studies on the stability of simple derivatives of sialic acid*

Studies on the Stability of Simple Derivatives

of Sialic Acid*

From the Cell Ch.(Jmistry Laboratory, Department of Biochemistry, College of Physicians and Surgeons, Columbia Uniurrsity, New York 32, New York

(Received for publication, May 8, 1963)

The sinlic acids, whose chemistry and metabolism have been reviewed repeatedly in the recent past (l-5), are important constituents of many biological heteropolymers. It is quite clear that the multiple types of linkage in which a nonulosaminic acid can engage may influence the outcome of reactions designed to demonstrate its presence, cstimatc its quantity, or determine its modes of attachment, in different polymers. Even the pre- treatment of the polymer, in the course of its purification, may not bc without effect on the attachment of sialic acid to one or the other of the constiturnts through its various functional grouljs. This problem of stability was brought to our attention in the course of studies in this laboratory on the mucolipids of brain (6-13).

l’he literature is not abundant in information on this point. It is generally recognized that the free sialic acid is unstable towards acid and especially towards alkali, and that it is re- leased readily, when linked as a glycoside, by very mild acid treatment and even by autohydrolysis. A qualitative study of the stability of sialic acids and their methyl esters at different pH values has been published (14), and occasional mention of I-arious observations on this subject will be found in many pnpcrs dealing with other aspects of the chemistry of sialic acid. -1 detailed investigation, spcc*ifying the exact conditions under whirh the ester or the glycositlic link of sialic acid is ruptured, however, has not come to our attention. The present study describes tsperimmts that may serve as a first step in this direction. Some of the findings have been mentioned very briefly in a preliminary form (15).

‘l’hc stnbility of sialic acid and of three of its simplest functional derivatives was investigated under a variety of conditions, mainly by means of calorimetric techniques. The three derira- tivcs represent models of the three principal ways in which sinlic acid (I) ma-y be linked to another molecule: (a) through an ester bond (sialic acid methyl ester, II); (b) through a glycosidic bond (methosysialic acid, III); and (c) through a combina- tion of these links (methosysialic acid methyl ester, IV). The structures of the four coml)ormds under investigation, in their currently accepted configurations (16), are shown in Fig. 1.

‘The majority of thr experiments were carried out with a

* These studies were supported by a research grant from the Sational Institutes of Health, United States Public Health Service.

t This report is from a dissertation submitted by John D. Karkas in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Faculty of Pure Science, Columbia LVnirersity. During part of the t.ime, J. D. K. held a fellowship from the Greek State Fellowship Foundation.

preparation of sialic acid isolated from ox serum proteins and composed of 37 y0 N-ncetyl- and 63% N-glycolylneuraminic acid. In addition, a few orienting studies were performed with a specimen of N-acctylneuraminic acid secured from human serum proteins. Experiments with sialic acid from sheep serum are omittrd for the sake of brevity, as the results only duplicated those presented here.

ESPERIMENT.4L PROCEUURE

Analytica. Methods

The direct Ehrlich reaction was carried out as described in the literature (17), but the volumes were reduced so that a final reaction mixture of 3 ml resulted. The procedure for the thioburbiturate reaction has been described before (18), as has been that for the resorcinol reaction (19) ; in the latter instance, heating at 100” was applied. For the hydroxamic acid reaction, a modification developed for sugar esters (20) was followed; the reaction time was reduced to 30 minutes, as preliminary experiments showed that with the methyl cstcrs of sialic acid and methosysialic acid the maximal nbsorbances were attained within a few minutes.

The oxidation exprrimc~nts with periodate were ljerformed in unbuffered aqueous 0.01 M solutions of SnIOh at room temperature in the dark. In samples removcad after various intervals, the remaining periodate was determined, after dilution with an equal volume of phosphate buffer (12 g of Na2HPOI. 12H20 and 20 ml of N HtSOl in 100 ml), pH 6.5, by the addition of 1 drop of a concentrated I<1 solution and titration of the liberated iodine with a 0.01 JI solution of sodium thiosulfate in the presence of starch.

Gilmont microburettes srrved for the removal of all samples and for the titrations. The melting points, reported without correction, were determinctl on an elcctricnlly h&cd stage (Fisher-Johns).

Materials

Siulic i4cids (I)-Preparations were isolated from serum proteins of man, sheep, and ox by hydrolysis and ion exchange chromatogral)hy on Dower; 50 and Dowcs l-X8, essentially as described in the literature (21). The c,rude substsnccs were recrystallized three times from water-methanol-ether. The melting point (with decomposition) of all three preparations was 184-185” (uncorrected). The ratios of Ar-acetyl- to W-glyco- lyhleuruminic acid, as found by microestimation of glycolic acid

cm, were: bovinr, 37:63; ovine, 87.5: 12.5; in human sialic

949

by guest on April 4, 2018

http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

950

‘COOH

%-OH

COOCH3

r r &OH H&H H-&H

0 H&OH 0 H-&OH

I- AcHN%-H

L- A&N-&H

6-H &ti HI&-OH H-&OH H%-OH H-A-OH

%H~OH 6~~0~

I It III m

FIG. 1. Structures of sialic acid (I), sialic acid methyl ester (II), methoxysialic acid (III), and methoxysialic acid methyl ester (ZV). AC = acyl.

Sialic Acid Derivatives Vol. 239, ?\‘o. 4

?OOH COOCHJ

r C-OCH3

H-6-H &OCHs

H-6-H 0 H-&OH

r I

0 H-&OH AcHN-L-H

I- AcHN-k-H

6-H 6-H H-k-OH H-b-OH H-&OH H-k-OH

LH~OH 6~~0~

acid, the reading for glycolic acid was essentially negative (less than 2%). These findings are in remarkable agreement with the literature (compare p. 33 of a recent monograph (3)).

