CONSTITUTION OF THE METHYL PENTOSES. PART I. 73X.-A Conti-ibutiorz to the Study oj' the Constitutionof the Merhyl! Peiltoses. Port I. Synthesisof an i-Methyl Tetrose und an i-&fethyl Tct7itol.By ROBERT GILMOUR.THE exact constitution of the known members of the methyl pentoseseries has not been fully established as yet, except in the cases ofrhamnose and isorhamnose, the configurations of which haverecently been determined by Fischer (Bey., 1912, 45, 3761). In theremaining members a doubt exists as to the position of thehydroxyl group contiguous to the methyl group. It seemed, there-fore, of some importance to attempt the synthesis of active methyltetroses of known constitution from which methyl pentoses mightbe prepared. The importance of the subject seemed enhanced bythe fact that the active methyl tetroses might be made to serveas standards with which the degradation products of the naturallyoccurring methyl pentoses could be compared, and thus enable theirexact configuration to be determined.The most suitable starting material appeared to be the dihydroxy-valerolactone, which Thiele obtained by oxidising Aha-angelicalactonewith potassium permanganate (AnnaJen, 1902, 319, 194) and con-sidered to possess the structure :0CW;&H *(:El (OH).C'H (OH).hO.On reducing this compound with sodium amalgam in the usualway a solution was obtained which contained an amount of reducingsugar equivalent to 6 per cent. of the lactone used, and as theresult of some twenty experiments this was found to be themaximum yield of sugar in any one reduction.A series of quantitative experiments was carried out with theobject of tracing the exact course of the reduction.I n each casethe unaltered lactone was separated from the syrupy reductionproduct and weighed, and the amount of sugar present in the reduc-tion product was estimated by titration with Fehling's solution..___-The result^ of three experiments were as follows:TotalLactone Sodium Lactone reduction Per cent.grams. grams. grams. grams. granis. lactone.used, amalgam, recovered, product, Sugar, of(1) . 28 250 16.5 10 containing 1.6 = 5 . 7160 11.0 4'8 ,, 1.0 = 6.0 1;; ;l":o" 200 5.7 4.5 ,, 0.7 = 6-8It will be seen that the percentage of sugar formed remain'14 GILMOUR: A CONTRTBUTION TO THE STUDY OF THEconstant.This pointed t o the fact that the reduction was proceed-ing too far, and that the sugar-alcohol was being formed from thesugar almost as quickly as the latter from the lactone, and thiswas found to be the case as the product proved to be mainly sugar-alcohol.From the reduction product an osazone was isolated which provedto be a methyl tetrosazone melting a t 140-142O. Similarly, bytreatment with phenylbenzylhydrazine in alcoholic solution a methyltetrosep?knylb enzylhydrazone melting at 99-looo was obtained.From the aqueous solution from the preparation of the hydrazonea methyl tetritol was isolated as a yellow spup. On benzoylatingthe latter a tetrabenzoylmethyltetm'tol melting a t 136-137O wasobtained.The methyl tetrose was obtained in the free state by decomposingthe phenylbenzylhydrazone with formaldehyde.It proved to be apale yellow, strongly reducing syrup. As the yield of pure sugarwas very small, amounting to only 3 per cent. of the lactoneused, an attempt was made to obtain an additional quantity of itby oxidising the methyl tetritol with Fenton's reagent. In thisway a small amount of the sugar was isolated as its phenylbenzyl-hydrazone.Dimet?