Extraction and characterization of lipopolysaccharide from Pseudomonas pseudomallei

FEMS MicrobiologyLetters 96 (Iq92) 129-134 © 1992Federation of European MicrobiologicalSocieties(1378-1097/92/$115.1,11 Publishedby Elsevier

129

FEMSLE 05019

Extraction and characterization of lipopolysaccharide from Pseudomonas pseudomallei Kazuyoshi K a w a h a r a a, S u r a n g Dejsirilert b, H i r o f u m i D a n b a r a "'~ a n d T a k a y u k i Ezaki c a Departmentof Bacteriology, The KitasatoInstitute, Tokyo, Japan. i, National Institute of Health, Nonthaburi, Thailand, and r Department of Microbiology, Gifu Unit'ersitySchool of Medicine. Gifu, Japan

Received I i May 1992 Revisionreceived I June 1992 Accepted 2 June 1992 Key words: Pseudomonas pseudomaUei; Melioidosis; Lipopolysaccharide

1. SUMMARY

similar to that of P. cepacia, which has close taxonomic relationship with P. pseudomallei.

The best yield of lipopolysaccharide (LPS) of Pseudomonas pseudomallei G I F U 12046 was ob-

tained by extraction of defatted cells by phenol/chloroform/petroleum ether. The LPS showed a smooth character on SDS-polyacrylamide gel electrophoresis and contained D-glucose, L-glycero-D-manno-heptose, and D-glucosamine as the main sugar components, and 3-hydroxypalmitic acid as an amide-linked fatty acid. The growth conditions did not affect the electrophoresis profile and chemical composition of LPS. 2-Keto-3-deoxyoctonic acid was not detectable, and mild acid hydrolysis could not liberate free lipid A, suggesting that the linkage between inner core and lipid A was stable against acid hydrolysis, and the structure of this region is

Correspondence to: K. Kawahara, Departmentof Bacteriology,

The Kitasato Institute, 5-9-1, Shirokane, Minato-ku, Tokyo 108, Japan. i Present address: School of Pharmaceutical Sciences, Kitasato University,Minato-ku,Tokyo 108,Japan.

2. INTRODUCTION Pseudomonas pseudomaUei is a causative agent of melioidosis, which is endemic in tropical areas, especially in Southeast Asia [1,2]. Because of this geographic specificity, the bacteriological properties, including virulence factors, of P. pseudomallei have not been intensively studied. Lipopolysaccharide (LPS) is one of the main virulence factors of Gram-negative bacteria [3], and at the same time, a major surface antigen [4], which is very important for the diagnosis of infectious diseases. In spite of such significance, few studies have been reported until now on the LPS of P. pseudomaUei. Rapaport et al. [5] reported that crude cell extracts of P. pseudomallei showed endotoxic properties but no chemical properties of the endotoxin or LPS were described. In the course of our study of the virulence of P. pseudomallei, we

130 found it necessary to characterize the LPS. In addition, LPS can provide chemotaxonomic data for this bacterium, as required for the proper grouping of Pseudomonas species.

3. M A T E R I A L S A N D M E T H O D S

3.1. Bacterial strains and growth conditions P. pseudomallei G I F U 12046 was grown on Heart Infusion agar (Difco, Detroit, MI) plates at 37°(2 for 7 days for the large scale preparation of LPS. Alternatively, the bacteria were grown on the same plates at 37°C for 4 days, or in Brain Heart Infusion broth (Difco) at 37°C for 8 days. Salmonella minnesota R60 (kindly provided by Dr. G. Schmidt, Forschungsinstitut Borstel, F R G ) was used to prepare standard rough-type LPS.

