1,2-Orthoacetate intermediates in silver

1,2-Orthoacetate intermediates in silver trifluoromethanesulphonate promoted KoenigsKnorr synthesis of disaccharide glycosides' JOSEPHBANOUB'A N D DAVIDR. B U N D L E

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 74.121.191.133 on 06/06/13 For personal use only.

Dieision of Biologict~lScietlcc.r, Ntrtiot~c~l Resetr1.c.h Co~rtzcilc?f'Crrt~rrrirr, Ottir~t,rr,Otzt., Cntzrrcltr KIA OR6 Received December 2 1 , 1978 JOSEPHBANOUB and DAVIDR. BUNDLE.Can. J. Chem. 57,2091 ( 1979). The three disaccharides lactose, cellobiose, and maltose in the form of their acetylated glycosy1 bromides have been reacted with 8-ethoxycarbonyloctanol to provide the 1,2-trans-glycosides. Conventional Koenigs-Knorr and Helferich conditions provided these glycosides in poor yield but silver trifluoromethanesulphonate, N,N-tetramethylurea gave the disaccharide glycosides in 50-60% yield. Use of 2,4,6-trimethylpyridine as proton acceptor provided the corresponding 1,2-orthoacetates in 60-70% yield. These 1,7--orthoesters were rearranged by stannic tetrachloride to the 1,2-tratzs-glycosides. The isolation of acetylated 8-ethoxycarbonyloctanol from Lewis acid catalysed isomerisation of 1,2-orthoesters and from Koenigs-Knorr reactions in which N,N-tetramethylurea was the proton acceptor is discussed in terms of a reaction mechanism proceeding from glycosyl halide to glycosidic products via a 1,2-orthoester intermediate. This proposal is supported by 'H nmr evidence for the presence of 1,2-orthoacetate in Koenigs-Knorr reaction mixtures, and by 1,2-orthoacetate isomerisation to glycoside by the conjugate acid of N,N-tetramethylurea. JOSEPHBANOUB et DAVIDR. BUNDLE. Can. J. Chem. 57.2091 (1979). Les trois disaccharides, lactose, cellobiose et maltose sous forme de bromure de glycosyles peracetyles ont ete traitts avec le ethoxycarbonyl-8-octanol pour produire les glycosides tratzs-1,2. Les rtactions conventionelles type Koenigs-Knorr et conditions d'Helferich ont produit ces glycosides, dans des pauvres rendements, mais la glycosidation en presence du trifluoromethanesulphonate d'argent et la tetramethyluret-N,N conduit aux disaccharides glycosides dans des rendements de 50-60%. L'utilisation de la trimethylpyridine-2,4,6 comme accepteur de proton conduit aux orthoesters-1,2 correspondants dans des rendements de 60-70%. Ces orthoesters-1,2 sont rearrangis par le tetrachlorure d'etain pour former les glycosides tratls-1,2. L'isolation de I'acetate du ethoxycarbonyl-8-octanol obtenu A partir de l'isomerisation catalysee par l'acide de Lewis des orthoesters-1,2 et des reactions de Koenigs-Knorr dans lequelles la tetramethyluree-N,N est l'accepteur de proton, est discutt en terme de mechanisme de reactions procedant A partir des halogenures de glycosyles pour former les glycosides trans1,2 via les orthoesters-1,2 comme intermediaires. Ceci est verifit par le mise en evidence par rmn de ' H de orthoesters-1,2 dans les melanges obtenus par les rkactions de Koenigs-Knorr et par I'isomerisation de orthoacetates-1,2 en glycosides par l'acide conjugte de la tetramethylurte-N, N.

promotor of glycosylation reactions between alcoIntroduction -hols and glycosyl halides. Consequently it was deDisaccharide glycosides were required in order to investigate serological cross reactions between cided to prepare the disaccharide glycosides in quespathogenic bacteria and antibody prepared against tion by this method. In the course of this work it was artificial carbohydrate antigens (1, 2). Previous at- found that 1,2-orthoacetates are formed exclusively tempts to prepare the 1,2-trans-linked glycoside of when the acceptor is 2,4,6-trimethylpyridine (collilactose and the alcohol, 8-ethoxycarbonyloctanol dine), whereas 1,2-trans-glycosides are the major pro(the alcohol used to prepare artificial antigens, via duct if the proton acceptor is N,N-tetramethylurea. 'H covalent attachment to proteins (3, 4)) by standard nmr and chemical evidence and a consideration of the Koenigs-Knorr or Helfrich conditions resulted in products of the latter reaction suggest that 1,2very poor yields (2). Following its initial use by orthoacetates are intermediates in Koenigs-Knorr Kronzer and Schuerch (5) it has been demonstrated reactions of acetylated glycosyl halides. independently in two laboratories (6,7) that silver Results trifluoromethanesulphonate (triflate), in conjunction G l ~ c o s ~ l a t i o of n 8-ethox~carbon~10ctano1 in diwith an appropriate proton acceptor, is an excellent chloromethane with either acetobromolactose (2) or acetobromocellobiose (1) (8,9) was accomplished 'NRCC No. 17339. in 60-65x yield by silver triflate, N,N-tetramethyl'MDS Health Group Limited, Rexdale, Ontario. 0008-4042/79/16209I-07$0 I ,0010 @ 1979 National Research Council of CanadalConseil national de recherches du Canada

