The Isolation and Structure Identification of a Disaccharide Containing Mairano-muramic Acid from Micrococcus lysodeikticus Cell Wall*

SUMMARY The dialyzable fraction obtained after degradation of Micrococcus lysodeikficus cell wall with egg white lysozyme was fractionated on columns of Dowex 50 and Dowex 1 ion exchange resins. In addition to the di- and tetrasaccharides and disaccharide peptides previously reported, a new tetrasaccharide and a new disaccharide were isolated in low yields, about 3 to 4 % and 2 %, respectively, of the total carbohydrate chain of the peptidoglycan. The tetrasaccharide contains 2 residues of 2-acetamido-2-deoxy-D-glucose (N-acetylglu-cosamine), 1 residue of N-acetyhnuramic acid, and 1 residue of muramic acid, probably as the internal amide. residue - 1 residue


SUMMARY
The dialyzable fraction obtained after degradation of Micrococcus lysodeikficus cell wall with egg white lysozyme was fractionated on columns of Dowex 50 and Dowex 1 ion exchange resins.
In addition to the di-and tetrasaccharides and disaccharide peptides previously reported, a new tetrasaccharide and a new disaccharide were isolated in low yields, about 3 to 4 % and 2 %, respectively, of the total carbohydrate chain of the peptidoglycan.
Despite the observation that a mirroheterogeneity of the components and of the chemical structure is well established for complex carbohydrates, no other carbohydrate components have been detected in all bacterial cell walls inr-esti-Rated until now, except for derivatives resulting from t,he deacttylation of N-acet~~lmuramic acid to give muramic acid having a free amino group (6) or internal amide (7).
Extensive invest+ tion of the low molecular weight oligosaccharides obtained by degradation with egg white lysozyme of Xicrococcus Zysodeik~icus cell wall showed the presence of additional oligosaccharides, besides the known di-and tetrasaccharides previously described (3). The isolation and structure identification of one of these oligosaccharides is the subject of the present paper.

METHODS AKD NATERIAL
General-Optical rotations were determined in semimicro or micro (for amounts smaller than 3 mg) tubes with lengths of 100 or 200 mm, with the use of a For two-dimensional combined paper chromatography and electrophoresis, the samples (0.5 to 0.6 mg) were deposited according t,o the method of Hoshino (8).
The compounds were detected on paper chromatograms and electrophoretograms by the benzidine-trichloroacetic and the ninhydrin reagents, and, additionally on paper chromatograms, by the silver nitrate and the benzidine-periodate reagents, or by treatment with alkali and observation under ultraviolet light, as described by Sharon and Seifter (9 Sharon and Jednhz (15) in a 400.ml stainless steel mixing chamber. The homogenization was performed for 90 min, and the yield of c&e11 walls was 4.6 to 4.8 g. The dialyzable fraction of the lysozyrne digest was obtained according to the method of Sharon et al. (5) starting with batches of 10 to 20 g of lyophilyzed cell XlllS.
The yield was 2.5 to 3.0 g of dialyzable material and 6.0 to 6.5 g of nondialyzable material from 10 g of cell wall.

Separation OJ Oligosaccharides and Oligosaccharide
Peptides by Chromatography of Dialyzable Fraction on Dowex 50-To a column (1.6 cm x 50 cm) of I)owex 50 (Hf form) was added a solution of the dialyzable material (2.90 g) in water (6 to 8 ml). The elution was started with 2 liters of water.
Fractions of 18 ml were collected at a rate of 72 ml per hour.
After 100 fractions had been collected, the column was connected to a mixing chamber conlaming 2 liters of water, to which a reservoir containing 2 liters of 0.4 RI hydrochloric acid was attached. Volume and rate were kept the same, and 200 fractions were rollected.
The reducing power was determined OIL aliquots (0.1 ml) of alternate tubes, while aliquots (0.5 ml) were analyzed with the Reissig, Strominger, and Leloir (16) modification of the Rlorgan-Elson test, with a heating time of 35 min (5). Free amino groups were determined on aliquots of 0.1 to 0.2 ml with the ninhydrin method (11). The hydrochloric acid concentration was determined, on a l-ml sample of every 10th tube, by titration with 0.1 M sod,um hydroxide in the presence of phenolphthalein as indicator.