The sialic acid from ox serum served for the preparation of the ester and methoxy derivatives investigated in this study. N-Acetylneuraminic acid:

CnH1gNO9 (309.3)

Calculated: C 42.7, H 6.19, N 4.53

N-Glycolylneuraminic acid:

GIHISNOIO (325.3)

Calculated: C 40.6, H. 5.89, N 4.31 Calculated for ox serum sialic acid: C 41.4, H G.00, N 4.39 Found (bovine) :’ C 40.80, H 6.17, N (Dumas) 4.69 Found (human) :* C 42.62, H 5.99, N (Dumas) 4.28

&a& Acid ,%ZethyZ Ester (II)-This derivative was prepared by heating, under reflux for 30 minutes, sialic acid in dry methanol in the presence of a small quantity of Dowex 50 which had previously been twice refluxed with dry methanol for 1 hour. The filtrate from the resin then was passed through a column of Dowex I-formate in order to remove sialic acid that had not reacted. The solvent was removed, and the residue was dried over Pz05 in a vacuum. The ester, recrystallized twice from methanol-ether, melted with decomposition at 169-171” (uncorrected). The N-acetyl to N-glycolyl ratio was 35:65. N-Acetylneuraminic acid methyl ester:

C,ZH2,NOg (323.3)

Calculated : N 4.33

N-Glycolylneuraminic acid methyl ester:

CrzHrrNO,n (339.3)

Calculated : N 4.13 Found :r N (I)umas) 3.94

Methoxysialic Acid Methyl Ester (IV)-This compound was prepared according to Blis et al. (23) by heating sialic acid, under a reflux, with three changes of dry methanol in the presence of Dowex 50 for a total of 15 hours. The product, recrystallized twice from methanol-ether-petroleum ether, melted at 195” (uncorrected). The N-acetyl to N-glycolyl ratio was

1 These analyses were performed by Micro-Tech Laboratories, Skokie, Illinois.

2 These analyses were performed by Schwarzkopf Microanalyti- cal Laboratory, Woodside, New York.

36 : 64. Methosy-N-acetylncuraminic acid methyl ester:

C,oH?xNOy (337.3)

Calculated: C 46.3, H 6.87, N 4.15

Methoxy-N-glycolylneuraminic acid methyl ester:

Calculated: C 44.2, H 6.5G, N 3.97 Found :I C 45.0, H 6.75, N (Dumas) 4.11

XethoxysiaZic Acid (III)-This derivative was prepared from methoxysialic acid methyl ester by mild alkaline hydrolysis. An aqueous solution of the ester (IV) was maintained at pH 10 and room temperature for 4 hours by means of an automatic titrator (Radiometer, Copenhagen), when 1 molar equivalent of NaOH had been consumed and the hydroxamic acid reaction for ester had become negligible. The solution was passed through a column of Dowex 50 (H+ form) and evaporated in the frozen state in a vacuum. Attempts at crystallization having failed, a semicrystalline powder was obtained by precipitation with ether from aqueous methanol. This product gave almost no color in the thiobarbiturate reaction; however, after hydrolysis in citrate buffer, pH 3, at 100” for 45 minutes, the absorbance was the same as that shown by free sialic acid. No hydroxamic acid reaction was observed. The extent of the resorcinol test was the same as that afforded by free sialic acid. The N-acetyl to N-glycolyl ratio was 37:63. Methoxy-N- acetylneuraminic acid:

Cr?H2rN09 (323.3)

Calculated: C 44.6, H 6.55, N 4.33

Methoxy-N-glycolylneuraminic acid :

C,tHz,NOIo (339.3)

Calculated: C 42.5, H 6.24, N 4.13 Found :3 C 43.5, H 6.32, N 3.80

RESULTS AND DISCUSSION

Color Reactions

The molar absorbances of the sialic acids (human, sheep, ox) and of the three derivatives of sialic acid from OS serum under investigation in the direct Ehrlich, resorcinol, and thiobarbiturate reactions, performed by the techniques mentioned before, are presented in Table I.

The thiobarbiturate reaction is by far the most sensitive reaction for sialic acids, as long as the hydroxyl on the anomeric carbon is unsubstituted. Its usefulness for glycosidically bound sialic acid is limited, however, because of the dependence on an effective hydrolysis. As the results of the experiments described below indicate, the conditions for such a hydrolysis must be chosen carefully, with the application of correction factors for destruction, according to the type of binding of sialic acid in the particular compound-information not always easily secured.

The resorcinol reaction is quite sensitive and very repro- ducible. Our results also indicate that, in contrast to the findings with the direct Ehrlich reaction, the absorbance is independent of the presence of substituents on the carbosgl and

3 These analyses were performed by Analytica Corporation, New York.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

April 1964 J. D. Karkas and E. Chargaff 951

carbonyl carbons of sialic acid. The interference by many substances, and especially by galactose, however, renders its use rather difficult in the case of natural materials.

In the direct Ehrlich reaction, on the other hand, which appears to be the most specific reaction in the case of comples mistures (compare IL 60 of (3)), the extent of absorbance was found to depend on the substitution of the hemiacetalic hydrosyl. Thus, whereas sinlic acid methyl e.+r exhibited the same absorbance as the free sialic acid, the molar absorbances of methosysialic acid and its methyl ester were 85 and 8Or/,, respectively, of the absorbance of the free acid.

The results of a time study of the color development in the direct Ehrlich reaction are illustrated in Fig. 2. As can be seen, the production of color increases linearly with time, with slopes almost identical for the four compounds, after a characteristic initial period of retardation. The linear rise in extinction continues much beyond the Ijeriod shown in Fig. 2.

