Eur. J. Biochem. 58, 621 -626 (1975)

The Isolation of Two Different Lipopolysaccharide Fractions from Various Proteus mirabilis Strains Jobst GMEINER Fachbereich Biologie, Mikrobiologie, Technische Hochschule, Darmstadt (Received June 13 / July 22, 1975)

Four distinct Proteus mirabilis strains were extracted by the phenol/water procedure. After ultracentrifugation of the dialyzed water phase, the pelleted lipopolysaccharide was purified and analyzed. The sugar composition of this lipopolysaccharide fraction I was similar for all four strains, containing only small amounts of strain-specific constituents. A second lipopolysaccharide fraction was isolated from the supernatant above (termed L1 fraction) after removal of nucleic acids. DEAE-cellulose chromatography indicated that this material is not a polysaccharide but rather a water-soluble lipopolysaccharide containing strain-specific constituents such as uronic acids, amino acids, amino sugars, neutral sugars, ethanolamine and phosphate, depending on the strain from which lipopolysaccharide I1 was isolated.

Lipopolysaccharides are main constituents of the outer membrane of gramnegative bacteria. A large body of knowledge has been accumulated during the last decade concerning the chemical structure, biosynthesis and biological function of lipopolysaccharide from the Escherichia coli- Salmonella group (for review see [ 1,2]). The lipopolysaccharide from Proteus mirabilis has not been characterized very well. Kotelko and coworkers [3] reported the isolation of two different polysaccharides by acidic extraction procedures. They also showed that in a number of Proteus mirabilis strains uronic acids are constituents of the lipopolysaccharide rather than of a capsular or K-like substance [4,5]. Dmitriev et al. [ 6 ] extracted the polysaccharides from various Proteus strains with hot acetic acid and classified these strains as smooth or rough strains according to the elution pattern of the extracted polysaccharides on a Sephadex G50 column. In this paper a convenient procedure is reported for the isolation and separation of two different lipopolysaccharide fractions from each of four distinct Proteus mirabilis strains. Whereas one of these fractions resembles the lipopolysaccharide extracted from smooth strains of Salmonella, the other resembles the lipopolysaccharide extracted from rough strains of Salmonella. The smooth type lipopolysaccharide from Proteus mirabilis contains either uronic acids or phos-

phate esters as 0 side chain constituents, in addition to neutral sugars and amino sugars or amino acids. MATERIALS AND METHODS

Organisms and Growth Conditions Strains D52, VI and 19 were obtained from Dr H. H. Martin (this laboratory); strain 19 Q was obtained from Dr G. Schmidt (Max-Planck-Institut fur Immunbiologie, Freiburg i. Br., Germany) and had originally been received from Dr F. Qrskov (Statens Serum Institute, Kobenhavn, Denmark). Bacteria were cultivated in complex medium as described previously [7]. Some batches were kindly grown by Dr S. Schlecht (Max-Planck-Institut fur Immunbiologie, Freiburg i. Br., Germany) [8].

Isolation of L@opolysaccharide

Whole cells (20- 30 g dry weight) were extracted by the hot phenollwater procedure of Westphal and Jann [9]. The water phase was dialyzed extensively against water, concentrated under reduced pressure and centrifuged in the 42.1 rotor at 40000 rev./min for 120 min in a Beckmann L50 centrifuge. The supernatant, or L1 fraction, was lyophilized. The pellet containing the crude lipopolysaccharide I was suspended in Tris/HCl buffer, pH 7.5, 0.025 M Enzymes. Ribonucleate 3’-pyrimidino-oligonucleotidohydroand centrifuged again. The pellet was resuspended in lase (EC 3.1.4.22); deoxyribonucleate 5’-oligonucleotidohydrolase Tris/HCl buffer, pH 7.5, 0.001 M, containing 3 mg (EC 3.1.4.5); D-amino-acid : oxygen oxidoreductase (dedminating) ribonuclease and 3 mg deoxyribonuclease and in( E C 1.4.3.3).

