Eur. J. Biochem. 58. 627-629 (1975)
Identification of Ribitol Phosphate as a Constituent of the Lipopolysaccharide from Proteus mirabilis, Strain D52 Jobst GMEINER Fachbereich Biologie, Mikrobiologie, Technische Hochschule, Darmstadt (Received June 13 /July 22, 1975)
A polyol was released from the lipopolysaccharide of Proteus mirabilis, strain D52, during alkaline hydrolysis and its phosphate ester was isolated after acid hydrolysis. This polyol has been identified as ribitol by comparison of the free polyol, its phosphate ester and its anhydro derivative formed after acid treatment with authentic xylitol, D- and L-arabitol, ribitol and their corresponding derivatives on paper and gas-liquid chromatography.
In 1956, Baddiley et al. first isolated CDP-ribitol from Lactobacillus arabinosus [ 11. Subsequent work from several groups has shown that polyolphosphates, mainly polyglycerol phosphates or polyribitol phosphates, are found in the cell wall of grampositive organisms. These teichoic acids are characteristically absent from gramnegative bacteria (for review, see '
PI).
On the other hand, glycerol phosphate has been found as a constituent of the lipopolysaccharide from an E. coli strain [3]. In this communication I describe the isolation and identification of ribitol phosphate as a constituent of of the lipopolysaccharide from a strain ,of Proteus mirabilis, a gramnegative organism.
Gas Chromatography Lipopolysaccharide and reference substances were hydrolyzed in 0.1 N hydrochloric acid for 48 h at 100 "C [5].Aldoses and polyols were converted to the corresponding alditol acetates according to the method of Sawardeker et al. [6] with xylose as internal standard. Gas-liquid chromatography was carried out in a Varian Aerograph 2701 gas chromatograph with a flame ionisation detector on a glass column (200 by 0.318 cm) containing 3 % (w/w) of ECNSS-M on Gas-Chrom Q (100- 120 mesh) (Applied Science Lab.) at an oven temperature of 190 "C or 170"C and 30 ml N2 per min as carrier gas. Chemicals and Preparation of Ribitol5-Phosphate
MATERIALS AND METHODS Strain and Isolation of Lipopolysaccharide Proteus mirabilis, strain D52 was obtained from Dr H. H. Martin (this laboratory). Cultivation of bacteria and isolation of lipopolysaccharide I1 were described [4].
Xylose, ribitol, D-arabitol and L-arabitol were from E. Merck (Darmstadt). Ribose 5-phosphate was purchased from Boehringer (Mannheim) and reduced with sodium borohydride. After removal of sodium ions by Dowex 50 (H+form) and of complexed boric acid by repeated evaporation to dryness with methanol, the ribitol phosphate was quantitated by phosphate determination according to Lowry er al. [7].
Paper Chromatography
Descending paper chromatography on Whatman paper Nr. 1 or 3MM was carried out in the following solvent systems : (A)1-butanol/pyridine/water (6/4/3); (B)phenol/water (4/1); (C) upper phase of I-butanol/ acetic acid/water 4/1/5); (D) 1-butanol/pyridine/ 0.05 M morpholinium tetraborate, pH 8.6 (7/5/2). Sugars and polyols were detected with aniline phthalate or alkaline silver nitrate, respectively.
RESULTS Proteus mirabilis, strain D52, was found to have a lipopolysaccharide fraction containing N-acetylglucosamine (16.1 %), galactose (8.0 %), glucose (7.0 %), ethanolamine (2.6 %) and phosphor (4.3 %) in addition to a polyol (14.5%) as 0 side-chain constituents [4]. After hydrolysis of this material in 1 N hydrochloric acid for 4 h, evaporation to dryness and paper chro-
Ribitol Phosphate in a Proteus mirabilis Lipopolysaccharide
628
?
