Studies on the Meningococcal Polysaccharides

Both Group B and C meningococcal polysaccharides were shown to be pure homopolymers of sialic acid. The two polysaccharides are, however, quite different chemically and immunologically. Acetyl determinations indicated that the C-polysaccharide contains both Nand 0-acetyl groups, whereas the B-polysaccharide contains N-acetyl but not Oacetyl groups. Acid-catalyzed methanolysis in anhydrous methanolic KC1 at 65” showed that the C-polysaccharide is much more susceptible to the cleavage than the B-polysaccharide. Under the conditions used, only 12 % of the B-polysaccharide was hydrolyzed while complete hydrolysis of the C-polysaccharide was achieved in 22 hours. On the other hand, treatment of the C-polysaccharide with two preparations of neuraminidase (Mbrio cholerae and Clostridium perfringens) resulted in the release of less than 2 to 3 % of free sialic acid. The B-polysaccharide was completely hydrolyzed by the enzymes under identical conditions. Possible implications of these observations on the structures of these two polysaccharides are discussed. A new procedure for the determination of sialic acid as methyl ester-methyl glycoside of neuraminic acid (methyl(methyl D-neuraminid)ate) or as methyl glycoside of neuraminic acid (methoxy neuraminic acid) on the amino acid analyzer has been developed and is discussed.


SUMMARY
Both Group B and C meningococcal polysaccharides were shown to be pure homopolymers of sialic acid. The two polysaccharides are, however, quite different chemically and immunologically.
Acetyl determinations indicated that the C-polysaccharide contains both N-and 0-acetyl groups, whereas the B-polysaccharide contains N-acetyl but not Oacetyl groups.
Acid-catalyzed methanolysis in anhydrous methanolic KC1 at 65" showed that the C-polysaccharide is much more susceptible to the cleavage than the B-polysaccharide. Under the conditions used, only 12 % of the B-polysaccharide was hydrolyzed while complete hydrolysis of the C-polysaccharide was achieved in 22 hours.
On the other hand, treatment of the C-polysaccharide with two preparations of neuraminidase (Mbrio cholerae and Clostridium perfringens) resulted in the release of less than 2 to 3 % of free sialic acid. The B-polysaccharide was completely hydrolyzed by the enzymes under identical conditions. Possible implications of these observations on the structures of these two polysaccharides are discussed.
A new procedure for the determination of sialic acid as methyl ester-methyl glycoside of neuraminic acid (methyl-(methyl D-neuraminid)ate) or as methyl glycoside of neuraminic acid (methoxy neuraminic acid) on the amino acid analyzer has been developed and is discussed.
The Group C meningococcal polysaccharide was first isolated by Watson and Scherp (2) and was shown by Watson,Marinetti,and Scherp (3) to be a polymer consisting primarily of sialic * A preliminary report of this work has been published (1). $ This research was carried out at Brookhaven National Laboratory under the auspices of the United States Atomic Energy Commission and by a contract with the United States Army Medical Research and Development Command, Office of the Surgeon General (MIPR 9959).
5 This research was also carried out at The Rockefeller University under a contract with the United States Army Medical Research and Develonment Command, Office of the Surgeon General (DADA 17-70-C-0627).
The advantage of this method is that the Group C polysaccharide isolated is of higher molecular weight and is immunogenic in man (5). Recently &tenstein et al. (6) have shown that immunization with this polysaccharide is able to prevent Group C meningococcal meningitis among young adults. It, t,herefore, is imperative to study the chemistry of this antigen in detail.
By the use of the cetavlon procedure it has been 1:ossible to isolate a protein free polysaccharidc from Group B meningococci (4). Inasmuch as this mat,erial may well be develolled as a vatcinc for human use, knowledge conccrrring its chemistry is important.
The present report describes investigations which have shown that both the B-and ('-polgsnccharides arc homopolymers of sialic acid, but that they are chcmicnlly and immunologically distinct. illaterials p-Toluenesulfonic acid, reagent grade, was purchased from the J. T. Baker Chemical Company, Phillipsburg, New Jersey, and recrystallized, chloride free as described by Liu and Chang (7 For the analyses of carbon, hydrogen, and nitrogen, a Perkin-Elmer model 450 gas chromatography machine was used. Phosphorus was det)ermined by the method of Chen,Toribara,and Warner (8) with slight modifications (9). Sialic acid u'as determined by the modified Ehrlich procedure of Werner and Odin (10) and by the thiobarbituric acid assay procedure of Warren (11). Reducing groups were determined by a modified (12) Park-Johnson procedure (13) and by the sodium borohydride procedure (9). N-Glycolyl group content of the polysaccharides were determined by a modification (14) of the method of Klenk and IJhlenbruck (15).

