain the significantly higher proportion of the unsaturated compound 6. Raising the tempera- ture to the boiling point of acetic acid (entries 6 and 23) slightly increases the proportion of the elimination product. Carrying out the reactions at room temperature in a 1:1 mixture of acetic acid and water (entries 2, 7,15, 18, and 24) practically brings about no change.

In boiling 2-propanol (entries 8 and 25) the elimination is favored, however, the glycosyl cyanides are still present in the mixtures in considerable amounts. Since crystallization of the glycosyl cyanides

but not of the 1-cyanoglycals can be achieved in some cases, and chromatographic separation of the mixtures is also unsuccessful because of the very similar mobilities of the elimination products and the 1,2-cis-glycosyl cyanides, these reactions are not suitable to reach our goal.

Glycals are assumed to be formed via carbanionic inter- mediates"¹, but transfer of electrons from zinc might be a stepwise rather than a synchronous process. According to the concept of capto-dative substituent effects'⁰¹, we assume that a radical which may be generated during the formation of the carbanion at the anomeric centre of glycosyl cyanides would be more effectively sta- bilized than the ionic intermediate. Therefore, it has been anticipated that these species can be trapped by efficient hydrogen donors, and an approach to the synthesis of 1,2-cis-glycosyl cyanides, which we were also interested in", could be worked out. Thus, some reactions have been carried out (Table 1) with zinc dust in refluxing benzene in the presence of an excess of triethylamine (entries 12,20, and 27), which has been used as a hydrogen source'", and these experiments again result in the three known products. If no triethylamine is added (entry 9) no reaction occurs within 2 days, and similarly no transformation is observed with triethylamine in the absence of zinc.

With less than one equivalent of triethylamine (entry 10) a partial reaction takes place, i.e. after 2 days the starting material can still be detected together with the reaction products by TLC. By using one equivalent of this base we obtained rather uniform mixtures (entries 3, 11,16, 19, and 26) containing mainly 1-cyanoglycals and traces of glycosyl cyanides. When pyridine is used as the additive (entries 4, 13, 21, and 28), NMR spectroscopically almost pure 1- cyanoglycals are formed, from which 7 and 9 are isolated by crys- tallization, while 6 and 8 must be purified by column chromatog- raphy.

'H-NMR spectra of compounds 6—9 show resonances for the same number of protons as the starting materials but one less for the acetyl groups. In the ^l3C-NMR spectra resonances at 8 = 110—113 and 130— 133 characteristic of double bonds appear, from which the latter are assigned to C-2 on the basis of spin-echo ex- periments. Absorption bands in the 1640—1650 c m^-' region of the IR spectra also reveal the presence of double bonds. Their Im- positions have been proved by the appearance of the v(C = N) stretching vibration, because the cyano group in glycosyl cyanides is only Raman-active'", and 2,6-anhydro-3,4-dideoxyhept-3- Liebigs Ann. Chem. 1990, 1265-1268 ©VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990 0170- 2041/90/1212-1265 S 3.50+.25/0

1-4

R' R² R^J R⁴ R⁵ R⁶ R⁷

1, 6, 100 . 1 0 a H OAc OAc H H OAc CH2OAC 2. 7, 110 , 1 1 A H OAc OAc H OAc H CH2OAC 3. 1 2 0 , 1 2 a OAc H OAc H H OAc CH2OAC 4, 8. 130 . 1 3 a H OAc OAc H H OAc H

9. 1 4 0 . 14 a OAc H H OAc H OAc H

Table 1. Reactions of 1 -bromo-D-glycosyl cyanides 1—5 with zinc dust

Entry Solvent, Temp. Additive

[equiv.] Starting

compd. Products

(relative percentage)