CAN. J. CHEM. VOL. 57, 1979

R~&A=O~ n

AcO


+

OAc O

AcO

disaccharide heptaacetates

3 R1 = H , R Z = OAc 4 R L =0 A c , R 2 = H

+ ROH OAc

+

\B

\

1 R1= H, R Z = OAc 2 R1= 0Ac,R2= H

O

AcO

+ AcOR +

+ ROH +

disaccharide

rnal tose heptaacetate

ROH

A

c

O

~

o

~

o

,+ AcOR R + acetates rnal tose

AcO

Reaction condition A : silver trifluorornethanesulphonate, collidine in CH2C12 Reaction condition B: silver trifluoromethanesulphonate, N , N-tetrarnethylurea in CHzCIz SCHEME 1. Synthesis of cellobiose, lactose, and maltose 1,2-orthoacetates and 8-ethoxycarbonylocty1 glycosides.

urea promoted Koenigs-Knorr reaction (6). Acetobromomaltose (7) (8, 9) reacted with 8-ethoxycarbonyloctanol under these conditions to provide glycoside (9) in 58% yield. However, when the bron~o sugars 1, 2, and 7 were reacted with 8-ethoxycarbonyloctanol in the presence of molar quantities of silver triflate and collidine, 60-80% yields of the corresponding 1,2-orthoacetates (3, 4, and 8) resulted (Scheme 1). Under these latter conditions, which do not use the solvent quantities of hindered pyridines used in standard literature (10, 1 I) methods of orthoester synthesis, the only side products were unreacted alcohol and a reducing disaccharide heptaacetate. Reaction with silver triflate and N,N-tetramethylurea (rather than collidine) produced acetylated 8-ethoxycarbonyloctanol, glycoside hexaacetate(s), and reducing disaccharide hexaacetate in addition to the glycoside heptaacetates 5, 6, and 9. The glycoside hexaacetate(s) exhibited tlc mobilities slightly lower than the heptaacetyl glycosides 5, 6, and 9, -and in-

tegration of 'H nmr spectra indicated the presence of six acetate groups. The disaccharide hexaacetate contained no aglycon and from 'H nmr spectra clearly contained no acetate at C-1. Although the product(s) have been assigned the 3,6,2'3'4'6'-hexa0-acetylglycosyl-glucose structure(s) the 13C nmr evidence for this must be considered tentative. Related products were obtained from the glycosylation reactions conducted with acetobro~nocellobiose(2) . and acetobromolactose (1). Similar products to those mentioned above were observed when the 1.2-orthoacetates 3, 4. and 8 were isomerised by acid. This isomerisation was also performed with stannic tetrachloride (12), which unlike the more general conditions (13) performed at elevated temperatures with Lewis acid or proton catalysis, proceeds at low temperature without additional quantities of alcohol (14) (Scheme 2). Stannic tetra;hloride isomerisation of 1,2-orthoacetates 3, 4, and 8 to the glycosides 5, 6, and 9 gave acetylated ,

,

,

2093

BANOUB A N D BUNDLE: I 1

R


OAc 3 R 1 = H , R Z = OAc 4 R1 = OAc, R2= H

5, CHzCIZ Ac R 2

' OAcO AcO & O ~ o / ' OAc OAc 5 R1 = H, RZ= OAc 6 R' = OAc, RZ= H

+

acetates disaccharide

+

, ,

+ AcOR +

maltose acetates

ACO-"~\

AcO

8 Suggested mechanism

SCHEME 2. Lewis acid catalysed rearrangement of a-D-disaccharide 1,2-orthoacetate to 8-ethoxycarbonyloctyl-p-Dglycosides.

