U. S. Department of Commerce Research Paper RP2018 National Bureau of Standards Volume 43, August 1949

Part of the Journal of Research of the National Bureau of Standards

Mechanisms for the Formation of Acetylglycosides and Orthoesters from Acetylglycosyl Halides By Horace S. Isbell and Harriet 1. Frush

To obtain information concerning the stereomeric factors that affect the course of replacement reactions, a study has been made of the replacement of the halogen of certain acetyglycosyl bromides by methanol in the presence of silver carbonate (the Koenigs-Knorr reaction). The investigation shows that the solvent and the temperature greatly influence the course of the reaction when the configuration is trans but have little influence when the configuration is cis. The experimental data support the concept that under definite stereo­ meric conditions a solvated orthoester intermediate takes part in the reaction at low temper­ atures. At higher temperatures part of the reaction appears to take place t hrough a free or thoester ion. 'By selection of suitable experimental conditions, it was pos ible to direct the orthoester reaction in large measure to either t he methyl orthoacetate or t he methyl acetyglycoside wi th apparent retention of configuration.

1. Introduction ing acetyl group, with the addition of methanol One of the most useful reactions for the syn thesis at the carboxylate carbon and the elimination of a of glycosidic compounds consists of the inter­ proton from the resultant orthoester intermediate action of a sub tance containing a free hydroxyl (eq 1). The acetylglycoside of like configuration group wi th an acetylglycosyl halide in the presence to the parent halide is formed by a secondary of silver carbonate or other substance capable of reaction of the orthoe tel' intermediate with removing Lhe halide ion. The reaction, com­ me thanol at the glycosidic carbon (eq 2). The monly known as the Koeni gs-Knorr reaction acetyl glycoside of unlike configuration is formed [1], l leads to the formation of a number of products. by an intermolecular, opposite-face attack by the When the halogen and the neigh boring acetyl methanol from the environment upon the gly­ group of th e acetylglycosyl halide are trans, the cosidi c carbon wi th the elimination of a proton reaction with methanol in the presence of silver (eq 3). The last reaction is in competition with carbonate produces an orthoacetate I and the reactions involving the orthoester intermediate. anomeric acetylglycosides II and III. It was When the halogen and the neighboring acetyl pointed out in previous publications from this group of the acetylglycosyl halide are cis, intra­ laboratory [2, 3, 4] that the orthoester is formed molecular reaction is imposs ible, and no orthoes ter from trans acetylglycosyl halides by an intra­ is formed. R eplacement takes places by the molecular, opposite-face attack upon the glycosidic mechanism of eq 3, to yield the acetylglycoside carbon by the nucleophilic oxygen of the neighbor- of configuration opposite to that of the parent

I Fi ~ urc s in brockets indicate the literature references at thc clld of this halide. p"pcr. H C H aO H "'C/ CH3 0-6 I'" "" c/ 1""0 1 0 C H 3- O / "" 0 - 1 1 -1 Aco -r-I I III

Orthoester Reactions 161 H Br "'C/ H H I I ,0 "'0 O-C CHa O- C CHa-C ,/' 1 - --> CHa- +C / 0 '" / I '" (1) I'" / C'" 1 0 ROH/ " '" O- C- : '"O -C- I RO 1 I ,I I O-y- R01H I

(R = H, CHa, CH3C= 0, etc.) R o H / H H OR I It ' O- C '" C / (2) CHa- +C / I'" 0-----+ o i "'0 , '" 1 II : O- C- CH3 CO- y- i : I I ROH H Br RO H R0-- - '"-+C / "'c/ H 1"'0-----+ \"'0 (3) AcO-YI -1I ACO-y-i

Shortly after the publication [2] presenting these this paper show that at low temperatures the mechanisms for the replacement reactions of the reaction products vary as required by the solvated acetylglycosyl halides, Winstein and Buckles [5] orthoester intermediate rather than by the fref' advanced similar mechanisms to account for cer­ orthoester ion. tain reactions ou tside the carbohydrate field . In reactions of the orthoester type, they em­ phasized the independent existence of the orthoester intermediate in the form of the ion IV. Although their mechanism does not differ markedly from that already presented, the question as to whether or not the orthoester intermediate exists in soluti.on as a free ion or a solvated ion is of importance. In a recent publication , Vinstein, Hanson, and Grunwald [6] claim that results of IV a study of the acetolysis of trans-2-acetoxycyclo­ hexyl-p-toluenesulfonate rule out the mechanism II. Role of the Solvent in the for orthoester formation symbolized by V, and remark that this mechanism was envisioned by Koenigs-Knorr Reaction one of us for orthoester formation from acetohalo­ Presumably, the role of the solvent in the reac­ gen sugars. We believe that the evidence of tions under consideration is somewhat similar to Winstein and coworkers docs not rule out the its role in some of the more common ionization orthoester mechanism cited. In order to test the reactions. The ionization of an acid in water, for concept of a solvated orthoester in termediate, a instance, involves coordination of the acidic hydro­ quantitative study has been made of the Koenigs­ gen with water, to form an oxonium ion, accom­ Knorr reaction, and especially of the effect of panied by separation of the anion. Similarly, the solvents and temperature upon the composition ionization of an acetylglycosyl halide may involve of the reaction products. The results given in coordination of the solvent at a reaetive center,