i o p 10a

1 A c O H , r o o m temp. _ ₃₇ 23 4 0

2 A c 0 H / H20 , room temp. — 39 22 38

3 C6H6, reflux 1 Et3N 95 traces

4 Q H « , reflux 1 Py > 9 5 traces

2 7 + ^{l i p} + ^11a

5 A c O H , r o o m temp. _ 21 2 0 59

6 A c O H , reflux ^— 31 19 49

7 A C0H / H20 , room temp. — 32 2 6 51

8 i P r O H , reflux — 8 0 7 12

9 C6H6, reflux ^— no reaction

10 C6H6, reflux < 1 E t j N partial reaction

11 CIHJ, reflux 1 Et,N 95 traces

12 Q H f c reflux 30 Et,N 6 7 4 29

13 C6H6, reflux 1 Py > 9 5 traces

3 6 + ^{1 2 P} + ^12a

14 A c O H , r o o m temp. 91 traces 8

15 A C0H / H20 , room temp. ^— 93 6 traces

16 C6H6, reflux 1 Et.,N 95 traces

4 8 + ^13p + ^13a

1 7 A c O H , r o o m temp. 50 17 33

18 A c 0 H / H20 , room temp. ^— 50 13 37

19 Q H6, reflux 1 E t j N 95 traces

20 C6H6, reflux 10 E t j N 77 8 15

21 C6H6, reflux 1 Py > 9 5 traces

5 9 H 14a + ^{14 ß}

22 A c O H , r o o m temp. ^— 45 27 28

23 A c O H , reflux ^— 56 18 26

24 A c 0 H / H20 , room temp. ^- 50 17 33

25 i P r O H , reflux ^— 85 6 9

26 C6H6, reflux 1 E t j N 95 traces

27 C6H6, reflux 72 E t j N 74 14 12

28 C6H6, reflux 1 Py > 9 5 traces

Liebigs Ann. Chem. 1990, 1 2 6 5 - 1 2 6 8

Preparation of 2,6-Anhydro-3-deoxyhept-(or hex-)2-enononitriles enononitriles¹³' exhibit only very weak bands around 2200 cm^{- 1}. The presence of the double bond to the carbon bearing the cyano group counteracts quenching by the alkoxy substituent, thus mak- ing the cyano stretching vibration IR-active in compounds 6—9.

In accord with this compounds of the type PhCH = C(OR)CN^l"

exhibit v(C=N) bands in the region 2215 - 2240 cm^{- 1}. Preferred conformations of 6—9 have been determined by CD spectrosco- py^,!)-

The zinc/nitrogen base reaction of 1-bromo-D-glycosyl cyanides constitutes a new, simple, and efficient approach to 1-cyanoglycals.

The method can be extended to acetobromo sugars, and results of this research will be published elsewhere¹".

Experimental

Melting points were measured in open capillary tubes or on a Kofler apparatus and are uncorrected. — 'Optical rotations: Perkin- Elmer 241 polarimeter. — IR: Perkin-Elmer 283B spectrophoto- meter. - NMR: Bruker WP 200 SY spectrometer ('H: 200 MHz, ¹³C:

50.3 MHz); TMS as internal standard. - TLC: DC-Alurolle, Kie- selgel 60 F ^ (Merck). Benzene/ether/hexane (6:3:1) was used as eluent throughout this work. TLC plates were visualized by gentle heating. — Column chromatography: Kieselgel 60 (Reanal) and Kieselgel 40 (Merck). Anhydrous magnesium sulfate was used for drying the solutions, evaporations were carried out in vacuo below 50°C. Solvents were purified and dried as usual, benzene was stored over sodium wires.

Materials: Zinc dust was purchased from Reanal and used with- out any further purification or activation. 2,3,4,6-Tetra-O-acetyl-l- bromo-D-gluco-^s" (1) and -D-galactopyranosy! cyanide⁵" (2), 2,3,4- tri-0-acetyl-l-bromo-D-xylo-^sb) (4) and -D-arabinopyranosyl cyan- ide^sbl (5) were prepared according to literature methods.

(IR)-2J,4.6-Tetra-0-acetyl-l-bromo-D-mannopyranosyl Cyanide (3): 2,3,4,6-Tctra-O-acetyl-a-D-mannopyranosyl cyanide¹" (12a) (0.425 g, 1.19 mmol) was dissolved in absolute tetrachloromethane (15 ml) and barium carbonate (0.5 g) and then bromine (0.3 ml) were added. The reaction mixture was kept at room temp, in am- bient light for 3 d, then the solids were filtered ofT, and the solution was decolorized and neutralized by washing with 10% sodium hy- drogen sulfite and saturated sodium hydrogen carbonate solutions.

After drying and evaporation of the solvent the syrupy residue was purified by column chromatography with benzene/ether/hexane (6:2:2). Fraction A (R, = 0.39) contained 0.304 g (49%) of 3;

[otjg³ = +96 (c = 1.14, CHClj). - 'H NMR (CDClj): 8 = 5.74 (d, J_u = 3.2 Hz, 1H, 2-H), 5.68 (dd, J_ht = 10 Hz, 1H, 3-H), 5.34 (dd, J_iS = 10 Hz, 1 H, 4-H), 4.35 (dd, J_u- = 13.2 Hz, 1H, 6'-H), 4.17 (ddd, y_S6. = 1.5 Hz, 1H, 5-H), 4.17 (dd, J_s,₆ = 5.5 Hz, 1H, 6-H), 2.02,2.09, 2.15, 2.18 (4s, 12H, 4 x OAc). - ^l3C NMR (C₆D₆): 8 = 113.76 (CN, y,._H._CN = 0 Hz), 78.32 (C-l), 75.51, 72.84, 67.99, 64.14 (C-2-C-5), 60.52 (C-6), 169.68,169.08,168.92,168.56 (C = 0), 20.00, 19.92, 19.80, 19.46 (CH3).

C_{l s}H„BrNO, (436.2) Calcd. Br 18.32 N 3.21 Found Br 18.11 N 3.15 Fraction B (R, = 0.28) contained 0.165 g (38%) of starting ma- terial 12a.