Tablc 1. Chromatographic mobilities of the polyol in comparison with reference substances RF values are based on glucose
Blue dextran
Substance
Polyol Ribitol D-Arabitol L-Arabitol
Fraction number
Fig. 1. Gel filtration on Sephadex G25 of' lipopolysaccharide after alkaline hydrolysis. 200 mg lipopolysaccharide I1 were hydrolyzed in 10 ml 1 N NaOH for 1 h at 100"C.After centrifugation and neutralization the supernatant was concentrated under reduced pressure and applied to a column of Sephadex G25 (86 by 2.5 cm). Elution was performed with water, the effluent was monitored by a differential refractometer from Winopal (Hannover, Germany) and fractions of 5 ml were collected. Fractions 40- 50 and 71 - 80 were pooled. Whereas the former contained all the specific sugars, the polyol was exclusively found in the latter
matography in solvent A and C, one spot was observed with an Rglucose of 1.17 and 1.36, respectively, which stained with alkaline silver nitrate but not with aniline phthalate. Neutral sugar analysis after hydrolysis, reduction and acetylation by gas-liquid chromatography at 190 "C oven temperature gave two unidentified peaks with aretention time of 0.23 and 0.65, relative to xylitol. Ribitol and arabitol behaved almost identically under those conditions.
Isolation of the Polyol and Paper Chromatographic Analysis The polyol was quantitatively released by alkaline hydrolysis in 1 N NaOH from the high molecular weight lipopolysaccharide I1 which was nucleic-acidfree [4] and recovered in tubes 71 -80 after gel filtration on Sephadex G25 (Fig.1). From this fraction, small amounts of ethanolamine and inorganic phosphate were removed by passage through ion exchange resins Dowex 50 (H' form) and Amberlite IRA 410 (HCO; form). After evaporation to dryness, the material was chromatographed on Whatman paper 3MM in solvent A. The polyol-band as visualized with silver nitrate on side strips was eluted with water, evaporated to dryness and taken up in 0.5 ml water. Aliquots thereof were chromatographed on Whatman paper 1 in solventA - D. The Rglucose-values are given in Table 1. The pol yo1 shows identical chromatographic behaviour as ribitol in systems which differentiate ribitol and arabitol. The antipodal compounds D-arabitol and
R F with solvent ~
A
B
c
D
1.20 1.18 1.20 1.17
1.41 1.40 1.48 1.48
1.36 1.38 1.32 1.31
1.25 1.2s 0.89 0.83
Table 2. Relative retention time of the polyol and itsanhvdro derivatirw in comparison with rejhrence substances after gas-chromatographic analysis The relative retention time is based onxylitol. For details see Methods Substance
Relative retention time ~~ ~
polyol
anhydro derivative
0.63 0.63 0.69
0.21 0.22 0.17 0.17 0.22 0.22
~~
Anhydroderivative formed after acid treatment
x Polyol Ribitol D-Arabitol L-Arabitol Polyol phosphate Ribitol5(1)-phosphate
0.69
0.63 0.63
15.8 18.0 1.5 2.0 65.5 64.7
L-arabitol have slightly different Rg,ucose-values in solvents A and D which is likely to be due to the particular experiment rather than to an actual separation.
Gas-Chromatographic Analysis and Anhydro Derivative Formation The isolated polyol was acetylated and subjected to gas chromatographic analysis (Table 2). Again, the retention time of the unknown compound is identical to the retention time of ribitol. A further characteristic for identification is anhydro derivative formation under acidic conditions. Baddiley and coworkers [8] have reported that acid treatment converts ribitol and its 5-phosphate rather easily to 1:4-anhydroribitol, whereas arabitol and xylito1 are only slightly affected. The isolated material, as well as authentic ribitol, D-arabitol and L-arabitol were separately treated with 0.1 N hydrochloric acid for 48 h at 100 "Cas described for the analysis of neutral sugars [5]. After acetylation and gas chromatography at an oven temperature of 170 "C, an additional peak emerged from the isolated polyol as well as from ribitol, with the retention time of 0.22 relative to xylitol whereas arabitol yielded a different additional peak with a relative retention time
J. Gmeiner
of 0.17. After estimating the response factors empirically relative to xylitol = 1, the percentage anhydro compound formed was calculated and is given in Table 2. Further proof for the identity of the isolated polyol with ribitol came from the behaviour of its phosphate ester. During the isolation of oligosaccharides from the lipopolysaccharide fraction after hydrolysis in 0.1 N hydrochloric acid for 100 min at 100"C the polyolmonophosphate was obtained in pure form (Gmeiner, in preparation). Analysis of this material together with authentic ribitol phosphate showed that about 65 from both substances were converted to the anhydro compound upon acid treatment, in accordance with the results from Baddiley's group [8] (Table 2).
