SYNTHESIS OF 2 3 5 (OR 2 3 NAL TRIM ETHYL GLUCOSE. 2729 CCCLXX1V.-Synthesis of 2 3 5 (or 2 3 4)-Tri-methyl Glume.* By JAMES COLQUHOTJN IRVXNE and JOHN WALTER HYDE RECENT developments in the constitutional study of carbohydrates demand that from time to time the structure of the simple methyl-ated sugars utilised as reference compounds in such work should be brought under review. In the present case we have selected 2 3 5-trimethyl glucose for further examination as this sugar is the key to the const.itution of both maltose and p-glucosan and acquires further importance in comexion with the chemistry of starches. Obviously the sigmficant feature of this particular form of trimethyl glucose is the non-reducing hydroxyl group as the identification of its position indicates the linkage of the two glucose residues in maltose and also the attachment of the anhydro-ring in p-glucosan.In order to give a clear view of the present position, it may be recalled that 2 3 5-trimethyl glucose was first obtained, in the form of the corresponding methylglucoside (Purdie and Irvine J. 1903 83 1021; Purdie and Bridgett ibid. 1037) by methylating methylglucoside in the presence of excess of methyl alcohol. The free sugar was also studied by the above workers, who converted it into 2 3 5 6-tetramethyl glucose and tentatively ascribed to it the structure still in use. Their views were after-wards supported by Irvine and Dick (J. 1919 115 593) who showed that on oxidation by nitric acid the sugar is converted into a trimethyl saccharo-lactone.This observation was con firmed by Haworth and h i t c h (ibid. 809) who isolated the m e sugar from fully methylated maltose and by Irvine and Oldham (J. 1921 119 1744) in studying the structure of p-glucosan. Although there seemed no reasonable doubt that the methylated glucose in question was correctly formulated the evidence obtained by subjecting sugars to the oxidising action of nitric acid should not be accepted as final so long as other and more diagnostic tests are available. The divergent results obtained by Irvine and Hogg (J. 1914 105 1386) and by Levene and Meyer ( J . Bio2. Chem., 1922 54 805) in oxidising monomethyl glucose by means of nitric acid may be quoted in illustration of this point. The constitution of 2 3 5-trimethyl glucose has now been In a letter to Nature 19th September 1925 Haworth states that he has obtained evidence leading to the conclusion that normally the oxidic ring in glucose couples positions 1 and 5.Should this be substantiated the methyIated glucose which f o m the subject of the present communication should be indexed as 2 * 3 4-trimethyl glucose. OLDHAM 2730 IsVINE AND OLD-: confirmed by the following synthetical scheme which was designed to produce a methylated glucose with 'unsubstituted hydroxyl groups definitely in the terminal positions 1 and 6. Triacetyl glucosan was converted by Karrer's process into triacetyl dibromoghcose identical with that previously obtained by Fischer from penta-acetyl glucose. Stage 11. The above dibromo-derivative when treated with methyl alcohol in the presence of silver carbonate gave triacetyl methylglucoside bromohydrin and the subsequent reactions were therefore designed to replace acetyl by methoxyl and thereafter to introduce the hydroxyl group in place of bromine.The acetyl groups were removed by the action of alcoholic ammonia giving methylglucoside bromohydrin. It may be mentioned that although the physical constants mere other-wise in good agreement the melting point of this compound was found to be several degrees higher than that quoted by Fischer. Methylation of methylglucoside bromohydrin under conditions which would have the minimum effect on the bromine atom was achieved by the silver oxide reaction and gave a mixture of (a) trimethyl methylglucoside bromohydrin (800/) and (t) the corresponding enolic anhydride (20%).From this mixture pure trimethyl methylglucoside bromohydrin mas isolated. Stage V . The above compound when heated at 150" with alcoholic potassium acetate gave a 72% yield of pure crystalline trimethyl p-methylglucoside identical with that obtainable from the