8-ethoxycarbonyloctanol, glycoside hexaacetates, and reducing disaccharide hexaacetates. This product distribution is also reported for the single-step synthesis of the 1,2-tratzs-glycosides5,6,and 9 from the corresponding1,2-trans-disaccharide octaacetates (I S), a process known to occur via a 1,2-acetoxonium intermediate (16). The presence in Koenigs-Knorr reaction mixtures of acetylated aglycon is consistent with 1,2-orthoacetate intermediates during such reactions and therefore attempts were made to isomerise orthoacetates under these conditions. Glycosylations conducted with silver triflate were observed both in this and other work (17) to give a precipitate of silver halide at temperatures between -40 and -30°C. Since the net result of halide abstraction and attack of alcohol on an acetoxonium ion is formation of the conjugate acid of either N,N-tetramethylurea or collidine, it was of interest to examine the effect of these acids on 1,2-orthoacetates at low temperatures. A crystalline orthoacetate available from other work (17), 3,4-di-0-benzyl-P-L-rhamnose 1,2-(methyl orthoacetate) was particularly suitable for study by 'H nmr, since isomerisation to glycoside would provide new signals uncomplicated by other acetate or ethoxy signals. This would not have been the case for ortho-

acetates 3, 4, or 8. Isomerisation with N,N-tetramethylurea - triflic acid (2: 1) did not occur at -70°C and only very slowly at -30°C. At - 10°C a molar or 0.1 molar equivalent of this acid-base mixture gave a near quantitative yield of methyl 2-0-acetyl3,4-di-0-benzyl-a-L-rhamnopyranoside from the 1,2orthoacetate in 3 h. At 20°C the same isomerisation was complete after 5 min. No isomerisation of 1,2orthoacetates was observed after several hours when a molar equivalent of triflic acid - collidine 1.02: 1.0 was added to the rhamnose 1,2-orthoacetate or the cellobiose 1,2-orthoacetate (3). When a Koenigs-Knorr reaction between tri-0acetyl rhamnopyranosyl bromide and S-ethoxycarbonyloctanol in the presence of silver triflate and N,N-tetramethylurea was quenched at -20°C, it was possible to observe by 'H nmr of the crude mixture the presence of glycoside and 1 ,Zorthoacetate in approximately equimolar proportions. The complex 'H nmr spectrum was simplified to that of S-ethoxycarbonyloctyl-tri-0-acetyl-a-L-rhamnoside by addition of a catalytic amount of N,N-tetramethylurea triflic acid to the contents of the nmr tube. A similar result was obtained when the 1,2-cis-glycosyl halide, tetra-0-acetyl-a-D-glucopyranosyl bromide was reacted in the same manner.


C A N . J. CHEM. VOL. 57. 1979

+ ROAc

& -O R OR

OAc

OAc

SCHEME 3. Possible mechanism for 1,2-orthoacetate intermediates leading to glycosidic products in Koenigs-Knorr reactions.

Discussion We have demonstrated that whereas silver triflate, N,N-tetramethylurea promotes reaction between acetobromo sugars and an alcohol to yield 1,Ztransglycosides (route B, Scheme I), silver triflate, collidine yields the isomeric 1,2-orthoacetates (route A , Scheme 1). Furthermore, the conjugate acid of the proton acceptor, which results when alcohol reacts with intermediates formed from acetobromo sugars, is capable in the case of N,N-tetramethylurea, but not collidine, of catalysing the rearrangement of 1,2-orthoacetate to 1,2-trans-glycoside. It is generally accepted (18) that glycosyl halides

possessing a participating group at C-2 react under Koenigs-Knorr conditions to yield initially a 1,2acyloxonium intermediate. The essential elements of these processes have been succinctly summarised by Lemieux (19) and are abbreviated in Scheme 3. The glycosyl halide under the driving force of ring oxygen participation forms an oxocarboniuil~-halide ionpair 11. Participation from an acetate at C-2 leads to an acetoxonium ion 12. The mechanistic interpretation of the steps leading from 12 or 11 to 1,Zorthoester or glycosidic products are numerous (18, 20, and references cited therein), as are the empirical modifications used to achieve the transformation of