In this reaction, cyclization to yield a pyrrolic compound has been postulated as the first event (3, 14, 24, 25). The differences observed in the shapes of the curves in Fig. 2 do not contradict such a mechanism. The free rarbonyl required for the cyclization is present in sialic acid and its methyl ester, but must first be produced in methosysialic acid and methosysialic acid methyl ester by the removal of the glycosidic methosyl, an event occurring more easily in methosysialic acid than in its methyl ester, as will be shown below.

The curves in Fig. 2 are constructed in terms of the optical density. I f the molar absorbances recorded in this experiment after a treatment of 30 minutes, the time required by the standard technique (17), were computed, they would be found to be aI)preciably lower than those listed in Table I; moreover, the differences in color development between the free carbonyl compounds, sialic acid and its methyl ester, and the glycosides, methoxysialic acid and its methyl ester, are more pronounced. These discrepancies are obviously due to the different esperi- mental arrangement required by the time study: the size and

TABLE I

Molar absorbances of sialic acids and dekvatives in various color reactions

A-Acyl Molar absorbances (X 10-a) substituents read at specified wave lengths

Sialic acid source and derivative

Acetyl GI - co yl r

Direct Ehrlich

- 565 ITI@

Human serum.. 100 Sheep serum. 87.5 Ox serum. 37

Sialic acid methyl ester 35 Methoxysialic acid. 37 Methoxysialic acid

methyl ester. 36 Methoxysialic acid,

after hydrolysist.

5’ 0

12.5 63

65

63

2.20 8.97

2.24 9.50

2.35 11.0

2.33 11.0 2.01 10.8

ii-l 1.90

2.30

Average deviat,ion.. 0.15 -

-7 RCSOT- cinol

-I-

580 m/I

Thiobar- biturate

549 m&l

68.1

63.8

57.8

58.0

(2.2)*

10.9 (2.8)*

10.7 56.8

0.1 1.8

* These readings are, in view of the average deviation, not sig- nificant.

t At pH 3 and 100” for 45 minutes

IO 20 30 40 MINUTES

Fro. 2. IXrect Ehrlich reaction: color development by sialic acid (from ox serum) and its derivatives. A solution of 10 p~noles of the compound in 00 ml of N HCl containing 0.83$; of p-di- methylarninobenzaldehgde was heated in a boiling water bath. At the times indicated, 3-ml samples were withdrawn, cooled in ice water, and brought to room temperature, and their absorbance was determined spectrophotometrically at 565 mp. A, sialir acid; A, sialic acid methyl ester; 0, methosgsialic acid; l , methoxysialic acid methyl ester.

I I I 1 ,

I I I I I I

IO 20 30 40 50 60 +7ihk-& MINUTES HOURS

FIG. 3. Oxidation of sialic acid (from ox serum) and derivatives by sodium metaperiodate. To a solution of 20 Mmoles of the compound in 10 ml of water, the same volume of a 0.01 M aqueous solution of NaIO, was added. Equal samples (1 ml) were removed at different intervals and mixed with 0.3 ml of phosphate buffer, pH 6.5, and 1 drop of a concentrated KI solution. Five minutes later, the liberated iodine was titrated with 0.01 M Na&;zOz solution in the presence of starch. A, sialic acid; A, sialic acid methyl ester; 0, methoxysialic acid; l , methoxysialic acid methyl ester.

type of the vessels and the heating bath used, the volumes of the assay mixture, the rate of temperature equilibration, etc., all of which are not without infuence on the color value.

Periodate Oxidation

Uptake of Oxidant-The periodate osidation of sialic acid (10, 23, 26) and of methosgsialic acid methyl ester (23) has been investigated before. We show, in Fig. 3, a reinvestigation of these compounds together with sialic acid methyl ester and methospsialic acid under uniform conditions. An initial veq rapid uptake of 2 moles of the osidant per mole of substance is observed with all compounds. In the case of the glycosidi- tally substituted derivatives, methosgsialic arid and methosysialic acid methyl ester, the curves level off at this point, with


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

952 Sialic Acid Derivatives Vol. 239, 1-o. 4

I I,, I I I! I I Id

IO 20 30 40 50 MINUTES

I I I I

FIG. 4. Alkaline hydrolysis of the ester link of sialic acid methyl ester (A) and methoxysialic acid methyl ester (B). Solutions of 40 pmoles of compound in 10 ml of COz-free water were used to determine the required equivalents of alkali, added in the form of 0.01 N NaOH, in order to maintain the indicated pH. An automatic titrator (Radiometer, Copenhagen) was employed.

COOCH3

r &OH OOCH3

H-+-H q o

$OOCH3

H-G-H

o@li;T- A$~~“=

H-k-OH H-C-OH

H-F-OH H-C-OH

bH20H H-q-0 H H-q-0 H

CH20H CH20H

II II A Fro. 5. Tautomeric structures of sialic acid methyl ester

very little additional uptake observable in the course of 24 hours. Sialic acid and sialic acid methyl ester, on the other hand, continue to consume periodate, although at a much slower rate, 1 additional molar equivalent being taken up in the course of 20 hours.

The initial rapid consumption of 2 moles of oxidant is obviously due to scission between carbon atoms 8 and 9 and carbon atoms 7 and 8 (compare Fig. 1). The uptake of a third mole by the two compounds possessing a free hemiacetalic hydroxyl (sialic acid and its methyl ester) points to the opening of the pyranose ring as a requisite for the second phase of the oxidation, which presumably concerns the rupture between carbon atoms 6 and 7. This assumption would explain the slower rate at which the third equivalent of oxidant is consumed, since the opening of the ring would now become the rate-limit- ing step. A slight degree of overoxidation beyond 3 equivalents is indicated by the curves for sialic acid and sialic acid methyl ester (Fig. 3), which presumably are attributable to a slow hydroxylation (27) of the methylenic carbon 3.