DISCUSSION The results obtained strongly indicate that the polyol isolated from a lipopolysaccharide fraction of Profeus mirabilis, strain D 52, is identical with ribitol. Ribitol phosphate is known to be a constituent of teichoic acids in cell walls of some grampositive bacteria where it is polymerized by phosphodiester links. In the lipopolysaccharide of Proteus mirabilis, strain D 52 ribitol seems to be linked by a phosphodiester bond to the sugar polymer as a side branch. This conclusion is based on the mode of isolation by alkaline hydrolysis which releases free ribitol quantitatively in contrast to acid hydrolysis which gives mainly ribitol5( 1)-phosphate. Except for E. coli 0100 [ 3 ] phosphate esters have not been reported as 0 side-chain constituents of lipopolysaccharides from gramnegative bacteria. On the other hand, uronic acids are frequently found in lipopolysaccharides of Proteus mirabilis ([4,9], but see [lo]). It seems that Proteus mirabilis can satisfy its
629
need for anionic groups in the 0 side chain of the lipopolysaccharide either by uronic acids or by phosphodiester groups. This is reminiscent of the polyglucosyl- glycerolphosphate-containing teichoic acids of Bacillus subtilis var. niger which, under phosphate-limiting conditions, switches to the production of uronic acid containing teichoic acid [ill. It would be interesting to investigate whether a real need for anionic groups in the lipopolysaccharide exists, i. e. whether phosphate limitation of strain D 52 causes a similar replacement of ribitol phosphate by uronic acid in its lipopolysaccharide. I thank Dr H. H. Martin for stimulating discussions and generous support and Mrs I . Diener for excellent technical assistance. This work was partly supported by the Deutsche Forschungsgemeinschafi and the Stifmng Vulkswagenwerk.
REFERENCES 1. Baddiley, J., Buchanan, J. G., Carss, B. & Mathias, A. P. (1956) J . Chem. Sac. Lond. 4583 - 4588. 2. Archibald, A. R. (1974) Adv. Microbial Physiol. 11, 53-95. 3. Jann, B., Jann, K . , Schmidt, G., Qrskov, I. & Orskov, F. (1970) Eur. J . Biochem. 15,29- 39. 4. Gmeiner, J. (1975) Eur. J . Biochem. 58, 621-626. 5 . Schmidt, G., Fromme, I. & Mayer, H. (1970) Eur. J . Biochem. 14, 357- 366. 6. Sawardeker, J. S., Sloneker, J. H. & Jeanes, R. (1965) Anal. Chem. 12, 1602- 1604. 7. Lowry, 0. H., Roberts, N. R., Leiner, K. J., Wu, M. & Farr, A. L. (1954) J . Biol. Chem. 207, 1-17. 8. Baddiley, J., Buchanan, J. G. & Carss, B. (1957) J . Chem. Soc. Lond. 4138 -4139. 9. Kotelko, K., Gromska, W., Sidorczyk, Z. & Zwolinski, J . (1968) Bull. Acad. Pol. Sci. Ser. Sci. Biol. 16, 739- 744. 10. Dmitriev, B. A., Hinton, N. A,, Lowe, R. W . &Jones, J. K. N. (1971) Can. J . Microhiol. 17, 1385- 1394. 11. Ellwood, D. C. & Tempest, D. W. (1969) Biochem. J . I l l , 1 - 5.
J. Gmeiner, Fachbereich Biologie, Mikrobiologie der Technischen Hochschule Darrnstadt, D-6100 Darmstadt, Schnittspahnstrak 9, Federal Republic of Germany
Note Added in Proof(0ctober 10, 1975). After this paper had been submitted I became aware of a report by Kotelko et al. [Kotelko, K., Fromme, I. & Sidorczyk, Z. (1975) Bull. Acad. Pul. Sci. Ser. Sci. B i d . 23,249 - 2561, in which an unidentified compound X, inseparable from threitol by gas liquid chromatography, is described as a constituent of the lipopolysaccharide from two Proteus miiahilis strains belonging to serogroups 016 and 033. Dr H. Mayer (Max-Planck-Institut fur Immunbiologie, Freiburg i. Br.) kindly compared the mass spectrogram of compound X with that of anhydro ribitol isolated from ribitoL5(l)-phosphate described in the present paper and found that the spectra are in accordance with X being anhydro ribitol. Since anhydro ribitol has a very similar retention time with threitol under the gas chromatographic conditions at an oven temperature of 170°C used in the present paper (Rxy,i,o, = 0.220 and 0.225, respectively) the actual constituent of the lipopolysaccharide from Proteus mirabilis, serogroup 016 and 033 (Kotelko et al.) might be ribitol.
