Aug. 20, 1954
ISOMER
DISTRIBUTION DURING FISCHER METHYLGALACTOSIDE FORMATION 4103
vessel were transferred to a 125 ml. flask with 15 ml. of water and acidified by addition of 0.1 ml. of 0.5 N sulfuric acid. After five minutes, 1.5 g. of sodium bicarbonate, 5 ml. of 0.03 M sodium arsenite and 1.0 ml. of 10% potassium iodide were added in succession.lO Twenty minutes later excess arsenite was titrated with 0.01 N iodine. A glycogen sample of 10.8 mg. required 12.9 ml. of iodine, which corresponded to a periodate consumption of 0.97 mole per mole. The pH was measured by transferring the contents of a vessel after equilibration, or a t the end of a run, t o a pHmeter having glass electrodes. A fine stream of carbon dioxide gas was passed into the solution during the measurement making it difficult to obtain an accurate reading. The pH values quoted are therefore approximate. Estimation by Titration of the Formic Acid Produced in Periodate Oxidation (Figs. 2 and 3).-In a typical experiment 100 mg. of glycogen, dissolved in 25 ml. of water, was
[CONTRIBUTION FROM
THE
treated with 0.2 M sodium metaperiodate (25 ml.), in the dark a t 16.6'. At suitable intervals aliquots (10 ml.) of the solution and of a reagent blank were withdrawn and, 20 minutes after the addition of 0.5 ml. of ethylene glycol and 2.0 ml. of 10% potassium iodide, were titrated with 0.02 N sodium thiosulfate.18
Acknowledgment.-The technical assistance of Mr. J. Giroux is gratefully acknowledged. The author is indebted to Prof. F. Smith, University of Minnesota, for his kind interest in this work and for the gift of glycogen samples. Sincere gratitude is also expressed to several colleagues in this Division for advice in the use of the Warburg respirometer and for providing apparatus. OTTAWA,CANADA
DEPARTMENT O F CHEMISTRY, TRINITY COLLEGE]
Chromatographic Adsorption. 111. Investigation of the Isomer Distribution during Fischer Methyl D-Galactoside Formation BY DWIGHTF. MOWERY, J R . , ~ AND GERALD €2. FERRANTE~ RECEIVED JANUARY 4, 1954 The changing distribution of isomers during the formation of methyl D-galactosides by the Fischer method a t room and reI t was found t h a t &isomers are formed first and change t o a-isomers, principally a-D-galactopyranoside, the change being accelerated a t higher temperatures or hydrogen chloride concentrations. The change of furanosides t o pyranosides which takes place simultaneously is, contrary to the commonly accepted view, a less important reaction, at least in this case. Conditions under which maximum yields of the various isomers may be expected are included.
flux temperatures using 0.5 and 4% hydrogen chloride was studied by means of chromatography and polarimetry.
Introduction aration of sugar mixtures' and that with the proper pretreatment* essentially 100% recovery of the inIn the past, study of the rearrangements taking place during glycoside formation by the Fischer dividual sugars can be expected. It was thought method has been seriously handicapped by the t h a t if methyl galactosides could be separated by lack of easy reliable methods by which the various this adsorbent, a quantitative method for their deisomers can be quantitatively determined. An termination might be realized. I n addition, it early investigation of this reaction by Levene, was hoped that data might be obtained for a comRaymond and Dillon3 indicated t h a t furanosides parison of the relative capacities of Florex and were formed first and changed slowly to pyrano- cellulose columns for the separation of methyl galacsides as the reaction continued. Until very recently tosides. I n the present investigation, methyl galactoside no good method for the quantitative determination mixtures formed under various conditions of hydroof a- and p-isomers has been known and therefore changes from a- to ,&forms or vice versa could not gen chloride concentration and temperature were be followed. I n 1946 Binkley and Wolfrom4 re- analyzed by passage in methanol through a Florex ported the chromatographic separation of the XXX column by a procedure previously6 described. anomeric forms of the penta-0-acetyl-D-glucopyran- The p-isomers, with negative rotations in methanol, oses. More recently Hough, Jones and Wadman5 were eluted from the column first, followed by the reported the chromatographic separation of methyl a-isomers, with positive rotations. Although sepa- and 0-L-rhamnopyranosides. Very recently aration of furanosides from pyranosides was not Augestad, Berner and Weigner6 have reported the possible with this adsorbent, a complete separation chromatographic separation of methyl fructosides of a- and /3-isomers was obtained. The distribuand methyl galactosides using a powdered cellulose tion of furanoside and pyranoside in each fraction column. It has been found t h a t Florex XXX, a can be calculated from the specific rotations of the fuller's earth type of adsorbent, can be used for sep- two components, the weight of the fraction, and the m1.-degree area under the elution curve provided (1) Ripon College, Ripon, Wisconsin. only that no more than traces of other substances (2) Taken in part from a thesis submitted by Gerald R. Ferrante in are present. This is believed to be the case, as the partial fulfillment of the requirements for the M.S. degree. most important by-products of the reaction, D(3) P. A. Levene, A. L. Raymond and R. T. Dillon, J. B i d . Chcm., 96, 699 (1952). galactose dimethyl acetal and the methyl galacto(4) W. W. Binkley and hI. L. Wolfrom, THIS JOURNAL, 68, 1720 septanosides would be expected in very small (1 946). amounts only. D-Galactose dimethyl acetal, for (5) L. Hough, J. K. N. Jones and W. H. Wadman, J . Chcm. Soc., 1702 (1950). (6) I. Augestad, E. Berner and E. Weigner, Chemistry & Indurlry, 376 (1953).