622

Two Different Lipopolysaccharides from Proteus mirahilis

Table 1. Yield and sugar composition of lipopolysaccharide I Yield is based on dry weight of bacteria. KDO, 3-deoxy-~-manno-octulosonic acid; m-Hep = L-glycero-D-mannoheptose;DD-Hep = Dglyycero-D-mannoheptose; RibOH = ribitol; EtN = ethanolamine; LipA = lipid A ; n.d. = not determined Strain

Yield

P

KDO

LD-Hep DD-Hep UA

1.8

3.4 3.9

12.2 10.3 10.5 14.6

Glc

Gal

RibOH GlcN

GalN

Lys

Ala

EtN

LipA

4.7 5.2 14.7 4.3

3.1 2.6 3.5 4.5

-

3.2

0.7 1.6

2.9

0.9

n.d. 38.0 38.6 32.5

O 4

19 VI D52 19Q

3.1 3.6 1.5

1.3

1.6 2.4 1.7

5.2

4.1

8.1 6.3 0.7 4.3

10.1 14.4 9.3 9.6

cubated at room temperature overnight. The lipopolysaccharide was recovered by ultracentrifugation and washed twice with Tris/HCI buffer. The final pellet was suspended in water, dialyzed against water and finally electrodialyzed at 4 "C in an electrodialysis apparatus from Sartorius (Gottingen, Germany). After changing the water in the outer chambers, the poles were reversed several times in order to avoid clogging of the membranes by precipitated lipopolysaccharide. When the pH in the outer chambers remained constant and no further ultraviolet-absorbing material was released, the purified lipopolysaccharide suspension was collected and lyophilized. The L1 fraction, containing the lipopolysaccharide 11, was freed from the bulk of nucleic acids by perchloric acid precipitation as described by Yuasa et al. [lo]. Residual nucleic acid was removed by incubation with nucleases, dialysis and electrodialysisas described above. Analytical Methods Total phosphate was determined according to Lowry et al. [l11. 3-Deoxy-D-manno-octulosonicacid was determined with the thiobarbituric acid method [12]. Amino sugars, ethanolamine and amino acids ,were determined after hydrolysis in 4 N HC1 for 15 h at 100 "C with a BioCal 200 amino acid analyzer. Uronic acids were quantitated as described by Blumenkrantz et al. [13] with the following modification: the samples were read 90 s after addition of the reagent because I did not observe the reported stability of the developing colour. Neutral sugars and polyols were determined after hydrolysis in 0.1 N HCl for 48 h at 100 "C as alditol acetates by gas-liquid chromatography according to Sawardeker et al. [14] using a Varian Aerograph 2701 [15]. The configuration of alanine was determined by incubating a lipopolysaccharide hydrolyzate with D-amino acid oxidase together with known amounts of D- or L-alanine and subsequent quantitation using the amino acid analyzer. Lipid A was liberated from lipopolysaccharide by hydrolysis in 0.1 N HC1 for 60- 90 min at 100 "C. The precipitate was washed once, suspended in water,

~

1.1 -

5.9 5.1 4.8 5.4

5.8

3.2 3.7

-

-

0.9 1.1 0.7

lyophilized and weighed. DEAE-cellulose column chromatography was carried out essentially as described previously [16].

Enzymes All enzymes were purchased from Boehringer (Mannheim, Germany). RESULTS AND DISCUSSION Characterization of LipopolysaccharideI These lipopolysaccharides were isolated by the phenol/water extraction procedure from four different Proteus mirabilis strains, sedimented by ultracentrifugation and further purified by treatment with nucleases, recentrifugation and electrodialysis. During the last step ultraviolet-absorbing material, i.e. nucleotides, as well as polyamines and ions were removed very effectively [17]. Yields and compositions are shown in Table 1. Preparations from all strains contain phosphorous, 3-deoxy-~-manno-octulosonic acid, L-glycero-D-manno-heptose, D-glycero-D-mannoheptose, glucose, galactose, uronic acids, glucosamine,galactosamine and ethanolamine in comparable quantities except for only small amounts of D-glycero-D-mannoheptose and higher amounts of glucose in strain D52. Lysine was found in strains 19 and VI to be a strainspecific component. This has already been reported to be a lipopolysaccharide constituent of another Proteus mirabilis strain [3]. Other strain-specific components found were L-alanine in strain 19 Q> and a polyol characterized as ribitol [15] in strain D52. Protein accounted for about 1- 3 "/, calculated according to amino acid analysis. In strain 19 8, however, more than 10 % protein was found in two different lipopolysaccharide preparations. A second phenol/water extraction removed this protein although not without considerable loss of lipopolysaccharide. All four preparations contain relatively high amounts of heptose, 3-deoxy-~-manno-octu~osonic acid and lipid A compared to small amounts of the strain specific components. This is reminiscent of the composition of lipopolysaccharides from Ra mutants