3.4. Chemical analysis Neutral and amino sugars were determined by GLC as described in a previous paper [10]. Fatty acid determination was also carried out by GLC after hydrolysis (4 M HCI, 100°C, 5 h), and methyl-esterification. Chemical-bonded capillary column, CBP-I and CBP-10 (Shimadzu, Kyoto, Japan) were installed in a GC-14A (Shimadzu) apparatus, and used for the entire experiment. GC-MS analysis of sugars and fatty acids was done on JMS-DX 300 (JEOL, Tokyo, Japan), if necessary. Total hexosamine was determined after hydrolysis with 4 M HCi at 100°C for 5 h by the method of Strominger et al. [11], and total phosphorus by the method of Lowry et al. [12]. 2-Keto-3-deoxyoctonic acid (KDO) was determined by the thiobarbituric acid (TBA) assay [13] after hydrolysis in 0.1 M sodium acetate buffer (pH 4.4) at 100°C for 1 h.

3.2. Extraction of LPS LPS was extracted by the following three procedures from harvested biomass killed by 1% phenol and dried as described by Galanos et al. [6]: dried cells were first extracted with p h e n o l / w a t e r [7], and crude LPS was purified by ultracentrifugation (LPS-1), Secondly, p h e n o l / chloroform/petroleum ether (PCP) extraction [6] was carried out, and the crude LPS recovered by acetone precipitation from condensed phenol solution. The hydrophobic contaminant was removed from the precipitate by c h l o r o f o r m / methanol ( C / M ) (2:1, v / v ) extraction, and dialysed against triethylamine-containing water (pH 8.5). Insoluble material was removed, and the supernatant was lyophilized (LPS-2). In the improved method, cells were first extracted with C/M (2:1, v / v ) , and C / M (1:3, v/v). Defatted cells were then extracted with PCP, the LPS precipitated with acetone and lyophilized after dialysis as LPS-2 (LPS-3).

3.3. SDS-polyacrylamide gel electrophoresis LPS was pretreated by the method of Tsai and Frasch [8], and SDS-polyacrylamide gel electrophoresis (SDS-PAGE) performed on 4 - 2 0 % gradient gel (TEF, Nagano, Japan) with the buffer system of Laemmli [9]. The gel was oxidized with periodate and silver stained as [8].

4. R E S U L T S

4.1. LPS profiles and chemical composition Classical p h e n o l / w a t e r extraction resulted in very low yield (Table 1) of LPS (LPS-1). LPS-1 Table 1 Chemical composition of LPS and acid-degraded LPS of P. pseudomalleiGIFU 12046 Components GIc Hep Gal Rha U- 1 U-2 GIcN GaIN KDO Phosphate Ci4:o C16:0 2-OH-CI4 3-OibCt4 3-OH-CI~ Yield (rag/5 g)

L P S - I LPS-2 LPS-3 Ppt of acidtreated LPS-3 69 76 105 26 33 20 25 23 136 3 4 3 48 8 8 5 0 52 64 7 0 15 23 3 10 23 26 44 15 4 12 II 1 1 1 1 27 24 18 31 2 3 2 15 1 4 I 6 4 2 2 7 12 7 9 28 17 10 11 36 6

28

49

The amount o f each component was indicated as × 1 0 -2 /~mol/mg. The abbreviations are described in the text.

131

a:

1

2

3

4

b:

1

2

3

Fig. 1. SDS-PAGE analy.~is of LPS extracted from P. pseudomallei GIFU 12046. (a) Lane I, S. minnesota R60 LPS (0.2 p.g); lane 2, LPS-I (0.5 ,~tg): lane 3. LPS-2 (0.5 p.g): lane 4. LPS-3 (I /.tg). (b) Lane L culture on agar for 7 days (LPS-3) (0.5 /zg); lane 2, culture on agar for 4 days (0.5 p,g): lane 3, culture in broth for 8 days (0.5/tg).