Color Reactions of Oxidation Products-Attempts to isolate the aldehydes expected to result from the action of periodate on the four compounds studied here have until now not been success-

ful. The reactivity of the crude oxidation products in the color reactions normally applied to sialic acid and its derivatives could, however, be examined. The compounds were subjected to the action of NaI04 under the conditions outlined above. and the excess of the oxidant was reduced by means of ethylene glycol to iodate, which was precipitated by the addition of the required quantity of barium acetate. The filtrates were assayed directly. Because of the sensitivity of color tests to admixtures, control solutions were prepared in which the periodate reagent was, with the omission of the substrate, put through the opera- tions mentioned here (reduction, precipitation, etc.); the resulting solution then served as the solvent for the untreated sialic acid derivative, whose absorbance was compared with that of the corresponding oxidation product.

The following observations were made on the oxidation products. (a) The absorbance in the thiobarbiturate reaction n-as not altered. This was to be espected, as periodate oxidation actually is the first step in this reaction. (b) The absorbance of sialic acid in the resorcinol reaction was almost doubled, with a shift of the absorption maximum from 580 to 600 mp. (c) The absorbance in the direct Ehrlich reaction decreased by about 657, for methoxysialic acid and its methyl ester and by about 80% for sialic acid and its methyl ester.

Alkaline Hydrolysis of Ester Link

The ester link of sialic acid is known to be cleaved under very mild alkaline conditions (21, 23). In the course of the present studies, the hydrolysis, at constant pH, of the ester bond in sialic acid methyl ester was compared with that of methoxysialic acid methyl ester. Aqueous solutions of the esters were kept at a constant pH by means of an automatic titrator. The equivalents of alkali consumed in these esperiments are plotted against time in Fig. 4. A parallel set of esperiments, in which the gradual disappearance of the ester was followed at constant pH by means of the hydroxamic acid reaction, confirmed that the alkali had actually been consumed by the saponification process.

As shown in Fig. 4, the ester link of sialic acid methyl ester is broken at room temperature and pH 8 within 20 minute*. whereas methoxysialic acid methyl ester requires a higher l)H and a much longer time for complete saponification. The following reaction rate constants were found for the conditions specified in Fig. 4: sialic acid methyl ester, lc = 2.7 x lop3 se0 ; methoxysialic acid methyl ester, k = 3.1 x lo-’ set?. The substitution of the neighboring hemiacetalic hydroxpl thus appears to render the ester link more resistant to alkali. Al- though steric hindrance could be invoked, it is more likely that another factor is of importance. In contrast to free sialic acid, the methyl ester has been reported to exhibit, in weakly alkaline solution, a transitory absorption in the ultraviolet region (maximum at 265 mM) which has been attributed (14) to the enol form of the open chain structure (11 B in Fig. 5), in analogy to observations on ethyl acetoacetate (28). In the open structure, sialic acid methyl ester would be expected to behave as an LY- keto acid ester, whereas methoxysialic acid methyl ester, owing to the stabilization of the pyranose ring (1Z in Fig. 5) by the substitution of the hemiacetalic hydroxyl, would be expected to show the reactivity of an cr-methoxy acid. Esters of a-keto acids are known to be saponified by alkali at rates much higher than those exhibited by their aliphatic counterparts or by the substituted a-hydroxy acids: the rate of alkaline hydrolysis of


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

April 1964 J. D. Karkas and E. Chargaff 953

ethyl pyruvate was found to be 18,000 times higher than that of ethyl propionate and 840 times higher than that of methoxy- glycolic acid ethyl ester (29).

Treatment with Acid

Free Sialic Acid-This substance is known to be decarbosyl- ated when heated in strong acid (23) but has been considered, on the basis of the resorcinol reaction, to be relatively stable under less drastic condit,ions (30). A similar apparent stability of sialic acid was observed in the present study when the direct Ehrlich reaction was employed (Fig. 6). The same figure, however, shows that, under identical experimental conditions, the color values recorded with the thiobarbiturate reaction decrease considerably in a linear fashion, approximately one-half of the initial absorbance being found at the end of 1 hour. A similar observation has been reported recently (31). ;\s is also shown in Fig. 6, Wacetylneuraminic acid (from human serum) gave identical results; this excludes the possi- bility that the two different .V-acyl substituents in the sialic acid from ox serum contributed in different manners to the behavior of the latter in the color reactions.

It is obvious that the action, even of weak acid, on sialic acid has complex consequences (14, 32), but the products cannot yet be identified conclusively. That some of the degradation products react with both the Ehrlich and the resorcinol reagents has been pointed out before (14). The treatment with acid obviously results in a modification of the molecule that depresses the formation of the thiobarbiturate chromogen, but does not interfere materially with the other color reactions.

Dccarboxylation does not seem to play a role, as shown by the behavior of the degradation product towards the periodate- thiobarbiturate reagent. The treatment of sialic acid with HI04 in a strongly acidic environment presumably affords B-formylpyruvic acid, in analogy to the 2-keto-3-deosy sugar acids (33, 34), whereas previous decarboxylation would lead to the formation of malondialdehyde (35). These two aldehydes form, in the reaction with thiobarbituric acid, colored compounds whose absorption spectra can be distinguished through the position of the centers of absorption and the respective molar absorbances: the product formed with P-formylpyruvic acid has a maximum at 549 rnM, and that formed with malondialdehyde, at 532 rnp; the molar estinction shown by the latter product is about twice as high as that of the first (34). As we have observed, sialic acid heated in 0.1 N HCl at 100” for 1 hour (Fig. 6) exhibits no shift in the position of the absorption maximum (549 mp), and this, together with the decrease in the molar absorbance, speaks against the occurrence of decarboxylation.