162 Journal of Research the effect of which is to facilitate release of the 1. 2 The ether-containing intermediate, however, halogen anion. unable to become stabilized by elimination of a In a trans acetylglycosyl halide, coordination of proton, will exist in solution until decomposed a nucleophilic solvent can occur either at the by thermal agitation, or by reaction with eiLher glycosidic carbon (eq 3) or at the carboxylate methanol or water at the glycosidic carbon, in the carbon of the neighboring acetyl group (cq 1). manner depicted in eq 4. Thus the addition of Presumably, coordination of the solvent with the the solvent Y should favor the formation of either carboxylate carbon releases electrons to the acetyl the glycoside or the acetyl sugar having the same oxygen and thus increases its nucleophilic prop­ configuration as the parent acetylgyl cosyl halide. erties, and its capacity to combine with the glyco­ On the other hand, if the reaction takes place sidic carbon, with release of the halogen. When through the free orthoester ion of Winstein and the coordinating substance possesses an ionizable coworkers, the probability of the formation of hydrogen, the solvated orthoester ion can yield a either the orthoester or the acetylglycoside having neutral substance by elimination of the proton. the same configmation as the parent halide should If the coordinating solvent Y (in place of ROH) be substantially independent of the presence of a does not have an ionizable hydrogen, the orthoester solvent such as ether , whi ch supposedly serves intermed iate cannot yield a neutral substance by only as a diluent. elimination of a proton; hence it exists in solution The foregoing considerations apply only to until decomposed by thermal agitation or by a trans acetylglycosyl halides. We have already secondary reaction, as for instance with methanol noted that if the acrtylglycosyl halide has a cis (eq 4). The secondary reaction yields the glyco­ configuration for the substituents of carbons sid e that has the same configuration as the original 1 and 2, the orthoester mechanism is not appli­ acetylglycosyl halide, and the proportion of this cable, and one would expcr,t no appreciable effect glycoside should therefore be increased by the of the solvent upon the character of the product pre ence of the coordinating olven t, Y. formed. In a cis acetylglyco yl halide, stereomeric con­ ditions make reaction between the glycosidic carbon and the neighboring acetyl group impossi­ III. Effect of the Byproduct, Water, on the ble. H ence, the mechanism of eq 4 is not appli­ Koenigs-Knorr Reaction cable, and the presence of the coordinating sol­ vent, Y, should have no direct effect on the com­ In addition to the alpha and beta acetylgly­ position of the product. cosides and the orthoestcr, which are formed in the Koenigs-Knorr reaction, two other products C R 3 must be taken into consideration. These are the R 0 alpha and beta free acetylsugars, which are formed I ,//H from water inevitably present as a byproduct. CH 3 O- C Theoretically, 1 mole of water is produced from '"+C / I~ 0 2 moles of methanol (eq 5), and 1 mole of water Y/ '" O- C-I1 produces 2 moles of the free acetylsLlgar (eq 6). 1 If all of the water were to react with the acetyl­ glycosyl halide, the product would contain equal In the event that s(-'veral nucleophili c substances proport.ions of the free acetylsugar and the are available for coordination with a trans acetyl­ methoxy compounds. 3 The extent of the water glycosyl halide, one can envision a competitive reaction depends in part on the manner in which system in which several solvated orthoes ter the replacement is conducted. When the alcohol intermediates are present. For instance, in a mixtme of ether, methanol, and water, solvated 2 Decomposition o[ t he orthoacid formed [rom water would yield the free orthoester intermediates can be formed from acetylsugar, CH,OAc.CH.(CH .OAc)"CH,OH. 1_-0 -_1 each constituent. The orthoester intermediates 3 In tbe preparation of orthoesters, other workers have round it advantage. derivrd from methanol and water yield neutral ous to co nduct tbe reactiou in the presence of a drying agent, ordinarily Drierite (7] . In spite o[ tbe use o[ t he drying agent, reaction with water compounds by elimination of a proton, as in eq co uld not be entirely elim inated in the present study.

Orthoester Reactions 163 is added slowly to the acetylglycosyl halide in metrically that the beta modification, as antici­ another solvent, considerable opportunity is pated, predominates in the reaction product. provided for the byproduct, water, to enter into the reaction, because it is in competition with a IV. Discussion of Experimental Results relatively small quantity of methanol. For this 1. Effect of the Solvent in the Koenigs-Knorr reason, more of the free acetylsugar should be Reaction and Evidence for a Solvated Orthoester formed when a large part of the methanol is Intermediate replaced by ether or benzene. The present The results given in table 1 show that the com­ investigation has shown that this is indeed the case. position of the products from tetraacetylman­

