THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 244, No. 18, Issue of September 25, pp. 4921-4928, 1969 Printed in U.S.A. Galactomuramic Acid CHEMICAL SYNTHESIS, PROPERTIES, ASSAY, AND SURVEY IN SEVERAL BACTERIAL SPECIES* (Received for publication, April 3 1969) ROBERT W. WHEAT, SHARED I<TJLKARNI,$ ALEXANDROS COSMAT.OS,$ ELEANOR R. SCHEER, AND RICHARD S. STEELE From the Department of Microbiology and Immunology and Department of Biochemistry, Duke University Medical Center, Durham, North Carolina 67706 SUMMARY Synthetic D-galacto- and D-glucomurarnic acids were pre- pared from the condensation of L-cr-chloropropionicacid with benzyl-2-acetamido-4,6-O-benzylidene-2-deoxy-cr-D-galacto- or glucopyranoside. The configuration of the D( -) lactate moiety (i.e. 3-O-(D-l-carboxyethyl) -) obtained by alkaline B elimination, was proven by use of specific D( -)- and I.( +)-lactic acid dehydrogenases. Standard methods conunoniy used for the detection of muramic acid were found to be unsuitable for distinction between the two prod- ucts. Thus, although the Barker and Snmmerson, Morgan- Elson, Park- Johnson, and Rondle-Morgan tests were shown to detect both gluco and galacto epimers of muramic acid (color yields differed slightly), such calorimetric assayscould not differentiate between the analogues. Ionophoresis did not distinguish the forms of muramic acid, and thin layer chromatography was found to be suitable for distinction only under carefully controlled conditions. Cation exchange chromatography, according to Garde& was used successfully in the separation of the glucomnramic and gaiactomuramic acids, but it was found that glucosamine co-elutes with galactomuramic acid. The most rapid and precise detection method was that in which a commercial amino acid analyzer was used. Detection of nanomole quantities of material was possible. The muramic acid content of 16 bacterial species was determined. Only the glucomuramic acid analogne was found. Muramic acid hasbeen detected by alkaline silver staining or ninhydrin reactivity on chromatograms of hydrolyzates of the * This study was supported by Grant AI01659 from the National Institute of Allergy and Infectious Diseases, the United States Public Health Service, and a Duke University Endowment Re- search Grant. An abstract of a preliminary report, made at the American Society for Microbiology Meeting in Detroit, 1968, has appeared (1). $ Present address, Department of Chemistry, Pennsylvania State University, Wyomissing, Pennsylvania. 8 Present address, Cycle Chemical Company, Los Angeles, Cali- fornia. semirigid cell wall component (murein) of all bacteria so far ex- amined (24). Identification of isolatedmuramicacid is usually established by relative Rp values or on the shift from the usual red color (X II1aLL at 530 mp) to the characteristic orange color &l, at 505 mp) due to C-3 substitution (5, 6) in the Elson- Morgan reaction for a-aminosugars. However, noneof the pro- cedures so far used is specific for glucomuramicacid, and they shouldalsodetect other configuration analogues suchas galacto- muramic acid, mannomuramicacid, etc. Thus, it is possible that “muramic acid,” which is considered to be only the gluco isomer, could be a family of muramic acid analogues in nature, all of which could be detected but not differentiated by these procedures. In this regard it is of interest to note that glucomur- amic acid has been isolated and characterized as 2-amino-3-O- (D-1’-carboxyethyl)-2deoxy-D-glucose from only a few gram- positive bacteria (7-10) and from a singlegram-negative species (11). Also, we have occasionally observed smallamounts of un- known materials which exhibit chromatographic mobilities slightly different from glucomuramicacid in cell wall hydroly- zates of various bacteria.’ It appearedpossible that such un- knowns might include galactomuramic acid, which might be produced from UDP-N-acetylglucomuramic acid by C-4 epi- merase reactions analogous to the known systems which inter- convert UDP-glucose and UDP-galactose, and UDP-N-acetyl- glucosamineand UDP-N-acetylgalactosamine. Therefore, to examinethe possibility that galactomuramic acid might occur in nature and exhibit properties enough like thoseof glucomuramic acid to have gone undetected, we undertook the chemical syn- thesis of D-galactomuramicacid, i.e. the galacto analogueof glucomuramicacid, in order to compare properties of the two compounds and to establish referencecriteria by which galacto- muramic acid might be distinguished from glucomuramic acid. With the information obtained, an assay procedurewas devel- oped which has beenappliedto severaldiierent bacterial species. The results are reported in the presentcommunication. MATERIALS AND METHODS Glucosamine hydrochloride was obtained from Pfanstiehl Laboratories. Galactosamine hydrochloride, prepared from chondroitin sulfate (12) was characterizedas chemically pure by ninhydrin degradation, which gave only lyxose as a product on 1 R. W. Wheat, unpublished observations. 4921 by guest on May 22, 2018 http://www.jbc.org/ Downloaded from
Transcript
THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 244, No. 18, Issue of September 25, pp. 4921-4928, 1969
Printed in U.S.A.