The partial lactonization of sialic acid (32) could account for the lack, or the retardation, of the response to periodic acid and for the diminution of the color value in the thiobarbiturate reaction. Dimerization through the formation of a glycosidic bond between the carbonyl group of one molecule and a hydroxyl of another could also be invoked. No direct evidence of either step, however, is available.

The cyclization to a pyrrole structure, as discussed before, would also block the oxidative formation of formylpyruvic acid. There is, in fact, some evidence that such an event takes place: we have observed that after sialic acid has been heated in 0.1 N HCl at 100” for 30 minutes, the solution forms a purple color in the cold when mixed with the p-dimethylaminobenzaldehyde reagent; this is characteristic of pyrroles, whereas untreated

/

I I I I I I I

IO 20 30 40 50 60 MINUTES

FIG. 6. Action of 0.1 N HCl at 100” on sialic acid preparations from ox serum (A) and human serum (B). Solutions of 10 pmoles of the compounds in 10 ml of 0.1 N HCI were heated in a boiling water bath under a reflux. Measured samples (1 ml) were removed at the indicated times and neutralized with an equal volume of 0.1 N NaOH; 1 ml of the mixture was used in the direct Ehrlich reaction (Curve 1), 0.1 ml in the thiobarbituric acid reaction (Curve 8)) and 0.2 ml in the resorcinol reaction (Curve 3).

sialic acid develops the color only after being heated with the reagent for some time.

It is not unlikely, therefore, that several events occur simul- taneously: lactonization, pyrrole formation, and further degradation; the extent to which each contributes to the final result, however, cannot yet be assessed.

Ester Link of Sialic and Methoxysialic .lcidsThe effect of the treatment of the methyl esters of these rompounds with 0.1 N HCl at 100” for 1 hour is illustrated in Fig. 7. l-rider these conditions, the ester link of sialic acid methyl ester is cleaved rapidly (Curve S, Fig. 78); the results of the direct Ehrlich and thiobarbiturate reactions resemble, as expected, those obtained with sialic acid (Fig. 6). The ester link of methoxysialic acid methyl ester, on the other hand, appears to be much more stable (Curve S, Fig. 7B): the hydroxamic acid values observed in the course of 1 hour diminish only little. The thiobarbiturate reaction, negative for methoxysialic acid methyl ester in the beginning of the treatment, increases with time, owing no doubt to the liberation of the carbonyl group. For the same reason a slight increase in the direct Ehrlich values is recorded (compare Table I). That the absorbance values in the thiobarbiturate test increase more steeply than the hydroxamate values drop (Fig. 7B) could be taken to indicate that the removal of the glycosidic methoxyl precedes that of the ester methoxyl.

Glycosidic Link of Methoxysialic Acid and Its Methyl Ester- The acid lability of the glycosidic link of sialic acid was recognized very early; the first isolation of the rompound made use of the


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

954 Sialic Acid Derivatives Vol. 239, No. 4

IO 20 30 40 50 60 MINUTES

GO 1 I

I I , lQ0 0

-

$ 0.2- 0’12 - 0 ?. 2 @ 0.08-

0.2

- 0.1 0.04

I .--I I IO 20 30 40 50 60

MINUTES

FIG. 7. Action of 0.1 N HCl at 100” on (A) sialic acid methyl ester and (B) methoxysialic acid methyl ester. Solutions of 20 pmoles of the compounds in 5 ml of 0.1 N HCl were heated in a boiling water bath under a reflux. Measured samples (0.5 ml) were removed at the indicated times and neutralized with an equal volume of 0.1 N NaOH. Of the mixture, 0.7 ml was used in the hydroxamic acid reaction (Curve 9); 0.25 ml was diluted to 3 ml, and 2.5 ml of this dilution were used in the direct Ehrlich reaction (Curve 1)) and 0.2 ml in the thiobarbituric acid reaction (Curve 2).

occurs more rapidly, but the free sialic acid evidently is degraded further.

The stability of the glycosidic link towards acid is heightened by the substitution of the neighboring carboxyl. No cleavage of methoxysialic acid methyl ester is apparent at pH values as low as 3. The compound is hydrolyzed slowly in 0.1 N HCl and more rapidly in 0.5 N HCl, but the liberation of sialic acid is overtaken by its destruction. Hence, the complete hydrolysis and quantitative recovery of the liberated sialic acid appear unattainable if it is doubly linked through its hemiacetalic hydroxy and carboxy groups.

The hydrogen of the carbosyl plays an important role in the removal of the glycosidic methoxyl. This intramolecular cata- lytic effect suggests the formation of a bond between this hydrogen and the glycosidic oxygen,

Such hydrogen bonds, resulting in the formation of a &membered ring, are not as strong as those of their 6-membered counterparts, but are still capable of influencing the rate of chemical reactions (37). In the present case, the removal of the methosyl, which first involves the addition of a proton to the methosyl oxygen atom, would be facilitated if the proton already were hyclrogen- bonded to it-an internal catalysis that is ljrevented in the doubly substituted methosysialic acid methyl ester, which consequently requires a higher acidity for hydrolysis.