2RBr+ Ag2COa+ 2CHaOH---? 2ROCHa+ 2AgBr+ nosyl bromide (a trans acetylglycosyl halide) H 20 + COZ' (5) depends on the character of the solvent and on the experimental conditions. When the solvent 2RBr+ Ag2COa+ H 20 ---? 2ROH + 2AgBr+ CO2• (6) is methanol, the product of the intramolecular reaction at 20° C is largely the methyl ortho­ The mechanisms that have been advanced on acetate; no alpha acetylglycoside is formed. The page 162 for the reactions of methanol with cis production of the or tho ester without the alpha and trans acetylglycosyl halides apply equally acetylglycoside shows that the orthoester inter­ well to the reactions with water. An inter­ mediate decomposes only by the reaction of eq molecular opposite-face attack on the glycosidic 1, and there is no evidence for reaction at the carbon of either cis or trans acetylglycosyl halides glycosidic carbon as would be expected if the yields the acetylsugar of opposite configuration to mechanism involved the free ion. The absence that of the parent halide, as shown in eq 3. Thus, of a measurable quantity of the alpha acetylgly­ an alpha acetylglycosyl halide yields the beta coside indicates that the orthoester intermediate modification of the free acetylsugar. The trans loses a proton rapidly, before reaction can occur at acetylglycosyl halide reacts also by the intra­ the glycosidic carbon by eq 2. The results also molecular orthoester mechanism of eq 1, but the show a small amount of the beta acetylglycoside, or tho acid is unstable and rearranges to give the free acetylsugar (eq 7).4 Therefore this mecha- presumably formed by the competing inter­ nism yields the same product as that produced molecular reaction of eq 3. When the methanol by the intermolecular, opposite-face reaction, is largely replaced by either ether or benzene, the namely, the beta acetylsugar. That is, the alpha proportion of the orthoester decreases, and a halide yields the beta acetylsugar by two mecha­ substantial quantity of the alpha acetylglycoside nisms. In the presence of a solvent such as appears. If the reaction took place through the ether, which cannot release a proton, decomposi­ free orthoester ion, one would not expect this tion of the solvated intermediate can yield the striking change. Thus, it appears that a solvated alpha acetylsugar by the mechanism of eq 4. orthoester intermediate enters into the reaction, as indicated in eq 4. The results in table 1 show HO H that the effect of benzene in causing formation of the alpha acetylglycoside is less than that of C / '" ether, but nevertheless considerable. The marked -----? (7) effect of ether upon the composition of the 1 "'0 CH3-CO-c- 1 reaction product is in accord with its well-known ~ I tendency to coordinate. Its basic properties arise from the unshared electrons of the oxygen. Since the free acetylsugars are interconvertible Although benzene is less basic than ether, its by the mutarotation reaction, the quantitative ability to coordinate likewise accounts for the determination of the alpha and beta modifications changes observed. The marked increase in the is difficult. In this work, the sum of the two proportion of the free acetylsugar when the amount modifications has been determined by a titration of methanol is low and the reaction is conducted method; and it has been ascertained polari- in either ether or benzene shows the effect of the

' The rearrangement of the orthoaeid to the hydroxy compound is essen. competition of the byproduct, water, for the tially tbe same as that outlined in [5]. reactive centers, as mentioned on page 163.

164 Journal of Research TABLE 1. Composition of products from reaction of acetylglycosyl bromides with methanol in th e presence of silver carbonate at 20° C

Optical rotation of Molecular composition of product I prod uct· Experi- ment SO lvent Addition of methanol la1'8 a t llumber Acetyl- Acetyl- Beta Alpha la] ~ at acetyl- acetyl- 60 min eQ uilib- sugar orthoester glycoside glycoside riUlll (MeOB) (CllCh)

P roduct from tetraacetyl-a-D-mannosyl bromide

Percent Percent Precent Percent L ___ __ 20 ml of methanoL ______Used as solvent. 6.9 78. 5 14 .6 0 -32.20 -27.70 IL ____ 20 111 1 of ether + 0.5 ml methanol. Included in solvent. 25.9 41. 0 11. 4 21. 7 -23.3 - 0. 7 IlL ___ 15 m! of ether ______0.5 ml of methanol in 5 ml of 32. 1 18.5 18.7 30. 7 - 16. 8 +7. 5 ether added dropwise over 2 min. I V _____ 15 ml of ether ______0.5 m! of methanol in 5 ml of 35.4 7.5 23. I 34. 0 -7. 3 +10.6 ether added dropwise over 18 min. V ______20 ml of benzene + 0. 5 m! of Included in solvent. 16.0 64.5 8. 1 11. 4 -32. 7 - 12. 3 methanol V1. ____ 15 ml. of benzene ______0.5 m! of metbanol in 5 ml of 28.7 36. 1 12. 9 22. 3 -16.0 + 0. 7 benzene added dropwisc ovel' 20 min.

Product from tetraaeetyl-a-D-glucosy! bromide

IX ____ 20 111 1 of methanol. ______Used as solvent. 5.7 ------93.8 0.5 - 17. 1 - 12. 5 X _____ 20 ml of eLher + 0.5 ml of meth- Included in solvent. 28.1 ------71. 9 0 -5.7 +8. 3 ano1. XL ___ 20 ml of benzene +0.5 ml of Inclu ded in solven t. 25.2 ------74.8 0 -9.9 +5.5 methanol I • S pecific rotation, based on the weight of Lh e acetylsuga rs, the acetylglycosides, and the orthoester formed from 2 g of the acetylglycosyl bromid e.