Galactomuramic Acid
CHEMICAL SYNTHESIS, PROPERTIES, ASSAY, AND SURVEY IN SEVERAL BACTERIAL SPECIES*
(Received for publication, April 3 1969)
ROBERT W. WHEAT, SHARED I<TJLKARNI,$ ALEXANDROS COSMAT.OS,$ ELEANOR R. SCHEER, AND RICHARD S. STEELE
From the Department of Microbiology and Immunology and Department of Biochemistry, Duke University Medical Center, Durham, North Carolina 67706
SUMMARY
Synthetic D-galacto- and D-glucomurarnic acids were pre- pared from the condensation of L-cr-chloropropionic acid with benzyl-2-acetamido-4,6-O-benzylidene-2-deoxy-cr-D-galacto- or glucopyranoside. The configuration of the D( -) lactate moiety (i.e. 3-O-(D-l-carboxyethyl) -) obtained by alkaline B elimination, was proven by use of specific D( -)- and I.( +)-lactic acid dehydrogenases. Standard methods conunoniy used for the detection of muramic acid were found to be unsuitable for distinction between the two prod- ucts. Thus, although the Barker and Snmmerson, Morgan- Elson, Park- Johnson, and Rondle-Morgan tests were shown to detect both gluco and galacto epimers of muramic acid (color yields differed slightly), such calorimetric assays could not differentiate between the analogues. Ionophoresis did not distinguish the forms of muramic acid, and thin layer chromatography was found to be suitable for distinction only under carefully controlled conditions.
Cation exchange chromatography, according to Garde& was used successfully in the separation of the glucomnramic and gaiactomuramic acids, but it was found that glucosamine co-elutes with galactomuramic acid. The most rapid and precise detection method was that in which a commercial amino acid analyzer was used. Detection of nanomole quantities of material was possible. The muramic acid content of 16 bacterial species was determined. Only the glucomuramic acid analogne was found.
Muramic acid has been detected by alkaline silver staining or ninhydrin reactivity on chromatograms of hydrolyzates of the
* This study was supported by Grant AI01659 from the National Institute of Allergy and Infectious Diseases, the United States Public Health Service, and a Duke University Endowment Re- search Grant. An abstract of a preliminary report, made at the American Society for Microbiology Meeting in Detroit, 1968, has appeared (1).
$ Present address, Department of Chemistry, Pennsylvania State University, Wyomissing, Pennsylvania.
8 Present address, Cycle Chemical Company, Los Angeles, Cali- fornia.
semirigid cell wall component (murein) of all bacteria so far ex- amined (24). Identification of isolated muramic acid is usually established by relative Rp values or on the shift from the usual red color (X II1aLL at 530 mp) to the characteristic orange color &l, at 505 mp) due to C-3 substitution (5, 6) in the Elson- Morgan reaction for a-amino sugars. However, none of the pro- cedures so far used is specific for glucomuramic acid, and they should also detect other configuration analogues such as galacto- muramic acid, mannomuramic acid, etc. Thus, it is possible that “muramic acid,” which is considered to be only the gluco isomer, could be a family of muramic acid analogues in nature, all of which could be detected but not differentiated by these procedures. In this regard it is of interest to note that glucomur- amic acid has been isolated and characterized as 2-amino-3-O- (D-1’-carboxyethyl)-2deoxy-D-glucose from only a few gram- positive bacteria (7-10) and from a single gram-negative species (11). Also, we have occasionally observed small amounts of un- known materials which exhibit chromatographic mobilities slightly different from glucomuramic acid in cell wall hydroly- zates of various bacteria.’ It appeared possible that such un- knowns might include galactomuramic acid, which might be produced from UDP-N-acetylglucomuramic acid by C-4 epi- merase reactions analogous to the known systems which inter- convert UDP-glucose and UDP-galactose, and UDP-N-acetyl- glucosamine and UDP-N-acetylgalactosamine. Therefore, to examine the possibility that galactomuramic acid might occur in nature and exhibit properties enough like those of glucomuramic acid to have gone undetected, we undertook the chemical syn- thesis of D-galactomuramic acid, i.e. the galacto analogue of glucomuramic acid, in order to compare properties of the two compounds and to establish reference criteria by which galacto- muramic acid might be distinguished from glucomuramic acid. With the information obtained, an assay procedure was devel- oped which has been applied to several diierent bacterial species. The results are reported in the present communication.