General Procedure (I) for the Reactions of 1—5 with Zinc Dust in Protic Media: Compounds 1—5 (1 mmol) were each dissolved in the solvent (15 ml) given in Table 1, and zinc dust (0.3 g) was added. Crystalline copper(II) sulfate (0.03 g) and sodium acetate (1.0 g) were also added when an acetic acid/water (1:1) mixture was used. The reaction mixture was stirred at the given temperature (Table 1) and followed by TLC. After completion of the reaction

1267 the solids were filtered ofT, the filtrate was diluted with water, ex- tracted with chloroform, the organic phase was neutralized with saturated sodium hydrogen carbonate solution, then dried, and evaporated to dryness. The residue was investigated by 'H-NMR spectrometry (Table 1).

General Procedure (11) for the Reactions of 1—5 with Zinc Dust under Aprotic Conditions: Compounds 1—5 (1 mmol) were each dissolved in absolute benzene (15 ml), and zinc dust (0.3 g) was added. The mixture was stirred and heated to boiling temp. An additive (Table 1) was given to the refluxing suspension, and boiling was continued until the starting material disappeared (TLC). After cooling to room temp, the solids were Altered off, the nitrate was washed with saturated potassium hydrogen sulfate solution, dried, decolorized with charcoal, if necessary, and evaporated to dryness.

The residue was investigated by 'H-NMR spectrometry or crystal- lized from absolute ethanol to give 7 and 9, or purified by column chromatography to give 6 and 8.

4,5,7-Tri-0-acetyl-2.6-anhydro-3-deoxy-D-arabino-liept-2-enono- nitrile (6): Eluent benzene/ether/hexane (6:2:2), yield 0.184 g (62%) of syrup, R, = 0.41, [a]? = - 3 9 (c = 1.041 in CHClj). - IR (nujol): v = 2220 cm^{- 1} (C=N), 1650(C=C). - 'H NMR(CDCIj):

8 = 5.75 (d, = 4 Hz, 1H, 3-H), 5.38 (dd, Jt i = 5.5 Hz, 1H, 4- H), 5.23 (dd, Ju = 6 Hz, 1H, 5-H), 4.46 (ddd, Jt.v = 5 Hz, 1H, 6- H), 4.45 (dd, Jv = 5 Hz, 1H, 7-H), 4.20 (dd, •/,.,. = 15 Hz, 1H, 7'- H), 2.05, 2.08, 2.10 (3s, 9H, 3 x OAc). - ^,3C NMR (C6D6): 8 = 112.58 (CN), 130.44 (C-2), 112.10 (C-3), 75.60,65.63,65.48 (C-4-C- 6), 60.05 (C-7), 168.78, 169.27, 169.81 (C = 0), 20.14 (CHj).

C,iH„NOi (297.3) Calcd. N 4.71 Found N 4.61 4.5.7-Tri-0-acetyl-2,6-anhydro-3-deoxy-D-lyxo-hept-2-enononi- trile (7): Yield 0.223 g (75%), m.p. 113 — 115 =C (ethanol), [a]? = - 5 2 (c = 1.321 in CHClj). - IR (KBr): v = 2230 c m^-' (CsN), 1650 (C = C). - 'H NMR (C₆D₆): 8 = 4.87 (dd, J_ht = 2.5 Hz, Ja = 1.7 Hz, 1H, 3-H), 5.14 (ddd, J,< = 4.5 Hz, J_iA = 12 Hz, 1 H, 4-H), 5.20 (ddd, J,f = 1.5 Hz, 1 H, 5-H), 3.56 (m, J6.i = 6.2 Hz, Jt.r = 6.2 Hz, 1 H, 6-H), 3.96 (2H, 7-, 7 -H), 1.55, 1.58, 1.63 (3 s, 9 H, 3 x OAc). - ^l3C NMR (C6D6): 8 = 113.40 (CN), 130.62 (C-2), 113.40 (C-3), 75.24, 64.11, 62.38 (C-4-C-6), 61.49 (C-7), 169.28, 169.70, 169.78 (C = 0), 19.84, 20.02 (CHj).

C,₃H_ISNO, (297.3) Calcd. N 4.71 Found N 4.63 4,3-Di-0-aceiyl-2,6-anhydro-3-deoxy-D-threo-ltex-2-enononilrile (8): Eluent benzene/ether/hexane (6:2:2), yield 0.175 g (78%) or syrup, R, = 0.41,[a]? = -274(c = 1.375 in CHClj). - IR (nujol):

v = 2218 cm^{- 1} (C = N), 1640(C = C). - 'H NMR(C6De):8 = 5.35 (dd, /j.4 = 5.1 Hz, Jis = 1.4 Hz, 1H, 3-H), 4.82 (dddd, Jti = 2.6 Hz, JiM = 1.8 Hz, J,}. = 0.6 Hz, 1H, 4-H), 4.65 (dddd, JiM = 1.8 Hz, 1H, 5-H), 3.76 (ddd, JSM = 3 Hz, 1 H, 6e-H), 3.36 (ddd,