In contrast to the results obtained with a -D-glucosyl bromide mutarotate in the dextro tetraacetyl-a-D-mannosyl bromide, th e reaction direction. In other words, the beta modification products from tetraacetyl-a-D-glucosyl bromide of the free acetylsugars predominates in the prod­ (a cis halide) show considerably less variation with du cts, as required by the reaction mechani sms change of solvent. The eff ect of competition of previously discussed. the methanol and water for the glycosidi c carbon is evidenced in the increase of the free acetylsugar 2 . Effect of Temperature on the Koenigs-Knorr Reaction wh en the amount of the methanol is small. The absence of the methyl orthoacetate and the alpha In the preceding section it was emphasized that acetylglycoside in the reaction products clearly the intramolecular reaction of tetraacetylmannosyl supports the prior concep t that cis acetylglycosyl bromide with methanol at 20° C takes place halides are incapable of reaction by the orthoes ter through a solvated ortboester intermediate with mechanism. That is, th e r eaction takes place the exclusive formation of the methyl orthoacetate. almost entirely by an opposite-face attack on the Inasmuch as this result is at variance with the glycosidic carbon by nucleophilic substances in conclusion of Winstein and coworkers, based on a the environment (eq 3). Since the alpha acetyl­ study of the acetolysis of trans-2-acetoxycyclo­ glycoside is not found in the product, there is no h exyl p-toluenesulfonate at 75° and 100° C, it evidence for the coordination of ether or benzene seem ed desirable to inves tigate th e effect of tem­ with the glycosidic carbon, since such coordination perature on the course of th e Koenigs-Knorr would yield the alpha acetylglycosid e by a second reaction.5 The data of table 2 show tbat th e opposite-face attack. composition of th e product from tetraacetyl- In all cases, the reaction products from , It was observed previously that tbe form ation of orthoesters is favored tetraacetyl-a-D-mannosyl bromide and tetraacetyl- by low tem peratures 18].

Orthoester Reactions 165 TABLE 2. Ej)'ect of temperatUl'e (In composition of products from reaction of acetylglycosyl bj'omides with methanol m the presence of silver carbonate

Molec ular com position of product Optical rotation of prod uct

Tcmperature Experiment num bel' [a]jj1 at [a]jj1 at Acetylsugar Acetylortho· Beta aeetyl· Alpha acetyl· acetate glycoside glycoside 60 miu equilibrium (M e OR) (CR C!,)

Product from tetraaeetyl·a·D·manllosyl bromide

° C Percent Percent P ercent Percent VII ...... -15 6.2 78. 1 15.7 0 -35. 7° -29.0° I...... 20 6.9 78.5 14.6 0 -32. 2 -27.7 VIII...... 50 13.5 53.4 25.1 8. 0 -22.2 - 19. 8

Product from tetraacctyl·a·D·glucosyl bromide

X II ...... - 15 5.0 94. 3 0.7 -16.5 - 12.9 IX ...... 20 5.7 93.8 .5 -17. 1 -12.5 XIII...... 50 6.7 90.8 2.5 - 12.6 -8.6

glucosyl bromide did not change significan tly in from 0 to 8 percent, the latter from 16 to 25 the range from - 15 0 to 20° C; tbe reaction prod­ percent. The increase in the beta acetylmanno­ uct, aside from the acetylsugar ,6 was almost side is evidence that at the higher temperature the exclusively the beta acetyglucoside. Since an intermolecular r eaction is favored at the expense intramolecular reaction is stereomerically impos­ of the intramolecular. This may be due to the sible for tetraacetyl-a-D-glucosyl bromide, the fact t11at less favorable conditions exist at 50° amount of the beta acetylglucoside may be taken than at 20 ° C for the intramolecular reaction to as a measure of the intermolecular reaction with take place, because of alteration in the positions methanol. At 50° C, 2.5 percent of the alpha of the reactive groups within the molecule. Cer­ acetylglucoside was formed; the incr ease over tain conformations of the pyranose ring place the that at 20° C may be due to recemization through halogen and the acetyl group in stereomerically a carbonium ion (the SN I mechanism of Gleave, favorable positions for intramolecular reaction by Hughes, and Ingold [9]) . the opposite-face mechanism; others do not. At In contrast to the r esults obtained with tetra­ the higher temperature, increased kinetic dis­ acetylglucosyl bromide, the results with tetra­ turbance of the sugar molecule may alter ring acetylmannosyl bromide show a striking eff ect of conformation and thus affect the probability of temperature on the co urse of the reaction at reaction by th e ortboes ter mechanism. 50° C/ although the composition of tbe product The increase in the alpha acetylmannoside at was substantially th e same at - 15° and 20 0 C. high temperatures is proportionately greater than Since no alpha acetylmannoside was found at the increase in the alpha acetylglucoside. H ence low temperatures, the amount of the orthoester an intermolecular reaction, with an SN 1 mechanism can be taken as a measure of the intramolecular such as that postulated for the formation of the reaction at 20° C or below. Pres umably the alpha acetylglucoside, does not adequately account amount of the beta acetylmannoside, by analogy for the facts. It seems more probable that th e to the glucose series, is a measure of the inter­ marked increase in the yield of the alpha acetyl­ molecular reaction. When the temperature was mannoside at the higher temperature is caused by raised to 50° C, the yield of orthoester dropped a modification of th e intramolecular course. One from 78 to 53 percent, and the yields of the alpha would expect. the solvated orthoester intermediate, and beta acetylmannosides increased; the former which in the mannose series yields the methyl • T he formation of the acetyl sugar as a byproduct was discussed on page 163 . orthoester almost exclusively at low temperatures, and will not be co n s id ~ re d here. to dissociate at higher temperatures with the I To show that the differences in the amounts of orthoester a nd a lpha acetylmannoside found at 20° a nd 50° C were real and not due to a secondary formation of the free orthoester lOn of Winstein decomposition of t he orthoester after its formation. a sample of triacetyl· and coworkers. The free ion, by reaction with manuose 1,2-(methyl ortboacctate) was heated at 50° C with methanol under the cond itions u sed iu the replacement reactions. No change was fo und. methanol, would produce the ol'thoester and the