~ox~vaZeroZacton~e has also been obtained by methylatingdihydroxyvalerolactone. It melts at 59-60O. An attempt wasmade to reduce the methylated lactone, but without success.The constitution of the methyl tetrose described has not beenestablished yet, but it must be represented by either formula (I)or (11), according to the positions assumed by the hydroxyl groupsduring the oxidation of Aa-angelicalactone.The yield was, however, very poor.FHO YHOA study of the methyl tetronic acids obtainable from methylglycerose is being undertaken with a view to settle the question ofthe constitution of the tetrose described.EXPERIMENTAL.AB-A rtgelicdactoite.The method followed in the preparation of this compound wasAs, however, the preparation was that due to Thiele (Zoc.cit.)CONSTITUTION OF THE METHYL PENTOSES. PART I. '75on a very much larger scale, and differed in one or two particulars,i t may with advantage be described.Two hundred and fifty grams of commercial lmmlic acid wereadded t o 300 grams of acetic anhydride, and the mixture was heatedto looo during six hours.The product was then distilled under apressure of 300 rnm., and the distillate collected in four approxi-mately equal fractions. The first fraction consisted entirely ofacetic acid, the others of acetic acid with increasing amounts ofAD-angelicalactone. Each fraction as it distilled over was immedi-ately neutralised with sodium carbonate, and the oily layerextracted with ether. (ThieIe recommends fractionation of thedistillate at a low pressure as the best means of removing the aceticacid, but in this case a quantity of acetyl-lzevulic acid is invariablyformed, which materially decreases the yield of AB-angelicalactone.Indeed, if the acid distillatm are allowed t o remain for any timewithout neutralising, reversion to acetyl-laevulic acid t.akes place toa large extent.) The combined ethereal extracts were dried, theether distilled off, and the lactone fractionated under 15 mm.pres-sure. It distilled almost completely between 6 4 O and 70°/15 mm.,and consisted of approximately pure AB-angelicalactone.I n all, 2 kilograms of lmmlic acid were worked up, and a totalyield of 1100 grams (70 per cent. of the theoretical) of AB-angelica-lactone was obtained.Conversion of AB- into A-Angelicalactone.From 1100 grams ofthe AB-lactone 835 grams of Pa-angelicalactone were obtained in apure condition.Thiele's method was employed (loc. cit.).Oxidation of Aa-Angelicdactone.The method employed was ewentially that of ThieIe (loc.cit.).In all, 820 grams of Aa-angelicalactone were oxidiaed; the totalyield of crude material being 170 grama, from which, on crystallia-ing from ethyl acetate, 160 grams of the pure product melting atlooo were obtained. This amounts to about 17 per cent. of thetheoretical yield.Reduction of aS-DiiEydroxyvaZeTolac tone.Ten grams of the lactone were dissolved in 100 C.C. of water, andthe solution was acidified with 20 per cent. sulphuric acid. Thesolution was then cooled to -5O, and 100 grams of 24 per cent.sodium amalgam were added in five portions, with continual vigor-ous shaking, and addition from time to time of 20 per cent76 GILMOUR: A CONTRIBUTION TO THE STUDY OF THEsulphuric acid in portions of 2 C.C. (5 C.C. €or each of the first four20 gram-portions of amalgam and 3 C.C.for the last portion).The solution was rendered faintly acid to Congo-red paper, andtested with Fehling’s solution. One volume waa found to reducetwo volumes of the solution.The aqueous solution was now concentrated as far as possibleon the Fater-bath. The residue was extracted several times withhot alcohol, and the extract filtered from sodium sulphate andcomelitrated to a syrup on the water-bath. The syrup was foundto have only a slight reducing action on Fehling’s solution. How-ever, after boiling with dilute hydrochloric acid, a vigorous reduc-tion was obtained, which justifies the conclusion that a compoundof the nature of an acetal or alcoholate had been formed duringthe extraction with alcohol.The syrup (9.5 grams) crystallisedalmost completely on nucleating with the lactone. The process wasthen repeated on the product, the sugar formed being protectedfrom further reduction by the fact that it had been convertedinto an alcoholate or acetal, or possibly glucoside. The second,third, and fourth