U. S. Department of Commerce Research Paper RP1910 National Bureau of Standards Volume 41, August 1948 Part of the Journal of Research of the National Bureau of Standards

Barium 2-Ketolactobionate and the Corresponding Barium Bromide Double Salt 12 By William W. Walton and Horace S. Isbell

As a preliminary step in the development of methods for the preparation of 2-ketoaldo­ bionic acids for use in the synthesis of glycosidi c derivatives of ascorbic acid, t he preparation of 2-ketolactobionic acid has been studied. By oxidation of lactose osone with bromine in the presence of barium carbonate, barium 2-ketolactobionate was obtain ed and separated in the crystalline state. In addition to the normal barium salt, a double salt con taining barium bromidc was found. This sa lt is uniq ue in that i t is the first a nd onl y crystalline barium bromide salt of a ugar acid reported in the literature. It crystalli zes freely and is useful for the eparation and identifi cation of 2-ketolactobionic acid .

I. Introduction development of methods for their production. The most generally used method consists of oxidation In consideration of the relfttionship between the of ketoses with potassium permanganate. With carbohydrates and vitamin C and the prevalence this method, it is necessary to protect the hydroxyl of glycosides in natural products, it seemed de­ groups other than those attached to carbons one su'able to undertake the preparation and study of and two, as for instance by preparing acetone de­ some glycosidic derivatives of ascorbic acid. The rivatives. It has been found, however, that L-sor­ most natural course for the synthesis of the glyco­ bose [4] and D-fructose [5] can be oxidized directly sIdic derivatives of ascorbic acid is through the with nitric acid to 2-keto-L-gluconic acid and 2- intermediate production of 2-ketoaldobionic acids. keto-D-gluconic acid, respectively. In some cases The esters of cer tain 2-keto acids undergo lactoni­ the aldonic acids or their esters can be oxidized, zation and enolization with the formation of a - and the keto acid or its ester can be obtained di­ cOl'bic acids [1 , 2, 3].3 Thus on e might expect rectly. This bas been accomplished by several that methyl 2-ketomelibionate on treatment wi th methods, notably by anodic oxidation of a salt of sodium methylate in methanol would give the the aldonic acid [6] and by oxidation of the aldonic following reaction: acid, it.s lactone, or es ter with a chlorate in the R H R OR presence of vanadium pentoxide [7] . The latter ·1 1 1 1 NaOCH 3 method requires the presence of an acid and is C6H II06- C-C--C--C--C- C- O C l h -----> 1 1 1 I , II II CH30H not suitable for the preparation of keto acids con­ R OR OR ROO taining easily hydrolyzable glycosidic groups. Methyl 2-ketomelibionate A method that do es not require the presence of an acid con sists of the production of the sugar H H R OH OR osone followed by oxidation with bromine in sub- 1 1 1 1 1 C6H II 0 6-C- C--C- C= C--C= O ~ 6R ~ o -_I 1 Abstracted rrom a t hesis submittcd by William W . Wa lton in June 194 7 to the raculty of the Gra<;l uatc School or the Uni ve rsity of :Ma ryland, repre­ se nted by Nathan L . Drake, in parLia l fulfillment of t he requirements for the 6-galactosylisoascorbic acid degrce of Doctor of P hiloso pby. ' Presented before the D ivision of Sugar Chemistry and Tech nology or tho The facility with which esters of 2-ketoaldonic American Chemical SOC iety at NelV York, Sept. J947. 3 l?igurcs in brackets indicate the literature TC'fercn ces at Lbo end of this acids yield a cOl'bic acid derivatives has led to the paper.