166 Journal of Research :alpha acctylglycoside in proportions different from made for the contributions of the acetylsugar and those characteristic of the solvated intermediate the orthoester, is due to the mixture of the alpha in which methanol is already in po ition to yield and beta methyl acetylglycosides. After the de­ the 01'Lhoeste1'.8 In this connection it should be termination of the acetylsugar and orthoester, the pointed out again that the experiments that Win­ total yield of the acetylglycosides can be ob.tained stein and coworkers claim to rule out the solvated by difference, and the proportions of the two orthoe tel' mechanism were conducted at relatively acetylglycosides can be calculated from the 1'0 id­ high temperatures. The results given in this ual rotation and the known optical rotations of the paper indicate that at low temperatures the intra­ two substances. molecular reaction takes place exclusively through By the procedure outlined in the preceding the solvated orthoester intermediate, but at high paragraph the yields of the acetylsugal', the methyl -temperatures part of the reaction takes place orthoacetate, and the alpha and beta methyl through the free ion. acetylglycosides from tetraacetyl-a-D-mannosyl bromide were determined. Application of the V. Experimental Procedure same method to the product from tetraacetyl-a­ 1. Discussion of the Analytical Method D-glucosyl bromide showed the absence of the The m ethod for the analysis of the reaction methyl orthoacetate and gave the yields of the products is based on the premise that the only acetylsugar, and the alpha and beta methyl substances present are the alpha and beta methyl acetylglueosides. tetraacetylglycosides, the methyl orthoacetate, The procedure was standardized by measme­ and th e free acetylsugar. It has beell found ments made with the pure constituents as controls. possible to determine the amount of th e acetyl­ The method for determining the orthoester, sugar by modification of the conventional method although convenient, is not very satisfactory for the iodimetric determination of aldoses. The because of the difficulty in maintaining a chloro­ acetylsugar initially formed in the reaction is form solution of hydrogen chloride of uniform gradually converted to the alpha-beta rquilibrium str ength. Because of variation in this reagent, it m ixture; in the presence of a catalyst this eq uilib­ was necessary to standardize the procedure for rium may be establish ed quickly. From the each set of determination by use of pure tl'iacetyl­ knovvn optical rotation of the pure acetylsugar in mannose 1,2-(methyl orthoacetate). Since the equilibrium, and the known amount of this sub­ quantity of the methyl acetylglycosidcs was ob­ stance in the mixture, it is possible to calculate tained by difference, the values recorded for these the contribution of the acetylsugar to the equilib­ ubstances are subject to the combined errors in rium rotation of the r eaction product. To deter­ the values for the methyl ol'Lhoacetate and the mine the amount of the orthoester, advantage is free acetylsugar. In the experiments in which taken of the observation [101 that methyl ortho­ the percentage of the methyl acetylglycoside is acetates react rapidly with hydrogen chloride in shown as zero, the optical rotation for the methyl chloroform and that the reaction is accompanied acetylglycoside fraction was calculated to be at by a large change in optical rotation. Normal least as great in the levo direction as that required methyl acetylglycosides and the free acetylsugars for the beta acetylglycoside. are substantially inert to this reagent. Thus the 2 . Experimental Details difference betw'een the rotation at eq uilibrium and (a) Materials the rotation after the addition of hydrogen chloride The ether and benzene used as solvents were of is a measure of the amount of orthoester present. analytical grade, dried by sodium, and freshly The contribution of the orthoester to the equilib­ distilled. The chloroform was USP grade. The rium rotation can be calculated from the amount methanol was dried by reftuxing with barium oxide of orthoester in the reaction mixtW'e and the and distilling immediately before use. known optical rotation of the pure substance. The hydrogen chloride-chloroform r eagent was The residual optical rotation, after allowance is prepared by passing dry hydrogen chloride into 0 8 Both the solvated and the free orthoester ion can yield the orthoester USP chloroform to the point of saturation at 20 C. and the all)l1 a acetylglycoside. It is noteworthy that both products were The reagent was protected from atmospheric se parated by appli cation of the Koenigs·Knorr reaction to heptaacetyl-4- glucosido·manTIosyl bromide at 0° C [10]. moisture an~ dispensed from a buret.