= 12.4 Hz, 1H, 6a-H), 1.55, 1.60 (2s, 6H, 2 x OAc). -

13C NMR (C6D6): 8 = 113.74 (CN), 133.05 (C-2), 110.73 (C-3), 66.02, 62.31 (C-4, C-5), 65.14 (C-6), 168.56, 168.99 (C=0), 20.01, 20.10

( C H j )' CioH||NOs (225.2) Calcd. N 6.21 Found N 6.12 4,5-Di-0-acetyl-2,6-anhydro-3-deoxy-o-erythro-hex-2-enononi- trile (9): Yield 0.169 g (75%), m.p. 94-95°C (ethanol), [«]£? =

+ 176 (c = 1.215 in CHClj). - IR (KBr): v = 2218 cm^{- 1} (C = N), 1650 (C = C). - 'H NMR (C₆D₆): 8 = 4.98 (dd, 7J.₄ = 4 Hz, J_ls = 0.7 Hz, 1H, 3-H), 5.15 (ddd, = 4 Hz, = 1.5 Hz, 1H, 4-H), 4.88 (dddd, J_SM = 6.7 Hz, 1 H, 5-H), 3.26 (ddd, J_SM = 2.7 Hz, 1H, 6e-H), 3.53 (ddd, J^u = 11.7 Hz, 1 H, 6a-H), 1.55, 1.64 (2s, 6H, 2 x OAc). - ^{I 3}C N M R (C6D6): 8 = 113.52 (CN), 132.07 (C-2), 111.80 (C-3), 63.81, 62.36 (C-4, C-5), 65.45 (C-6), 169.04, 169.11 (C = 0 ) , 19.96, 20.03 (CHJ).

C,0H„NOs (225.2) Calcd. N 6.21 Found N 6.15 Liebigs Ann. C h e m . 1990, 1 2 6 5 - 1 2 6 8

CAS Registry Numbers

1: 82469-74-7 / 2: 83497-42-1 / 3: 120199-66-8 / 4: 83497-43-2 / 5:

89158-09-8 / 6: 120085-62-3 / 7: 120085-63-4 / 8: 120085-64-5 / 9:

120085-65-6 / 10 (a anomer): 129849-45-2 / 10 (ß anomer): 52443- 05-7 / 11 (a anomer): 73014-31-0 / 11 (ß anomer): 52443-07-9 / 12 (a anomer): 84856-51-9 /12 (ß anomer): 86651-66-3 /13 (a anomer):

129849-46-3 / 13 (ß anomer): 52443-06-8 / 14 (a anomer): 82266- 99-7 / 14 (ß anomer): 89158-08-7

" ^1,1 R. J. Ferrier in Carbohydrates: Chemistry and Biochemistry (W. Pigman, D. Horton, Eds.). 2nd ed., vol. IB, p. 843, Academic Press, New York 1980. - "» R. J. Ferrier, Ado. Carbohydr.

Chem. Biochem. 24 (1969) 199. - ^lc| W. Roth, W. Pigman. Meth- ods Carbohydr. Chem. 2 (1963) 405.

21 F. Baumberger, D. Beer, M. Christen, R. Prewo, A. Vasella, Helu.

Chim. Acta 69 (1986) 1191.

31 R. J. Ferrier, R. H. Furneaux, P. C. Tyler, Carbohydr. Res. 58 (1977) 397; J. F. Cassidy, J. M. Williams, Tetrahedron Lett. 27 (1986) 4355.

41 H. M. Deitinger, G. Kurz, J. Lehmann, Carbohydr. Res. 74 (1979) 301.

F. W. Lichtenthaler, P. Jarglis, Angew. Chem. 94 (1982) 643;

Angew. Chem. Int. Ed. Engl. 21 (1982) 625. - ^5b| L. Somsák, Gy.

Batta, I. Farkas, Carbohydr. Res. 106 (1982) C4; ibid. 124 (1983) 43. T. Huynh-Dinh, C. Gouyette, J. Igolen, Tetrahedron Lett. 21 (1980) 4499.

R. Blattner, R. J. Ferrier, P. C. Tyler. J. Chem. Soc.. Perkin Trans, t. 1980, 1535. - "» T. Chiba, P. Sinay, Carbohydr. Res.

151 (1986) 379.

L. Somsák, Gy. Batta, I. Farkas. Tetrahedron Lett. 27 (1986) 5877.

L. Somsák, Carbohydr. Res. 195 (1989) CI.

H. G. Viehe, Z. Janousek, R. Merényi, L. Stella. Acc. Chem. Res.

18 (1985) 148; H. G. Viehe, R. Merényi, L. Stella. Z. Janousek.

Angew. Chem. 91 (1979) 982; Angew. Chem. Int. Ed. Engl. 18 (1979) 917.

S. G. Cohen, A. Parola, G. H. Parsons. Chem. Reo. 73 (1973) 141.

R. W. Myers, Y. C. Lee. Carbohydr. Res. 132 (1984) 61; A. T. Tu, W. K. Liddle, Y. C. Lee, R. W. Myers, ibid. 117 (1983) 291.