Barium 2-Ketolactobionate 119 stantially neutral solution. The procedure is suitable for use with osones derived from both monosaccharides and disaccharides, and it has been applied to a number of oson es including maltosone [8]. The product from maltosone was reported to be amorphous and was undoubtedly an impure salt of 2-ketomaltobionie acid. As mentioned before, the enolization and lactoniza­ tion of certain 2-ketoaldobionie aClds would be expected to yield ascorbic acid derivatives of the desired type. 2-Ketomaltobionic acid and 2-keto­ lactobionic acid, however , contain substituents on the fourth carbon atom that prevent ring closure to that carbon atom. For this reason they cannot yicld ascorbic acid derivatives. N evertheless, on account of the ready availability of lactose it was selec ted for preliminary study of oxidation methods.

II. Barium 2-Ketolactobionate and Its Barium Bromide Double Salt

As an in termediate in the preparation of 2-keto­ lactobionic acid, it was necessary to prepare lac­ .\ tose osone. Ordinarily osones are obtained by reaction of phfmylhydrazine with the sugar fol­ lowed by ,treatment of the r esulting phenylosazone with b e r\ '~al'd. ehyd e [9] to remove the phenylhydra­ zine . . Ip some cases osones have beon .' ob tain ed by dU'ett oxidation of the sugar with copper acetate [10] ; but this method is not applicable to disaccharides, b ecause the reaction is slow and considerable hydrolysis takes place. H ence, the first method was adopted. ' The lactose oson e was oxidized with bromine in the presence of barium carbonate, the r esulting barium bromide was re­ moved by treatment with silver sulphate, and barium 2-ketolactobionate was crystallized from the solution. Barium 2-ketolactobionate is the F IGURE 1. A , Barium 2-ketolactobionate dihydrate, first crystalline metallic salt of a 2-ketoaldobionic [a Jb' = - 3 7.9°, B , Barium 2-ketolactobionate-barium bro­ acid to be prepared. It is quite soluble and does mide tetrahydm te, [aJb' = -28 ,2°.(X250). not cl:ystallize easily . It was found, however, that barlum 2-ketolactobionate forms a double The double salt is particularly suitable for the salt with barium bromide, which crystallizes separa'Lion of 2-ketolactobionic acid from the readily. The salt appears to be somewhat anal­ mixture oMained by the oxidation of lactose osone ogous to calcium lactobionate-calcium bromide wit h bromine because it crystallizes fr eely from [11] . Nevertheless, it is unique in that it is th e the impure mixture, and the barium bromide first and only crystalline salt of a sugar acid to necessary for the formation of tbe salt is a con tain barium bromide. Photomicrographs of byproduct of the reaction. The salt is charac­ the normal salt and the double salt are shown in teri stic of th e 2-ketolactobi onic acid structure figure 1. and should be useful for identification purposes.

120 Journal of Research III. Structure of Barium 2-Ketolactobio­ draw a conclusion concerning the ring structure nate and its Barium Bromide Double of the crystalline barium salt by analogy to imiJar Salt compounds. It has been shown previously that an equilibrium To establish that tbe normal barium 2-keto­ state involving both pyranose and furanose modi­ lactobionate and the barium bromide double salt fi cations is particularly sensitive to changes in are derivatives of the same 2-ketoaldobionic acid, temperature, and that this equilibrium can be the normal salt was prepared from the double studied by measurements of optical rotation at salt by r emoval of the barium bromide with various temperatures. A compari on of the silver sulfate followed by crystallization. Further­ equilibrium rotations of barium 2-ketolactobionate, more, by the addition of a mole of barium bromide, lactulose, and levulose at two temperatures is given the double salt was regenerated from the normal in table 1. In each case the optical rotation is salt. appreciably altered by a change in temperature, Since the synthesis began with lactose, (4-{3-n­ and in the same direction. The like behavior of galactosyl-n-glucose), in order to establish the the compounds is evidence that they establish structure of the new compound it was necessary similar equilibrium states. H ence, barium 2- (1) to prove that the biose linkage was still ketolactobionate, like fru ctose and lactu10se, present ltnd (2) to identify the products of hydrol­ establishes an equilibrium tate containing sub­ ysis as D-galactose and 2-ke .;0-D-gluconic acid. stantial amounts of both pyranose and furanose Enzymatic hydrolysis of the substance with a modifications. The lack of a more extensive sample of lactase 4 es tablished the existence of mutarotation for the salt must be explained by the biose linlmge. Galactose was separated from changes too rapid to be detected, or by the pres­ the products of hydrolysis and identified by its ence of more than one modification of the sugar melting point and optical rotation . Its presence entity in the barium salt at hand. in the hydrolytic mixture was also shown by the