reductions were carried out with 55 grams ofsodium amalgam, and the fifth with 50 grams of amalgam. Thesyrupy product now weighed 7 grams, and showed no signs ofcrystallisation on nucleation with the lactone, and very little actionon Fehling’s solution. On titration with baryta it was found stillto contain about 5 per cent. of unaltered lactone.Hydrolysis of the Methyl Tetrose (Acetal) 9The syrup (7 grams) wits boiled with 60 C.C.of 5 per cent. hydro-chloric acid for half an hour. The acid was removed with bariumcarbonate, and the filtrate decolorised with animal charcoal, andconcentrated in a vacuum at 4 5 O to dryness. The residue waOextracted three times with boiling alcohol, and the extract concen-trated in a vacuum a t 35O to a syrup. The product was a yellowsyrup, and weighed 5 grams.0.116 Gram of the syrup reduced 2 C.C. of standard Fehling’ssolution, which gives 10 per cent. as the sugar content of thesyrup (calculated as dextrose).Methyl T e trosazone.Two grams of the syrup prepared as above were mixed with1 gram of phenylhydrazine, 1 C.C. of 60 per cent. acetic acid, and2 C.C. of water. The mixture was heated on the water-bath foran hour, when the osazone separated as a black oil.After somehours the mother liquor was poured off, and the oil dissolved iCONSTITUTION OF TEE METHYL PENTOSES. PART 1. 77alcohol. About twenty volumes of water were added to the alcoholicsolution, which was then allowed to remain overnight. The osazoneseparated as a flocculent, yellow precipitate, which wm collected,dissolved in alcohol, and again precipitated with water. Thecompound ~7as then collected and dried in a vacuum, the yieldbeing 0.2 gram.The analytical figures obtained showed that it was still impure;consequently it was still further purified by dissolving the remainder(0.1 gram) in a very small quantity of a mixture of alcohol andether, t o which 100 C.C. of light petroleum were then added. Afterremaining for some time with occasional stirring the osazoneseparated suddenly as a pale yellow, microcrystalline precipitate.It was collected, washed with light petroleum, and dried in avacuum.The compound forms yellow, microcrystalline needles,very sparingly soluble in water or ether; but readily so in alcohol,and melts at 140-142O:O.OG52 gave' 0.1652 CO, and 0-0408 H,O. C= 66.05 ; H = 6.69.C,,H,,O,N, requires C = 65-38 ; €€= 6-41 per cent,As the yields of sugar by this method were still far from satis-factory a further variation of the method of reduction was adopted,which obviated any risk of the methyl tetrose alcoholate or acetalbeing further reduced by the action of sodium amalgam in acidsolution.Twenty-eight grams of dihydroxyvalerolactone were dissolved in270 C.C.of water, acidified with 20 per cent. sulphuric acid, andreduced with 250 grams of 24 per cent. sodium amalgam in fiveportions, exactly as in the experiment previously described. (Onevolume of the resulting solution reduced one volume of Fehling'ssolution. Volume of solution= 340 c.c., therefore sugar in solu-tion=1-6 grams, o r 5.7 per cent. of lactone used.)The solution was worked up as usual. and the product separatedfrom sodium sulphate by extraction with alcohol. The solvent wasremoved at 35O,/15 mm. The product was a crystalline mass ofunchanged lactone mixed with Byrupy material, which was quiteinsoluble in ether. The material was treated with a mixture ofethyl acetate and ethyl alcohol (3:1), which dissolved the syrup,but very little of the crystallins lactone.After removing the latter,the filtrate was again concentrated, and the process repeated threetimes. In this way the lactone was almost completely separated,the total amount recovered being 16.5 grams. The filtrate wasconcentrated, and the residual syrup was dried in a vacuum. Itweighed 10 grams.A second reduction wit^ carried out on the 16.5 grams of recoveredlactone. One hundred and sixty grams of amalgam were used, a 78 GlLMOUR: A CONTRIBUTfOK TO TBE STVDY OF 'IHEthe solution obtained reduced 1.1 volumes of Fehling's solution.