(7) B.W. Lew, M. L. Wolfrom and R. M. Goepp, Jr., TEISJOURNAL, 68, 1449 (1946). ( 8 ) D. F. Mowery, Jr., ibid., 73, 5047, 5049 (1951).
4104
DWIGHT F.nlOK'ERY,
JR., .\ND
GERALD It. FERRANTE
Vol.
i(j
ringlo and is as sensitive t o acid hydrolysis as a fivemembered ring. l1 None of these substances have ever been isolated from reactions producing glycosides by the Fischer method.
Experimental
t \
c
Reagents.-The D-galactose was Pfanstichl C.P. material. Hydrogen chloride was obtained from a cylinder and dried by passage through a column of calcium chloride. The weakly basic ion-exchange resin, Deacidite, WLS obtninrtl from the Permutit Company. Apparatus.-The chromatographic coluinn was ii 1 .X x 122 cm. column of air blown Florex XXX constructed and operated i n a tnanner previously8 described. %3th absolute methanol as developer, pressures of about 1 atni. wcrc sufficient t o produce an effluent of 20-25 ml. per minute from this size column. n'ith pressures of this magilituric :t 3.5-gallon Pyres bottle provided a convenieiir i o l v c ~ i t reservoir. Determination of Column Capacity.-T!venty-s of D-galactose, suspended in one liter of methanol contaiiiing 0.5% of dry hydrogen chloride, was stirred a t room or rcflux temperature for about 70 hours. The clear solutiori was neutralized with silver carbonate, filtered and evaporated t o a thick sirup. The sirup ivas divided into five portions of 2, 3, 5, 7 and 10 g. Each portion was run through the chromatographic column and the usual plot of olitical activity 18.7. mi. of effluent was made. Assuming coinplcte s-paration into positively and negatively rotating fractioii.; for the 2 g. charges, the areas t o be expected upon rotnpletc scparation of the other charges should be proportioiiatcl!. larger and the % separation of the larger charges can thcrefore be calculated as 100 X observed arealcalculated a r w , I t had previously been determined that the specific rotation of methyl a-D-galactopyranoside in methanol changes v c r y little with concentration or small temperature changeq and it \vas assumed t h a t the other methyl galactosides would behave similar]>-. The results of these experimeuts are summarized in Table I and indicate t h a t g . of a predominantly negative mixture of methyl galactosides and 3 g. of a predominantly positive mixture can be complctcl!. selxirated b y the chromatographic column used in this work.
4
/
_'
TABLE I DETERMINATIOS OF C o ~ r CAPACITY ~ s Conditions
Charge, fi.
Room temp.
2 :i 1
I
c
I
1
10
I
-.
0 t
I
v'
-
19
t
I
2 3
5 I
in
0
IO
1000
Reflux temp.
1500 Eluant, ml.
2000
Fig. 1.-Polarimetric curves sho\ving elution of methyl a- and P-D-galactosides from t h e Florex XXX Column.
instance, has been found9 to be very rapidly converted to methyl galactosides in acidified methanol and a seven-membered ring has considerably less tendency to form than 3 five- or six-membered ( 9 , H. A Campbell a n d K P Link. J Bud C h c > n . , 122, iiRZ (19881; 11. I, \Volfrum and S \V. \Vxi>brut, T H I SJOLRNAI.. 61, 1-408 (lU391.
Area, sq. 111. Ohsd. Cnlcd.