action of a gaJactose-fermenting yeast. The TABLE 1. Compm'ison of optical Totalion 2-ketogluconic acid was identified by separation of t be characteristic phenylhydrazine salt of the Equilibrium rotation phenylbydrazone. Substance This experimental work shows that the new compound is the barium salt of 4-{3-D-galactosyl-2- Barium 2·kctolaciobiona te ______"-42. 3 -37.9 keto-D-gluconic acid (2-ketolactobioni c acid). The I,actulosc ______. _____ .___ -24.8 -23.0 Lcvulosc _____ .______- 46.6 -42.7 freshly dissolved salt gave a specific rotation of -36.4°,4 minutes after dissolution, which changed Ii Tcmoeraturc 0.6 0 c. to - 37. go in 45 minutes. The Jack of a more extensive mutarotation is peculiar in light of the similarity of the substance to fructose and lactu­ IV. Experimental Details lose. 2-Ketolactobionic acid, like lactulose, has 1. Preparation of Lactose Phenylosazone the fructose structure; it differs from lactulose (4-{3-D-galactosyl-n-fructose) in having a carboxyl One hundred and eighty grams of lactose was group in place of the terminal hydroxymethyl dissolved in 1 liter of water, and 164 ml of phenyl­ group. Fructose crystallizes in the pyranose form hydrazine and 328 m1 of a 50-percent aqueous but establishes an equilibrium containing a sub­ solution of acetic acid were added. The mixture stantial proportion of the furanose modification. was heated in a water bath for 2 hours at 80 00, Lactulose cr ystallizes in the furanose form, but cooled, and stored in a refrigerator overnight. its equilibrium state is very similar to that of The osazone was removed by filtration, washed fructose [12] . Since crystalline ugars havin g the fru ctose structure are thus known to exist in both with dilute acetic acid, water, and finally with furanose and pyranose forms, it is not possible to ether. The product was recrystallized from 20- percen t aqueous alcohol. The yield was 38 • T he lactase wa' ki ndly supplied by tbe r",carch labo miorics of Rohm & B aas, PhilHdclphia. Pa. percent.

Barium 2-Ketolactobionate 121 2 . Preparation of Lactose O sone bromide was obtained. Analysis: Calculated for Seventy grams of lactose phenylosazone was dis­ Ba(C12H 190 12)2.BaBr2. 4H20 : 0 , 23.7 ; H, 3.8 ; Ba, solved in 6 liters of hot water, 56 g of benzaldehyde 22'.6; Br, 13. 1. Found: C, 23.9 ; H , 3.8 ; Ba, 22 .6; was added, and the mixture was stirred vigorously Br, 13.2 . At equilibrium, the specifi c rotation while being heated on the steam bath in an atmos­ [aj ~ =- 2 8 .2 ° (water c= 2). phere of nitrogen for 4 hours. The solution was 4. Preparation of Barium 2-Ketolactobionate cooled and filtered and the filtrate extracted several times with ether to remove excess benzaldehyde. Barium 2-ketolactobionate was prepared by The aqueous solution was then treated with a oxidation of lactose osone in a manner analogous decolorizing carbon, filtered and con centrated to the preparation of tbe double salt. However, under reduced pressure to a volume of 800 ml. after treatment of the oxidized solution with car­ The yield on several preparations was about bon , the barium bromide was removed from the 65 percent. filtrate by treatment with silver sulfate and fil tra­ tion . The solutioI' was con centrated to about 3 . Preparation of Barium 2-Ketolactobionate­ 50 ml under reduced pressm e, and 50 ml of Barium Bromide methanol was added. The solution was seeded The aqueous solution of lactose osone obtained wi th crystals of barium 2-ketolactobionate and above, 800 ml in volume, was saturated with car­ allowed to crystallize for 2 days. bon