I I I I 1 1 I

zo.5

s Fo.4 a

0 IO 20 30 40 50 60 MINUTES

FIG. 8. Hydrolysis of the glycosidic link in methoxysialic acid and its methyl ester at 100” and at different pH values. Solutions of 5 rmoles of the compounds in 25 ml of 0.05 M solutions of citric acid, sodium citrate, or mixtures of both, at the indicated pH values, and in 0.1 N and 0.5 N HCl, were heated in a boiling water bath under a reflux. Measured samples (0.2 ml) were withdrawn at the indicated times and mixed immediately with the periodate solution for the thiobarbiturate reaction. 0, methoxysialic acid; l , methoxysialic acid methyl ester.

autohydrolysis of submaxillary mucin (36). The results of our study of the stability of substitution of the hemiacetalic hydroxyl at pH values between 2 and 7 and in 0.1 and 0.5 N HCl are illustrated in Fig. 8. Between pH 5 and 7, methosysialic acid was not hydrolyzed appreciably; moderate cleavage took place at pH 4. The optimal condition for the hydrolysis of the glycosidic link seems to be at pH 3, since the rupture is com- pleted in a relatively short time, without the degradation of the liberated sialic acid. At higher degrees of acidity, cleavage

Treatment with Sialidase

The enzyme was preljared from culture filtrates of 17iixio cholerae and used in the purification stage referred to as Stage 6 in a previous publication (9). When methosysialic acid or its methyl ester, in concentrations of 10 or 100 pg per ml of 0.01 M Tris-acetate buffer, pH 6.6 (4 m&i with respect to Ca++), containing 100 sialidase units (9), was incubated for 24 hours at 37”, no liberation of sialic acid was observed. I-rider the same conditions of ensymic assay, ox brain mucolipid (8) and sialyl- lactose4 (38) showed considerable cleavage even after 1 hour of incubation. The liberation of sialic acid was followed by means of the thiobarbituratc test. It is likely that the resistance of methosysialic acid and its methyl ester to attack by sialidase, when compared to the susceptibility of the naturally occurring sialic acid glycosidcs, is due to the a&cone rather than to differences in the stereochemistry around carbon 2.

Treatment with Glacial ;lcetic -4cid and Pyridine

Mucolipids and related derivatives of sialic acid are, in the course of their preparation, often exposed to solvents such as pyridine or glacial acetic acid. It was of some interest to examine the stability of the simple derivatives studied here to treatment with these solvents.

Solutions of the four compounds in either pyridine or acetic acid were heated at 100”. The samples were tested by the direct Ehrlich and thiobarbituric acid reactions. The experimental conditions and results are summarized in Figs. 9 and 10.

4 We are very grateful to Dr. Karl Meyer for a sample of this compound.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

April 1964 J. .!I. Karlcas and B. C’hargaff

The treatment with glacial acetic acid results in the rapid

degradation of sialic acid (Fig. 9-l), as shown by both the

direct Ehrlich and the thiobarbiturate reactions. In the case

of methoxysialic acid (Fig. 9C), considerable degradation is

shown by the direct Ehrlich reaction, while the thiobarbiturate

values fail to indicate an accumulation of free sialic acid. To

what extent the glycosidic bond is broken in this case cannot be

assessed, as most of the sialic acid liberated would be further

degraded. The methyl ester of sialic acid (Fig. 9B) is degraded

slowly; it will be noticed that there is an initial, as yet unex-

plained, increase of the absorbance in the thiobarbiturate reap-

tion. N-hcetylneuraminic acid has a higher molar absorbance

than N-glycolylneuraminic acid (18), but a transamidation,

effecting the replacement of :V-glycolyl by N-acetyl groups, is

not likely under the experimental conditions. Methosysialic

acid methyl ester (Fig. 9D) appears to be rather stable; the small

initial increase of absorbance in both color reactions could

indicate some cleavage of the glycosidic link.

It should be noted that the most severe degradation is suffered

by the two compounds possessing a free carbosyl group, sialic

acid and methoxysialic acid. Whether the degradation is

caused by a specific action of the acetic acid or is due in part to

the acidity of sialic acid itself cannot be decided without further

experimentation.

The treatment with pyridine results in a rapid degradation

of sialic acid methyl ester (Fig. lOB), which was confirmed b3

both color reactions. With sialic acid (Fig. lOA) and mcthoxy-

-I

FIG. 9. Treatment with glacial acetic acid at 100” of (A) sialic acid, (B) sialic acid methyl ester, (c’) methoxysialic acid, and (I)) methoxysialic acid methyl ester. A 0.5 rn~ solution of sialic acid (80 ml) and 16 ml of 2.5 lllM solntions of sialic acid methyl ester, methoxysialic acid, and methoxysialir arid methyl ester in glacial acet,ic arid were heated in a boiling water bath under a reflux. Measured samples (10 ml for sialir arid and 2 rnt for the other derivatives) were removed at the indicat,ed times, cooled in ice water, evaporated immediately under reduced pressure at 35”, and dried overnight in a vacuum desiccator over KOH. The residues were dissolved in water, and samples corresponding to 0.45 and 0.036 finrote of the starting compound were used for the direct Ehrlich (--) and t,he thiobarhiturnte (- - -) reactions, respectively.

**- -.-----O.-----o h I I I I I *w--.------c -----. I I I I 1 I

I 2 :, U R’S

2 3 H

FIG. 10. Treatment with pyridine at 100” of (A) sialic acid, (B) sialic acid methyl ester, (C) methosysialic acid, and (D) methoxysialic acid methyl ester. Solutions of each compound in pyridine (16 ml; 2.5 11lM) were heated in a boiling water bath under a reflux. Measured samples (2 ml) were removed at the times indicated, cooled in ice water, evaporated immediately under reduced pressure at 35”, and dried overnight in a vacuum desiccator over concentrated H?SOl. The residues were dissolved in water, and samples corresponding to 0.45 and 0.036 rrnole of the starting compound were used for the direct Ehrlich (--) and the thiobarbiturate (- - -) reactions, respectively.

sialic acid (Fig. lOC), on the other hand, a disagreement between

the two color tests was observed, similar to that mentioned

before in the treatment with HCI: absorbance in the direct

Ehrlich reaction remained practically unchanged whereas the

thiobarbiturate reaction indicated a marked degradation. I’yr-

role formation, which is more pronounced under alkaline condi-

tions, could account for this discrepancy, as discussed before.