G. Grynkiewicz, J. N. BeMiller, Carbohydr. Res. 108 (1982) 229.

M. Cariou, G. Mabon, G. le Guillanton, J. Simonét, Tetrahedron 39(1983) 1551.

I. Farkas, G. Snatzke, L. Somsák, Liebigs Ann. Chem. 1989,599.

L. Somsák, I. Németh, unpublished results.

G. D. Kini, C. R. Petrie, W. J. Hennen, N. K. Dalley, B. E.

Wilson, R. K. Robins, Carbohydr. Res. 159 (1987) 81.

[142/90]

Liebigs Ann. Chem. 1990, 1 2 6 5 - 1 2 6 8

J. CARBOHYDRATE CHEMISTRY, 11(3), 201-216 (1992)

RADICAL-MEDIATED HALOGENATIONS OF ANOMERICALLY N-SUBSTITUTED GLUCOPYRANOSYL DERIVATIVES.

J.-P. Praly³, L. Somsak^b, S. H. Mahmoud^b, Z. El Kharraf³, G. Descotes³ and

I. Farkas

.

a. University Lyon I - URA CNRS 463 43, Boulevard du 11 Novembre 1918

69622 - Villeurbanne - Ranee b. University of Debrecen Department of Organic Chemistry H- 4010, P. O. B. 20, Debrecen - Hungary

Received April 2, 1991 - Final form December 31, 1991

ABSTRACT

Reactions of acetylated N-aryl-D-glucosylamines with N-bromosuccinimide- benzoyl peroxide or with sulfuryl chloride-azoisobutyronitrile gave aromatic halo derivatives. The corresponding N-acetylated compounds were mostly inert towards halogenation. Bromination of acetylated cellobiosylpiperidine resulted in the formation of acetobiomocellobiose, while the acetylated 2,6,8-trichIoro-9-(p-D-glucopyranosyl)-puriiie was transformed into its 5'-bromo derivative. In contrast, peracetylated glucopyranosyl isothiocyanate or azides, when treated by N-bromosuccinimide under free-radical conditions, essentially undergo an initial homolysis of the anomeric C-H bond which is faster for p-anomers. This initiates a new, simple and efficient free-radical transformation of such sugar azides into an unprecedented bromimino lactone (92% isolated yield).

INTRODUCTION

Radical-mediated brominations at ring positions of carbohydrates¹ have been known for about 15 years. Brominations of uronic acid derivatives, peracylated aldoses, anhydro sugar derivatives, glyculoses and glycosuloses, C-glycosyl heterocycles and

201 Copyright© 1992 by Marcel Dekker, Inc.

glycosyl cyanides, glycosyl halides, glycopyranoside esters, 1-thio-glycoside esters, glucopyranosyl phenyl sulfoxides and sulfones and several disaccharides were performed either with N-bromosuccinimide (NBS) or bromine (Br2> in refluxing carbon tetrachloride in the presence of chemical initiators such as benzoyl peroxide (BZ2O2) or azoisobutyronitrile (AIBN) and/or by photoinidation. Some chlorinadons were carried out with sulfuryl chloride (S02Cl2)-AIBN. Most reactions took place at C-l or C-5 of the pyranoid ring, and the halides contained an axial carbon-halogen bond.

Stabilizing effects of substituents of different character (c = acceptor: CN, CC^Me, COR, NO2, etc.; d = donor: OR, NR2, SR, etc.) at carbon radical centres and especially joint actions of pairs of substituents (££, dd, £d) attracted considerable attention recently.²*³ Application of these considerations to explain the outcomes of the carbohydrate brominations was also attempted.¹'⁴ It can be briefly concluded that £d substituted centres react readily to give high yields of the bromo derivatives. The effect of oxygen-oxygen dd pairs at the anomeric centre depends upon the substituent on the glycosidic oxygen atom. Methyl glycoside derivatives react at C-l, but C-S substitution is favoured for glycosyl esters and aromatic glycosides.

The very good stabilization of carbon radicals by amino substituents is well- known.⁵'⁶ Among the investigated carbohydrate derivatives with a nitrogen substituent at the anomeric centre, adenosine pentabenzoate is the only compound described to give a 4'-bromo product⁷ by photobromination. Therefore we have undertaken to brominate some glucosylamines and other N-substituted glucopyranosyl derivatives.

RESULTS AND DISCUSSION

The readily available⁸*⁹ acetylated N-aryl-D-glucosylamines (1-4) as well as their N-acetylated derivatives (5-8) (SCHEME 1) were brominated with NBS in refluxing carbon tetrachloride in the presence of catalytic benzoyl peroxide (method A).