Orthoester Reactions 167 The silver carbonate was prepared by the addi­ reaction product was evaporated to dryness in a tion of aqueous sodium bicarbonate to an aqueous current of dry air, and the last trace of alcohol was silver nitrate solution. It was collected on a filter removed by the addition of 5 ml of benzene and and thoroughly washed, first with water, and fi­ re-evaporation. The residue Wfl,S taken up in nally .with acetone. It was then dried at room 5 ml of peroxide-free dioxane, and the free acetyl­ temperature in the absence of light in a vacuum sugar was determined by oxidation with iodine desiccator containing phosphoric anhydride. by the method of Kline and Acree [17]. Control The Drierite was finely powdered and was re­ experiments showed that neither the methyl dried at 230° C immediately before use. The acetylglycosides nor triacetylmannose 1,2-(methyl following substances were prepared by the methods orthoacetate) reacts with iodine, whereas tetra­ given in the corresponding literature references: acetylmannose and tetraacetylglucose react stoi­ tetraacetyl-a-D-mannosyl bromide [11], tetraacetyl­ chiometrically when an excess of 5 to 10 ml of mannose [12], methyl tetraacetyl-a-D-mannoside O.I-N iodine solution is used. The amount of [13], methyl tetraacetyl-i1-D-mannoside [14], tri­ free acetylsugar in the entire reaction product acetylmannose 1,2-(methyl orthoacetate) [13], was calculated, and expressed in t.ables 1 and 2 as tetraacetyl-a-D-glucosyl bromide, page 500 of [15], percentage of the acetylsugar theoretically equiva­ tetraacetylglucose [16], methyl tetraacetyl-a-D­ len t to the acetylglycosyl halide employed. glucoside [1], methyl tetraacetyl-i1-D-glucoside (2 ) Optical rotation oj the reaction mixture at equi­ [1] and page 516 of [15]. librium.- Preliminary experiments showed that the equilibrium of the free aceLylsugars can be es­ (b) Method of Conducting the Replacement Reactions tablished readily by the addition of sodium ace­ The replacement reactions were conducted in tate, without alteration of the rotation of the 100-ml glass-stoppered flasks, carefully dried methyl acetylglycosides or of triacetylmannose before use, and maintained at the temperature of 1,2-(methyl orthoacetate). Hence, sodium ace­ the experiment by use of a suitable constant­ tate was employed as catalyst to establish equilib­ temperature bath. Silver carbonate (2 g), Dri­ rium. Ten milliliters of the solution containing erite (2 g), and the solvent were placed in the the reaction product was evaporated to dryness reaction flask and brought to temperature. The in a 25-ml volumetric flask containing 0.05 g of acetylglycosyl halide in a glass boat (2 g) was crystalline sodium acetate. Sufficient USP chlor­ dropped into the flask, and the mixture was shaken oform was added to make a volume of 25 ml at for 15, 30, or 60 min 9 according to whether th e 20° C. After 18 hI' at room temperature, the temperature was 50°, 20° or -15° C, respectively. chloroform solution was filtered, and the optical In some of the experiments, the methanol was rotation was read in a 4-dm tube. The readings introduced into the flask as the solvent or as a are recorded in table 3, and the corresponding constituent of the solvent prior to the addition of specific rotations are given in tables 1 and 2. the halide. In other experiments, 15 ml of a (3) Optical rotation oj the reaction mixture at 60 solvent was added at the beginning; after intro­ minutes.- Since the solu tion exhibits mutarotation duction of the silver carbonate, Drierite, and because of the presence of th e free acetylsugar, acetylglycosyl halide, 5 ml of the solvent contain­ measurements of optical activity were made on ing 0.5 ml of methanol was added dropwise from a the methanol solution of the reaction product at buret with exclusion of moisture by use of a rubber convenient times and were extrapolated to obtain dam. After the reaction was completed, the mix­ the values at 60 min from the beginning of the re­ ture was filtered, the residue was washed with action. These values were converted to specific methanol, and the filtrate and washings were made rotations based on the weight of the reaction to a volume of 50 ml with methanol. Portions of products.1O this solution were used for the measurements that follow. 10 Sincc the acetylglycosyl bromide (of molecular weight 411.21 ) is con­ (c) Analysis of the Reaction Product verted to the free acetylsugar (of molecular weight 348 .30) and to the methyl acetylglycosides and methyl acetylorthoester (of molecular weight 362.33), (1) Determination oj the jree acetylsugar.­ the weight of reaction product from 2 g of acetylglycosyl bromide is 2 [ 348.30 A+3~2.33 (I-A) ] , Twenty milliliters of the solution containing the 4i1.21 411.21 , The reaction times were so ,elected the t the filtered reaction products where A is the fraction of the acetylglycosyl bromide con verted to the free were halogen-free. acetylsugar, as determined by titration.

168 rournal of Research

~--. ~-~= ---_. TABLE 3.- Experimental data on p roducts derit'ed from acetvlulVcosy l bromides at fO o C

Optical rotation Change in optical rotation F ree acetyl· I----~----I------I Experiment number Oriil oesier psruodgaurct in• Afier NaO_' c I After n Cl- I Orthoesier ~ Cn Cl t t P roduct treatment b m e~t r~a - control

P roduct from ietraaeetyl-a ·D-mannosyl brom ide

(molecular) (molecular) P ercent ° S oS oS oS P ercent

1. ______. ___ . ______. _. ______._ 6. 9 -4.50 + 9.39 + 13.83 +17.61 78.5 I L . ___ . ______. _. ___ . ___ . ___ . ______._ 25. 9 - 0. 11 +6. 45 6. 35 15. 47 41. 0 III. ______32. 1 + 1. 20 + 4.32 2.86 15. 47 18.5 I V ______. ______35.4 + 1.71 + 3. 16 1. 16 15. 47 7.5 V ______16. 0 - 1. 99 + 8. 12 Q. 98 15. 47 64.5 v I. ______28. 7 + 0. 12 + 5.95 5. 59 15. 47 31l.1 VII. ______. ______6. 2 - 4. 72 + 9. 09 13. 76 17. 61 7 . 1 VIIL ______13.5 -3.21 + 6.30 9. 40 17. 61 53. 4