Noistw e Defermination
Moisture content was determined by the Carl-Fisher titration method (16).

Acetyl Determination
Four different methods were used for the determination of acetyl groups present in the polysaccharides.
Chromic Acid Procedure (17)-Samples containing 0.03 to 0.05 pmole of acet'yl were placed into a 25-ml round bottom flask which was attached to a reflux condenser.
Chromic acid solut.ion, 2.0 ml, prepared by mixing 4 parts of 5 N chromic acid and 1 part of concentrated H&0( were added to the flask and the mixture was heated in an oil bath maintained at 170 f 10". Two drops of concentrated H#OI were placed at the glass joint between the condenser and the flask to check for the tightness and to seal the joint.
The reaction was allowed to proceed for 30 min during which period the condenser was maintained with efficient cooling.
After cooling, the contents of the flask were quantitatively transferred to a micro-Kjeldahl distillation apparatus with the aid of IO-ml rinse of H20. Sat*urated K$O,, 2.0 ml, was added to the solution and the acetic acid was distilled for titration in the usual manner.
Toluenesulfonk Scid Procedure-A modification of the Eleck and Harte (18) procedure was used. Samples of polysaccharide (4 to 5 mg) were hydrolyzed with 2.0 ml of a 257, solution (w/v) of p-toluenesulfonic acid in sealed heavy wall glass tubes (Corning 9860, 18 x 150 mm) at 110" for 2 hours. The content of the tubes were quantitatively transferred to a micro-Kjeldahl distillation apparatus with the aid of rinses (2 x 2.0 ml) of water. During the sealing and opening of the glass tubes the tubes were kept frozen in a Dry Ice-acetone bath to prevent the escape of acetic acid. The acetic acid was distilled for titration with 0.01 N NaOH in the utiual manner.
NtZ,OIf Procedrtre-The method of Hestrin (19) was used for the reaction of 0-acetyl groups with hydroxylamine in alkali to form hydroxamic acid. The hydroxamic acid formed was measured by the formation of a colored complex with Fe+++ in acid solution.
A standard of P-n-glucose pentaacetate was run simult,aneously and the quantity of 0-acetyl in the sample was calculated by comparison with the standard curve obtained from the analyses of ,&n-glucose pentaacetate.
Hydrolysis wifh 0.01 N NaOH (go)-The procedure is described in the previous paper (9).

Ion Exchange Procedure
The amino acid and amino sugar content of acid hydrolysates was determined by ion exchange chromatography with the automatic recording equipment described by Spackman, Stein, and Moore (21).
For the determination of amino acids and acid-stable amino sugars, samples of polysaccharides were subjected to 6 N HCl hydrolysis.
Polysaccharide samples, 3 to 5 mg, were dissolved in 10.0 ml of H,O containing 1.0 mole each of norleucine and c+ amino-/3-guanidopropionic acid used as internal standards. A l.O-ml aliquot of this solution was mixed with an equal amount of 12 N HCl and hydrolyzed in a sealed evacuated glass tube at 110" for 2, 6, 12, and 24 hours. The hydrolysates were evaporated to dryness in a rotary evaporator at 40" and the residues were dissolved in 3.0 ml of the pH 2.2 buffer used with the amino acid analyzer.
The solutions were millipore filtered and 1.0.ml portions were added to the 7-and 60.cm columns of the amino acid analyzer with the aid of a sample injector (Chromatronix model SV-8031).
The recoveries of amino acids and amino sugars found on the chromatogram for each hydrolysis were calculated by reference to the internal standards norleucine and cr-amino-P-guanidopropionic acid present in the sample.

Methanolysis with Anhydrous Methanolic p-Toluenesulfonic Acid
Anhydrous p-toluenesulfonic acid, recrystallized and purified chloride free (7), was used for this purpose.