MATERIALS AND METHODS
Glucosamine hydrochloride was obtained from Pfanstiehl Laboratories. Galactosamine hydrochloride, prepared from chondroitin sulfate (12) was characterized as chemically pure by ninhydrin degradation, which gave only lyxose as a product on
paper chromatography. Naturally occurring muramic acid (glucomuramic acid) was isolated from Micrococcus lysodeikticus (8). Glucomuramic acid was also chemically synthesized by use of both the or-benzyl and a-methyl glucoside derivatives by using established methods and routes (10, 13-17). A sample of syn- thetic glucomuramic acid was furnished by Dr. J. T. Park, Department of Microbiology, Tufts University School of Med- icine. D ( -)-Lactic dehydrogenase from Leuconostoc mesen- teroides and a D( -)-lactic acid standard were the kind gifts of Drs. Rebecca C. Garland and N. 0. Kaplan, Brandeis University. The n(+)-lactic acid standard was purchased from Sigma. 3- Acetylpyridine DPN+ was obtained from P-L Biochemicals, and L( +)-lactic dehydrogenase was purchased from Worthington.
Acetone powdered cells of Pseudomonas saccharophila were ob- tained from Dr. A. Hu, University of Kentucky, Lexington; Rhizobium japonicum cells were donated by Dr. Gerald Elkan, North Carolina State University, Raleigh; a Brucella abortus cell wall preparation was generously given by Dr. David Platt, Uni- versity of Pittsburgh. Veillonella parvula cells were received from Dr. Stephen Mergenhagen, National Institute of Dental Research, National Institutes of Health, Bethesda, Maryland; Mkrococcus lysodeilcticus and Clostridium acetobutylicum cells were purchased from Worthington. The other organisms were grown and harvested and acetone powders and cell walls were prepared as described previously (8, 11, 18-21).
Partition Chromatography-Ascending thin layer cellulose chromatography and descending paper chromatography were carried out with the following solvents: (a) l-butanol-pyridine- water (6 :4 :3) ; (b) 1-butanol-acetic acid-water (5: 1:2) ; (c) ethyl- acetate-pyridine-acetic acid-water (5: 5 : 1: 3) ; (cZ) phenol-water (70:30). Spots were visualized with ninhydrin and alkaline silver nitrate.
Ion Exchange Chromatography-Amino sugars were separated according to Garde11 on Dowex 50-H columns, 1 X 60 cm, eluted with 0.30 N HCl (22). Separation of galactomuramic acid by use of a Beckman/Spinco 120C amino acid analyzer was achieved on a column, 55 x 0.9 cm, of Dowex 50-H form Aminex A4 resin (Bio-Rad Laboratories, Richmond, California), eluted at 68 ml per hour at 50”, first with citrate buffer, pH 3.15, 0.2 N
Naf, and then with a citrate buffer, pH 4.25, 0.20 N Na+. The second buffer contained 3% n-propanol and 2.3% benzyl alcohol. The buffer change made at 75 min was detected on recorder tracings just after alanine. The amino acid analyzer was mod- ified to split the column effluent and analyze for amino groups by ninhydrin reactivity and reducing groups by a coupled ferri- cyanide-arsenomolybdate assay, details of which will be pub- lished separately.*
Electrophoresis-Ionophoresis was carried out on Whatman No. 1 paper in 0.05 M pyridine acetic acid buffer, pH 6.4, at 8 volts per cm with a Savant flatbed electrophoresis apparatus.
Analyses--Reducing sugars were determined by the ferri- cyanide assay of Park and Johnson (23). The Rondle-Morgan (24) modification of the Elson-Morgan assay was used for the determination of a-amino sugars. Amino sugars and 2-aceta- mido sugars were also determined by a modification of the Morgan-Elson procedure (25). Lactic acid was assayed chem- ically by the method of Barker and Summerson (26), with the use of lithium lactate as standard. Spectra were recorded on a Cary 14 spectrophotometer. Following /? elimination by alka-
2 R. S. Steele, K. Brendel, and R. W. Wheat, unpublished ob- servations.
line treatment of muramic derivatives according to the procedure described by Tipper (27), the enzymatic assay of D( -)- and L(+)-lactic acid was utilized as described by Dennis (28) and modified by Drs. N. 0. Kaplan and Rebecca Garland3 as follows to establish the configuration of the I-carboxyethyl (lactate) moi- ety. D( -)-Lactic acid assays were run with 1.0 M Tris buffer at pH 9.0, while glycine buffer, 1.0 M at pH 10.0, was utilized for L(+)-lactic acid assays. Enzyme assays were followed at 365 rnp with a Gilford model 222 photometer with recorder and 210 D cuvette positioner. Melting points were obtained with a Thomas Hoover capillary melting point apparatus and are not corrected. Optical rotations were obtained with a Rudolph and Sons, Inc., model 80 polarimeter. Microanalyses were done by Galbraith Laboratories, Inc., Knoxville, Tennessee; the ana- lytical samples were dried over P,Oh in vacua for 2 hours at 80”.