B A N O U B A N D B U N D L E : I1

glycosyl halide (10) to glycoside (14). The results presented in this paper illustrate important aspects of these mechanisms and are interpreted on the basis of a 1,2-orthoester intermediate. Attack of alcohol on the ainbidentate cation 12 may occur at either the dioxolenium carbon atom, leading to 1,2-orthoacetates, or at the anomeric centre, leading to 1,2-trans-glycosides. Since the nature of the proton acceptors, N,N-tetramethylurea or collidine, seems unlikely to influence the site of attack when present directly in only molar ratios, the different products which result in each case are attributed to the ability of protonated N,N-tetramethylurea to isomerise the initially formed 1,2-orthoacetate. Under kinetic control (- 30°C) 1,2-orthoacetates were observed, and isolated, and shown to isomerise when exposed to catalytic amounts of protonated N,N-tetramethylurea. Protonated collidine was demonstrated to be incapable of catalysing the same isomerisation even when present in molar proportion. The observation of acetylated alcohol (8-ethoxycarbonyloctanol), a frequently observed side product of Koenigs-Knorr reactions (2, 21, 22), is most convincingly explained via a I ,2-orthoacetate intermediate (cf. ref. 18). We propose, therefore, that acetylated glycosyl halides react under Koenigs-Knorr conditions via 1,2-orthoacetate intermediates, which isomerise under proton or Lewis acid (e.g., in the case of mercuric cyanide promoted glycosylations) catalysis to yield the thermodynamic product, 1,2-trans-glycosides. A schematic representation of this argument is presented (Scheme 3) which is based on recent mechanistic proposals of Garegg and Kvanstrom (14, 23) for proton catalysed I ,2-orthoester isomerisation. In brief, the essentials of the three pathways a, b, and c are as follows. Protonation via pathway a leads by way of an acetoxium ion 12 to glycoside 14. Pathway b may result after intramolecular rearrangement of 15 in ci-glycosides, or alternatively via 17 with loss of acetylated alcohol and either ci- or P-glycoside (19) formation. The glycosyl cation (18) which results from pathway c ultiillately yields a similar product, the 2-hydroxy ci- or P-glycoside (19). In the absence of sufficient alcohol, 18 may presumably react to give products during work-up such as the disaccharide hexaacetates observed in these studies (these products appear to be related to products obtained from 1,2-orthoesters in acidic methanol (24)). In polar solvents or at higher temperature (20°C), the direct reaction of 11 or 12 to glycoside cannot be excluded. However, when the Koenigs-Knorr reaction is performed at low temperature (-20-0°C) in weakly polar solvents, the products of kinetic control, 1,2orthoesters, are formed initially and subsequently

2095

isomerise or 'break down' to 1,2-trans-glycosides and side products. The empirical modifications of the Koenigs-Knorr reaction are many and varied (see refs. 18, 20) and several factors affect the stereochelnical outcome of the reaction. Consequently an all embracing mechanism which satisfactorily explains all aspects of these reactions is still lacking. However, it has been suggested (25) in the more recent literature that 1,2orthoester intermediates enjoy at least transient existence in such reactions, and the results reported here support this prediction. Irrespective of mechanistic aspects, silver triflate - N,N-tetramethylurea provides an excellent route to 1,2-trans-linked glycosides especially for alcohols which have been shown to be unreactive in the more conventional Koenigs-Knorr or Helfrich conditions. The use of silver triflate collidine gives excellent yields of the 1,2-orthoacetates of disaccharides, which are sometimes difficult to prepare by more conventional routes (10, 11). Experimental Thin-layer chroniatography was performed with Merck precoated silica gel 60 F-254 plates, and the detection of cornpounds was achieved by quenching of uv fluorescence and by charring after spraying with 5% sulphuric acid in ethanol. Silica gel G60 (70-230 mesh) and redistilled solvents were used for column chromatography. The loading on all columns was 1 :50 unless otherwise indicated. Skellysolve B refers t o hexane supplied by Getty Refining and Marketing Cornpany, Tulsa, OK. Solvents were purified and dried according to standard procedures (26). Processed solutions were dried over anhydrous sodium sulphate and solvent was removed at bath temperatures 40°C or lower unless otherwise stated. Melting points were determined on a Fisher-Johns apparatus and are uncorrected. Optical rotations were measured at 589 nm in a 1 dm cell at rooni temperature (20-23°C). Carbon-13 and 'H nnir spectra were recorded at 20 and 79.9 MHz respectively in the pulsed Fourier transform mode on a Varian CFT-20 spectrometer. Proton chemical shifts are expressed relative to 1% tetramethylsilane (TMS) in deuteriochloroforrn. Carbon-13 shifts are expressed relative to internal TMS in deuteriochloroform. 3,6-Di-0-acetyl-4-0- ( t e t r a - O - n c e t y l - ~ - ~ - g / r t c o p ~ ~ r n-~ ~ o s y l ) a-D-ghtcopyrar~ose1,2-(8-Ethoxycnrbor1)~/oc1y/ Orthoacetate) (3) A solution of 8-ethoxycarbonyloctanol (1.61 g, 8 rnmol) in 15 mL of dry dichloromethane containing silver triflate (2.57 g, 10 mmol) and collidine (1.21 g, 10 mmol) was cooled t o - 40°C. Hepta-0-acetyl a-D-cellobiosyl bromide (1) (6.4 g, 9.15 mniol) (8) in 30 mL of dichloromethane was added dropwise with stirring. After 30 min at -40°C the reaction temperature was allowed to reach 20°C and the reaction was left for 3 h at rooni temperature. The reaction mixture was filtered through charcoal and Celite and the filtrate was washed with 3% hydrochloric acid, saturated sodium bicarbonate, and water. After evaporation, 7.4 g of crude syrup was obtained and purified by chroniatography on silica gel with the solvent mixture, ethyl acetate - Skellysolve B 1 : 1. Pure orthoester 3 (5.8 g, 78%) crystallised from ether - Skellysolve B, mp 8586"C, [aID 10.3 (c 1.1, CHCI,); 'H nmr (CDCI,) 6: 1.151.45 (rn, 15H, -(CH2)6-, CH3-CH,), 1.69 (s, 3H, CH3-C