The concentration of p-toluenesulfonic acid used was 1.0 N in anhydrous methanol. Polysaccharide samples and LY-and P-methoxy neuraminic acid methyl ester (2 to 3 mg) were dissolved in 2.5 ml of 1 N p-toluenesulfonic acid in anhydrous methanol and 0.5-ml aliquots were distributed into four glass tubes (Kimble 45066A, 1 X 150 mm) equipped with Teflon-coated screw caps. The tubes were flushed with a stream of nitrogen, sealed with the cap, and immersed into a heating block (Exacta-Heat, model 218, Techn; Laboratory Instruments, Pequannock, New Jersey, hole depth 50 mm) maintained at 65 f 0.1" for 4, 8, 16, and 24 hours. At the end of methanolysis, the tubes were attached with a short section of Tygon to the condenser of a rotary evaporator which can be operated with the condenser axis at a downward tilt of about 30". The methanol is removed in about 20 min at 40". Alternatively, the solvent can be removed by evaporation in a stream of nitrogen at 40".
For this purpose the methanolysate was treated with 1.10 ml of a 0.5 N NaOH for 60 min at 25" (pH should be 12 to 13). The solutions were quantitatively transferred to a 5.0.ml volumetric flask and made to volume with water.
One-milliliter aliquots were used for analysis on the 60-cm column of the amino acid analyzer with the pH 3.25 buffer as eluent.
The amino acid analyzer constant for methoxy neuraminic acid determined with an authentic crystalline preparation was 5.12 for an instrument for which the aspartic acid constant is 8.84. The elution volume of methoxy neuraminic acid and aspartic acid are 40 and 65 ml, respectively, on this instrument. methanol. The concentration of HCl in methanol was adjusted to 1.0 N. Polysacchnride samples and a-and P-methoxy N-acetylneuraminic acid methyl esters (2 to 3 mg) were dissolved in 1.0 ml of Hz0 and 200-111 aliquots were placed into four glass tubes (12 x 150 mm) equipped with Teflon-coated screw caps. The contents of the tubes were dried over YsOs in a vacuum. Methanolyses of the samples were performed with 1.0 ml of the 1 K methanolic HCI for 4, 8, 16, and 24 hours at 65" as described above. The methanolysate was then exposed to 5.0 ml of 0.01 N NaOH for 60 min at 25". One-milliliter aliquots were used for analyses on the 60.cm column of the amino acid analyzer as described in the previous section.

Gas Chromatography
Kinhydrin-negative components of the polysaccharides were determined by gas chromatography on a stainless steel column, approximately 10 feet in length and 0.125 inch in diameter, packed with Chromsorb W coated with 3% OV-17 polymer. The column was equilibrated at 120" in a Perkin-Elmer model 900 gas chromatograph equipped with a flame ionization detector. The procedure used was the one described by Reinhold et al. (22) which utilizes anhydrous methanolic HCl for the cleavage of the polysaccharide and analyses of the trimethylsilylated carbohydrate derivatives on the column of the gas chromatograph.
Periodate Oxidation l'eriodate oxidations and formaldehyde determination were performed on the polysaccharides essentially according to the method of Suzuki and Strominger (23) and as described by Liu et al. (9).

Digestion of Group B and Group C Polysaccharides with Neuraminidase
Group B, and C-polysaccharide, de-O-acetylated C-polysaccharide, and partially hydrolyzed C-polysaccharide were treated with neuraminidases from V. cholerae, C. perfringens, and influenza virus, at 37" essentially according to the method of Cassidy, Jourdian, and Roseman (24). Thiobarbituric acid assays were performed on aliquots of the digestion mixture to measure the release of free sialic acid.

Optical Rotation
Optical rotations of the polysaccharides were measured with a Zeiss photoelectric precision polarimeter capable of an accuracy of rtO.005".

Isolation of N-ilcetylneuraminic
Acid from Group B Potysaccharide A sample of B-polysaccharide (50 mg) was dissolved in water (5 ml) and 0.09 N NaOH was added to the solution until the pH remained constant at 11.0 to hydrolyze any internal ester which Group B and Group C Meningococcal Polysaccharides Vol. 246,No. 15 might exist (25). The pH was then reduced to 5.0 with 0.1 N HCl.
The volume was brought to 20 ml with 0.1 M acetate buffer, pH 4.9. C. perfringens neuraminidase, 2 ml, was added and the solution was incubated at 37" for 72 hours.
Thiobarbituric acid assay performed on an aliquot of the incubation mixture indicated 38 mg of free sialic acid was present.
The enzymic hydrolysate was adjusted to pH 8 to 9 with 1 M NH,013 and placed on a Dowex l-X8 (formate form, 100 to 200 mesh) column (2 x 24 cm). The column was washed with 100 ml of Hz0 and then eluted with dilute formic acid by gradient technique.