EXPERIMENTAL RESULTS
Synthesis of o-Galactomuramic Acid (%Amino-%@(D-I- carboxyethyl)-d-deoxy-o-galactopyranose)
The galacto analogue of muramic acid was synthesized via the route described by Strange and Kent (lo), as modified by Matsushima and Park (14) and Flowers and Jeanloz (16), for the synthesis of glucomuramic acid as follows. Benzyl-2- acetamido -4,6-O-benzylidene-2 -deoxy-a-D-galactopyranoside (I), (17) was condensed with L-or-chloropropionic acid (13-15) in the presence of sodium hydride to give benzyl-2-acetamido- 3 -0 - (D - 1’ - carboxyethyl) - 4,6 - 0 - benzylidene - 2 - deoxy - a! - D -
galactopyranoside (II). D Configuration was assigned to the carboxyethyl group in analogy to the fact that n-glucomuramic acid results by the use of n-cr-chloropropionic acid under similar conditions (13, 14, cf. Reference 29). Hydrolysis of II with 3 N HCl gave the galacto analogue of muramic acid, a product homogeneous chromatographically and electrophoretically. Re- moval of the benzylidene group from II by 60% aqueous acetic acid gave benzyl-2-acetamido-3-O-(o-l’-carboxyethyl)-2-deoxy- a-n-galactopyranoside (III). Hydrogenolysis of III yielded a compound (N-acetylgalactomuramic acid) which, on hydrolysis and treatment as described in the case of the hydrolysis of II, yielded free galactomuramic acid.
Benzyl-l-acetamido-S-0-(o-1 -carboxyethyl) -4, 6-0- benxylidene-%deoxy-a+galactopyranoside (II)
Benzyl-2-acetamido-4,6-0-benzylidene-2-deoxy-a-~-galac- topyranoside (1.10 g, 0.00275 mole) was dissolved in 80 ml of dry dioxane at 60”. To this solution 0.9 g of sodium hydride suspended in 1.5 ml of mineral oil was added, the temperature was raised to 95”, and the mixture was stirred for 1 hour. After cooling to 65”, a solution of 1.3 ml (0.018 mole) of L-ar-chloro- propionic acid in 20 ml of dry dioxane was added dropwise and the reaction mixture was stirred for an hour. Then 1.32 g of sodium hydride suspended in 2.2 ml of mineral oil were added, and the stirring was continued overnight at 65”. After cooling to room temperature, ice-cold water was added slowly until a clear solution was obtained. The dioxane was removed by dis- tillation under reduced pressure at 50”, water (50 ml) was added to the residue, and the aqueous solution was extracted with chloroform to remove mineral oil. The aqueous solution was then cooled to O”, and upon neutralization by addition of cold
Issue of September 25, 1969 R. W. Wheat, S. Kulkami, A. Cosmatos, E. R. Scheer, and R. S. Steele 4923
6 N HCl, a white solid separated which was extracted immediately with 50 ml of chloroform. The chloroform layer was washed with water and dried over anhydrous sodium sulfate, and the solvent was removed by distillation in a vacuum. The solid residue was recrystallized from aqueous alcohol. The yield was 0.7 g; the mother liquor yielded an additional 0.1 g (61% total). After two recrystallizations, the melting point was 20&210”, [LY]~~ +208” (c, 0.86, in methanol).
C&H2908N (471.49)
Calculated: C 63.69, H 6.20, N 2.97 Found : C 63.65, H 6.33, N 2.93
A solution of II (0.472 g, 0.001 mole) in 20 ml of 60% aqueous acetic acid was refluxed for 1 hour. The solvent was removed by distillation at 40” in a vacuum, water was added, and the mixture was again distilled in a vacuum to dryness. The residue was dis- solved in 25 ml of water and 2 ml of N NaOH were added. After 2 hours the solution was titrated with Dowex 50-H+ to remove Na+, filtered, and taken to dryness in a vacuum at 40”. The crystalline residue was recrystallized from methanol-ethyl ace- tate. The yield was 0.29 g (75%), m.p. 178” (shrinkage, lSS”), [& +175” (c, 1, in methanol).
C&H2508N (383.39)
Calculated: C 56.38, H 6.57, N 3.65 Found : C 56.28, H 6.64, N 3.78
~-Amino-~-~-(~-i’-carboxyethyl)-~-deOXy-a-D-gU~aetOSe (Galactosamine Analogue of Muramic Acid, or Galactomuramic Acid (IV))
In a sealed tube, 0.2 g of II was heated with 10 ml of 3 N hydro- chloric acid at 100” for 4 hours. The solution was evaporated to dryness under reduced pressure at 50”, and the residue was ex- tracted with water (5 ml) and shaken with ether (10 ml). Lac- tones were hydrolyzed as follows by a modification of the procedure described by Matsushima and Park (14) and Matsu- shima, Park, and Montague (15). The aqueous layer was ti- trated with Dowex 1 ccarbonate form) to pH 5.5 and then filtered. The filtrate was carefully titrated with 1 N sodium hydroxide to pH 8.5 to 9.0 and kept at room temperature for 2 to 3 hours. The solution was then carefully back-titrated with Dowex 50-H+ to pH 5.5 and filtered. The filtrate was lyophilized, yielding a white powder. It was crystallized from methanol-ethyl acetate (1:2) and recrystallized from the same mixture three times to give 21 mg (20% yield) of a hygroscopic compound (m.p. 130- 131”, with decomposition). The product was homogeneous chromatographically and electrophoretically. The Rp was 0.18 on Whatman No. 1 paper, solvent system B, with the use of silver nitrate and ninhydrin as staining agents, and [cy]i5 +138” at 3 min to +128” was constant after 6 min (c, 1, in water).