form of trimethyl glucose which is the subject of the investig-ation. In the glucoside h a l l s obtained the hydroxyl group occupies the position of the bromine atom in triacetyl methylglucoside bromohydrin. As Fischer succeeded in reducing the latter to triacetyl methylisorhamnoside (Ber. 1912 45 3761) the group is therefore in the 6-position. In consequence the methyl groups in the corresponding sugar must be in positions 2 3 and 5.Although we are unwilling to attach undue importance to colour tests the proof is strengthened by the following considerations. In the glucose molecule positions 1 2 3 and 5 are occupied by secondary alcohol groups and the only primary alcohol group is in the 6-position. If therefore the vacant hydroxyl group in a trirnethyl methylglucoside is primary the corresponding nitro-derivative should give a red colour by Meyer's test; otherwise tt blue solution will result. Trimethyl methylglucoside bromohydrin was therefore converted into t'he corresponding iodohydrin which was transformed with some diEculty into mononitro-trimethyl methylglucoside. This on treatment with nitrous acid gave as Stage I . Xtage I I I . Stage I V. This result supplies the evidence required SYNTHESIS OF 2 3 5 (OR 2 3 ~)-TRR~ETHYL GLUCOSE.2731 expected the characteristic red colour similar to that obtained from nitromethane. In verification of our former work on p-glucosan we have repeated the conversion of this anhydroglucose into trimethyl glucose and subjected the sugar to the following successive operations : I. Acetylation giving trimethyl glucose diacetate. 11. Bromination by the action of hydrogen bromide in glacial acetic acid giving trimethyl acetyl glucose bromohydrin. 111- Action of sodium methoxide giving trimethyl P-methyl-glucoside. The methylated glucoside finally isolated was identical with that obtained in the synthetical processes already synopsised thus confirming that P-glucosan is 1 6-anhydroglucose.In order to render the scheme of reactions intelligible the various changes involved in the two alternative methods of producing 2 3 5-tri-methyl methylglucoside may be represented structurally : I. From Penta-acetyl Glucose or from Triacetyl G1ucosa.n. rFHBr ryH*OMe b cjH*OAc YH-OAc 1 yH*OAc CH-OAc 'H*OAc r$!H*OMe rFH=OMe CH-OMe I YH-OMe A tH*OMe YE*oue LCH LYH hH*OMe YH-OMe Penta -ace tyl glucose or + I YH*OAc ~ g1ucosa.n T riacet yl LYH LYH Stage 1. hH2Br Stage 11. &H2Br Stage IV. bH,Br Stage V- CH,*OH 11. From Glucosan. p-1 rYH-1 l$JH*OH I &vH*OMe 1 [vE:Ege I YH'OH I CH*OMe I VH*OMe CH*OH 1 +H*OMe I $H*OMe dH2 -l CH,- CH,*OH ryHBr YH-OMe 4 YH-OMe LYH 0 --f L(!H 0 L(lH LPH -+ H*OMe YH-OMe CH*OMe XH2*OAc C1H,-OAc h2*O 2732 IRVINE AND OLDHAM: The collective results may be compressed into the statement that dibromotriacetyl glucose maltose and glucosan are all convertible into the same form of trimethyl glucose.We are engaged in attempts to s-vnthesise other partly methyl-ated sugars and the isomeric distinction between 2 3 5- and 2 3 6-trimethyl glucose having now been established the results will be utiLised in forthcoming papers from this laboratory. E X P E R I M E N T A L . The following account of experimental procedure is limited to the synthetical reactions described in the introduction. Other reactions to which reference is made were carried out by methods which are now standardised and their description is therefore omitted. T r i m t y l Dibromogluwse from Triacetyl p-G1ucosan.-The method recommended by Karrer (Helv.Chim. Acta 1922 5 124) was adopted minor variations being introduced as the reaction which is most successful when small quantities of material are manipul-ated requires careful control. Triacetyl p-glucouan (in lots of 5 g.) was heated on a boiling water-bath with phosphorus pentabromide (8.5 g.) the flask being fitted with a ground-in condenser. When effervescence had nearly ceased the contents were poured into finely-crushed ice and thoroughly disintegrated by a glass rod. Similarly the residue in the flask was mixed as rapidly as possible with small pieces of ice until all halides of phosphorus had been destroyed. Rise of temperature must