P roduct fr om tctraaceiy]-a-D-glu cosyl bromide

I X ______5. 7 -2. 03 - 1. 82 + 0. 21 X ______28. 1 + 1. 33 + 1. 60 + . 27 ::::::::: ::: I:::::::::::::: XL ______25. 2 + 0. 89 + 1.09 +.20 X I L ______5. 0 - 2. 09 -2.01 -.08 X IIL ______6.7 - 1. 39 -1.16 + .23 :::::::::::f::::::: :::::

• D etermined as described on page 168. b 0.000973 moles in a volume of 25 mI. read in a 4-d m tube. In order to evaluate the contributions of the cific rotations for the glycoside fractions of the several constituents of the product to the equili­ products from experiments I and VII in the man­ brium rotation, the measurements given in table nose series were calculated to be - 56.3 ° and 4 were made with the pure substances under con­ - 60.7 °, respectively. These are higher values in ditions like those used in the study of the reaction the levo direction than the specific rotation, products. To obtain the equilibrium rotation, a - 48.2° of the methyl tetraacetyl-{3-D-mannopy­ 0.000973-mole sample of the substance in a 25-ml ranoside. H owever, repetition of the experiments volumetric flask was dissolved in 10 ml of meth­ gave results in substantial agreement wi th those anol ; 0.05 g of sodium acetate was added, and the reported. Since the rotations of the glycoside mixture was evaporated to dryness. The mate­ fractions are based on only 15 percent of the prod­ rial was then dissolved in chloroform, the volume uct and are obtained by difference, they are was adjusted to 25 ml, and the optical rotation greatly affected by all errors of measurement. was measured in the manner described for obtain­ Better agreement was found in the glu cose ing the equilibrium rotation of the reaction prod­ series in which experiments IX, X , XI, and XII uct. The observed rotations of + 3.20° Sand gave specific rotations of - 17.9°, - 18.6°, - 18.5°, + 12.49° S for tetraacetylmannose and tetra­ and - 17.5°, respectively, for the glycosidic frac­ acetylglucose, respectively, and of - 4. 32° S for tions in comparison with - 18.6 ° for the specific triacetylmannose 1,2-(methyl orthoacetate) were rotation of methyl tetraacetyl-/3-D-glucopyrano­ used to calculate the contributions of these three side. su bstances to the equilibrium rotations of the re­ (4) Determination oj the orthoester.- T en milli­ action products. The values of + 7.95° S and liters of the methanol solution of the reaction - 7. 84° S were used to calculate the proportions product was evaporated in a current of dry air of the alpha and beta methyl tetraacetylmanno­ in order to remove the alcohol. The residue was sides from the residual rotations, as described on dissolved in the hydrogen chloride-chloroform re­ page 167. The values of + 21.20° Sand - 3.03 ° agent, and the volume was adjusted to 25 ml. S were used for similar calculations for the alpha The optical rotation was read in a 4-dm tube, and beta methyl tetraacetylglucosides. The spe- 30 min after the addition of the reagent, and

Orthoester Reactions 169 recorded in table 3. It was mentioned earlier halides with compounds containing a free hydroxyl that the method was standardized for each meas­ group (ROH) takes place both by an oppositr-face urement in the mannose series by a control experi­ intermolecular course and by an opposite-face ment with pure t.riacet.ylmannose 1,2- (methyl intramolecular course involving a solvated ortho­ orthoacetate). The change in optical rotation for ester intermediate. The former yields the gly­ t.he con trol applicable to each experimen t is also coside with inversion, the latter an or tho ester ion, given in table 3. which then stabilizes by one of two mechanisms The effect of the addition of the Hel reagent on to yield either the orthoester or the glycoside with the opticall'otations of tetraacetylmannose, tetra­ apparent retention of configuration_ In this paper acetylglucose, and the alpha and beta methyl it has bern shown that if the oI'thoester ion is tetraacetylglycosides of mannose and glucose was solvated, the course of the intramolecular reaction measured in separate experiments, and t,he results should be drastically affected by a solvent that is are recorded in table 4. The addition of the Hel able to coordinate in the formation of the ortho­ reagent did not cause an appreciable change in ester ion, but unable to eliminate a proton. On the rotations of the glycosides and of tetraacetyl­ the other hand, if the orthoester intermediate glucose. A small change was observed in the rxists as a free ion, such a solvent should act only rotation of tetraacetylmannose; in the eaIeulation as an inert diluent in the reaction. A quantitative of the amount of orthoester, allowance was made study has been made of the reaction with methanol for the corresponding change in the rotation of of tetraacetyl-a-D-glucosyl bromide (a cis halide) , the acetylmannose present in the product. In and tetraacetyl-a-D-mannosyl bromide (a trans each case the correction amounted to 0.8°8 X % halide) in the presence of silver carbonate (the of free acetylsugar/100. The changes in optical Koenigs-Knorr reaction). At -15°, 20°, and rotation for the product recorded in table 3 were 50° C, tetraacetyl-a-D-glucosyl bromide yielded corrected for this factor. The percentage of the almost exclusively methyl tetraacetyl-/3-D-gluco­ orthoester in the product was obtained from the side. At 20° C and below, tetl'aacetyl-a-D­ ratio of the corrected change and the change found mannosyl bromide yielded 78 percent of triacetyl­ for the pure substance. mannose 1,2-(methyl ol'thoacetate) and 15 percent of methyl tetraaeetyl-/3-D-mannoside. No methyl TABLE 4. ETperimental data obtained with pure compounds tetraacetyl-a-D-mannoside was formed, a result