122 Journal of Research (b) Identification of Galactose (c) Identification of 2-Ketogluconic Acid The hydrolyzed solution obtained in (a) was The olution from (b) in which the galactose had separated from the toluene and filtered. A small been partially de troyed by fermentation wa fLl­ quantity of a galactose-ferm.enting yeast and 1 ml tered and concentrated to a volume of 4 ml. Five of a nutrient medium wrre added, and the solu­ drop of acetic acid and 5 drops of phenylhydrazine tion was stored at 37 ° C for 24 hours. The were added, and the solution was seeded with optical rotation indicated the destruction of 18 crystals of the phenylhydrazine salt of the phenyl­ percent of the theoretical amount of galactose in hydrazone of 2-ketoglueonic acid. Crystals formed 1 day. This solution was reserved for part (c). readily, and were separated by filtration, washed In a separate experiment, 2 g of barium 2-keto­ with ether, and dried. The melting point was lactobionate was hydrolyzed by dissolving it in 95 0 to 97° C with decornpositlOn. The melting 25 ml of 7 .5-porcen t sulfuric acid and heating the point of the phenylhydrazine salt prepared from solution on the steam bath for 4 hours. About authentic 2-ketogluconic acid was 96° to 98 ° C 2.5 g of calcium carbonate was added to the cool with decomposition. A mixed melting point of solution , and the mixture was allowed to stand the two preparations was 95° to 97 0 C. ovornigh t. It was then filtered and passed suc­ cessively tlu-ough a cation exchange resin and an V. References anion exchange resin to remove calcium 2-1,eto­ glueonate. The filtrate wa evaporated to a sirup, [1] T. Reich tcin and A. Grilssner, H elY . chim. Acta, 17, 311 (1934) . 10 ml of methanol was added, and th e solution [21 K. Maurer and B. Sch ieclt, Ber. deut. chem. Ges. was seeded with galactose and allowed to crystal­ 66, 1054 (1933). lize overnight. The crystals were separated, [3] T . Reichstein, H elY. chim . Acta. 17, 996, 1003 (1934). dried over calcium chloride, an.d weighed. A [4] British Patent 443.901 ( ept. 1, 1934) . yield of 51 percent of the theoretical amount of [5] H . Ohle, H. Erlbach, and H. Carl s, Ber. deut. chem. . Ges. 67, 324, 555 (19 34) . galactose was obtained. The material was re­ [6] R. Pasternack and P. R egna, U. S. Patent 2,222, 155 crystallized from a water-methanol solution and (Noy. 19, 1940). the melting point found to be 165° C. Muta­ [7] R. Pasternack and P. Regna, U. S. Patent 2,207,991 rotation measurements were made, and the (July 16, 1940) . following valu es were found : initial rotation [8] T . Kitasato, Biochem. Z. 207, 217 (1929). [91 E. Fischer, Liebigs Ann. Chem . 288, 145 (1895) ; Ber. [al ~o = 140.5 °; fin al rotation [al~o = 80.2 0; muta­ deut. chem. Ges. 35, 3142 (1902) . rotation constan t 0.009. The corresponding [101 1. Stone, U. S. Patent 2,206,374 (July 2,1940). values reported for galactose are: melting point, [11] H. S. Isbell, J. Research NBS 17, 331 (1936) RP914. 165° to 167 ° C; initial rotation, [al~o = 150.7° ; [12] H. S. Isbell and W. W. Pigman, J . Research NBS final rotation [al ~o = 80.20; mutarotation con­ 20, 773 (1938) RP1104. stant 0.008 to 0.009 depending on the carbon dioxide content of the water. W ASHING'l'ON, April 22, 1948.

Barium 2-Ketolactobionate 123