A 500.ml Erlenmeyer flask containing 0.05 M formic acid served as the mixing vessel. To this was gradually added 2.0 N formic acid and 6.0.ml fractions of the eluate were collected.
Analyses were performed for sialic acid on 200+1 aliquots from each tube with the Ehrlich reagent according to the method of Werner and Odin (10). Fractions 330 to 402 ml contained N-acetylneuraminic acid, 22 mg. The peak was dried by lyophilization and the resulting material was crystallized by dissolving in 0.5 ml of H20 and diluting with 2.5 ml of glacial acetic acid as described by h4cGuire and Binkley (25) Isolation of N-Acetylneuraminic Acid from &oup C Polysaccharide Group C polysaccharide (50 mg) was dissolved in 5.0 ml of H20, 50 mg of Dowex 50-X8 (H+ form) was added and the mixture was heated with stirring at 100" for 40 min, cooled, the resin removed by filtration, and the filtrate m-as lyophilized. The resulting white powder was dissolved in Hz0 (2 ml) and placed on a Dowex l-X8 (formate form) column (2 X 24 cm) and the column was washed n-ith 100 ml of Hz0 and then eluted with a gradient of formic acid as described in the previous section.
The fractions containing N-acetylneuraminic acid were located by the Ehrlich reagent, combined, and lyophilized, yielding 28 mg of a white powder.
Preparation of De-0-Acetylated C-Polysaccharide C-Polysaccharide (33 mg) was treated with 10.0 ml of 0.02 N NaOH for 22 hours at 25". The solution was neutralized with 1.0 ml of 0.2 N acetic acid, dialyzed extensively against water to remove salt, and lyophilized.
The de-O-acetylated Cpolysaccharide (25 mg) prepared in this manner contains less than 0.12 residue of 0-acetyl groul) per residue of sialic acid present in the polymer (see Table IV) and was used for the studies of its susceljtibility to enzymic hydrolysis with neuraminidases.
Partial Hydrolysis of C-Polysaccharide C-Polysaccharide has weak reducsillg Ilroperties when tested with the Park-Johnson reagent. This may be caused by the presence of a free reducing groul) on one end of the polymer chain.
Upon heating Cl)olysnccharide in water between l)H 3 to 5, hydrolysis occurs, and the amount may be estimated by measuring the increase in reducing value. A 0.1% solution of C-polysaccharide in water was heated at 100" for 30, 60, and 90 min. In Fig.  3 is shown a hydrolysis curve of the C-polysaccharide under these conditions. A standard curve of reducing groups determined versus sialic acid concentration was constructed and this curve was used to determine the degree of hydrolysis of ('-1)olysaccharide. Liu, Gotschlich, Dunne, and Jonssen any glucose , galactose, mannose, xylose, fucose, or ribitol (less than 1% by weight).
The polysaccharide is also devoid of glycoyl group.
Acid catalyzed methanolysis of the polysaccharide with 1.0 N HCl in anhydrous methanol followed by the identification of the product by ion exchange chromatography as methyl(methy1 D-neuraminid)ate or methoxy neuraminic acid methyl ketoside (Compounds II and III in Fig.  1) or by gas chromatography as trimethylsilyl derivative yielded only 11 to 12% of the polysaccharide as sialic acid. However, when the B-polysaccharide was treated with anhydrous methanolic toluenesulfonic acid followed by analysis on the ion exchange column 74.8 to 85.2% by weight was recovered as sialic acid. The Group B polysaccharide could also be hydrolyzed with neuraminidases from V. cholerae and C. perfringens.
Thiobarbituric acid assay performed on an aliquot of the incubation mixture indicated that the sialic acid released accounted for 76 to 82% of the weight of the Group B polysaccharide added. The analytical data in Table II also indicate that nearly all the nitrogen in the preparation can be accounted for as sialic acid. The moisture content of the B-polysaccharide as determined by the Carl-Fisher titration was 10.07, for lot B-4. The results of these analyses indicate that the sum of the moisture, sodium, sialic acid, and N-acetyl accounts for about 88y0 of the weight of the material. Chemical Composition of Group C I'olysaccharide-The results of the analyses of Group C polysaccharide are presented in Tables I and III. In the case of the Group C polysaccharide the toluenesulfonic acid method for the determination of total acetyl seems to give too high an estimate whereas the chromic acid digestion gives a more reasonable value.