CsHl~07N (251.24)
Calculated: C 43.02, H 6.82, N 5.58 Found: C 43.19, H 5.98, N 5.23
Assay and ConJiguration of Lactic Acid (S-0-(0-1 ‘-carboxy- ethyl) -) Sztbstituent
Measurement of the lactic acid (3-O-(1-carboxyethyl) -) moiety of muramic acid by the Barker and Summerson procedure
TABLE I Determination of D(-) and L(+) lactic acid (i.e. I-carboxyethyl)
content of or-benzyl-N-acetylgalactomuramic and -glucomuramic acids
Aliquots of solutions (la) of cu-benzyl-N-acetyl-galactomuramic acid (532 nmoles per ml) and ol-benzyl-N-acetylglucomuramic acid (600 nmoles per ml) were analyzed for free amino sugar content (lb) by the amino acid analyzer assay as described in the text (see also Fig. 6) and for D(-)-lactic acid content (lc) with a n(-)-lactic dehydrogenase assay (no alkaline treatment prior to assay), according to Tipper (27). Also, treatment of aliquots of these derivatives with alkali (not shown in Table I) did not result in p elimination of n-lactic acid.
Aliquots from the a-benzyl-N-acetylmuramic acid samples were autoclaved at 123” for 75 min in 3 N HCl to convert the derivatives to free muramic acids. These hydrolyzates were then analyzed for amino sugars (2a) and lactic acid (2b) without alkaline treatment prior to lactate assay. Aliquots of the hy- drolyzates were then removed for alkaline p elimination treatment and assays as indicated below.
Samples of the acid hydrolyzates (containing the free muramic acids) were treated under p elimination conditions (phosphate buffer at pH 12.5, 37” for 2 hours) (e.g. cf. Reference 27). The alkali-treated material was then analyzed for amino sugars (3a) and for both L(+)- and D(-)-lactic acids (3b, 3~) with specific lactic dehydrogenases.
Treatment and assay Galac- tomur- amic acid
““iY
Yield
% 1. Prior to treatment
a. By weight.. . . . . . b. By amino acid analyzer assay
as in Fig. 6. . . . c. By D(-)-lactic dehydrogen-
ase....................... 2. After acid hydrolysis
a. By amino acid analyzer.. . . b. By D(-)-lactic dehydrogen-
ase....................... 3. After alkaline hydrolysis of
acid hydrolyzate
532 100
0 0
0 0
428 80.5
0 0
a. By amino acid analyzer. . 0 0 b. L(+)-lactic dehydrogenase. 0 0 c. D(-)-lactic dehydrogenase. . 436 82.0
-
n
--
I
-
GlllCO- mramit
acid Yield
moles/ ml %
600 100
0 0
0 0
380 63.3
0 0
0 0
420
0 0
70.0
(26), as suggested by Kent and Strange (30), indicated that the two muramic acid analogues gave approximately the same color yield. Standard curves showed that glucomuramic acid gave 95a/, and galactomuramic acid 105% the color values, per mole, of the lithium lactate standard.
The configuration of the l-carboxyethyl groups was deter- mined by the /3 elimination procedure described by Tipper (27). cu-BenzylX-acetyl derivatives of glucomuramic acid and galacto- muramic acid were hydrolyzed with 3 N HCl for 75 min at 123’. The hydrolysates were dried in vacua at 26” and re-evaporated several times after addition of water to remove HCl. Alkaline treatment at pH 12.5 for 2 hours followed, according to Tipper (27). The results of enzymatic lactic acid dehydrogenase assays of the alkaline-treated samples are shown in Table I. Somewhat greater quantities of product were detected enzymatically (i.e. n-lactate) than by means of the amino acid analyzer (i.e. free
muramic acids). Thii can be attributed to the enzymatic detec- tion of lactate produced from free muramic acid, acetamidomura- mic acid and lactam forms of muramic acid. The last two forms of muramic acid are not detected by ion exchange chromatog- raphy. The oarboxyethyl moiety of the synthetic glucomuramic and galactomuramic acids is thus shown to be in the D configura- tion.