Identification of Ribitol Phosphate as a Constituent of the Lipopolysaccharide from Proteus mirabilis, Strain D52 Jobst GMEINER Fachbereich Biologie, Mikrobiologie, Technische Hochschule, Darmstadt (Received June 13 /July 22, 1975)
A polyol was released from the lipopolysaccharide of Proteus mirabilis, strain D52, during alkaline hydrolysis and its phosphate ester was isolated after acid hydrolysis. This polyol has been identified as ribitol by comparison of the free polyol, its phosphate ester and its anhydro derivative formed after acid treatment with authentic xylitol, D- and L-arabitol, ribitol and their corresponding derivatives on paper and gas-liquid chromatography.
In 1956, Baddiley et al. first isolated CDP-ribitol from Lactobacillus arabinosus [ 11. Subsequent work from several groups has shown that polyolphosphates, mainly polyglycerol phosphates or polyribitol phosphates, are found in the cell wall of grampositive organisms. These teichoic acids are characteristically absent from gramnegative bacteria (for review, see '
PI).
On the other hand, glycerol phosphate has been found as a constituent of the lipopolysaccharide from an E. coli strain [3]. In this communication I describe the isolation and identification of ribitol phosphate as a constituent of of the lipopolysaccharide from a strain ,of Proteus mirabilis, a gramnegative organism.
Gas Chromatography Lipopolysaccharide and reference substances were hydrolyzed in 0.1 N hydrochloric acid for 48 h at 100 "C [5].Aldoses and polyols were converted to the corresponding alditol acetates according to the method of Sawardeker et al. [6] with xylose as internal standard. Gas-liquid chromatography was carried out in a Varian Aerograph 2701 gas chromatograph with a flame ionisation detector on a glass column (200 by 0.318 cm) containing 3 % (w/w) of ECNSS-M on Gas-Chrom Q (100- 120 mesh) (Applied Science Lab.) at an oven temperature of 190 "C or 170"C and 30 ml N2 per min as carrier gas. Chemicals and Preparation of Ribitol5-Phosphate
MATERIALS AND METHODS Strain and Isolation of Lipopolysaccharide Proteus mirabilis, strain D52 was obtained from Dr H. H. Martin (this laboratory). Cultivation of bacteria and isolation of lipopolysaccharide I1 were described [4].
Xylose, ribitol, D-arabitol and L-arabitol were from E. Merck (Darmstadt). Ribose 5-phosphate was purchased from Boehringer (Mannheim) and reduced with sodium borohydride. After removal of sodium ions by Dowex 50 (H+form) and of complexed boric acid by repeated evaporation to dryness with methanol, the ribitol phosphate was quantitated by phosphate determination according to Lowry er al. [7].
Paper Chromatography
Descending paper chromatography on Whatman paper Nr. 1 or 3MM was carried out in the following solvent systems : (A)1-butanol/pyridine/water (6/4/3); (B)phenol/water (4/1); (C) upper phase of I-butanol/ acetic acid/water 4/1/5); (D) 1-butanol/pyridine/ 0.05 M morpholinium tetraborate, pH 8.6 (7/5/2). Sugars and polyols were detected with aniline phthalate or alkaline silver nitrate, respectively.
RESULTS Proteus mirabilis, strain D52, was found to have a lipopolysaccharide fraction containing N-acetylglucosamine (16.1 %), galactose (8.0 %), glucose (7.0 %), ethanolamine (2.6 %) and phosphor (4.3 %) in addition to a polyol (14.5%) as 0 side-chain constituents [4]. After hydrolysis of this material in 1 N hydrochloric acid for 4 h, evaporation to dryness and paper chro-
Ribitol Phosphate in a Proteus mirabilis Lipopolysaccharide
628
?
Tablc 1. Chromatographic mobilities of the polyol in comparison with reference substances RF values are based on glucose
Blue dextran
Substance
Polyol Ribitol D-Arabitol L-Arabitol
Fraction number
Fig. 1. Gel filtration on Sephadex G25 of' lipopolysaccharide after alkaline hydrolysis. 200 mg lipopolysaccharide I1 were hydrolyzed in 10 ml 1 N NaOH for 1 h at 100"C.After centrifugation and neutralization the supernatant was concentrated under reduced pressure and applied to a column of Sephadex G25 (86 by 2.5 cm). Elution was performed with water, the effluent was monitored by a differential refractometer from Winopal (Hannover, Germany) and fractions of 5 ml were collected. Fractions 40- 50 and 71 - 80 were pooled. Whereas the former contained all the specific sugars, the polyol was exclusively found in the latter
matography in solvent A and C, one spot was observed with an Rglucose of 1.17 and 1.36, respectively, which stained with alkaline silver nitrate but not with aniline phthalate. Neutral sugar analysis after hydrolysis, reduction and acetylation by gas-liquid chromatography at 190 "C oven temperature gave two unidentified peaks with aretention time of 0.23 and 0.65, relative to xylitol. Ribitol and arabitol behaved almost identically under those conditions.