If this later figure (17.2%; 410 pmoles/lOO mg) is used and the 0-acetyl (5.8%; 138 pmoles/lOO mg) is subtracted, the N-acetyl content is 272 moles, in agreement with the content of sialic acid. The major constituents sialic acid, acetyl, sodium, and moisture account for 96%.
Hydrolyses with 6 K HCI at 110" and subsequent analyses of the hydrolysates on the amino acid analyzer revealed that there were no amino acids or muramic acid in the preparation indicating that the material is free of protein and mucopeptide. By gas chromatographic analyses the C-polysaccharide contained no ninhydrin-negative monosaccharides such as glucose, galactose, mannose, xylose, fucose, or ribitol.
The material is also devoid of glycoyl group.
Digestion of the C-polysaccharide with neuraminidases resulted in the release of less than 3% of free sialic acid as determined by the thiobarbituric acid test. Optical Rotatory Properties and Infrared Spectra-The optical rotation of a 1.03% aqueous solution of the B-polysnccharide Group B and Group C Meningococcal Polysaccharides Vol. 246,No. 15 Table IV).
Infrared spectra of Group B and Group C polysaccharide, de-O-acetylated Group C polysaccharide, colominic acid, and N-acetylneuraminic acid were taken in potassium bromide pellets with 1 to 2 mg of the polysaccharide mixed with 150 mg of the salt. The spectra are recorded in Fig. 2.
Reducing Group Analyses-On reduction with NaBH4 sialic acid consumes 1 mole of hydrogen, the keto group obviously being reduced to a secondary alcohol group.
The reduction was allowed to proceed at room temperature for 18 hours. Glacial acetic acid (500 ~1) was added to destroy the excess sodium borohydride.
Methanol, Since it was found that even after liberation of the excess of tritium with acetic acid and many subsequent treatments of the residue with methanol to form the volatile methyl borate, n detectable amount of radioactivity still remained; a blank was carried out containing the same amount of NaWH4 as the sample. This blank correction never exceeded 10% of the tritiurn incorporated into the sample and in our experience is more satisfactory than purification of the reduced product by column chromatography since in some instances the radioactive contaminant moved alol!g with the reduced products.
A standard curve was cotlstructed with the values obtained from the N-acetylneuraminic acid experiment (Fig. 4). The number of reducing groups present per g of the polysaccharides and the oligosaccharides were computed from the amount of radioactivity incorporated into each sample with the standard curve (Fig. 4). It was found that there is one reducing group per 74 residues of NAN in the ll-polpsaccharide and the value for the C-polysaccharide is one per 143 residues.
Periodate Ozidalion- Table   IV and Fig. 5 show the results obtained from the oxidation of colominic acid, N-acetplncuraminic acid, and the Group I< and C polyxaccharides with l)eriodate. The uptake is less than 0. Fro. 4. Incorporation of radioactivity ("IT) into N-acetylneuraminic acid by treatment with tritiated sodium borohydride. This c~6r~e was used ill computing t,he number of reducing groups in B-and C-polysaccharides.
hours per pmole of NAN in Group B polysaccharide and less than 0.04 pmole of formaldehyde per I.cmole of NAN is produced. The uptake of periodate for the C-polysaccharide is less than 0.6 Fmole per /.mlole of NAN after 22 hours, and the formation of formaldehyde is also less than 0.02 pmole per pmole of NAN.
Since the C-polysaccharide was found to bc 0-acetylated, and that this group may interfere with periodate oxidation, the periodate oxidation was also performed with C-polysaccharide that had been treated with base to remove 0-acetyl group.
The results were essentially identical with the native material.

DISCuSsION
In the early stages of our work, WC had used the standard conditions (0.1 N HzS04, 80", 1 hour, see Reference 26) to liberate sialic acid from our polysaccharides.
We soon found out, however, that the results obtained from such studies were misleading and inaccurate.
Sialic acid is in general unstable in aqueous acidic conditions (26,27). The standard mild condition widely used (26) is good perhaps only for the liberation of sialic acid from glycoproteins because of its terminal positions in these molecules.
When the method was applied to polysaccharides with molecular weight in excess of 100,000 such as B-and C-polysaccharides, the method failed to yield more than 2O"r of the sialic acid content in 60 min. Prolonged incubation (3 to 4 hours) resulted in higher recovery of sialic acid (40 to 50%).
But the yield never exceeded 55'3 from both B-and C-polysaccharides.
The failure of the "standard" procedure to yield more than 55% of the sialic acid from these polysaccharides is most likely caused by compensating factors; continuous release and destruction of the released sialic acid during the hydrolysis in 0.1 N HzS04 at 80".