0 81 1.22 2.01
2.79 3.76 0 92 1.41 2.10 2.92 4.03
1.21 2.03 2.83 4.07
1.38 2 30 3.24 1 GO
SCJIi1r:i
lion,
5;
101 100
08 $12 102 $11 00
8S
Preparation of Methyl Galactoside Mixtures.- hlcth!.I galactoside mixtures were made by dissolving 3.0 g. of I)galactose in 400 ml. of absolute methanol a t room teinl)er:tture or 200 ml. :it reflux temperAture. One-h:llf or 4 % of dry hydrogen chloride w:ii then introduced and the nlisturi3 allowed t o stand or refluxed until the reducing sugLtr : t ~ i a l y sisi*indicated that only traces of galactose remained. SOITIV of the mixtures were worked up a t this point while others were allowed t o continue for longer periods of time. The reaction was stopped b y neutralization of the hydrochloric acid by means of silver carbonate or pas5agi. through a. I)?acidite column. The clear solution was evaporated undcr vacuum a t R O O t o a thick sirup, which \vas weighed to thc nearest 0.1 g. The sirup wa? then tr;tnsfcrred rjumititiitively t o the top of the chromatographic colulnn by mentis of absolute methanol a n d run don-n into thr. ntlsorbcnt utit1c.r ( I O ) G. W.W'heland. " A d v a n c s d Organic C h e m i s t r y , " J c ~ i 1 1 1 \\.iI?> a n d Sons, Inc., S e w Y o r k . S.I-,1$14q, pi) 3 7 2 . i l l ) F Micheel a n d F Suckfull, B o . , 66B, 1!137 ( l ! l A ~ . (12) L. J . H e i d i and I- \V. Southam, 'lriis J O U K N Z I . , 7 2 , 390 < I ' ! l X J ) .
Aug. 20, 1934
ISOMER
DISTRIBUTION DURING FISCHER hlETHYLGALACTOSIDE FORMATION 4105 TABLE I1
Run no.
DISTRIBUTION OF METHYLGALACTOSIDE ISOMERS PRODUCED UNDERYARIOCSCOSDITIONS Pos. fract. or-Isomers, % Keg. fract. @-Isomers,R T o t a l , c/c Conditions
Wt.a
Areab
Total
Fur.
Pyr.
Wt a
Areab.
Total
Fur.
Pyr.
1 2 3 4
0 . 5 % HCl r.t. 6 hr. 48 240 940
0.5 0.9 1.7 2.3
60 126 344 606
17 29 49 79
17 29 33 30
0 0 16 49
2.5 2.2 1.8 0.6
320 270 220 66
83 71 51 21
33 26 19 7
50 45 32 14
5 6
0 . 5 % HC1 refl. 3 hr. 87
1.6 2.3
354 564
52 70
30 33
22 37
1.5 1.0
220 94
48 30
23 7
7
4% HC1 r.t. 5 hr.
75
1.2 2.6
214 624
38 87
30 43
8 44
2.0 0.4
258 54
62 13
25
8
14 9 0.6 73 54 0.6 39 40 Area in m1.-degrees = 200
18 21
4yoHC1 refl. 3 hr. 20
82 2.8 440 79 2.3 554 1(I W t . of fraction in E, after evaDoration in vacuum. from plot. $1
I
1 a t m . pressure. Methanol was then run through the column and the rotation of the effluent, determilied a t 50mi. intervals, was plotted in Fig. 1. The effluent was separated into positively and negatively rotating fractions and each fraction evaporated under vacuum and weighed. Recovery of material from the column was 9770 or better. Compositions of Chromatographic Fractions.-Methyl p-~-galactopyranoside'~ of specific rotation 0' (c 0.69)14 in water and - 17" ( c 1.52) in methanol was crystallized from the negatively rotating fractions of the effluent. Methyl a-D-galactopyranoside monohydrate'5 of specific rotation +178" (c 0.73) in water and f171" (c 1.20) in methanol was crystallized from the positively rotating fractions. Methyl 8-D-galactofuranosidel" of specific rotation - 110" (c 0.68) in water and - 137" (c 0.71) in methanol was crystallized from a negatively rotating fraction from the chromatographic column using as charge a furanoside mixture prepared from galactose diethyl mercaptal b y the method of Pacsu and Green.le The fraction from which this isomer was crystallized appeared at the same point in the effluent as the negative fractions from the Fischer reactions and it is therefore assumed t h a t in all cases the negative fractions consist of the two negatively rotating methyl p-D-galactosides. In addition, the specific rotations of all the sirups obtained from the negative fractions indicated the presence of a n isomer having a larger negative rotation than the methyl $-Dgalactopyranoside actually isolated. The above crystalline methyl galactosides were obtained from absolute ethanol or ethyl acetate solutions after removal of methanol and water by several vacuum evaporations with absolute ethanol. The specific rotations of the positive fractions indicated the presence of an isomer of specific rotation less than +70° in methanol. This was assumed t o be the methyl a-D-galactofuranoside which, although not isolated, could be calculated by means of Hudson's isorotation rules t o have a specific rotation of +87" in water and +68' in methanol. It has been found t h a t Hudson's rules hold very exactly for the methyl glucosides, ethyl glucosides and ethyl galactosides and they should therefore hold also for the methyl galactosides. A specific rotation of +104' has recently been reported6 for methyl or-D-galactofuranoside in water, but in view of the divergence of this figure from that predicted by Hudson's rules its validity is questionable until further substantiated. Calculations of Isomer Distributions.-The distributions of methyl a- and P-D-galactofuranosides and pyranosides produced b y 0.5 and 4% hydrogen chloride in methanol (13) C. N. Riiber, J. Minsaas a n d R. T. Lyche, J . Chem. Soc., 2173
(1929). (14) All specific rotations in this paper are given for t h e o-line of sodium a t 20'; c = g. of sugar per 100 ml. of solution. (15) M. Voss, A n n . , 486, 297 (1931). (16) E Pacsu a n d J. W . Green, THISJ O U R N A L , 69, 1205 (1937), 6 0 , ?05Ci (1938)
Fur.