623

J . Gmeiner Table 2. Yield and sugar composition of lipopolysaccharide 11 isoLated from L, fraction Yield is based on dry weight of bacteria. For abbreviations see Table 1 Strain Yield L, frdCtion

Yield lipo-

P

KDO

Lu-Hep DD-Hep UA

Glc

Gal

RibOH GlcN

0.9

1.0 1.0 1.3 1.4

2.5 2.3 3.6 6.0

1.4 1.4 7.0 5.1

0.5 0.3 8.0 26.0

14.5 -

GalN

Lys

Ala

EtN

LipA

15.8 15.0 1.0 1.0

8.5

8.2

0.2 n.d.

n.d.

2.6

7.0 9.4

POlY-

saccharide I1

'"

VI D52

n.d. 6.7 5.0

199

5.9

19

3.1 3.3 1.8 2.2

0.7 4.3 1.1

0.7 0.6 0.4 0.9

18.3 22.4 3.6 21.8

of Salmonella which have no 0 side chain material or from SR mutants which have only one O-specific repeating unit attached to the core moiety. However, no lipopolysaccharide could be extracted from Proteus strains with phenol/chloroform/light petroleum, a procedure by which specifically R lipopolysaccharide from Salmonella and E. coli is extracted [lS]. Characterization of Lipopolysaccharide II It has been already reported by Kotelko and coworkers [3] that the L1 fraction obtained in the supernatant after ultracentrifugation of the phenol/water extract contained heptose and all other lipopolysaccharide specific sugars. The rather large amount of material found as L1 fraction (Table 2) led me to a closer investigation of this material. Before analysis, nucleic acids were removed by precipitation with perchloric acid, enzymatic cleavage and electrodialysis. The purified material gave clear solutions in water and could not be sedimented by ultracentrifugation. From the analytical data (Table 2) one can see that the strainspecific components found in the lipopolysaccharide I preparations were greatly enriched in the L1 fraction, i.e. lysine, galactosamine and uronic acids for strain 19 and VI, ribitol, glucosamine, ethanolamine and phosphorous for strain D52 and L-alanine, glucosamine, galactose and uronic acids from strain 19 8. On the other hand, lipid A and the core-specific compoacid and heptose nents 3-deoxy-~-manno-octu~osonic were reduced to about 1/5 to 1/3 of the amount found in lipopolysaccharide I fractions. Protein is virtually absent in these preparations, even from strain 19 Q (see above). Variation of the Isolation Procedure Whereas lipopolysaccharide I can be separated and purified very easily from lipopolysaccharide I1 by ultracentrifugation, the question arises whether the fraction designated as lipopolysaccharide I1 is really a lipopolysaccharide containing long acidic 0 side