A new procedure for the analysis of sialic acid in polysaccharides has been developed.
The method utilizes p-t,oluene- sulfonic acid instead of I-ICI as the catalyst for methanolysis and the amino acid analyzer for the quantitative estimation of sialic acid as methoxy neuraminic acid (Compound III, Fig. 1). Under the acid-catalyzed methanolysis condition used, the product is presumably essentially the /3 anomer (27). However, when authentic samples of LY-and P-methoxy neuraminic acid were analyzed on the amino acid analyzer, the two anomers coeluted from the column.
The color constants for the two anomers are similar.
The method has been applied to several glycoproteins and cell wall carbohydrates and was also found to be successful in the determination of sialic acid content in these samples.* Cleavage of Group B and C polysaccharides with 1 K anhydrous methanolic HCl at 65" for 22 hours (a procedure commonly used for the methanolysis of polysaccharide (22)), followed by conversion of the methyl glycosides of sialic acid to their trimethylsilyl derivative and analyses by gas chromatography gave sialic acid content for the I%-polysaccharide as 11.2% (Table II, Lot B-4) and the C-polysaccharide as 77.7% (Table III, Lot C-2). When the products of methanolysis were saponified and analyzed directly on the 60-cm column of the amino acid analyzer, the yield of methoxy neuraminic acid from these polysaccharides was 12.3% for the B-polysaccharide (Table II, Lot B-4) and 76.2oj, for the C-polysaccharide (Table III, Lot C-2). It appears, therefore, that the analyses of sialic acid after its cleavage from the polysaccharide with Group B ad Group C dieningococcal Polysacclzaricles Vol. 246,No. 1;5 anhydrous methanolic HCl by either gas chromatography as the trimethylsilated derivative or by the amino acid analyzer as methosy ueuraminic acid will give almost identical results. Enzymic hydrolysis of the polysaccharides with neuraminidases followed by the Warren test on aliquots of the digestion mixture revealed the release of free siahc acid from the B-polysaccharide to be 76.4y0 (Table II, Lot. B-4), but from the C-polysaccharide, the yield of free sialic acid was less than 37, (Table III, Lot C-Z).
The conclusion to be drawn from these studies are as follows: 1. Acid-catalyzed methanolysis of polysaccharide with 1 x anhydrous methanolic HCl at 65" for 22 hours, sometimes may not be sufficient to cause the complete release of sialic acid from the polysaccharide as is shown in the case of Group B polysaccharide.
2. Enzymic digestion of polyeaccharide with neuraminidases likewise may not cleave all the sialic acid from the polysaccharides as is shown in the case of the C-polysaccharide.
Evidently, these neuraminidases are not capable of hydrolyzing certain ketosidic linkages of sialic acids or its 0-acetylated derivative.
3. When anhydrous methanolic 1 N p-toluenesulfonic acid is used as a catalyst in methanolysis, the release of sialic acid from the polymer has been consistently higher as is shown here with the B-and C-polysaccharides and with other sialic acid-containing glycoproteins and cell wall polysaccharides.2 For the estimation of sialic acid content in polysaccharides and glycoproteins, the advantages of using p-toluenesulfonic acid as a catalyst for methanolysis, therefore, are a-fold. First, the reagent is effective in causing more complete cleavage of sialic acid from the polysaccharides or the glycoproteins.
Second, the product of methanolysis after removal of solvent, can be analyzed directly on the ion-exchange column of the amino acid analyzer without further derivatization as is required for the gas chromatographic proccdurc.
The use of 3 x aqueous p-toluenesulfonic acid in the hydrolysis of proteins aild the quantitative estimation of t,ryptophan by the amino acid analyzer has recently been described (7).
The results of the present studies indicate that both the Band the C-polysaccharides of meningococcus are essentially pure polymers of sialic acid. The high yield of sialic acid or its derivative obtained either by methanolysis with toluenesulfonic acid as catalyst, or by enzymic hydrolysis of the polysaccharides, and the absence of other carbohydrates, proteins, and nucleic acid support this conclusion. This concept is supported by the finding that the infrared absorption spectra of NAN isolated from B-and C-polysaccharides are similar to an authentic sample of NAN ( Fig. 2A). Furthermore, the infrared absorption spectra of these polysnccharides are quite similar to colominic acid (Fig. 2B), a polymer known to consist mainly of NAN. The existence of sialic acid homopolymer in nature was first shown by Barry and Goebel (28). They prepared a pure polymer of sialic acid from the culture fluid of Escherichia co& K235-L + 0 and named this compound colominic acid. Subsequently, colominic acid-like compounds have been isolated from other strains of E. co& (12). The existence of sialic acid polymer in the culture medium of Group C meningococcus was first reported by Watson et al. (3). On the other hand, the nature of the group-specific polysaccharide from Group B meningococcus has hitherto not been identified.