+

2096

C A N . J . CHEM. VOL. 57, 1979


TABLE 1. Lewis acid catalysed rearrangement of 1,2-orthoacetates to 8-ethoxycarbonyloctyl glycosides 1,2-transGlycosides

Yield

1,2-Orthoacetates

(%I

[aIDz3in CHCI3

4 3 8

Lactoside 6 Cellobioside 5 Maltoside 9

60 68 58

-2.4 - 15 55.8

errdo), 1.85-2.13 (m, 18H, CH,CO-), 2.25 (t, 2H, CH,CO), 3.64 (t, 2H, CH,O). 5.59 (d, J,,, = 5.5 HZ, lH, H-1), 3.3-5.5 (remaining protons). Anal. calcd. for C37H56020:C 54.14, H 6.88; found: C 54.01, H 6.90.

Melting point ("C)

91-92 -

AcO(CHZ)~CO2Et

(%) 22 20 20

octanol gave the expected nmr spectrum; 'H nmr (CDCI,) 6: 1.23 (t, 3H, 0CHzCH3), 0.9-1.60 (b m, 12H, -(CH2)6-), 2.03 (s, 3H, CH,CO), 2.26 (t, 2H, CH2CO), 4.49 (t, 2H, 0 C H 2 ) , 4.59 (q, 2H, OCHZCH,).

Formntiorr of 8-E/Aosycnr.bor~yloctylGlycosirles Usirrg Silver. 3,6Di-O-nce/~~/-4-0-(/etrn-O-rrcetyl-~-~-galnctopyrnrro.s)~1) a-D-glucopyrnnose 1,2-(8-Et/1o.~)~carbony/octyl Ortlroncetnte) (4) The procedure was similar to that described for 3 except that alcohol (5.05 g, 25 mmol) in 30 mL of dichloromethane containing silver triflate (8.48 g, 32 mmol) and collidine (3.98 mL, 30 mmol) was reacted with hepta-0-acetyl-a-D-lactosyl bromide (2) (22.7 g, 32 mmol) (9). After chron~atography(Skellysolve B -ethyl acetate 1 : l), orthoester 4 (12.2 g, 62%) was obtained as a honlogeneous syrup, [a], = -2.4 (c 3.7, CHCI,); RI 0.62, that contained C-CH, endo and exo isomers in the ratio 11 :3; 'H nmr (CDCI,) 6 : 1.12-1.42 (m, 15H, -(CH&, CH3-CH2), 1.52, 1.69 (S, 3H, enclo and exo CH3-C), 1.96 (s, 3H, CH,CO), 2.03 (s, 3H, CHSCO), 2.09 (s, 6H, CH3CO), 2.12 (s, 3H, CH3CO), 2.14 (s, 3H, CHSCO), 2.25 (t, 2H, CHZCO),5.62 (d, J I v = 2 5.1 HZ, IH, H-I), 3.30-5.50 (remaining protons). Pure 4 (1.2 g) was de-0-acetylated in dry methanol (5 mL) containing a catalytic amount of sodium methoxide. After 18 h amorphous crystals (650 mg) were collected; n ~ p168-1 70°C 14.3 (c 3.6, methanol). Arrnl. calcd. for (recryst.), [a], C Z 4 H 4 2 0 1 4 . H 2 0C: 50.34, H 7.75; found: C 50.65, H 7.79. 3,6Di-0-rrcetyl-4-0- (tetrn-0-ncetyl-a-D-g/~rcopyr.arrosyI) -aD-gl~icopyrnrroseI,2-(8-Etlrox~~cnrboNy/octyl Ortlroncetnte) ( 8 ) Hepta-0-acetyl-a-D-maltosyl bromide 7 (8.8 g, 12.7 mmol) (8) was added to a solution of 8-ethoxycarbonyloctanol (2.02 g, 10 mn~ol)containing silver triflate (3.34 g, 13 mmol) and collidine (2.5 mL, 12 mmol) under the conditions described for the preparation of orthoester 3. After chromatography of the crude syrup (9 g) the orthoester 9 (4.9 g) was obtained as a syrup which resisted crystallisation and which by 'H nmr appeared to be a mixture of isomers CH3-C er7rlo:exo 9:4, [a], f 6 . 8 (c 7.5, CHCI,); IH nmr (CDCI,) 6: 1.12-1.45 (m, 15H, -(CH2)6-, CH3CH2), 1.58, 1.72 (s, 3H, eso and errdo CH,-C), 1.99-2.15 (m, 18H, CH,CO), 2.27 (t, 2H, CH2CO), 5.65 (d, J l , Z= 5.0 Hz, l H , H-1), 3.40-5.60 (remaining protons). Lervis Acid Renrrnrrge/rren/ of I,2-Ortlroncetotes3, 4, ancl9 to I,2-trans-Glycosidcs Gerrernl Procerl~rre Lewis acid, stannic tetrachloride (0.01 mL, 0.1 mmol) was added to a cooled solution of 1,2-orthoacetate (1 mmol) in dry dichloromethane 10 mL. The reaction temperature was maintained at - 10°C for 30 min and the mixture was then poured into 10 mL of saturated sodium bicarbonate solution. Extraction with 20mL of dichloron~ethane followed by washing of the organic phase provided crude glycosides 5, 6, or 9 after the usual work-up. Chromatography on silica gel with the solvent ethyl acetate - Skellysolve B I : 1 yielded pure glycosides and the acetate of 8-ethoxycarbonyloctanol (Table 1). The ' H nmr data for glycosides 5, 6, and 9 were exactly as reported later in this section. Acetylated 8-ethoxycarbonyl-