Methosysialic acid methyl ester (Fig. 1OD) also appears stable

by the direct Ehrlich test; the comparison, however, of the

results of the hydrosamir a&l reaction before and after the

treatment (3 hours) indicated a drop in the rxter content of

507.

.< mentioned beforr, the stability of the glyrosidic link to

acetic acid or pyridine could not be determined by treatment

of the simple derivatives, owing to further degradation. For this reason, an orienting experiment with a l)reparation of OS

brain mucolipid may be of interest; it shows that the pretrrat-

ment accorded these co~n~;les and delicate substances may not be

without consequence. .I specimen of mucolipid from OS brain

purified to “Stage III” (8) was found, by means of the direct

Ehrlich reaction with our l)rcparation of crystalline sialic acid

as the standard, to contain 32.4’;, of sialic acid.5 Portions

6 The preparation of ox brain mucolipid investigated previously in detail (8) had a sialic acid content of 2F.O%. Since that time, we have encountered numerous instances of preparations of the same degree of purity that contained considerably more sialic acid. For instance, in as yet unpublished studies in collaboration with I)rs. 0. W. Garrigan and N. %. Stanacer, 10 preparations were examined whose sialic arid rontent ranged from 27.3 to 32.4%, with a mean of 30.1% and a standard deviation of 1.4.


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

956 Sialic Acid Derivatives Vol. 239, X0. -I

(5 ml each) of 0.6% solutions of the lipid in pyridine and in glacial acetic acid were kept for 5 minutes at 100” and for 20 hours at room temperature, and then evaporated under reduced pressure at 35”. The aqueous solutions of the residues were subjected to dialysis against running water for 24 hours and against distilled water at 2” for the same period. The preparation that had been treated with pyridine was recovered by lyophilization in a yield of 80y0 of the starting material and contained 25.6c/;, of sialic acid; the product treated with glacial acetic acid was similarly recovered in a yield of 70yc and contained 20.4y, of sialic acid. An untreated control specimen, subjected to a parallel dialysis and recovery procedure, had a practically unchanged sialic acid content of 32.2%.

Concluding Remarks

The heteropolymers occurring in biological material are usually determined through the quantitative estimation of one or more of their characteristic monomeric constituents. This requires, among other things, the availability of procedures permitting the complete release and recovery of the intact mono- mer. For the proteins and the nucleic acids these requirements can, on the whole, be met, but this is less true of the lipids and the polysaccharides. The increasing interest in the biological function of various derivatives of sialic acid occurring in nature prompted the present study, which deals with the simplest representatives of the various types of links in which sialic acid is able to engage. Since the investigation of the quantity and the manner in which sialic acid is integrated into a polymer will have to make empirical use of reactions that mostly are not entirely specific, the results reported here may serve to place certain limitations on the interpretation of such studies.

Another cautionary reservation which may accrue from the present observations regards conclusions as to structure drawn from esperiments on the action of periodic acid on polymers containing sialic acid, such as were attempted in a previous study from this laboratory (10). The finding that the products of the action of periodate on the sialic acid derivatives studied here still give some direct Ehrlich reaction makes difficult the quantitative interpretation of oxidation experiments performed on an intact polymer.

The present studies also emphasize the necessity of gaining an insight into the types of links of sialic acid prevailing in the polymer to be investigated before the results of quantitative estimations can be evaluated properly. They may contribute to the choice of suitable means of release and determination of sialic acid and of the standards to be used in a given case. The great lability of the ester link in a sialic acid molecule that is not stabilized by glycoside formation (sialic acid methyl ester is completely hydrolyzed at pH 8 and 25” within 20 minutes) makes it evident how difficult the isolation of an intact macro- molecule may be under certain circumstances. It is not unlikely that many discrepancies in the literature are attributable to this sort of labilit,y having been overlooked. What this study underlines is that for results to be meaningful the appropriate standard must be chosen, the history of the sample must be fully known, and the hydrolysis method must be in keeping with the type or types of linkage in which sialic acid occurs in the polymer. Some of these polymers, in whose structure sialic acid participates, probably are to be counted among the most delicate and labile instances of macromolecules found in nature.

SUMMARY

The stability of sialic acid (isolated from ox serum proteins and composed of 63% of N-glycolylneuraminic acid and 375; of N-acetylneuraminic acid), of its methoxy derivative, and of the respective methyl esters of these compounds was studied under various conditions, mainly by ‘means of calorimetric techniques. A few orienting experiments were also performed with N-acetylneuraminic acid from human serum proteiny. The molar absorbance of sialic acid in the calorimetric testy was dependent on the presence of substituents on the carbonyl carbon atom. Periodate oxidation was found to modify, but

not to abolish, the reactivity in the color tests. The ester link of sialic acid methyl ester was cleaved rerl

rapidly at pH 8 and 25” and relatively fast in 0.1 N HCl at loo”, but simultaneous substitution of the neighboring hemiacetalic hydrosyl, as in methosysialic acid methyl ester, ren- dered the ester bond much more resistant to hydrolysis both by acid and by alkali. Treatment with 0.1 N HCl at 100” was found to cause a modification of the molecule of sialic acid, resulting in a loss of reactivity in the thiobarbiturate, but not in the direct Ehrlich, reaction. The hydrolysis of the glycosidic link of methoxysialic acid was investigated at various pH values; pH 3 was found to be the most suitable for a rapid hydrolysis without further degradation of the liberated sialic acid. Here again, the same effect of double substitution was observed: esterification of the neighboring carboxyl, as in methoxysialic acid methyl ester, increased markedly the resistance of the glycosidic link to acid hydrolysis.