Properties of Galactomuramic Acid (IV)
The properties of natural and synthetic glucomuramic acid and of the analogue, galactomuramic acid, were examined in the following ways. (a) Reactivities in several calorimetric assays were compared; (b) mobilities were compared with those of gluco- samine-HCl on thin layer cellulose and paper chromatograms and on ionopherograms; and (c) elution positions relative to gluco- samine were compared on various cation exchange elution sys- tems.
Calorimetry: Amino Sugar and Redwing Sugar Aseays
Rondle-Morgan Reaction-The expected orange color seen with glucomuramic acid was also observed with galactomuramic acid when assayed by the Rondle-Morgan modification of the Elson- Morgan assay. In Fig. 1, the absorption spectra of gluco- muramic and galactomuramic acids (X,, of both occurs at 505 rnp) are compared with that of glucosamine (X,, at 530 rnp to 540 mp) .
Morgan-Elson Reaction-Galactomuramic acid yielded about half the color value of glucomuramic acid under conditions opti- mal for glucosamine in a modiied Morgan-Elson reaction (25).
RONDLE MORGt4N METHOD
I 1 I 450 500 550 600
w
FIG. 1. Absorption spectra of amino sugars in the Rondle- Morgan reaction. Glueosamine-HCl (A), galactomuramic acid (B), and glucomuramic acid (C).
TABLE II Paper and thin layer chromatography Rcluoossmine values
0 PC, paper chromatography; TLC, thin layer cellulose chro- matography.
The heating time and pH are known to be critical (25,31). This assay, which involves N-acetylation with acetic anhydride in alkali before heating in a borate buffer, is known to give a 2:l color yield with other glucose and galactose derivatives (cf. Ref- erence 25)) respectively.
Reducing Sugar Assay-Both muramic acids reacted in the fer- ricyanide reducing sugar assay of Park and Johnson (23). How- ever, compared with a glucosamine standard (loo’%), galacto- muramic acid (55%) was much less sensitive than glucomuramic acid (85%) in this assay.
Partition Chromatography
As shown in Table II and Fig. 2, galactomuramic acid and glucomuramic acid can be separated in several solvent systems on paper or thin layer cellulose plates. However, as indicated in Fig. 2, mobilities are close enough in some solvents so that varia- bility in application of the compound to the chromatogram or un- fortunate choice of solvents could cause confusion between the two compounds.
lonophoretic Mobility
Fig. 3 shows the similarity of mobility of the two muramic acid analogues compared with glucosamine upon ionophoresis at pH 6.4.
The elution pattern on a Garde11 column of galactomuramic acid, glucosamine, and glucomuramic acid is shown in Fig. 4. Each compound was eluted separately under identical conditions from a Dowex 50-H (X8,200 to 400 mesh) column, 1 X 60 cm, with 0.30 N HCl collecting 2.5 ml fractions. Aliquots removed
FIG. 2. Thin layer cellulose chromatogram comparing mobili- ties of glucosamine-HCl (1) with glucomuramic acid @) and galac- tomuramic acid (3) ; solvent system B; ninhydrin stain.
Issue of September 25, 1969 R. W. Wheat, S. Kulkami, A. Cosmatos, E. R. Scheer, and R. S. Steele 4925
were neutralized with an equal volume of 0.3 N NaOH and as- sayed for reducing sugar by the ferricyanide assay. The same elution pattern, as seen in Fig. 4, of two peaks designated Peak I and Peak II, was observed when the three compounds were mixed together and then eluted. The two peaks were pooled sepa- rately, dried in a vacuum at 50”, and redissolved in 1 ml of water and then assayed by the Rondle-Morgan procedure. The pres- ence of galactomuramic acid in the usual glucosamine peak could be detected by a shift toward 505 rnp in the wave length of maxi- mum absorbancy which was dependent on the ratio of glucosa- mine and galactomuramic acid. When these two peaks were chromatographed on thin layer cellulose, the presence of galacto- muramic acid and glucosamine in Peak I was confirmed, as was the presence in Peak II of glucomuramic acid (Fig. 5). However, spots due to formation of lactones or lactams (14, 15, 29) were also observed in peaks containing either of the muramic acids which were evaporated in vacua to remove acid. These could be removed by careful de-ionization or both alkaline hydrolysis and de-ionization as outlined in the above section on synthesis, fol- lowed by lyophilization. The bulk of the glucosamine HCl could be removed from the mixed sample in Peak I by an initial treat- ment near pH 5.5 to 6.0 with Dowex 50-H+, leaving galacto-
FIG. 3. Electrophoretic mobilities of glucosamine-HCl (A), glucomuramic acid (Micrococcus Zysodeikticus) (B), synthetic glucomuramic acid (C and E), and galactomuramic acid (D). The conditions are as described in the text. Spots are detected with silver nitrate.