The anilino compound 1 gave in a relatively short time of reaction two products from which only one has been isolated by crystallization (TABLE 1). In its mass spectrum m/e = 580,582,584 (M +1)-*- peaks could be observed with an intensity ratio -

1 : 2 : 1 indicating the presence of two bromine atoms. Since the proton spectrum contained resonances for each sugar proton + NH and in the aromatic region two doublets and one double doublet characteristic for a 1,2,4-trisubstitution pattern appeared, aromatic dibromination was concluded. Having in mind that bromination of different aniline derivatives with NBS^1Qi11 preferably leads to substitution in positions 2 and/or 4 of the aromatic ring, we prefer to suggest structure 9 for this compound. This is

R A D I C A L - M E D I A T E D H A L O G E N A T I O N S 2 0 3

SCHEME 1

1 - S , 14

R' 9 - 13

R²

2

R¹ R² H H

H CH3

H OCH3

H N02

R¹ R² 5 Ac H 6 Ac CH3

7 Ac OCH3

8 Ac N02

14 Ac CHnBr( 3^ n = 0,1,2

9

1 0 11 1 2

13 R¹ Br

CI

Br Br

CI R

Br

CI

CH3

OCH3

NO2

15 R¹= -N^ , R2 =H 16 R¹ = H , R² = Br

17 R = H 18 R = Br

also supported by the comparison with the proton chemical shifts for 2,4-dibromoaniline showing about 0.1-0.15 ppm difference for each proton. The proton spectrum of the mother liquor contained resonances in addition to those of 9 at 6.99 (d, 9 Hz, 1H), 7.32 (dd, 1H) and 7.60 (d, 2Hz, 1H) indicating the presence of an isomeric dibromo but not isolated product. Chlorination of 1 with SO2CI2/AIBN (method F) resulted in the analogous dichloro product 10, together with an unidentified isomer.

Bromination of 2 gave two products each of which had m/e = 515, 517 (M)⁺ peaks in their mass spectra indicating a monobromination. The large difference in the chemical

Starting

Compound Isolated compound

Method³ Reaction time(h) Yield

(%)

Mp C Q

[ a ]2 0 b Formula Analysis

Caled Found

1 9 A 1 42 189-191 -52 C2QH23BR2N09 Br:

N : 27.49 2.40

27.09 2.30

1 10 F 0.5 28 163-165 -5 C20H23CI2NO9 CI:

N :

14.43

2.85 14.51 2.71

2 l i p + l l a l i p

A 2 83

34 138-140 4 6 C21H26B1NO9 Br:

N :

15.50

2.71 15.40 2.51

11a 9 syrup +121 C21H26B1NO9 BR

N :

15.50 2.71

16.20 2.62

3 12 A 1 78 167-168 -50 C2lH26BrNOio B R :

N :

15.03 2.63

14.83 2.63

4 4 A 16 68 153-154

(155 ref.9)

4 1 3 F 0.5 30 155-156 -94 C20H23CIN2O11 C I :

N :

6.97 5.57

7.04 5.42

5 5 A 9 57 106-107

(100ref.9)

7 7 A 9 90 186-187 (185-186)^c

8 8 A 10 62 160-161

(161 ref.9)

15 16 A 1 70 181-182 +94

(183 réf. 16) (+95.8)

17 18 A 3 50 syrup -64 Ci9Hi8Cl3BrN409 Br 12.63 12.52

CI : 16.80 16.68

N : 8.85 8.61

19 20 E 6 46

21 E 6 25

19 20 D 10 80

19 20 B 8 25

2 2 B 8 60 see réf. 40

19 2 3 F 12 60 syrup -97

C15H18O9CINS

N : 3.30 3.45

24 2 5 F 5 55 synip -88

C14H18O9CIN3

N : 10.31 9.99

24 27 B 0.1 92 syrup See ref. 20

a. See Experimental; b. For chloroform solutions; c. Mixed mp with the starting material.

O«

Compound NH H-l H-2 H-3 H-4 H-5 H-6 H-6' Arom. OAc

0NH.H1) (JH1.H2) (JH2.H3) 0H3.H4) (JH4.H5) (JH5.H6') (JH5.H6) (JH6.H6')

9 5.41 4.68 5.16^a 5.39^a 5.1 l^a 3.84 4.31 4.12 6.61 2.05

8.1 8.7 9.2 9.2 9.7 2.2 5.4 11.9 (d, 8.6Hz, 1H)

7.24 2.08

(dd. 1H) (12H)

7.53 (d, 2.1 Hz, 1H)

5.42 4.71 5.15^a 5.40^a 5.1 l^a 3.86 4.33 4.13 6.68 2.06

8 9 9.5 9.5 10 2.5 5.5 12.5 (d, 8.5Hz, 1H)

7.13 (dd, 1H)

7.30

(6H) 2.08 2.09 (d. 2.5Hz, 1H)

IIP 5.22 4.69 5.16^a 5.38^a 5.11^a 3.85 4.31 4.12 6.64 2.05

(d. 8Hz, 1H) 2.07

8.5 8.5 9 9.5 9.5 2.5 5 12 6.98 2.10

(dd, 1H) (12 H)

7.26 (d, 2Hz, 1H)