-

3.6 1.7 16.1 14.2

9.3 -

0.2

6.8

chains or whether it is an acidic polysaccharide contaminated with lipopolysaccharide I. It is known that capsular polysaccharides from E. coli and various other organisms are found in the corresponding L1 fraction after phenol/water and ultracentrifugation. They are freed of nucleic acids and purified by cetavlon fractionation [19]. With E. coli 0100 B. Jann et al. [20] could separate the acidic polysaccharide from the lipopolysaccharide by this method although both polymers had an identical composition, i.e. glycerol phosphate in addition to rhamnose, galactose and glucosamine as repeating unit constituents. Proteus mirabilis, on the other hand, is not known to possess a capsule or capsular polysaccharides. In fact, in a study on phenol/ water extracts from more than 30 Proteus mirabilis strains Sidorczyk and Kotelko did not succeed in precipitating any acidic polysaccharide with cetavlon [ 5 ] . Therefore, I first investigated whether nucleic acids or nucleotides influence the separation of both lipopolysaccharides during ultracentrifugation and electrodialysis. A new batch of strain 19 cells (16.7 g dry weight) was extracted with phenol/water. The extract was dialyzed and lyophilized giving 2.8 g. 750 mg of this material was processed as described under Methods for lipopolysaccharide I and II except that repeated ultracentrifugation for the purification of lipopolysaccharide I was omitted. The resulting lipopolysaccharide preparations IA and IIA served as control. Part of preparation I A was then suspended in water and purified further by two subsequent ultracentrifugations. The very soft pellet was resuspended and lyophilized giving preparation IA, , 1.5 g of the extracted material was dissolved in 150 ml Tris/HCl buffer, pH 7.5, 0.025 M, and incubated first with 3 mg ribonuclease and deoxyribonuclease, respectively, for two days at room temperature. The solution was then divided in two equal portions. One part was electrodialyzed prior to the separation step by ultracentrifugation. Here again, the resulting pellet was very soft and the separation of the supernatant from the pellet was more difficult. Both, supernatant and pellet were lyophilized giving lipo-

Two Different Lipopolysaccharides from Proteus mirabilis

624

Table 3. Yield and composirion of three dgjerent prepurutions of lipopolysaccharides I and IIfrom strain 19 Yield is based on phenoljwater extract (dry weight). For abbreviations see Table 1 Preparation

KDO

Yield

x

~.

...

LD-Hep

DD-Hep

Glc

Gdl

GlcN

GalN

.

.

Lys

~~

Lipopolysaccharide I A

A,

B C

Lipopolysaccharide TI A B C

5.5

2.7 2.9 2.8 3.1

1.8 1.9 2.0 2.1

4.4 4.1 5.5 5.8

5.6 4.8 9.5 7.0

1.8 1.5 2.1 1.8

1.7 1.5 0.9

2.3 3.6 1.5

1.o 0.8

1 .o

4.0 2.0 1.5

20.3 21.5 26.3

10.1 9.1

4.4

3.7

8.4 9.5 8.9 9.2

1.4 1.3 1.1

4.2 4.3 3.3

17.1 9.9 8.9 15.0

2.8 2.7 3.4

16.5 10.6 12.7

polysaccharide preparations IB and IIB. The second part was first ultracentrifuged and supernatant and sediment electrodialyzed separately giving lipopolysaccharide preparations IC and IIC after lyophilization. Yields and the percentage of the main constituents are given in Table 3. Whereas the standard procedure gives the best yields for both types of lipopolysaccharides the composition of the products is rather similar. In preparation B and C all nucleic acids had to be removed by electrodialysis. During this extensive treatment no degradation was observed, i.e. the rather labile ketosidic 3-deoxy-u-manno-octulosonicacid bond was not broken. On the other hand, the presence of nucleic acids during the ultracentrifugation seemed to facilitate the separation of the two lipopolysaccharide fractions. When nucleic acids where removed prior to the ultracentrifugation step the pellet was softer and the resulting lipopolysaccharide I was more contaminated with lipopolysaccharide I1 as can be estimated from the lysine content (compare IB and IC, Table 3). DEAE-Cellulose Chromatography Kent and Osborn [21] found that lipopolysaccharide from Salmonella remained irreversible bound to DEAE-cellulose whereas pol ysaccharides were elutable with water or buffer depending upon the presence of negatively charged groups [16]. If Proteus mirabilis lipopolysaccharides behave similarly, it should be possible to attack the question raised above as to whether lipopolysaccharide 11 is a mixture of polysaccharide and lipopolysaccharide. 1 mg each of the various preparations of the last experiments were dissolved in 1 ml water and 1 drop 1 N NH,OH was added. The samples were put on top of small DEAE-cellulose columns (0.8 x 10 cm) and eluted successively with 10ml water, 10 m10.3 M pyridinium acetate buffer, pH5.3 and 10ml 1 M pyridinium acetate buffer, pH 5.3. The recoveries of