+

Trifite nnd N,N-Te/rnnret/~ylurec~ 8-Etl~o,~ycnrbor~loctyl2,3,6-Tri-0-nce/~~/-4-0-(tetrn-Oacety/-~-~-g/ucopy,'nnosy/)-~-~-g/rrcopyr.nnoside (5) A solution of 8-ethoxycarbonyloctanol(1.01 g, 5 nlnlol) and N,N-tetramethylurea (0.6 mL, 5 nlmol) in 10 mL of dry dichloromethane containing silver triflate (1.03 g, 4 mmol) was cooled to -40°C. Hepta-0-acetyl-a-D-cellobiosyl, bromide (1) (2.8 g, 4 mmol) (8) in dry dichloromethane (20 mL) was added dropwise over 15 min. After 30 min at -40°C the cooling bath was allowed to warm to 20°C and the mixture was left for a further 3 h. The suspension was filtered through a pad of Celite and charcoal and the solids were washed with dichloromethane (50 mL). The filtrate was extracted with saturated sodium bicarbonate and water. After concentration of the dried solution the crude syrup was purified by column chromatography and crystallised from ether, yield 2.23 g, 68%; mp. 91-92"C, [a], - 15 (c 1.0, CHCI,); ' H nmr (CDCI,) 6: 1.11.7 (m, 15H, -(CH&,-, CH,CH,), 2.11 (m, 21H, CH,CO), 2.24 (t, 2H, CH2CO), 3.35-5.55 (remaining 18 protons). Arml. calcd. for C37H56020: C 54.14, H 6.88; found: C 53.96, H 6.92. Elution of the column after the glycoside was obtained yielded components with ' H and I3C nmr consistent with disaccharide hexaacetates. 8-E/hox~~cnrbonylo~yl2,3,6-Tri-O-nce/~~l-4-0(te11.n-0ncet)~/-~-~-gn/flctopyr.~~~rosy/)-~-~-g/ucop)'ronosir/e (6) The reaction was carried out on a 4 mmol scale as described for glycoside 5 using hepta-0-acetyl-a-D-lactosyl bromide (2) (9). Column chronlatography gave 6 as a pure syrup (2.03 g), 62%; [a], - 11.2 (c 2.0, CHCI,); l H nmr (CDCI,) 6: 1 .loCH,CH,), 1.93, 2.02, 2.05, 2.07 (s, 1.80 (m, 15H, -(CH,),-, 21H, CH,CO), 2.26 (t, 2H, CH2CO), 3.25-5.55 (remaining 18 protons). In addition, acetylated 8-ethoxycarbonyloctanol was isolated in 20% yield together with a lactose hexaacetate whose I3C nmr gave C-l(a) 90.0, C-1(B) 95.1, and C-l'(p) 101.0, which is consistent with a reducing. disaccharide hexaacetate lacking an acetate at C-2. 8-Etlroxycnrbonyloc/yl2,3,6-Tr.i-O-ncet~~l-4-0-(tetm-Oacetyl-a-D-glrrcop)~rnr~osyl) -a-D-gl~icop~~rnnoside (9) The bromide 7 (4 mmol) (8) was reacted with 8-ethoxycarbonyloctanol in the manner described for glycoside 5. The glycoside 9 (1.93 g) crystallised from ether and Skellysolve B, mp 63-64"C, [a],39.1 (c 1.0, CHCI,); ' H nmr (CDCI,) 6: 1.10-1.80 (m, 15H, -(CHZ)~-, CH3CH2), 1.97, 2.00, 2.02, 2.06, 2.10 (s, 21H, CH,CO), 2.25 (t, 2H, CHZCO), 3.25-5.55 (remaining 18 protons). Arrnl. calcd. for C37H560ZO: C 54.14, H 6.88; found: C 53.96, H 6.92. In addition, a compound with the properties of a hexaacetate was isolated by continued elution after 9 had been eluted from the silica colunln. The 13C nmr showed no aglycon signals and anomeric signals C-l'(a) 95.5, C-l(p) 94.8, and C-](a) 89.8 ppm. Thc intensity of these signals were respec-