muramic acid, because the latter compound exists as a dipolar ion at this pH, whereas glucosamine exhibits a net positive charge in acid solutions. As shown in Fig. 5, Spots 3 and 7, the same respective lactones could be produced from the chromatographi- tally homogeneous glucomuramic and galactomuramic acid sam- ples by merely evaporating them to dryness at 40-50” in 0.3 N HCl.
separation and Identification by Use of Commercial Amino Acid Analyzer
As shown in Fig. 6, the separation of galactomuramic acid and glucomuramic acid in the presence of expected or usual bacterial cell wall hydrolyzate components, i.e. glucosamine, alanine, glutamic acid, and diiminopimelic acid, as well as other protein component amino acids, can be achieved with a buffered elution system with commercially available amino acid analyzers. This separation is much faster than our previously reported system (32, 33). Galactosamine uranic acid, like galactomuramic acid, is eluted by 0.30 N HCl in the same peak with glucosamine from Dowex 50-H+ columns by the procedure of Garde11 (22). How- ever, under the conditions used for the separation shown in Fig. 6, galactosamine uranic acid is eluted with methionine. Galactosa- mine uranic acid therefore does not interfere with identification of either glucosamine or the muramic acids in this system.
Assay for Glucomuramic and Galactomuramic Acids in Bacteria
Cell wall preparations of varied purity and whole acetone pow- dered bacteria from species of several taxonomic groups were scanned for the presence of glucomuramic acid and galactomura- mic acid by the procedure shown in Fig. 6. Hydrolyzates of Citrobacter freundii 8090 and C. ballerup phenol-insoluble residues were used to compare recoveries from Whour 6 N HCl at 100” hydrolyzates with 2-hour 3 N HCl at 125’. Approximately equivalent amounts of glucomuramic acids were observed in both
0.30
i
!
^O_ z 8 ‘,
B C $
010 ,a,
A I : :: I :: :: :: 1 : 1 : : ’ : :
: :
I ; ‘\ .,. *.. ,-’ ‘L. . , , ,
00 20 40 60 00 loo 120 TUBE NUMBER
D
FIG. 4. Cation exchange chromatography of amino sugars ac- cording to Garde11 (22). Glucosamine-HCl (A), galactomuramic acid (B), and glucomuramic acid (C). The conditions are as de- scribed in the text.
FIQ. 5. Paper chromatogram of peaks from Garde11 column chromatography of amino sugars shown in Fig. 4. Glucosamine-HCl standard (I), Peak I @), galactomuramic acid heated in 0.30 M KC1 (8), galactomuramic acid standard (4), glucomuramic acid stand- ard (6)) Peak II (6)) glucomuramic acid heated in 0.3 M HCl (7) ; solvent system B; ninhydrin stain.
.7
.6
5
. 60
54 56 7 I <
9 IO I:
1 II I2 13.14 I5 16 I
20
I8 I9 20 1 21
180 ‘ b Time Min
FIG. 6. Beckman/Spinco model 120C amino acid analyzer trac- acid (IQ), isoleucine (16), leucine (16), norleucine (l7), tyrosine ing. The numbers below the peaks correspond to 0.1 pmole each (18)) phenylalanine (IQ), glucosamine (QO), and galactosamine (Qf ). of hydroxyproline (I), aspartic acid (Q), threonine (8), serine (4), The conditions are as described in the text under “Ion Exchange galactomuramic acid (6), glucomuramic acid (6), glutamic acid Chromatography” in “Methods.” Solid lines, 570 rm.~; dotted (7), proline (8), glycine (Q), alanine (IO), cystine (11), valine (IQ), lines, 440 rnp. 01, e-diaminopimelic acid (IS), methionine or galactosamine uranic
hydrolyzates, although amino acid recoveries were slightly greater in Table III. Values listed are results of integration of areas in the l&hour 6 N HCl procedure. With the shorter a-hour 3 N under the curves, with the use of Technicon integrator calculator HCl procedure, only glucomuramic acid was observed in hydrol- (model AAG). Serine and glutamic acid values are included for yzates of the various bacterial preparations examined, as listed comparison with those found for glucomuramic acid.
Synthetic D-galacto- and o-glucomuramic acids were prepared from the condensation of L-cY-chloroproprionic acid with benzyl- 2-acetamido-4,6-0-benzylidene-2-deoxy-a-D-gaIacto- or glucopy- ranoside. The configuration of the D( -)-lactate moiety (i.e. 3-O-(n-1-carboxyethyl) -) obtained by alkaline /3 elimination (4, 27, 34) was proven by use of the specific D( -)- and L( +)- lactic acid dehydrogenases. This had not been previously shown for muramic acid prepared by this route of chemical synthesis (10, 14, 16) and confirms expectations based on knowledge of chemical mechanisms (13). Standard methods commonly used for the detection of muramic acid were found to be unsuitable for distinction between the two configurational isomers of muramic acid. Thus, although the Barker and Summerson, Morgan- Elson, Park-Johnson, and Rondle-Morgan tests were shown to detect both forms of the synthetic muramic acid (color yields differing slightly), such calorimetric assays could not differen- tiate between the analogues. Ionophoresis did not distinguish the forms of muramic acid, and thin layer chromatography was found to be suitable for distinction only under carefully controlled conditions.