One outstanding feature of the B-and C-polysaccharide prepared by the cetavlon method (4) and used for the present studies is their high molecular size. The 1)olysaccharides prcpared in this manner have mean molecular weights in excess of 100,000 as has been shown both by the ultracentrifugal studies3 and gel filtration studies (4). The fact that these polysaccharides can be prepared free of mucopeptide, without the use of trichloroacetic acid, or hot formamide, which are required for the preparation of mucopeptide-free group-specific carbohydrates from many of the gram-positive bacteria, together with the finding that muramic acid phosphate was absent from the preparation, would support the idea that these polysaccharidcs are not cross-linked to the mucopel)tide in these gram-negative organisms (29).
As noted earlier, the B-polysaccharide contains one reducing group per 72 residues of NAN.
The formula weight of JSAX in the polymer is 291. The molecular weight of an average chain length of polymer with 72 residues of NAN is 291 X 72 = 21,000. Yet, the average molecular weight of the B-polysaccharide is in excess of 100,000. This might imply that the B-polysaccharide consists of cross-linked five to six chains each with an average length of 72 residues of NAN, or that the preparation contains a mixture of higher molecular weight polymer and lower molecular weight polymers that arc not separated.
For the C-polysaccharide, since there is one reducing group per 143 units of N4N, and since the molecular weight of the polymer is in esccss of 400,000,3 the molecule might consist of nine to ten cross-linked chairis with an average chain length of 140 residues of NAN or that the preparation is contaminated with low molecular weight polysaccharides that were not separated.
The nature of the cross-links that might exist in the B-and the C-polysaccharides have not been studied. Comparison of the uptake of periodate by sialic acid and by these polysaccharides may be instructive (see Table IV and Fig. 5). If N-acetylneuraminic acid retains the 1)yranose ring structure in the 1)olymer as has been suggested by McGuire and Binkley (25) for colominic acid, then the uptake of less than 0.1 or 0.2 pmole of llcriodatc per NAN urlit in l<-and C-polysaccharides, respectively, would be best explained by a two to eight ketosidic linkage between the NAN units in both 1)olysaccharides. Any other linkage would require the uljtake of at least 1 pmole per NAN unit. The formation of less than 0.1 pmole of formaldchyde per residue of sialic acid present in the polysaccharide even after 22 hours of periodate oxidation is in agreement with this proposal.
For C'-polysaccharide, which is partially 0-acetylated, the same conclusion is valid since the de-O-acetylated polysaccharide was similarly unaffected by the treatment with periodate.
The position of the 0-acetyl group iu the C-polysaccharide remains to be established.
Although B-and C-polysaccharides are both homopolymers of sialic acid some important differences do exist between the two polymers.
Chemically, the 1%polysaccharide is devoid of Oacetyl group, while the C-polysaccharide is partially 0-acetylated. As showu in Fig. 6B, the B-polysaccharide is much more rcsistant to the acid-catalyzed methanolysis in anhydrous Il( I than the C-polysaccharide. during methanolysis in 1 N anhydrous methanolic ptoluenesulfonic acid. The time of methanolysis is plotted against the number of residues foimd per mg of sample. B, rate of release exists between the I%-aud C-polysaccharides.
The two polysaccharides also differ in their susceptibility to neuraminidase. Inspection of Fig. 7 indicates that while the B-polysaccharide is almost completely split iuto monomeric NAN by the hydrolytic action of neuraminidasc, the C-polysaccharide and its de-Oacetylated derivative are resistant.
Under identical conditions, colominic acid is rapidly and completely hydrolyzed to NAN by the enzyme. From the study of the consumption of alkali before and after saponification and reaction with NH20H, PIlcGuire and Binkley (25) suggested that in colominic acid some of the carboxylic groups exist as ester linkages with the hydroxyl groups of NAN.