5.1

5.0

7.5

3-dcoxy-~-manno-octu~osonic acid and amino compounds are summarized in Table 4. No material was eluted with water, and only a small percentage of the lipopolysaccharide I constituents was recovered from the columns. From lipopolysaccharide 11, however, 30 - 60 % of the strain specific constituents lysine and galactosamine as well as the core-specific constituent 3-deoxy-~-manno-octulosonic acid was recovered. This indicates that the water-soluble material found in the L1 fraction is a lipopolysaccharide which is only partly retained by DEAE-cellulose. It seems that the high amount of the strain-specific constituents can overcome the strong absorption of the core-lipid A part to DEAE-cellulose. Since the molar ratio lysine acid (Table 4) deto 3-deoxy-~-rnanno-octulosonic creased with increasing buffer concentration the conclusion can be drawn that the lipopolysaccharide I1 fraction consists of molecules with varying length of polysaccharide chains. A small contamination by lipopolysaccharide I, on the other hand, cannot be ruled out by this method. This conclusion is further supported by preliminary results using a new method for the separation of smooth and rough lipopolysaccharides on sodium dodecylsulfate polyacrylamide gel electrophoresis. The development of this method is under way in Dr Jann's laboratory (Max-Plancklnstitut fur Immunbiologie, Freiburg i. Br., Germany) and will be published elsewhere. By means of DEAE-cellulose chromatography and gradient elution a further fractionation of lipopolysaccharide I1 might be achievable. In this connection it should be mentioned that the linkage between the polysaccharide and lipid A was more resistant to 0.1 N acetic acid than the same linkage in Salmonella lipopolysaccharides. Hydrolysis in 0.1 N HC1 for 60-90 min at 100 "C was necessary to precipitate lipid A. Under these conditions, however, the lipopolysaccharide became partly degraded. Hence, subsequent gel filtration did not give satisfying separation of the polysaccharides from both lipopolysaccharide fractions.

625

J. Gmeiner

Table 4. Recoveries from DEAE-cellulose columns of .some constituents of the two lipopolysaccharide fractions from preparations A , B and C from strain 19 ~