Pyr.
50 52 37
50 45 48 63
25 23
54 40
46 60
37 8
55
5
45 52
4 5
14 16
77
55
48
44
23 56
+ area in sq. in. obtained by planimeter
solutions a t room m d reflux temperatures in various times are presented in Table 11. The total percentage of a-isomers was calculated as 100 (wt. of positive fractions)/(wt. of positive fraction a t . of negative fraction). Using +68" and 171" as specific rotations of methyl or-D-galactofuranoside and pyranoside mc nohydrate, respectively, in methanol, the proportion of furanoside in this fraction can be calculated as (171 - ml. deg. area/2 X fract. wt.)/103. The percentage e-furanoside and a-pyranoside can then be determined. In exactly comparable manner the percentage p-furanoside and 8-pyranoside can be calculated. In this case specific rotations of - 137" and - 17' for the 8-furanoside and pyranoside, respectively, lead t o the following proportion of furanoside: (ml. deg. area/2 X fract. wt. 17)/120. The last two columns in Table I1 give'the total % furanoside and pyranoside, calculated as the sum of the a- and @-isomersin each of these ring forms. Reducing Sugar and Furanoside Analyses .-Tu o methods of analysis were employed in following the course of the glycoside formation. Both of these involved the determination of reducing sugar and were the hypoiodite method used by Levene, et al.,3 and the cupritartrate method recently refined by Heidt and S3utham.l2 The latter method was found t o be more reliable than the hypoiodite procedure and was therefore used in most cases. A heating time of one hour was found most satisfactory for galactose analyses. Total furanoside was determined by the method of Levene, et al.3
+
Discussion of Resu'ts General Course of the Reaction.--A study of Table I1 and the graphs of the corresponding runs, reproduced in Fig. 1, shows that in the formation of methyl galactosides by the action of acidified rrethanol upon galactose, &isomers are formed first. These may amount to as much as 83% for short reaction times under mild conditions (run 1). As the reaction continues B-isomers change to a , the change being more rapid the higher the temperature or the greater the hydrogen chloride concentration. For example, with O.jg/,hydrogen chloride a t room temperature p-isomers drop from 83 to 2lY0 in about 3 weeks time (runs 1-4), whereas the change requires less than 3 hours with 4yc hydrogen chloride a t reflux temperature (runs 9 and 10). Intermediate conditions such as O.rjCr, hydrogen chloride a t reflux temperature or 4yc hydrogen chloride a t room temperature require intermediate times (runs 5-8). At the same time that p-isomers are changing to a a smaller change of furanosides to pyrano-
DWIGHTF.hlOWERY,
4106:
JR., AND
sides is taking place, this change reducing the furanoside content from 50 to 377, in 3 weeks with 0.57, hydrogen chloride a t room temperature and to 447, in 20 hours with hydrogen chloride a t reflux temperature. About 4 days are required for a similar change at the intermediate conditions of 0.5% hydrogen chloride a t reflux temperature or 47, hydrogen chloride a t room temperature. The CYpyranoside content is seen to increase from 0 to 4970 in 3 weeks a t room temperature with 0.5y0 hydrogen chloride, accounting for most of the p-isomer decrease. The a-pyranoside increase is faster under the more drastic conditions of temperature or acid concentration. The formation of a-furanoside seems to be somewhat erratic, first increasing and then decreasing in 0.5T0 hydrogen chloride a t room temperature, remaining almost constant for 87 hours in O.5y0hydrogcn chloride a t reflux teniperature, increasing in 3yo hydrogen chloride a t room temperature, and decreasing in 4y0hydrogen chloride a t reflux temperature. The data here are too incomplete to prove a definite relationship but seem to indicate the a-furanoside content may pass through a maximum and then decrease, the maximum being attained faster under more drastic conditions such as 456 hydrogen chloride a t reflux tensperature. Maximum Yields of Various Isomers.-Conditions under which maximum yields of the various isomers may be expected are listed in Table 111. COSDITIONS FOR Isomer
TABLE 111 ~ ~ A X I M L W YIELDS " OF ISOMERS Time, HCI, hr. %
Yield, Temp.