be avoided in these operations. The fine white powder resulting from several experiments was united washed with water until free from phosphoric acid and thereafter with absolute alcohol until the washings were nearly colourless.Purification was effected by dissolving in a small quantity of chloroform and precipitating with light petroleum. The yield of crystalline product (m. p. 173") including the material obtained from the mother-liquors averaged 50% of the theoretical amount. As the rotation of the compound does not appear in the literature the following values were determined : Solvent. C. ra3,. Chloroform ....................................... 1-537 + 189-9O Glacial acetic acid .............................. 1.256 + 185.9 Conversion of Triacetyl Dibrmoglucose into Triacetyl Methgl-gluwside Bromohydrin.-This reaction was carried out exactly as described by Fischer (Ber.1902 35 857; 1920 53 873) the yield of glucoside being nearly quantitative ; after recrystallisation from absolute alcohol the product melted a t 126-127". As i SYNTHESIS OF 2 3 15 (OR 2 3 4)-TRIMETHYL GLUCOSE. 2733 this case also no optical data appear to have been published the specific rotation in the following solvents was determined : Solvent. C. [a],. Chloroform ....................................... 3.012 - 1.4" Methyl alcohol ................................. 3.014 -3.1 Glacial acetic acid .............................. 3.006 - 2.7 Methylglumside Bromohydrin.-The acetylated methylglucoside bromohydrin obtained as above was dissolved a t room temperature in methyl alcohol containing 5-lOyo of ammonia so as to form a 5% solution.It is unnecessary to saturate the liquid with ammonia as stated by Fischer and the use of a dilute solution enables the end-point of the reaction to be determined polari-metrically. When the specific rotation had diminished to - 19-3", the product was isolated by evaporating to dryness and extracting with chloroform to remove acetamide. The yield of crude bromo-hydrin was nearly quantitative and after recrystallisation from ethyl acetate the compound melted and decomposed a t 153-154" in place of 148" as quoted by Fischer. In aqueous solution under conditions identical with those used by Fischer the specific rotation wits - 33~6"~ the literature value being - 34.9". Tribenzoyl methyQ1ucoside bromohydm'n has no dirdct bearing on the main investigation but reference may be made to it.The normal procedure was followed the bromohydrin being acted on by a slight excess of benzoyl chloride dissolved in pyridine. The product was brought into solution in a minimum of glacial acetic acid and the pure tribenzoate precipitated by addition of absolute alcohol m. p. 160-162"; [a], in chloroform - 5.0" for c = 2.413; needles insoluble in water and light petroleum and readily soluble in organic solvents with the exception of alcohol and ether. Trimethyl Afeth ylglucoside Bromohydm'n.-For the particular object in view alkylation by silver oxide and methyl iodide is the only method applicable to methylglucoside bromohydrin. The usual procedure was followed methyl alcohol being added during the first methylation.After four successive treatments the refrac-tive index was constant and the liquid product wits distilled a mobile syrup being obtained (b. p. 140°/l mm. ; nD 1.4720; [.ID in methyl alcohol - 20.5" for c = 1). Examination showed that at least two compounds were present one of them containing no bromine. This constituent was present to the extent of 20% and was evidently dimethyl anhydromethylglucoside (Found : C 42.7 ; H 6-7 ; OMe 41.7 ; Br 21.3. Calc. for trimethyl methyl-glucoside bromohydrin C 40.1 ; H 6.35 ; OMe 41.4; Br 26.7%. Calc. for a mixture of 80% of the above with 20% of trimethyl anhydromethylglucoside C 42-7 ; H 6.6 ; OMe 42.2 ; Br 21.3%) 2734 SYNTHESIS OF 2 3 5 (OR 2 3 ~ ) - ~ I M E T H Y L GLUCOSE.The close agreement with the experimental figures particularly the resulf of bromine determinations confirms the composition ascribed to the mixture. The constituents were separated by solution in ether and repeated extraction with water a process which completely removed dimethyl