CH3: 2.25

TJ 7>

11a 4.99 5.31 5.18³ 5.51³ 5.12^s 4.17 4.32 4.04 7.02 2.07

2.5 5 10 10 10 2 4.5 12 (s,2H) 2.10 H

7.31 2.11 >

(s, 1H) 2.13

CH3: 2.27

12 5.02 4.65 5.16^a 5.38^a 5.11^a 3.84 4.31 4.12 6.71 2.05

8.5 8.5 8.5 9.5 9.5 2.5 5.5 12.5 ((1.9H.1H) 2.06

6.80 2.08

(dd, 1H) 2.09

7.05 (d,3Hz, 1H)

OCH3: 3.75

6.77 2.10

(d, 9Hz, 1H) 2.22

8.09 (12H)

(dd, 1H) 8.24 (d. 2.5Hz, 1H)

18 6.42 5.75 5.52 4.56 4.45 1.89

2H,brs 9.5 10 12 2.07

2.16(6«

23 5.58 5.22 5.43 5.41 4.45 4.22 2.0

9.1 9.2 9.9 12.2 2.2

25 5.21 5.03 5.51 5.39 4.45 4.26 1.99

9.1 9.3 9.7 12.2 2.07

2.09 2.12

13 6.17

8

4.80

9 5.17^a

9

5.43^a 10

5.11^a

10 3.90

2.5 4.32

5.5 4.13

12.5

a. Tentative assignments.

Compound C-l C-2 C-3 C-4 C-5 C-6 Others

18 74.75^a 70.84^a 67.84^a 67.84^a 97.35 64.91 153.61 169.42

153.21 Purin 168.74 C = 0 151.00

143.25 20.32

19.96 CH3

2 2 157.2 139.6 126.2 64.2^a 77.9^a 62.0 170.4 ,169.7 .168.2 : C = O

20.7 , 20.6,20.3 : CH3 t

2 3 80.9 69. l^a 69.8^a 71.3^a 99.0 64.8 169.8,169.7,169.1 : C = 0

20.6,20.5 : CH3 145.4 : NCS

2 5 86.9 68.3^a 69.9° 69.9^a 99.6 64.9 169.7,169.2,169.1,167.6 : C = O

20.6 , 20.5 . 20.4 : CH3

a. Tentative assignments.

RADICAL-MEDIATED HALOGENATIONS 209

shifts for H-l of these compounds as well as the different values of the 1,2-couplings together with a downfield shift of H-3 and H-5 suggest the formation of the epimeric monobrominated compounds l i p and 11a. It was not investigated whether this anomerization was due to any radical attack at C-l. Nevertheless, the known anomerization of per-0-acetylated N-aryl-glycosylamines¹² catalyzed by acids allows the assumption that the hydrogen bromide present at least in traces in the bromination mixtures can be responsible for this process. A somewhat similar epimerization was observed during radical-mediated bromination of an acetylated disaccharide.¹³

Bromination of 3 gave the aromatic bromo derivative 12 and the deactivated 4 was inert towards bromination, while its chlorination gave die aromatic chloro derivative 13.

N-Acetylation as in compounds 5-8 leads to further deactivation of the aromatic rings so that no bromination occurred for rather long times and compounds 5,7 and 8 could be recovered by crystallization in reasonable yields (TABLE 1). From the bromination mixture of 6, a crystalline substance was isolated. However, its proton spectrum, similar to that of 6, exhibited resonances for each sugar and aromatic protons, but the intensity of the CH3 signal (2.38) decreased and sharp singlets at 4.50 (CH2Br) and 6.68 (CHBr2) appeared indicating the presence of the corresponding compounds in a ratio 1 : 7 : 2 . Elemental analysis of this substance (calcd for C23H27Br2NOio: Br : 25.08, found 25.44) suggests that even tribrominated products can be found in the mixture.

Bromination of hepta-0-acetyl-D-cellobiosylpiperidine¹⁴ (15) gave clearly one product which according to its physical constants and spectral data proved to be the corresponding acetobromo disaccharide 16.¹⁶ Bromination of the purine gluco- pyranoside derivative 17¹⁵ took place at C-5 to give 18 as could be expected from a similar reaction of a benzoylated adenosine derivative,⁷ and an unselective transformation was observed in the case of 2', 3', 5'-tri-0-benzoyl-uridine.

In summary, radical halogenation of N-aryl-glucopyranosylamines mainly occurs at the aromatic position of the aglycone without any reaction at the anomeric carbon of the carbohydrate moiety. a-Bromination was only observed in an example of one cellobiosyl derivative by an undetermined mechanism of substitution.