usual LPS components, L-glycero-D-mannoheptose (Hep), and D-glucosamine (GIcN). However, almost no KDO was detected in any of the preparations of LPS by the method used in this study. Even when stronger hydrolysis conditions (1 M HCi, 100°C, 2 h) were used, the pink colour (maximum absorption at 549, 520 and 456 nm) observed in the TBA assay was very weak. Two neutral sugar peaks on GLC were unidentified. U-2 showed an El-MS profile and molecular mass identical to Rha, but had a slightly different retention time on GLC. U-1 was 28 mass units smaller than Rha in the acetyl alditol form, and could be a degradation product of some sugar component. All LPS preparations contained myristic acid (Ci4:0), 2-hydroxymyristic acid (2-OH-CI4), 3-bydroxymyristic acid (3-OH-C~4), and 3-hydroxypalmitic acid (3-OH-CI~). Only 3-OH-CI~ was not released by mild alkali hydrolysis (0.25 M NaOH, 60°C, I h), indicating that 3-OH-C1~ was bound to GIcN by an amide linkage.

4.2. Comparison of LPS under different culture conditions showed no ladder bands on SDS-PAGE, and distinct bands were observed at the similar position as S. minnesota R60 LPS (Fig. la, lane 2). LPS-1 contained rather high amounts of o-galactose (GAD, D-glucose (GIc), and rhamnose (Rha), whose configuration has not yet been determined. The amounts of th~:se sugars were lowered by repeating the ultracentrifugation, and in the supernatant after centrifugation the polysaccharide containing Gal, GIc and Rha was present, suggesting that not all of these neutral sugars were components of LPS-I. These results indicated that LPS-1 was rough-type LPS without O-antigenic polysaccharide, and the phenol/water method was not suitable for the extraction of LPS from this bacterium. LPS-2 and 3 extracted by the PCP method showed clear ladder bands at the high molecular mass region on SDS-PAGE (Fig. la, lanes 3, 4), suggesting that these LPS contained smooth-type LPS with the O-antigenic polysaccharide having repeating units in it. As shown in Table 1, they contained GIc as a main neutral sugar, and, as

LPS was extracted from the cells of 4-day culture on agar, or 8-day culture in broth by the same method as used for LPS-3, and compared with that of 7-day culture cells on agar. The ladder profile (Fig. lb) and the chemical composition (data not shown) were not distinguishable among these LPS. This result indicated that chemotypc of LPS was not easily altered by the growth pe~"iod or condition, although old cultures of this bacterium often showed rough-type morphology.

4.3. Effect of mild acid hydrolysis Since no KDO was detected by usual methods, LPS-1 and LPS-3 were submitted to mild acid hydrolysis (0.1 M HCI, 100°C, 1 h) in order to see the liberation of lipid A. As a result, only a small amount of lipid A-like precipitate was obtained. The hydrolysate was then dialysed, and the lipid portion collected by uitracentrifugation. The migration behaviors o~: t~ese precipitates were analysed on SDS-PAGE by comparison with lipid A liberated from Salmonella R60 LPS.

132

1

2

3

4

5

6

Fig. 2. Mild acid hydrolysisof LPS extracted from P. pseudomallei GIFU 12046in comparisonwith S. minnesota R60 LPS. Lane 1, R60 LPS (0.2 /.tg); lane 2, ppt of acid-treated R60 LPS (1 p,g); lane 3, LPS-I (0.5/zg); lane 4, ppt of acid-treated LPS-I (1 p,g);lane 5, LPS-3(I /.tg); lane 6, ppt of acid-treated LPS-3(I/~g).

Lipid A obtained from R60 LPS gave faint and broad bands at the smaller molecule region (Fig. 2, lane 2). In contrast with this, the mobility of the precipitate from LPS-1 was identical with the non-treated LPS-1 (Fig. 2, lanes 3, 4). The ladder bands of LPS-3 at the high-molecular-mass region disappeared after acid hydrolysis, but the bands in the low-molecular-mass region remained (Fig. 2, lanes 5, 6). The chemical composition of the precipitate from LPS-3 is shown in Table 1. It contained Gic and Hep with about half the amount of GleN in addition to the usual lipid A components: fatty acids, G l e n and phosphate. These results suggested that the linkage between the core oligosaccharide and lipid A was resistant to mild acid hydrolysis, resulting in little liberation of free lipid A, and the lipid portion obtained by ultracentrifugation still contained som¢ of the core sugars.