(Solution = 190 c.c., therefore sugar = 1 gram = 6 per cent. oflactone.)On working up, 11 grams of lactone were recovered, and thesyrup remaining weighed 4.8 grams.The 11 grams of recovered lactone were treated as before with200 grams of amalgam. (Sugar formed=0'7 gram=6.3 per cent.of lactone.) Lactone recovered = 5.7 grams ; syrup recovered =4*5grams.The combined syrups (19 grams) were hydrolysed as usual with5 per cent.hydrochloric acid, and the product, which weighed15 grams, was found to contain 2'5 grams of sugar (estimated asdextrose).Seventy grams of dihydroxyvalerolactone were worked up in asimilar manner, and the resulting 50 grams of syrup hydrolysed.Net h yl Tetrosephenylb entylhydrazone.The combined syrups from reduction experiments (60 grams)were dissolved in 200 C.C. of absolute alcohol, and 10 grams ofphenylbenzylhydrazine were added. After heating on the water-bath for tweiitIy minutes water was added, which precipitated thehydrazone as a brown oil.This was dissolved in benzene, and onallowing the benzene to evaporate spontaneously the hydrazoneremained as a sticky, brown, crystalline mass. It was washedseveral times with cold benzene, and recrystallised from a largeamount of boiling water, the yield being 6.5 grams. The compoundforms fine, colourless needles, is readily soluble in benzene, ether,or alcohol, but only very sparingly so in cold water, and melts a t99-1ooo :0.1199 gave 0.3012 C02 and 0.0746 H,O. C=68-51; H=6*95.0.1838 ,, 14.5 C.C. N, at 17O and 748 mm. N=9.10.C18H2203N2 requires C= 68.78 ; H = 7-00 ; N = 8-91 per cent.Methyl Tetritol.The aqueous solution from the above hydrazone preparation wasextracted thoroughly with ether in order to remove any excess ofphenylbenzylhydrazine or hydrazone, and concentrated a t 55O /33 mm.to a syrup. Alcohol was added, and some barium chloridewhich was precipitated was removed. The filtrate was then concen-trated, and the residue dried over sulphuric acid in a vacuum; theyield was 40 grams.Methyl tetritol so obtained forms a yellow syrup, which does notreduce Fehling's solution. It is readily soluble in water or alcohol,but insoluble in ethyl acetate or ether. It was not in a sufficientlCONSTIT'UTION OF THE METHYL PENTOSES. PdElT I. 79pure state for analysis, but was identified by the formation of itst e trabenzoyl derivative.Tetrab enzoylmethyltetritol.Two grams of the syrup were mixed with 14 grams (7 mols.) ofbenzoyl chloride and 100 C.C.of 2X-sodium carbonate solution. Thereaction was carried out. in the usual manner, and the gummy,caoutchouc-like product was washed with sodium carbonate andwater, and dissolved in ether. The ethereal solution, after beingdried, was allowed to evaporate spontaneously. A sticky, crystallinemass was obtained, which was rubbed with alcohol until quitecrisp, collected, and recrystallised twice from absolute ethyl alcohol.Tetrab enzoylmethyltetritol crystallises in hard, colourless, prismaticneedles, which are readily soluble in ether, moderately so in boilingabsolute ethyl alcohol, but very sparingly so in the cold, andinsoluble in water. It melts at 136-137O:0,1560 gave 0.4094 GO, and 0-0736 H,O.C=71*60; H=5.27.C,H,O, requireg C=71.73; H=5.07 per cent.Methyl Tetrose.4.5 Grams of the phenylbenzylhydrazone were heated on thewater-bath with 10 C.C. of 40 per cent. formaldehyde in about80 C.C. of water. After two hours' heating the formaldehydephenyl-benzylhydrazone which had separated was removed by extractionwith ether. The aqueous solution was then concentrated at 35O/30 mm. to a syrup, water was added, and the solution again concen-trated, the process being continued until no odour of formaldehydecould be detected in the product. The tetrose was obtained as aviscid, pale yellow syrup, which