anhydromethylglucoside. After evaporation of the ether the residual syrup was fractionated in a high vacuum giving a liquid distillate which slowly crystallised. Owing to the ready solubility of the compound in all solvents with the exception of water no suitable recrystallising medium could be found but after spreading on a tile the crystals were hard and crisp m. p. 24" ; nD 1-4735 ; [.ID in acetone - 5.8" (c = 3.851), in methyl alcohol - 4.7' in benzene - 7-7' in chloroform - 3.5' (Found C 40-3; H 6-3; OMe 41-3; Br 27.2.Trimethyl methylglucoside bromohydrin requires C 40-1 ; H 6-35 ; OMe, 41.4; Br 26.7%). Hydroxylation of Trimethyl Methylglucoside Bromohydrin.-Re-liminary experiments having shown that both aqueous and alcoholic sodium hydroxide react with the bromohydrin eliminating hydrogen bromide and giving unsaturated derivatives the hydrosylation was effected by means of potassium acetate. A 3% solution of the bromohydrin in methyl alcohol was heated with excess of potassium acetate at 150" for 3 days during which the laworohtion increased greatly. The crude product on isolation in the usual manner was obtained in nearly quantitative amount but was contaminated with an unsaturated impurity. On keeping how-ever the syrup solidified and after draining on a tile and recrystallis-ing from light petroleum a 72% yield of pure trimethyl p-methyl-glucoside was obtained in characteristic crystals.The melting point was 93-94' a value which was unaffected when the material was mixed with an authentic specimen of the compound. The specific rotation also was in close agreement ([aID in chloroform - 11.9" for c = 1.084). It was ascertained that owing to the alkalinity of potassium acetate the above hydroxylation was accompanied by a side-reaction giving rise to an unsaturated compound; this is being further examined. Trimethyl Methylglumside 1odohydrin.-The general method described by Finkelstein (Ber. 1910 43 1528) was employed a solution of trimethyl methylglucoside bromohydrin in acetone being heated a t 100" for 6 hours with twice the theoretical amount of sodium iodide.After removal of the acetone the residue was extracted several times with ether and the united extracts were washed with aqueous sodium thiosulphate solution allowed to stand over sodium sulphate and evaporated to dryness. The product crystallised readily in needles and when drained on a tile G~LCHRJST AND PVRVES GLYCE~OL GLUCOSIDE. 2735 melted at 31-34'. Owing to the excessive solubility in all solvents except water no recrystallising medium could be found but the compound was evidently pure (Found OMe 36-1 ; I 36.5. Calc., OMe 35.8 ; I 36-77/ ; n 1-4992 ; [a], + 8.6" in chloroform for c = 3.637 + 4.1" in acetone for c = 34397 and + 6.5" in methyl alcohol for c = 4.221).Trimethyl methylglucoside iodohydrin wtts converted into the corresponding nitro-compound by heating with dry silver nitrite at 100" for 2 days. After distillation under diminished pressure the nitro-derivative was isolated in the form of the sodium salt by the addition of the calculated amount of sodium methoxide dissolved in methyl alcohol. On removal of the solvent the salt was dissolved in dilute sulphuric acid and the solution extracted with ether. After being washed with sodium thiosulphate solution to remove a trace of iodine the ethereal layer was dried and evapor-ated and the mononitro-trimethylglucoside isolated by distillation as a colourless syrup (nD 14603). Beyond checking the methoxyl content (Found OMe 464. Calc. OMe 46-7%) the compound wm not fully analysed as it was prepared solely for the purpose of carrying out the colour test with sodium nitrite and dilute sulphuric acid. This treatment gave a bright red solution free from any shade of blue and whilst the colour was not so intense or so lasting as that obtained under parallel conditions with nitro-methane the result was characteristically positive. The thanks of the authors are due to the Carnegie Trust for a Fellowship which has enabled one of them to take part in the work. UNITED ~OLLECIE OF ST. SALVATOB AND ST. LEONARD, UNIVERSITY OF ST. ANDREWS. [Received October 7th 1925.