The pattern of elution is shown in Fig. 1 and the yields are reported in Table I. A similar experiment (Experiment 2, Table I) was performed on 4.26 g of dialysate using the same column of Dowex 50, but prolonging the water elution to 150 tubes and using a less steep gradient of hydrochloric acid. This resulted in a better separation of the third fraction (W,) eluted with water and of the fourth fraction (P4) eluted with hydrochloric acid, but the main fractions were essentially the same.
The propert.es of Fractions W2 and PZ are reported in Tables III and IV. Isolation of Oligoxaccharides by Chromatography on Dowex I-Acetate-The separation was performed as previously described (5). Fraction W, (1.28 g) obtained frorn the Dowex 50 column was dissolved in water (5 ml).
The solution was applied to a column of Dowex 1 (AcO-form; 2.2 cm x 60 cm; bed volume approximately 220 ml). Elution was performed with 0.8 M acetic acid introduced into a mixing flask containing 2 liters of water. Fractions of 18 ml were collected at a rate of 72 ml per hour.
The fractions were examined for reducing power and by the Morgan-Elson test.
The concentration of acetic acid was determined by titrating the solution of every 10th tube with 0.1 M sodium hydroxide in the presence of phenolphthale~n.
The elution pattern is shown in Fig. 2. The material from the six peaks was recovered by lyophilization of the solutions and analyzed by paper chromatography and paper electrophoresis. The yields are reported in Table II. The material recovered in the six fractions represented more than 70% of the material applied to the column. A similar experiment (Experiment 2, Table II) was performed on 1.93 g of Fraction W, and gave similar results, but the overloading of the column resulted in a relatively smaller Fraction Dd, contaminated in part by Fraction D5. Properties of Fractions Isolated from Dowex 1 Column-Only the fractions which were homogeneous on paper chromatography or paper electrophore.;is in various solvent systems and in various buffers using a wide pII range were investigated.
The properties of the fractions are reported in Tables III and IV and in Figs. 3 and 4.
The first peak (I),) consisting of 2-acetamido-2-deoxy-n-glucose and a small amount of neutral oligosucchnrides, and the second peak (Dz) consisting of a small amount of weakly acidic oligosaccharides, were not further investigated.
Fraction D,-The material obtained from the third peak (D3) showed only one spot on paper chromatography and electrophoresis (Table IV).
The fractions (18 ml) were eluted at the rate of 72 ml per hour.
The elution was performed with a gradient of acetic acid.
The reducing power (---) and the color formation in the Morgan-Elson test (----) were determined as described under "Methods and Materials" and "Results." The yields are reported in Table IT. The yields are reported in Table I.   No. 3MM paper, and the chromatogram was developed with Solvent E. The substances were localized by cutting thin strips on the edges which were tested with ninhydrin and the silver nitrat,e reagent. From the remaining part of the paper, the bands migrating at the same speed as muramic acid were cut off and eluted with water. The eluates were evaporated to give, each, about 4 mg of material.
An aliquot (2 mg) of each of the residues was suspended in 0.5 ml of dry dichloromethane, and the solution was cooled in an ice bath.
To the suspension was added 0.25 ml of boron tribromide cooled with ice.

Synthetic
(g&o)-muramic acid was also similarly treated as standard. The reaction mixture was kept for 48 hours at room temperature under moisture exclusion. The substances gradually dissolved, and, after 48 hours, the colored reaction mixture was evaporated in a vacuum.
Methanol (2 ml) was added to the glassy residue and evaporated, and this procedure was repeated twice.
Disaccharide D5 migrated at the same speed, different from that of the product obtained from the acidic component of Disaccharide l)?. The three spots, however, migrated at a speed faster than that of any of the known hesosamines, and it was assumed that the lactyl group had migrated, under the anhydrous conditions of the reaction, to the amino group or to one of the hydroxyl groups.
Consequently, the residue was hydrolyzed with 1 M hydrochloric acid (0.1 ml) for 1 hour at 100". After evaporation,    had ceased, the tube was sealed and kept for 4 hours at 100". The solution was evaporated at 5" in a vacuum desiccator in the presence of potassium hydroxide and concentrated sulfuric acid.