Cation exchange chromatography, according to Gardell, was used successfully in the separation of the glucomuramic acid and galactomuramic acid, but it was found that glucosamine co-elutes with galactomuramic acid. The most rapid and precise detection method was that in which a commercial autoanalyzer was used.
Detection of nanomole quantities of material was possible. The survey of muramic acid content of 16 bacterial species by means of an amino acid analyzer system revealed the presence of only the glucomuramic analogue under our conditions of assay. It is not unreasonable to believe that the galactomuramic acid ana- logue may yet be found in the peptidoglycan cell wall structure of unexamined species in analogy to the occurrence of genetic drift exhibited by the variety of substitutions in both the muramyl peptides and bridging peptides recently recognized in murein from various species (4, 35-37). The recognition of muramic acid analogues other than the gluco form has been shown feasible, and alert observations by investigators in the future may reveal a muramic acid of a configuration other than that of n-glucomura- mic acid. It may further be expected that the glucosamine moiety of murein might also be substituted with galactosamine or some other sugar or amino sugar.
Acknourledgmenta-We express appreciation to Drs. N. 0. Kap- lan and R. Garland, and to Dr. J. T. Park for generous gifts of D( -)-lactic dehydrogenase and glucomuramic acid, respectively. The technical skills of Mr. Winfred Clingenpeel are gratefully acknowledged. We wish to thank Dr. Roger Jeanloz, Harvard University, for criticism.
REFERENCES
1. WHEAT, R. W., KULKARNI, S., AND COSMATOS, A., Bacterial. Proc . .63 (1968).
2. SALTON, M: R. j., The bacterial cell wall, American Elsevier Publishine Comnanv. Inc.. New York. 1964. D. 293.
3. SALTON, M.-R. J.: Ark. Rev. Biochem:, 34, ‘i43 (1965). 4. GHUYSEN, J.-M., Bacterial. Rev., 32, 425 (1968). 5. JEANLOZ, &. W., AND TREMEGE, M., Fed. Proc., 16,282 (195(i). 6. CIFONELLI. J. A.. AND DORFMAN, A.. J. Biol. Chem.. 231, 11
(1958). ’ ’ ,
7. AGREN, G., AND DE VERDIER, C. H., Acta Chem. Stand., 12, 1927 (1958).
8. PARK, J. T., MATSUSHIMA, Y., WHEAT, R. W., AND ROLLINS, E., Biochem. Prep., 10,106 (1963).
9. STRANGE, R. E., AND DARK, F. A., iVature, 177,186 (1956). 10. STRANGE, R. E., AND KENT, L. H., Biochem. J., 71, 333 (1959). 11. WHEAT, R. W., Biochim. Biophys. Acta, 121,170 (1966). 12. WHEAT, R. W., AND DAVIDSON, E. A., Biochem. Prep., 10, 52
(1963). 13. Fu, S.-C. J., BIRNBAUM, S. M., AND GREENSTEIN, J. P., J.
Amer. Chem. Sot., 76,6054 (1954). 14. MATSUSHIMA, Y., AND PORK, J. T., J. org. Chem., 27, 3581
(1962). 15. MATSUSHIMA, Y., PARK, J. T., AND MONTAGUE, M. D., Bio-
them. Prev.. 10, 109 (1963). 16. FLOWERS, d. &~.,.AND ~EAN~OZ, R. W., J. Org. Chem., 28, 2983
(1963). 17. FLOWERS, H. M., AND SHAPIRO, D., J. Org. Chem., 30, 2041
(1965). 18. SMITH, E. J., AND WHEAT, R. W., Arch. Biochem. Biophys., 99,
109 (1962). 19. WHEAT, R. W., DORSCH, C., AND GODOY, G., J. Bacterial., 89,
539 (i965). 20. LUDERITZ, O., RUSCHMANN, E., WESTPHAL, O., RAFF, R., AND
WHEAT. R. W.. J. Bacteriol.,93,1681 (1967). 21. WHEAT, l% W., ‘in S. P. COL~WICK AND N. 0. KAPLAN (Edi-
tors), Methods in enzymology, Vol. VIII, Academic Press, New York, 1966, p. 60.
22. GARDELL, S., Acta Chem. Stand., 7, 207 (1953). 23. PARK, J. T., AND JOHNSON, M. J., J. Biol. Chem., 181, 149
(1949). 24. RONDLE, C. M. J., AND MORGAN, W. T. J., Biochem. J., 61, 586
(1955). 25. BROWNLEE, S. T., AND WHEAT, R. W., Anal. Biochem., 14,414