%
3 4 , 0 Reflux '73 Methyl a-D-galactofuranoside Methyl a-n-galactopyranoside 20 4 . 0 Refluxh 40+ 33 A 0 . 5 R.t. Methyl p-D-galactofuranoside 50 6 0 . 5 R.t. Methyl 8-D-galactopyranoside a For conditions investigated. Yield of this isomer appears t o increase with time and although all conditions eventually produce methyl a-D-galactopyranoside, these conditions produce a faster reaction and larger yields would be expected with longer reaction times than 20 hours.
GERALD R.FERRANTE
Vol. 76
able. If the acid treatment time is increased from 10 t o 20 minutes the furanoside analysis (45-hr. samples in Table IV) increases from 39 to 557, by the hypoiodite method and from 38 to 497, by the cupritartrate method. Apparently the hydrolysis time is so critical the method cannot be relied upon to give an accurate absolute furanoside estimation Galactose determinations of Levene, et al., could not be duplicated either by the hypoiodite method, which they used, or the cupritartrate method, used in the present work, although these two methods were themselves in good agreement (colurnns 2, 3, and 4 of Table IV). The discrepancy may lie in the initial concentration of galactose which Levene, et al., state to be 0 343 molal, whereas a t room temperature the solubility of galactose in methanol is about l, 4 of this concentration, indicating that they may have had some undissolved galactose a t the start of the reaction, which could account for their higher free sugar analyses. bIETHYL
Time, hr
TABLE 11. GALACTOSIDE FORMATIOY U S I Y G 0 52 ' CHLORIDE AT 25' Galactose, Levene Hypoel ai. iodite
0
101
I
75 46
3 7 24 48 48" 400
21 6 1
100 58
25 8 4
3
Cunritartrate
98 59 27 6
lIYDROGEU
Furanosides, % Levene, el a1 1
23 38
4
48 40
3
30
6 25 49 52 48
Cupri tartrate
0 15 42 50 44
39 55
38 49 0 19 a 20-minute hydrolysis instead of standard 10 miiiutc time.
Less satisfactory agreement is evident upon comparison of the results of the present work with a very recently published note by Augestad, Berner and Weigner,6 who used 0.0177, hydrogen chloride in methanol for 6 hours a t reflux temperature. They reported 15% methyl a-D-galactofuranoside, 20% methyl a-D-galactopyranoside, 507, methyl 6-Dgalactofuranoside and only a trace of methyl BAgreement of Results with Previous Work.The results obtained are in agreement with an early D-galactopyranoside. No run was made under observation by Jungius'7 that methyl P-D-galacto- exactly these conditions, run 5 using 0 5% hydrogen side in methanol containing hydrogen chloride chloride a t reflux temperature for 3 hours being the changes partially into methyl a-D-galactoside. most similar. The greatest discrepancy appears There is also agreement with the work of Levene, to be in the distribution of the p-isomers, the presRaymond and Dillon,3 who showed that with 0,5yo ent work indicating an approximately equal distrihydrogen chloride at room temperature the total bution between furanoside and pyranoside and the galactofuranoside passes through a maximum be- work of the above authors indicating principally tween 3 and 24 hours. Although the furanoside furanoside with very little pyranoside, a very unuanalysis by means of acid hydrolysis under care- sual situation in view of the present work. Since the fully standardized conditions as used by Levene, et above prepublication note lacks experimental detail and does not account for 15% of the glycoside a/., undoubtedly provides a satisfactory indication of the position of a furanoside maximum, the ab- mixture, a satisfactory reconciliation of the two solute values for furanoside content are question- methods cannot be made a t present. (17) C. L. Jungius, 2. p h y s i k . Chem.. 63, 97 (1905).
HARTFORD, CONN.