Photobromination of 2,3,4,6-tetra-0-acetyl-|)-D-glucopyranosyl isothiocyanate 19^{1 7} (SCHEME 2) has been studied in the presence of either bromine or N- bromosuccinimide. In the first case, 2,3,4,6-tetra-0-acetyl-a-D-glucopyranosyl bromide 20 (46 %) and 2,3,4,6-tetra-0-acetyI-D-glucopyranose 21 (25 %) were isolated after a 6 h irradiation with a medium pressure mercury lamp (method E). A comparable transformation occurred by irradiating the reaction vessel with a sun lamp (method D) yielding essentially the a-bromide 20 (80 %). Compound 20 was also produced (25 %)

SCHEME 2

-OAc

AcO I

2 0 ^{B r}

method B D E Z = p-NCS

25%

8 0 %

4 6 % 25%

O

2 2 60%

OAc

AcO AcO •z

z p- NCS P- N3

a - N3

Z=N3

method A or B

^OAc

A c O ^ N B r

27 92%

19 24 26

Z=p-NCS, N3

method F

^OAc

AcO

Z %

23 p-NCS 60

25 p-N3 55

^OAc

ACO X^^A —N

2 8 OCH3

after prolonged treatment (8 h) of 19 in the presence of NBS (method B), along with the unsaturated lactone 22 (60 %). However, boiling a carbon tetrachloride solution of 19 in the presence of sulfuryl chloride (method F) and AIBN led to the C-5 chlorinated¹⁸ compound 23 (60 %).

RADICAL-MEDIATED HALOGENATIONS ^{2 1 1}

Similar conditions (SO2CI2, AIBN, A) also brought about the transformation of 2,3,4,6-tetra-O-acetyl-p-D-glucopyranosyl azide 24¹⁹ into the C-5 chlorinated derivative 25 (55 %) within 3-4 h. In contrast, use of NBS in boiling carbon tetrachloride in the presence of a catalytic amount of benzoyl peroxide (method A) resulted in the very fast transformation (4-6 mn) of the starting p-azide 24. The presence of a single new spot, slightly less mobile than 24 and strongly UV light absorbing on a TLC plate, indicated a clean reaction which also occurred when the mixture was irradiated with a 250 W sun lamp (method B), in the absence of benzoyl peroxide. Though the product decomposed during attempted chromatographic purifications, it has been obtained in a pure form (92

% isolated yield) using the photolytic procedure (method B) and identified as being the new peracetylated bromimino lactone 27.²⁰

Use of molecular bromine instead of NBS does not provide alternative conditions for the preparation of 27 from 24. However, TLC monitoring showed that the afore- mentioned transformation occurred within 2-3 h with the less reactive a-anomer 26.²¹ No reaction was observed either for compound 28^{2 2} when submitted to the same treatment or for 24 when heated in the absence of radical initiators (light or benzoyl peroxide). Recent synthetic results also demonstrate that NBS itself does not alter azide groups.²³

The observed reactivities of the anomeric azides (p > a) suggest a free-radical abstraction of the anomeric hydrogen atom, as the initial step of the reaction, in connection with the easier homolysis of axially oriented acetalic C-H bonds, as compared to the equatorial ones.²⁴ This proposal is also in keeping with the stability of 28. The suggested initial homolysis of the anomeric C-H bond which calls to mind the observed hydrogen abstraction from the a-methylene of n-butyl azide by t-butoxyl radicals²⁵ should benefit from the specific features of the anomeric centre³'²⁶ and from the acetoxy group at C-2.²⁷ Whereas photolysis of the azide group proceeds through other ways,²⁸»²⁹ carbon-centered radicals bearing an azido substituent are involved in a few synthetic transformations.³⁰³ However, in the present case, decomposition of the initial radical into a lactone iminyl radical with release of molecular nitrogen should ensue, following a behaviour also encountered for alkyl azides.²⁵*³¹ The radical chain should be continued by a final bromine abstraction from either NBS or the in situ generated bromine.³² Hence, despite the absence of related literature data³⁴ and of a unique pattern for the free radical reactivity of organic azides, route A (SCHEME 3) best represents the probable reaction pathway.

The possible bromination of the initial azido alkoxy radical (route B) constitutes a less probable alternative. The variety of resonance forms of such a radical should result in a lowered reactivity.²⁶ In addition, the eventual bromoazido intermediate should be, in

SCHEME 3 24 or 26

R-

^ O A c

A c o X ^ X - N ^ + RH AcO \

route A route B

< T

light of theoretical data,³⁵ a short-lived species, prone to a C-Br bond heterolysis. Since derivative 28 failed to react when exposed to the same reaction conditions, a plausible rationale for the formation of 27 via a bromoazido intermediate is not straightforward.

In contrast to the azido group, the influence of an isothiocyanate substituent towards hydrogen abstraction is better appreciated through a variety of experimental studies.^{3 0 6}-^{3 6}-^{3 7} An initial homolysis of the axial C-H bond most probably initiates successive transformations of the p-isothiocyanate 19. Not surprisingly, the corresponding C-1 a-bromo adduct (SCHEME 4) remained elusive, since the stability of a-bromo isocyanate derivatives is enhanced by electron withdrawing substituents.³⁶ Then, a dehydrobromination step should lead to a 1,2-unsaturated compound prone to Ferrier's reaction while the final conversion to the unsaturated lactone 22 could arise from hydrolysis. Such a reaction pathway is reminiscent of the mechanism proposed for