It was also noted that the enzyme neuraminidase hydrolyzed the "base-treated" colominic acid much more rapidly than the untreated colominic acid. The data shown in Table I for the acetyl determinations of the IL and Cpolysaccharides seem to rule out the possibility for the existence of "internal ester bond" such as the one proposed to exist in colominic acid. In our experiment, treatment with 0.01 N NaOH followed by distillation and titration will give the value for the volatile acid, acetic acid (the existence of glycoyl group in the Rand C-polysaccharides has been ruled out by the calorimetric test of Klenk and Uhlenbruck (15)) which exist in the Cpolysaccharide as 0-acetyl groups. Reaction with NILOH on the other hand could be attributed to the presence of 0-acetyl groups, the presence of lnctone linkages, or ester linkages, between the carboxyl group of one unit and a hydroxyl group of a neighboring unit. Since our data show that the two values obtained by the 0.01 x NaOH procedure and the NHzOH are nearly identical, it must follow that lactone and ester linkages do not exist in the C-polysaccharide.
For the Group B polysaccharide, since the reaction with NHBOH is negative, neither 0-acetyl nor an internal ester can exist.
Slowness of the attack by the enzyme neuraminidase on some sialic acid containing glycoprotein has been attributed to the presence of 0-acetyl groups substituted on the sialic acid residues (30, 31). Confirmation of this hypothesis was offered in the case of bovine submaxillary mucin when almost complete hydrolysis was obtained after treatment of the mucin with alkali (30, 31). Cassidy et al. (24), however, cautioned against this interpretation for two reasons; first, in their studies with neuraminidase isolated from C. perjringens and from V. cholerae, both enzymes showed no significant difference in the rate of release of NAN from the acetylated and deacetylated substance, and secondly, the effects of alkali on mucins may be much more profound than simple hydrolysis of 0-acetyl linkages.
In the present studies, the C-polysaccharide was not affected by the enzyme even after it had been treated with base to remove 0.acetyl group; hence, the inactivity of the enzyme toward this by guest on March 22, 2020 http://www.jbc.org/ Downloaded from 471" Group B and Group C Meningococcal Polysacch,aritles Vol. 246,No. 15 polymer of sialic acid is unlikely to be caused by the presence and Mr. S. J. Tassinari for his expert performance of the microof 0-acetyl group.
analysis. In order to ascertain whether the resistance of the C-polysaccharide to the hydrolytic action of neuraminidases is the result of its large molecular size, we have prepared a mixture of oligosaccharide from the C-polysaccharide by partial hydrolysis in boiling water (see "Experimental Procedure"). The product was shown to have an average molecular size of 7 to 8 sialic acid units per one reducing group determined by the Park-Johnson procedure.
When this preparation was treated with neuraminidase from V. cholerae at pH 5.0 in the presence of serum albumin as described by Cassidy et a.Z. (24), free sialic acid was released to the extent of 11.7% from the oligosaccharide in 24 hours at 37". Under identical conditions, the release of monomeric sialic acid from the original unheated C-polysaccharide was 1.3 y0 and from the B-polysaccharide was 100%.
The results of these experiments would seem to suggest that the inactivity of the neuraminidases toward the C-polysaccharide is not likely the result of the presence of 0-acetyl group or to its large molecular size, although smaller fragments of the C-polysaccharide are hydrolyzed more extensively than the one with larger molecular size.
Some evidence in support of this hypothesis was offered by McGuire and Binkley (25) when they showed that the enzyme failed to release more than 30/O of one form of methyl glycoside of N-acetylneuraminic acid. This form of glycoside was suggested to be in p conformation. Subsequently, Yu and Ledeen (33) conclusively showed that the anomer which is resistant to neuraminidase is the P-ketoside and that the one susceptible to the enzymic hydrolysis is the Qketoside.
If this proposal is correct, the results presented in this manuscript would suggest that the B-polysaccharide consists of a-ketosidic linkages and that the C-polysaccharide might possibly have /3-ketosidic linkages.
The existance of P-ketosidic linkages involving sialic acids in polysaccharides or in the glycoprotein have so far not been reported in the literature.
It should be emphasized that our suggestion that the C-polysaccharide might involve P-ketosidic linkages is tentative especially in view of the fact that the specificity studies of the neuraminidases were performed on synthetic substrate, N-acetylneuraminate methyl ketoside rather than on di-or trisaccharides of sialic and, and that it was found that 3% of the P-methyl ketosides were hydrolyzed by the enzyme. It is not known from these studies whether its 3% cleavage is caused by the contamination of the substrate with the a-ketoside or is the result of the intrinsic property of the neuraminidases.
Studies are uiiderway in our laboratory to delineate t'hese points. WARREN,