5. DISCUSSION In the present study LPS of P. pseudomallei was first isolated and chemically analysed. The extraction was interfered with by the unknown hydrophobic substance, and it was necessary to remove this substance first by C / M extraction in order to obtain LPS in good yield and in a comparatively pure form. Although the physicochemical mechanism was not clear, smooth-type LPS was extracted by PCP extraction, which is usually applied to the extraction of rough-type LPS. The most remarkable character of this LPS was the absence of TBA assay positive substance when the usual method was used. This characteristic, including the absorption pattern of the TBA assay colour obtained by stronger hydrolysis, was identical to that of P. cepacia recently reported by Cox and Wilkinson [14]. In addition, both LPS contained 3-OH-C1~ as the main and amide-linked fatty acid. P. pseudomailei and P. cepacia belong to r R N A - D N A homology group 1I of genus Pseudomonas in the grouping proposed by Palleroni et al. [15], and this idea was supported by the recent sequence data of rRNA (unpublished results). The present study provides further evidence to suggest the close relationship of these species, but further chemical study is necessary on the structure of the inner core-lipid A region of both species.

ACKNOWLEDGEMENTS We are very grateful to Ms. E. Asaga and Prof. Y. Nakase (Kitasato University) for their cooperation and helpful discussions. We also thank Ms. A. Nakagawa and Ms. C. Sakabe (Kitasato University) for the excellent GC-MS analysis.

REFERENCES [1] Kanai, K. and Dejsirilert,S. (1988)Jpn. J. Med. Sci. Biol. 41, 123-157. [2] Leelarasamee,A. and Bovornkitti,S. (1989) Rev. Infect. Dis. 11,413-425.

133 [3] Stocker, B.A.D. and Miikel~i, P.H. (1986) Curr. Top. Microbiol. lmmunol. 124, 149-172. [4] Liideritz, O., Freudenberg, M.A., Galanos, C., Lehmann, V., Rietschel, E.T. and Schaw, D.H. (1982) In: Membrane lipids of procaryotes, Vol. 17. Current topics in membranes and transport, (Razin, S. and Rottem, S., Eds.), pp. 79-151. Academic Press, New York. [5] Rapaport, F.T., Millar, CJ.W. and Ruch, J. (1961) Arch. Pathol. 71,429-436. [6] Galanos, C., Liideritz, O. and Westphal, O. (1969) Eur. J. Biochem. 9, 245-249. [7] Westphal, O. and Jann, K. (1965) in: Methods in carbohydrate chemistry, Vol. V, (Whistler, R.L., Ed.), pp. 83-91. Academic Press, New York. [8] Tsai, C.-M, and Frasch, C.E. (1982) Anal. Biochem. 119, 115-119.

[9] Laemmli. U.K. (1970) Nature 22/. 680-685. [10] Kawahara, K., Brade, H., Rietschel, E.T. and ZAhringer, U. (1987) Eur. J, Biochem. 163, 489-495. [ll] Strominger. J., Park, LT. and Thompson, R. (1959) J. Biol. Chem. 234, 3263-3267. [12] Lowry, O.H., Roberts, N., Leiner, K., Wu, M and Farr, L. 0954) J. Biol. Chem. 207. i-17, [13] Brade, H., Galanos, C. and Lfideritz, O. (1983) Eur. J. Biochem. 131, 195-200. [14] Cox, A.D. and Wilkinson, S.G. (1991) Mol. Microbiol. 5, 641-646. [15] Palleroni, N.J., Kunisawa, R., Contopoulou, R. and Doudoroff, M. (1973) Int. J. Syst. Bacteriol, ~ , 333-339.