~

Preparation

~~~~~~

Molarity buffer

M

~

KDO

~~

~

GlcN

~

~

GalN

LYS

Molar ratio Lys/KDO

”/, - . -

Lipopolysaccharide I A

1.o 4.4 2.0

0.3 1.o

3.2 5.3

1.4 5.7

11.5 6.5

14.4 10.6

0.3 1.o

2.9

3.0

4.2

1.o 3.3

1.6

5.5 10.1

0.3 1.o

9.7 6.0

7.4 4.3

10.0 4.0

15.1 4.6

0.3 1.o

3.5 3.0

1.5 1.0

3.4 1.o

4.2 1.o

0.3

1 .o

193 9.1

35.7 13.1

21.0 ll.Y

21.0 12.5

12.2

0.3 1.o

40.1 27.0

23.7 21.3

42.1 13.4

40.9 13.6

12.3 12.6 6.2

0.3 1.o

44.8 28.2

22.6 20.9

42.9 17.2

41.6 16.0

14.7 13.7 8.3

0.9

B

C

Lipopolysaccharide I1 A

B

C

Concluding Remarks

2.0

2.1 1.3 2.0 1.o 0.8 1.0

-

8.9 9.4

I am thankful to K. Pfizenmaier for some of the earlier analyses and the determination of the alanine configuration and to Dr C. Galanos, Max-Planck-Institut fur Immunbiologie, Freiburg i. Br., Germany, for a gift of 3-deoxy-o-manno-octulosonicacid. For stimulating discussions and generous support I wish to thank Dr H. H. Martin. This work was partly supported by the Deutsche Forschungsgemeinschuft and the Stiftung Volkswwgenwerk.

The results obtained suggest that Proteus mirabilis produces two types of lipopolysaccharides which presumably differ in their amount of strain-specific polysaccharide linked to a core-lipid A structure. In contrast to Salmonella and related organisms, Proteus mirabilis contains mainly charged compounds as REFERENCES constituents of the strain-specificpolysaccharide. This might be the reason for the increased solubility in water 1. Luderitz, O., Westphal, O., Staub, A. M. & Nikaido, H. (1971) and the decreased aggregation of lipopolysaccharide Microbial Toxins (Weinbaum, G . ,Kadis, S. & Ajl, S. J., ed.) vol. 4, pp. 145-233, Academic Press, New York. 11, thus allowing the simple separation of both frac2. Nikaido, H. (1973) Bacterial Membranes and Walls (Leive, L., tions by ultracentrifugation. There are some indicaed.) pp. 131-207, Marcel Dekker, Inc., New York. tions, however, that lipopolysaccharide I1 itself is 3. Kotelko, K., Luderitz, 0. & Westphal, 0. (1965) Biochem. 2. a heterogeneous fraction. 343, 227 - 242. From the amount of 3-deoxy-~-manno-octulosonic 4. Kotelko, K., Gromska, W., Sidorczyk, Z. & Zwolinski, J . (1968) Bull. Acad. Pol. Sci.,Ser. Sci. Biol. 16, 739-744. acid recovered with lipopolysaccharide I and lipo5. Sidorczyk, 2.& Kotelko, K. (1973) Arch. Immunnl. Ther. Exp. polysaccharide 11, respectively (Tables 1 and 2), one 21,829-838. can calculate that only about 1/4 to 1/3 of the core6. Dmitriev, B. A,, Hinton, N. A,, Lowe, R. W. & Jones, J. K. N. lipid A moiety found in P. mirabilis carries long chains (1971) Can. J . Microbiol. 17, 1385-1394. 7. Martin, H. H., Heilmann, H. D. & Preusser, H. J. (1972) Arch. of the strain-specific polysaccharide and is therefore Mikrobiol. 83, 332 - 346. recovered as lipopolysaccharide 11. Similar hetero8. Ring, K. & Schlecht, S. (1970) Zentralhl. Bakteriol. Parasitenk. geneity of lipopolysaccharides isolated from other Infektionskr. Hyg. Abt. I. Orig. 213, 103- 119. organisms has been reported (for review and discus9. Westphal, 0 . & J a m , K. (1965) Methods Carbohydr. Chem. 5, sion see [2]). 83-91.

J. Gmeiner : Two Different Lipopolysaccharides from Proreus rnirubilis

626 10. Yuasa, R., Levinthal, M. & Nikaido, H. (1969) J . Bucteriol. 11.

12.

13. 14. 15.

100,433 -444. Lowry, 0. H., Roberts, N. R., Leiner, K. Y . , Wu, M. & Farr, A. L. (1954) J . Biol. Chem. 207, 1-17. Waravdekar, V. S. & Saslaw, L. (1959) J . B i d Chem. 234, 1945- 1950. Blumenkrantz, N. & Asboe-Hansen, G. (1973) Anal. Biochem. 54,484 - 489. Sawardeker, J. S., Slonecker, J. H. &Jeanes, A. R. (1965)AnaL Chem. 37, 1602- 1604. Gmeiner, J. (1975) Eur. J . Biochem. 58, 627-629.

16. Gmeiner, J. (1975) Eur. J . Biochem. 51, 449-457. 17. Galanos, C. & Liideritz, 0.(1975) Eur. J . Biochern. 54. 603610. 18. Galanos, C., Liideritz, 0. & Westphal, 0. (1969) Eur. J . Biochem. 9,245 -249. 19. Jann, K., Jann, B., Qrskov, F., grskov, 1. & Westphal, 0. (1965) Biochern. Z . 342, 1-22. 20. Jann, B., Jann, K., Schmidt, G., Qrskov, I. & Qrskov, F. (1970) Eur. J. Biochem. 15,29- 39. 21. Kent, J. L. & Osborn, M. J. (1968) Biochemistry,8,4396-4408.

J. Gmeiner, Fachbereich Biologie, Mikrobiologie der Technischen Hochschule Darmstadt, D-6100 Darmstadt, Schnittspahnstralle 9, Federal Republic of Germany

The isolation of two different lipopolysaccharide fractions from various Proteus mirabilis strains.

Four distinct Proteus mirabilis strains were extracted by the phenol/water procedure. After ultracentrifugation of the dialyzed water phase, the pelle...
468KB Sizes 0 Downloads 0 Views