BANOUB AND BUNDLE: I1


tively 8:5:3; 'H nmr gave acetate signals and ring protons only in the ratio 18: 16.

2097

(eso and erlclo) With collidine - triflic acid the two C-CH, signals at 6 1.61 and 6 1.69 remained unchanged after 4 h. N,N-~etramethylurea - triflic acid caused these signals to disappear after 5 min at room temperature. In addition the H-1 doublet 6 J I , ,= 5 Hz also disappeared indicating reaction of the orthoester.

Observatior~of 1,2-Ortl1oesterlnfern~erliafesby H Nriclear Magnetic Resor~mlce 8-Ethoxycarbonyloctanol (600 mg, 3 mmol) was dissolved in dry dichloromethane (7 mL) containing silver triflate (770 Acknowledgements mg, 3 mmol) and N,N-tetramethylurea (0.72 mL, 6 mn~ol). The mixture was cooled to -40°C and 3 mL of a dichloroWe wish to thank Mr. J. Christ for technical assismethane solution containing 2,3,4-tri-0-acetyl-u-L-rhamnotance. The larger portion of this work was made pospyranosyl bromide (1.02 g, 2.89 mmol) (16) was added dropsible by a collaborative research project of the wise over 15 min. The reacrion was maintained between -20 and -30°C for 3 h and a mixture of water (1 mL) and triethyl- National Research Council of Canada and MDS amine (1 mL) were added to quench the reaction. The reaction Health Group Ltd. through the Pilot Industry was poured into dichloromethane (10 mL), filtered through a Laboratory Program. We would like to thank Dr. pad of Celite, and then washed with sodium bicarbonate and water. After drying the syrup under high vacuum the l H nrnr Klaus Bock for constructive criticism of our original manuscript. was recorded; 'H nrnr (CDCI,) 6: 1.08-1.50 (m, -CH2),-, CH3CH2,H-6) 1.64 (s, erirlo CH3C), 1.90 (s, CH,CO) 1.97 (b s, 1. D. R. B U N D L EAbstracts . of Papers. IXth International CH3CO), 2.03 (b S, CH3CO). Symposium on Carbohydrate Chemistry. London, EngT o the nrnr tube containing the crude syrup, one drop of a land. 1978. p. 373. solution containing N,N-tetramethylurea - triflic acid 2: 1 2. D. R. BUNDLE. Can. J. Biochem. 57,367 (1979). (the solution was 1 M with respect to acid) was added and the D.. R. B U N D L Ei~nd , D. A. B A K E RJ.. Am. 3. R. U . L E M I E U X 'H nmr was rerun. The orthoester C-CH, peak 6 1.64 was no Chem. Soc. 97.4076 (1975). longer evident and the multiplet complex due to acetate signals 4. R. U. L E M I E U X D., A. B A K E Rand , D. R. BUNDLE. Can. J . had simplified to 3 singlets. The chen~icalshifts of these signals Biochem. 55,507 (1977). were identical to those obtained with an authentic sample 5. F. J. KRONZER and C. SCHUERCH. Carbohydr. Res. 27,379 of 8-ethoxycarbonyloctyl 2,3,4-tri-0-acetyl-a-L-rhamnopyrano(1973). side (27); ' H nrnr (CDCI,) 6: 1.08-1.80 (m, -(CH2),-, andJ. BANOUB. Am. Chem. Soc. Symp. Ser. 6. S. HANESSIAN CH3CH2, H-6) 1.94 (s, CH3CO), 2.00 (s, CH,CO), 2.10 (s, 39. 36(1976); Carbohydr. Res. 53, C13 (1977). CH3CO), 2.25 (t, CH,CO), 3.30-5.50 (remaining protons). 