Isolation of the Polyol and Paper Chromatographic Analysis The polyol was quantitatively released by alkaline hydrolysis in 1 N NaOH from the high molecular weight lipopolysaccharide I1 which was nucleic-acidfree [4] and recovered in tubes 71 -80 after gel filtration on Sephadex G25 (Fig.1). From this fraction, small amounts of ethanolamine and inorganic phosphate were removed by passage through ion exchange resins Dowex 50 (H' form) and Amberlite IRA 410 (HCO; form). After evaporation to dryness, the material was chromatographed on Whatman paper 3MM in solvent A. The polyol-band as visualized with silver nitrate on side strips was eluted with water, evaporated to dryness and taken up in 0.5 ml water. Aliquots thereof were chromatographed on Whatman paper 1 in solventA - D. The Rglucose-values are given in Table 1. The pol yo1 shows identical chromatographic behaviour as ribitol in systems which differentiate ribitol and arabitol. The antipodal compounds D-arabitol and
R F with solvent ~
A
B
c
D
1.20 1.18 1.20 1.17
1.41 1.40 1.48 1.48
1.36 1.38 1.32 1.31
1.25 1.2s 0.89 0.83
Table 2. Relative retention time of the polyol and itsanhvdro derivatirw in comparison with rejhrence substances after gas-chromatographic analysis The relative retention time is based onxylitol. For details see Methods Substance
Relative retention time ~~ ~
polyol
anhydro derivative
0.63 0.63 0.69
0.21 0.22 0.17 0.17 0.22 0.22
~~
Anhydroderivative formed after acid treatment
x Polyol Ribitol D-Arabitol L-Arabitol Polyol phosphate Ribitol5(1)-phosphate
0.69
0.63 0.63
15.8 18.0 1.5 2.0 65.5 64.7
L-arabitol have slightly different Rg,ucose-values in solvents A and D which is likely to be due to the particular experiment rather than to an actual separation.
Gas-Chromatographic Analysis and Anhydro Derivative Formation The isolated polyol was acetylated and subjected to gas chromatographic analysis (Table 2). Again, the retention time of the unknown compound is identical to the retention time of ribitol. A further characteristic for identification is anhydro derivative formation under acidic conditions. Baddiley and coworkers [8] have reported that acid treatment converts ribitol and its 5-phosphate rather easily to 1:4-anhydroribitol, whereas arabitol and xylito1 are only slightly affected. The isolated material, as well as authentic ribitol, D-arabitol and L-arabitol were separately treated with 0.1 N hydrochloric acid for 48 h at 100 "Cas described for the analysis of neutral sugars [5]. After acetylation and gas chromatography at an oven temperature of 170 "C, an additional peak emerged from the isolated polyol as well as from ribitol, with the retention time of 0.22 relative to xylitol whereas arabitol yielded a different additional peak with a relative retention time
J. Gmeiner
of 0.17. After estimating the response factors empirically relative to xylitol = 1, the percentage anhydro compound formed was calculated and is given in Table 2. Further proof for the identity of the isolated polyol with ribitol came from the behaviour of its phosphate ester. During the isolation of oligosaccharides from the lipopolysaccharide fraction after hydrolysis in 0.1 N hydrochloric acid for 100 min at 100"C the polyolmonophosphate was obtained in pure form (Gmeiner, in preparation). Analysis of this material together with authentic ribitol phosphate showed that about 65 from both substances were converted to the anhydro compound upon acid treatment, in accordance with the results from Baddiley's group [8] (Table 2).