Optical rotation· that indicates that conversion of the orthoester intermediate to the orthoester is the preferred Reference substance After NaOAc After RCI­ course when methanol alone is present. In the treatment CRCl, treat­ ment presence of ether and small quantities of methanol, os os tetraacetyl-a-D-mannosyl bromide yielded 7.5 per­ T etraacetyl-D-mannose ______+3.20 +4.02 cent of the orthoester, 23 percent of the beta Tetraacetyl-D-glu cose .... ______~ ~ + L2.49 +12.31 Metbyl tetraacetyl-a-D-mallilopyrallosidc + 7.95 +8.01 acetylglycoside and 34 percent of the alpha Methyl tetraacetyl~/I~D-manllopyrallo- acetylglycoside. ThE' effect of benzene on the s ide~ _~ _~ ~ ~ ~ ~ ~ ~ ~ ~ _~ ~ ~ ~ ~ __ ~ ~ ~ ~ ~~ ____ ~ _~ ~ -7.84 -7.80 Methyl tetraacetyl-a-D~g lil copyralloside ~ +21. 20 +21. 2L proportion of the products was similar to that of Methyl tetraacetyl-/l~D-glucopyranoside ~ -3.03 -3.02 ether but less pronounced, in accord with its Triacetylrnallllose 1,2-(mcthyl ortho~ acetate) ~ ~ ~ _~~~ ~ ~ ~ ~~ __ ~_~~~ ______~~ ~ ~~~ -4.32 b+I1. 16 smaller tendency to co-ordinate. These striking Triaeetylrnannose 1,2-(rnetbyl ortho- changes support the mechanism involving a sol­ acetate) ~ _~ ~ ~ ~ ~ _~ ~ _~ ~ ~ ~ ______~ ~ ~ ~ ~ ~ ~ ~ ~ ~ -4.32 b+13.30 vated orthoester intermediate rather than a free lOn. a A O.OOO973~mole samrle in a volume of 25 ml, read in a 4-dm tube at 20° C. When the temperature of the reaction was b The "alueR lI'ere obtained with different preparations of the RCI·CRCIa reagent. raised to 50° C, the amount of the orthoester VI. Summary dropped from 78 to 53 percent, and 8 percent of the alpha acetylmannoside was formed. This A study has been made of the effect of solvent shift in the proportions of the products of the and temperature on the course of certain organic intramolecular reaction seems to indicate that the replacement reactions. It was formerly postulated solvated orthoester ion is somewhat thermally [2,3,4] that the reaction of trans acetylglycosyl unstable, and that part of the reaction at the higher

170 Journal of Research tempera ture takes place through the free 01'­ [6] S. Winstein, C. Hanson, and E. Grun\\'ald, J. Am. thoester intermediate. This would react to yield Chem. Soc. 70, 812 (194S). the orLhoestcr and the alpha acetylglycosidc in [7] D. D. R eynolds and W. L. Evan~, J . Am. Chern . Soc. 60, 2559 (193S). different proportions from the solvated ion. The [S] V.,T . W . Pigman and H . S. Isbell , J. Hesearch NBS 19, reaction of tetraacetyl-a-D-glu cosyl bromide was 199 (1937) RP1021. only slightly affected by a change of solvent or of [9] J. L. Gleave, E. D. Hughes, and C. 1(, Ingold, J . temperature. This is compatible with the stcreo­ Chern. Soc. 1935, 236; E. D . Hughes, J . Chem. m eric inability of the alpha glucose configuration Soc. 1946, 968. [10] H . S. Isbell , BS J. Research 7, 1115 (1931) RP392. to form an orthoester intermediate. [11] F. Micheel and H . Micheel, Ber . deut. chem. Ges. 63, 3S6 (1 930); D . H. Brauns, BS J . Research 7, 573 The authors express their appreciation to Nancy (1931) RP35S. [12] P. A. Levene and R. S. Tipson, J . BioI. Chem. 90, B. Holt for her assistance in conducting the 89 (1931) . analytical work reported in this paper. [131 J. K Dale, J . Am. Chem. Soc. 46, 1046 (1924). [1 4] H. S. Isbell and H . L. Frush, J. R esearch NBS 24, VII. References 125 (1940) RP1274. [II W. Koeni gs and E. Knorr, Ber. deut. chem. Ges. 34, [15] F. J. Bates and as ociates, Polarimetry, saccharimetry 957 (1901). and the sugars, NBS Circular C440 (1942). [21 H. S. Isbell , Ann. R ev. Bioehem. 9, 65 (1940). [161 E. Fischer and K. H ess, Ber. den t. chem. Ges. 45, [3] H . L. Frusn and H . S. Isbell , J. llesearch NBS 27, 413 912 (1912). (1941) RP1429. [17] G. M. Kline and S. F. Acree, BS J. Research 5, lOS] [4] H. L . Frush and H . S. Isbel l, J . R esearch NBS 34, (1930) RP247. III (1945) RP]663. [5] '. Winstein and R . E. Buckles, J . Am. Chem. Soc. 64, 2780. 2787 (1942). WASHINGTON . May 12, 1949.

Orthoester Reactions 171