Sialidase from Vibrio cholerae was inactive against the meth- oxyglycosides of sialic acid and its methyl ester. Heating with glacial acetic acid or pyridine caused degradation of sialic acid and, to a lesser extent, of its monosubstituted derivatives, whereas the doubly substituted methosysialic acid methyl ester appeared rather stable. Treatment with glacial acetic acid or pyridine of an OS brain mucolipid, used as a model of more complex sialic acid derivatives, resulted in a considerable loss of sialic acid from the preparation.

REFERENCES

1. Br,rx, G., in M. L. WOLFROM (Editor), Proceedings of thefourth international congress of biochemistry, Vol. I, Pergamon Press, London, 1959, p. 94.

2. ZILLIKEN, F., AND WHITEHOUSE, M. W., in M. L. WOLFROY AND R. S. TIPSON (Editors), Advances in carbohydrate chemistry, Vol. fd, Academic Press, Inc., New York, 1958, p. 237.

3. GOTTSCHALK, A., The chemistry and biology of sialic acids and related substances, Cambridge University Press, Cambridge, England, 1960.

4. SMITS, G., Psychiat. New-ok Neurochir., 64, 9 (1961). 5. ROSE&AN; S.,Federation Proc., 21, 1075 (1962). 6. ROSENBERG. A.. HOWE. C.. AND CHARGAFF. E.. Nature. 177.

234 (1956): ’ ’ ’ I

7. ROSENBERG, A., AND CHARGAFF, E., Biochim. et Biophys. Acta. 21, 588 (1956).

8. ROSENBERG, A., AND CHARGAFF, E., 1. Biol. Chem., 232, 1031 (1958).

9. ROSENBERG, A., BINNIE, B., AND CHARGAFF, E., J. Am. C’hem. Sot., 82, 4113 (1960).

10. ROSENBERG, A., AND CHARGAFF, E., Biochim. et Biophys. Arta, 42, 357 (1960).

11. KARKAS, J. D., AND CHARGAFF, E., Riochim. et Biophys. Acta, 42, 359 (1960).

12. STANACEV, N. Z., AND CHARGAFF, E., Biochim. et Biophys. Acta, 69, 733 (1962).


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

April 1964 J. D. Karkas and E. Charga$ 957

13. GARRIGAN, 0. W., AND CHARGAFF, E., Biochim. et Biophys. Acta, 70, 452 (1963).

14. BERGG~~RD, I., AND ODIN, L., Arkiv Kemi, 12, 581 (1958). 15. KARKAS, J. D., AND CHARGAFF, E., Federation Proc., 21, 80

(1962). 16. KUHN, R., AND BRCISSMER, R., Angew. Chem., 74, 252 (1962). 17. WERNER, I., AND ODIN, L., Acta Sot. Med. Upsaliensis, 67, 230

(1952). 18. WARREN, L., J. Biol. Chem., 234, 1971 (1959). 19. SVENNERHOLM, L., Biochim. et Biophys. Acta, 24, 604 (1957). 20. WEISSMAN, B., AND MEYER, K., J. Am. Chem. Sot., 76, 1753

(1954). 21. SVENNERHOLM, L., Acta Sot. Med. Upsaliensis, 61, 75 (1956). 22. M~RTENSSON, E., RAAL, A., AND SVENNERHOLM, L., Biochim.

et Biophys. Acta, 30, 124 (1958). 23. BLIX, G., LINDBERO, E., ODIN, L., AND WERNER, I., Acta Sot.

Med. Upsaliensis, 61, 1 (1956). 24. HIYAMA, N., TGhoku J. Ezptl. Med., 61, 319 (1949). 25. KLENK, E., AND FAILLARD, H., 2. physiol. Chem., 298, 230

(1954).

26. YAMAKAWA, T., AND SUZUKI, S., J. Biochem. (Tokyo), 39, 175 (19521.

27. SPRINSON, D. B., AND CHARGAFF, E., J. Biol. Chem., 164. 433 (1946).

28. GROSSMANN, P., 2. physik. Chem., 109,305 (1924). 29. SKRABAL, A., PFAFF, F., AND AIROI~DI, H., Monatsh. Chem., 46,

141 (1924). 30. SVENNERHOLM, L., Actu Chem. Stand., 12, 547 (1958). 31. AMINOFF, D., Biochem. J., 81, 386 (i961). 32. GIBBONS. R. A.. Biochem. J.. 82. 32~ (1962). 33. WEISSBA&, A.,’ AND HURW&, j., J.‘Bioi. Chem., 234, 705

(1959) . 34. SRINIVASAN, P. R., AND SPRINSON, D. B., J. Biol. Chem., 234.

716 (1959). 35. WARAVDEKAR, V. S., AND SASLAW, L. D., Biochim. et Biophys.

Acta, 24, 439 (1957). 36. BLIX, G., 2. physiol. Chem., 240, 43 (1936). 37. PAULING, L., The nature of the chemical bond, Ed. 3, Cornell

University Press, Ithaca, 1960, p. 495. 38. HEIMER, R., AND MEYER, K., Biochim. et Biophys. Acta, 27,

490 (1958).


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/

John D. Karkas and Erwin ChargaffStudies on the Stability of Simple Derivatives of Sialic Acid

1964, 239:949-957.J. Biol. Chem.

http://www.jbc.org/content/239/4/949.citation

Access the most updated version of this article at

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/239/4/949.citation.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at


http://ww

w.jbc.org/

Dow

nloaded from

http://www.jbc.org/content/239/4/949.citation

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;239/4/949&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/239/4/949.citation

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=239/4/949&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/239/4/949.citation

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/239/4/949.citation.full.html#ref-list-1

http://www.jbc.org/

studies on the stability of simple derivatives of sialic acid*

Documents