The residues were analyzed on an amino acid analyzer. In both disaccharides, the proportion of n-glucosamine was unchanged, whereas the proportions of muramic acid and of mannomuramic acid (calculated as n-glucosamine) was decreased to 11 y. and 36yo for Disaccharides DE and Dd, respectively. Iodine Oxidation of Disaccharides Dq and DS-To solutions of the Disaccharides Dd or D, (0.5 mg), respectively, in water (0.05 ml), were added at 0" 5 mM iodine (0.2 ml) and 0.05 ml of 0.1 M sodium carbonate-sodium hydrogen carbonate (1: 1). The solution was kept overnight at 0", and then saturated with carbon dioxide.
After acidification with concentrated hydrochloric acid Under these conditions, manno-and gluco-muramic acid emerge at the same volume.
(0.03 ml), the solution was saturated with carbon dioxide to remove the excess of iodine, the tube was sealed, and t,he content was hydrolyzed as just described.
For both disaccharides the proportions of n-glucosamine remained unchanged, whereas the proportions of (gluco)-muramic acid and manno-muramic acid (calculated as n-glucosamine) were decreased to 33% and 45yo for Disaccharides Ds and Dq, respectively.

Isolation of Acidic Component of Disaccharide Dq and Preparation of N-Acetyl
Derivative-A solution of Disaccharide D4 (40 mg) in 8 M hydrochloric acid (8 ml) was heated in a sealed tube for 4 hours at 100". After cooling, water (10 ml) was added, and the solution was evaporated in vucuo, the last traces of acid being removed in a desiccator, in the cold, in the presence of potassium hydroxide and concentrated sulfuric acid. The residue was dissolved in 0.3 M hydrochloric acid (2 ml) and the solution was fractionated on a Dowex 50 column according to the procedure of Garde11 (22).
The results are reported in Fig. 5. Fractions 63 to 79 were pooled and lyophilyzed.
The residue (16.8 mg) migrated on paper chromatography and paper electrophoresis as n-glucosamine.
It crystallized from a mixture of water-methanol-acetone, and gave, after degradation with ninhgdrin (23), a compound migrating as n-arabinose on paper chromatograms.
Fractions 85 to 98 were pooled and lyophilized. The residue was dissolved in water (2 ml) and the solution was treated with Amberlite In-45 (OH-form) until pH 6.7 was reached. The solution was lyophilyzed; the residue (7.0 mg) could not be crystallized, it became yellow above 130", and melted at about 145-147". C$,H19N09 (251.25) Calculated: N 5.28 Found : N 4.94 (Kjeldahl) The color obtained in the Elson-Morgan reaction had a masimum absorption peak at 523 nm. Degradation with ninhydrin (23) gave a product migrating on paper chromatogram at the same speed as the product obtained by ninhydrin degradation of muramic acid (Table VIII). Paper chromatography in Solvents E, 13, and C and electrophoresis in Solvent I-I are reported in Table VIII  was performed according to the procedure described for muramic acid (24). The resulting product could not be crystallized. It was examined by paper chromatography in Solvents E and C, and the results are report,ed in Table VIII.  1.75, 6.8, and 3.3, respectively (25).
The mixture of anomeric glycosides resulting from the methanolysis consists practically exclusively of the 01 anomers for the gluco and manno-muramic derivatives, and the small peaks corresponding to t,he @ anomers have not been numbered. After evaporation in vacua, traces of hydrochloric acid were removed by repeated addition of water and evaporation. The residues were examined by paper chromatography in Solvents A and E, and the results are reported in Fig. 6.
Since the separation of manno-muramic acid from (gluco)muramic acid was difficult to obtain by paper chromatography, the separation by gas-liquid chromatography of the N-acetyl-4, 6bis-O-(trimethylsilyl) methyl ester methyl Lu-n-glycoside was performed as follows.
Disaccharide Dq ( at 120" and the temperature was raised at the rate of 5" per min.
The results are presented in Fig. 7.
Treatment of Disaccharide Dg with Dowex 1 (Acetafe Form)-In order to test the possible epimerization of Disaccharide D5 into Disaccharide D4, a solution of pure Disaccharide Dh (28 mg) in water (0.1 ml) was adsorbed on Dowex 1 (acetate form) as described for the isolation of the oligosaccharides.
After being kept at room temperature (20-30") for 2 days, the disaccharide was eluted as previously described.