DISCUSSION The results obtained strongly indicate that the polyol isolated from a lipopolysaccharide fraction of Profeus mirabilis, strain D 52, is identical with ribitol. Ribitol phosphate is known to be a constituent of teichoic acids in cell walls of some grampositive bacteria where it is polymerized by phosphodiester links. In the lipopolysaccharide of Proteus mirabilis, strain D 52 ribitol seems to be linked by a phosphodiester bond to the sugar polymer as a side branch. This conclusion is based on the mode of isolation by alkaline hydrolysis which releases free ribitol quantitatively in contrast to acid hydrolysis which gives mainly ribitol5( 1)-phosphate. Except for E. coli 0100 [ 3 ] phosphate esters have not been reported as 0 side-chain constituents of lipopolysaccharides from gramnegative bacteria. On the other hand, uronic acids are frequently found in lipopolysaccharides of Proteus mirabilis ([4,9], but see [lo]). It seems that Proteus mirabilis can satisfy its
629
need for anionic groups in the 0 side chain of the lipopolysaccharide either by uronic acids or by phosphodiester groups. This is reminiscent of the polyglucosyl- glycerolphosphate-containing teichoic acids of Bacillus subtilis var. niger which, under phosphate-limiting conditions, switches to the production of uronic acid containing teichoic acid [ill. It would be interesting to investigate whether a real need for anionic groups in the lipopolysaccharide exists, i. e. whether phosphate limitation of strain D 52 causes a similar replacement of ribitol phosphate by uronic acid in its lipopolysaccharide. I thank Dr H. H. Martin for stimulating discussions and generous support and Mrs I . Diener for excellent technical assistance. This work was partly supported by the Deutsche Forschungsgemeinschafi and the Stifmng Vulkswagenwerk.
REFERENCES 1. Baddiley, J., Buchanan, J. G., Carss, B. & Mathias, A. P. (1956) J . Chem. Sac. Lond. 4583 - 4588. 2. Archibald, A. R. (1974) Adv. Microbial Physiol. 11, 53-95. 3. Jann, B., Jann, K . , Schmidt, G., Qrskov, I. & Orskov, F. (1970) Eur. J . Biochem. 15,29- 39. 4. Gmeiner, J. (1975) Eur. J . Biochem. 58, 621-626. 5 . Schmidt, G., Fromme, I. & Mayer, H. (1970) Eur. J . Biochem. 14, 357- 366. 6. Sawardeker, J. S., Sloneker, J. H. & Jeanes, R. (1965) Anal. Chem. 12, 1602- 1604. 7. Lowry, 0. H., Roberts, N. R., Leiner, K. J., Wu, M. & Farr, A. L. (1954) J . Biol. Chem. 207, 1-17. 8. Baddiley, J., Buchanan, J. G. & Carss, B. (1957) J . Chem. Soc. Lond. 4138 -4139. 9. Kotelko, K., Gromska, W., Sidorczyk, Z. & Zwolinski, J . (1968) Bull. Acad. Pol. Sci. Ser. Sci. Biol. 16, 739- 744. 10. Dmitriev, B. A., Hinton, N. A,, Lowe, R. W . &Jones, J. K. N. (1971) Can. J . Microhiol. 17, 1385- 1394. 11. Ellwood, D. C. & Tempest, D. W. (1969) Biochem. J . I l l , 1 - 5.
J. Gmeiner, Fachbereich Biologie, Mikrobiologie der Technischen Hochschule Darrnstadt, D-6100 Darmstadt, Schnittspahnstrak 9, Federal Republic of Germany
Note Added in Proof(0ctober 10, 1975). After this paper had been submitted I became aware of a report by Kotelko et al. [Kotelko, K., Fromme, I. & Sidorczyk, Z. (1975) Bull. Acad. Pul. Sci. Ser. Sci. B i d . 23,249 - 2561, in which an unidentified compound X, inseparable from threitol by gas liquid chromatography, is described as a constituent of the lipopolysaccharide from two Proteus miiahilis strains belonging to serogroups 016 and 033. Dr H. Mayer (Max-Planck-Institut fur Immunbiologie, Freiburg i. Br.) kindly compared the mass spectrogram of compound X with that of anhydro ribitol isolated from ribitoL5(l)-phosphate described in the present paper and found that the spectra are in accordance with X being anhydro ribitol. Since anhydro ribitol has a very similar retention time with threitol under the gas chromatographic conditions at an oven temperature of 170°C used in the present paper (Rxy,i,o, = 0.220 and 0.225, respectively) the actual constituent of the lipopolysaccharide from Proteus mirabilis, serogroup 016 and 033 (Kotelko et al.) might be ribitol.