7. R. U. L E M I E U XT.. T A K E D Aand , B. Y. C H U N GAm. . Chem. Soc. Symp. Ser. 39.90(1976). Renction of 1,2-Ortl~oesfers~vithN, N-Tetrntt~efl~~~liirco/ 8. J. K. DALE.J. Am. Chem. Soc. 38,2187(1916). Collidine- Trific Acid and A. K U N Z J. . Am. Chem. Soc. 47.2052 The orthoacetate, 3,4-di-0-benzyl-p-L-rhan~nose-l,2-(methyl9. C. S. HUDSON (1925). orthoacetate) (I mmol) in dichloromethane (5 mL) was reacted X A. R. MORGAN. Can. J. Chem. 43,2199 with I mL of N,N-tetramethylurea - triflic acid 2: 1 (the solu- 10. R. U. L E M I E Uand (1965). tion made up in dichloromethane was 1 M with respect to and A. S. PERLIN.Can. J. Chem. 43, 1918 triflic acid) at room temperature. After 5 min tlc (Skellysolve I I . M. MAZUREK ( 1 965). B -ethyl acetate 3: 1) showed complete reaction to the correand J. BANOUB. Carbohvdr. Res. 44. C14 sponding glycoside. The reaction was repeated on a small scale 12. S. HANESSIAN (1975). in a nnir tube using deuteriochloroform as solvent. The A. J. K H O R L I N and , A. F. BOCHKOV. 13. N. K. KOCHETKOV, glycoside, methyl 2-0-acetyl-3,4-di-0-benzyl-a-L-rhamnopyTetrahedron, 23,693 (1967). ranoside, had ' H nrnr (CDCI,) 6: 1.38 (d, J,,, = 6.0 Hz, 3H, and I. KVANSTROM. Act21 Chem. Scand. B, H-6), 2.17 (s, 3H, CH,CO), 3.37 (s, 3H, OCH,), 3.88 (d d, 14. P. J. GAREGG 30,655 (1976). J,,, = 4 H z , J,,, = IOHz, l H , H-3), 5.32 (d d, J,,, = ? H z , and D. R. BUNDLE. Can. J . Chem. Thisissue. 15. J. BANOUB J2,3= 4 HZ, l H , H-2), 7.34 (b S, 10H, +CHz), 3.35-5.00 2nd J . BANOUB. Methods Carbohydr. Chem. (remaining protons). Isomerisation of 1,2-orthoacetate to 16. S. HANESSIAN 8. In press. glycoside was essentially quantitative. Can. J. Chem. 57, The experiment described above was repeated with 0.1 mmol 17. D. BUNDLEand S. JOSEPHSON. 662 (1979). of N,N-tetramethylurea triflic acid with similar results. When this was repeated at - 10°C the reaction took 3 h to reach 18. G. WULFFand G. ROHLE.Angew. Chem. 13, 157 (1974). R. V. STICK,and K. completion. At -70 and -30°C no detectable rearrangement 19. R. U. L E M I E U XK., B. HENDRIKS, JAMES. J. Am. Chem. Soc. 97,4056(1975). had occurred after 4 h. In these instances reactions were 20. K. IGARASHI. Adv. Carbohydr. Chem. Biochem. 34, 243 quenched with excess triethylamine. (1977). The rhamnose orthoacetate (1 mmol) was reacted with Chem. . Can. 16. 14(1964). 1 mrnol of collidine - triflic acid 1 : I (I M in dichloromethane) 21. R. U. L E M I E U X Chem. Ber. 105, for 4 h at room temperature. After this period tlc showed only 22. G. WULFF,G. ROHLE,and U . SCHMIDT. 11 1 1 (1972). starting material, which was confirmed by a similar experiment and I. KVANSTROM. Acta. Chem. Scand. B, performed in a nrnr tube. The nrnr parameters for the 3,4-di- 23. P. J. GAREGG 31.509 (1977). 0-benzyl-a-L-rhamnose-1,2-(methyl orthoacetate) were 'H Can. J . Chem. 41,555 (1963). nrnr (CDCI,) 6: 1.31 (d, J5,, = 5.8 HZ, 3H, H-6), 1.73 (s, 3H, 24. A. S. PERLIN. X T . L. NAGABHUSHAN. Methods CarCH,CO), 3.29 (s, 3H, OCH,), 3.27-3.79 (m, 3H, H-3, H-4, 25. R. U . L E M I E Uand bohydr. Chem. 6,487 (1972). H-5), 4.40 (d d J Z , , = 3.7 HZ, J l V= z 2.4 HZ, l H , H-2), 4.58and D. R. PERRIN. , L. ARAMAREGO, 5.02 (m, 4H, +CHZ), 5.28 (d, J I , , = 2.4 Hz, l H , H-I), 7.34 26. D. D. P E R R I NW. Purification of laboratory compounds. Pergamon Press. (b S, 10H, (bCH2). London. 1966. The cellobiose 1,2-orthoacetate was treated with N,N-tetramethylurea - triflic acid 2: 1 and also collidine - triflic acid 27. S. JOSEPHSONand D. R. BUNDLE.J . Chem. Soc. Perkin Trans. I. In press. 1 : 1.02 in the manner described for the rhamnose orthoacetate.