could not be crystallised; the yieldwas 1 gram. The substance is readily soluble in water or absolutealcohol, and reduces Fehling's solution vigorously.After drying at 60°/50 mm.for several hours the substance wasanalysed :0.1405 gave 0.2301 CO, and 0.0932 H,O. C=44.65; Ei=7*42.C5HloO, requires C = 44-77 ; H = 7.46 per cent.Oxidatiorc of Methyl Tetm'tol.Ten grains of the methyl tet'ritol were dissolved in 60 C.C. of water,and a solution of 2.5 grams of ferrous sulphate in a few C.C. ofwater was added. Sixty C.C. of hydrogen peroxide (3 per cent.)were then added slowly. The temperature of the mixture rose toabout 40°. After some hours the solution was shaken with bariumcarbonate, filtered, and concentrated in a vacuum to a syrup.The product was treated with 2.5 grams of phenylbenzyl80 CONSTITUTION O F THE METHYL PENTOSES. PART I.hydrazine in the manner previously described, and the hydrazoneseparated by extraction with benzene.A red syrup was obtained,which partly crystallised. After washing the crystalline matterwith benzene, and recrystallising from hot water, the phenylbenzyl-hydrazone of the methyl tetrose was obtained, and identified by itsmelting point, 99-looo. The yield was, however, very small.a& Dim e t .32 o x y vale r olac tone .Two grams of dihydroxyvalerolactone (1 mol.) were dissolved in8.6 grams of methyl iodide (4 mols.). To the solution 7 grams ofdry silver oxide (2 mols.) were added, and the mixture was heatedon the water-bath for five hours. Ether waB then added, and thesilver residues were removed. On evaporation of the ether theproduct crystallised a t once.After recrystallising from a mixtureof acetone and ether it was obtained pure. It crystallises in hard,colourless prisms, which melt a t 59-60°:0.1982 gave 0.3792 CO, and 0.1312 H,O. C=52*22; H=7.40.C7H,,0, requires @= 52-50 ; H = 7.05 per cent.An attempt was made to reduce the methylated lactone in theusual way, but the resulting solution had no action on Fehling'ssolution, so evidently no reduction takes place. The solution wasneutralised with sodium carbonate, evaporated to dryness, and theresidue extracted with alcohol. The alcoholic extract on evapor%tion left a residue of sodium y-hydroxy-a~-dimethoxyvat?eric acid :0.0774 gave 0.0278 Na,SO,. Na= 11.62.C7HI3O,Na requires Na= 11.50 per cent.Brucine Salt of Methyl Tetronic Acid.Attempts were made to effect a resolution of the dihydroxy-valerolactone by means of brucine.The brucine salts of the d- andZ-forms of the acid unfortunately form a partially racemic com-pound, which cannot be resolved by crystallisation, and has theconstant specific rotatory power [a]:' - 29'6O.Proof that the salt is coniposed of one molecule of the d-saltand one molecule of the 2-salt is afforded by the fact that therotation of the solution obtained by boiling a solution containingone equivalent of lactone and one equivalent of brucine, is almostexactly the same, namely, [a]: -29'2O.The sctlt crystallises in fine needles, which are very soluble inwater, but only sparingly so in absolute alcohol, even at theboiling point. It melts at 180-181O:0.2335 gave 0.5256 CO, and 0.1374 H20. C=61*41; H=6.58.C,H1006,C23H3604N2 requires C = 61.76 ; H= 6-61 per centTHE ROTATORY~DISPEKSIVE-POWER OF ORGANIC COMPOUNDS. 8 II n conclusion, I desire to thank Professor J. C. Irvine for theinterest which he has taken in this work. I have also to thankDr. H. J. H. Fenton, of Cambridge, for the experimental detailsof his method of oxidising sugar alcohols; Professor Letts forextending to me the hospitality of his laboratory during the finalstages of this work; and the Carnegie Trust for a research grantwhich partly defrayed the expenses of the investigation.CHEMICAL RESEARCH LABOKATOKY,UNII‘ED COLLEGE OF d ~ . SALVATOR AND ST. LEONARD,UNIVERSITY OF ST, ANDREWS.THE S I R DONALD CUltBIE LAl:ORATOItIES,QUEEN’S UNIVERSITY, ~~J:LF-AS’I-