Each fraction (1 to 2 ml) of the symmetrical peak that corresponds to D5 was lyophilized, and then acetylated, methanolyzed, and per(trimethylsilyl)ated as just described, and finally examined by gas-liquid chromatography.
No evidence for manno-muramic acid was found.
This simplified conception was modified after the isolation of muramic acid devoid of N-acetyl substitution, as such, from M. Zysodeilcticus cell wall (6), or as internal amide from Bacillus subtilis spores (7), and after the recent isolations of the N-glycolyl derivative of muramic acid from mycobacteria cell walls (27,28).
Since various forms of bacterial cell walls are present and these, in addition, possess regions of growth, it may be assumed that variation in the physical structure could be reflected in changes in the chemical composition of the carbohydrate part of the backbone.
It is with this objective in mind that a study of the minor components of the carbohydrate moiety of the peptidoglycan backbone of the M. lysodeikticus cell wall was undertaken.
The cell wall was degraded with egg white lysozyme as in previous experiments.
Adsorption on a Dowex 50 column retained most of the peptide material.
From this material, two oligosaccharides were isolated in pure state; they are assumed to be identical with the two disaccharide peptides previously isolated by Mirelman and Sharon (6). The material not retained on the Dowex 50 resin was adsorbed on Dowex 1 resin in the acetat,e form.
A very minute fraction was not retained; it contained a trace of n-glucosamine, no more than 0.5y0 of the total n-glucosamine of the cell wall. This re-sult illustrates the high specificity of egg white lysozyme as an endoenzyme.
The first peak of substance (D,) to be eluted with a gradient of acetic acid consisted of a tetrasaccharide composed of an equimolar proportion of n-glucosamine and muramic acid as shown by its reducing power.
Comparison with the known Tet,rasaccharide Ds and determination of the presence of one free carbosylic group (instead of two), of three acetyl groups (instead of four), of the absence of a primary amide group, and of the resist'ance to further lysozyme degradation suggest the structure of a tetrasaccharide similar to De. In this structure, however, the carboxylic and amino groups of the muramic acid unit farther removed from the reducing end form an internal amide ring (see Scheme 1,Tetrasaccharide D,). This structure agrees with the resuits of the paper chromatography and electrophoresis, and with a preliminary determination of the nuclear magnetic resonance spectrum at 100 MHz. 1 It is probable that the structure of Tetrasaccharide Da is identical with that of the tetrasaccharide isolated in high yield from the spores of B. subtilis by Warth and Strominger (7). Complete identification of this structure will require larger amounts of material and comparison with synthetic internal amides of muramic acid (29). The very low proportion of those internal amide groups in a "normal" cell wall (as compared to the cell wall of a spore) suggests a definite biological role for this internal amide.
Since Mirelman and Sharon (6) have denonstrated the presence of free amino groups in the muramic acid units of the peptidoglycan it is possible that the internal amide isolated in this study result from a condensation between free carboxylic and amino groups. Such condensation has been shown to take place very easily (29), and could take place during the lyophilization process. Warth and Strominger (7), however, have presented good evidence for the presence of these groups in B. subtilis spore cell walls before isolation.
The second new oligosaccharide to be isolated (D,) was shown by determination of the reducing power to be a disaccharide composed in equimolar proportion of a 2-acetamido-2-deosy-nglucose unit and of a residue very similar to, but not identical with (gZuco)-muramic acid. This structure was in agreement with the results of the paper chromatography and electrophoresix, and with the elementary analysis. After hydrolysis the product' corresponding to muramic acid had a behavior very similar to that of (gluco)-muramic acid.
The first evidence of a manno structure for this compound was the long time necessary for reduction of the silver nitrate reagent, a condition typical for man nose derivatives.
The structure of a manno derivative of muramic acid for the acidic component of Disaccharide Dd was ascertained by hydrolysis, under strong acid conditions, of the lactyl ether chain and detection, in low yield, of 2-amino-a-deoxymannose (mannosamine) by paper chromatography and electrophoresis.
Since it was difficult to ascertain the chemical structure of this acid component, because of the small amount of unstable material available, the synthesis of manno-muramic acid [2-amino-3-O-(n-l-carboxyethyl)-2-deosy-n-rnannose] was accomplished (25). Comparison of the natural with the synthetic compound proved to be difficult.
In addition, manno-muramic acid gave, after hydrolysis, many secondary products, and in only two solvent systems was it