Vol. 16, No. 3 Printed in U.S.A.

INFECTION AND IMMUNITY, June 1977, p. 983-994 Copyright ©D 1977 American Society for Microbiology

Biological and Physicochemical Properties of the Lipopolysaccharide of Chromatium vinosum RONALD E. HURLBERT*

AND

IRIS M. HURLBERT

Department of Bacteriology and Public Health, Washington State University, Pullman, Washington 99164 Received for publication 26 January 1977

The lipopolysaccharide (LPS) of Chromatium vinosum has anticomplementary activity. This anticomplementary activity is destroyed by alkaline digestion

of the LPS and is suppressed by both Mg2+ and Ca2+ ions. Treatment of the LPS with ethylenediaminetetraacetic acid, sodium deoxycholate, or dimethyl sulfoxide did not affect its toxicity toward mice; however, alkaline-treated LPS was not toxic. Treatment of the LPS with sodium deoxycholate, dimethyl sulfoxide, or sodium dodecyl sulfate resulted in reversible dissociation into subunits. Aggregation of the subunits into the original form was achieved by removing the dispersing agent by dialysis against distilled water followed by freezing and thawing. Electron micrographs of phenol-extracted LPS showed long filaments. Electron micrographs of sodium deoxycholate- and sodium dodecyl sulfatetreated and dialyzed LPS showed a mixture of small subunits and short filaments, whereas dimethyl sulfoxide-treated and dialyzed LPS contained only small ovoid spheres. The LPS produced an ordered series of multiple bands on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A similar banding pattern was observed for Salmonella abortus-equi and Proteus mirabilis LPS. The C. vinosum LPS appears to be mitogenic for mouse spleen cells.

The cell wall lipopolysaccharide (LPS) of gram-negative bacteria is known to be responsible for many of the serological properties of these organisms (16). In addition, LPS has been shown to be a potent toxin (endotoxin) that is thought to be one of the virulence factors of gram-negative pathogens (30). Among the toxic responses elicited by LPS are lethal shock, pyrogenicity, general and dermal Schwartzman reactions, complement destruction, platelet damage, and vascular disturbances (6, 7, 21, 36). LPS have also been reported to cause tumor necrosis (18), to stimulate resistance to infection (3), and to have mitogenic activity (19, 28). However, not all LPS preparations possess these biological activities or have them to the same degree. Although it is known that the lipid A portion of LPS is responsible for some of the toxic (15) as well as the mitogenic characteristics of LPS (2), the relationship between these activities and the molecular structure of LPS is not understood. It is, therefore, of value to investigate the biological and chemical properties of various LPS so that those molecular characteristics that are responsible for the different biological activities can be distinguished. The present paper describes some biological and physicochemical properties of the LPS of the purple sulfur bacterium Chromatium vinosum.

MATERIALS AND METHODS Cultivation of bacteria and isolation of LPS. The bacteria were cultivated as previously described (9). The C. vinosum LPS was obtained from the phenol phase after extraction with hot phenol/water, as previously described (9). Thiocapsa (Lascelles strain) and Escherichia coli strain B LPS were also extracted by the hot phenol/water method (41) and were purified by three sequential ultracentrifugations (100,000 x g, 3 h). Electron microscopy investigations. LPS preparations were stained with 2% phosphotungstic acid and were carbon coated. Preparations were examined in a Hitachi H-8 electron microscope fitted with cooling equipment. Analytical ultracentrifugation. Sedimentation coefficients were determined by using a Spinco analytical ultracentrifuge (Beckman Instruments, Inc., model E) with a 12-mm cell. The centrifugations were carried out at 20°C in an analytical D-rotor on solutions containing 1% (wt/vol) of the sample. Complement fixation test. Pooled guinea pig serum was stored at -80°C. Complement fixation was

determined by the following modification of the method of Galanos et al. (5). An amount of freshly titrated serum containing 21.5 U of complement (0.07 to 0.09 ml) was added to graded amounts of LPS suspended in distilled water. The volume of the mixture was brought to 150 ,lI with distilled water, and the mixture was incubated for 60 min in a 37°C water bath. From each mixture, 7 ,ul was added to 1 ml of a Mg2+-saline solution (0.5 g of MgSO4 7H20 + 0.1 g of CaCl2 per liter of 0.85% NaCI). To this was added 0.5 ml of a suspension of sheep erythrocytes 983 -

984

HURLBERT AND HURLBERT

sensitized with amboceptor. This mixture was incubated at 370C for 1 h, 1.5 ml of the Mg2+-saline solution was added, and the mixture was centrifuged. The absorbance of the supernatant was read at 546 nm in a Bausch & Lomb Spectronic 20 spectrophotometer. Under these conditions, the complement control and a suspension of erythrocytes diluted to the same volume with distilled water had an absorbance of 0.38. Anticomplementary activity is expressed as the percent inhibition of hemolysis against the LPS concentration in milligrams per milliliter in the 150-,ul initial incubation mixture. Treatment of LPS. Alkali-treated LPS was prepared by suspending the LPS (12.5 mg/ml) in 0.25 N NaOH, followed by heating in a water bath at 56 or 1000C for 1 h. After cooling, the solution was neutralized by the addition of concentrated acetic acid or 6 N HCl. In some cases the neutralized material was used directly, whereas in other cases it was dialyzed against deionized water at room temperature overnight and then lyophilized. Sodium deoxycholate (SD)-treated LPS was prepared by dissolving LPS (10 mg/ml) in 0.5% SD in 0.1 M tris(hydroxymethyl)aminomethane (Tris)-hydrochloride buffer, pH 7.8. This solution was dialyzed at 80C against several changes of distilled water for 72 h and lyophilized. Ethylenediaminetetraacetic acid (EDTA)-treated LPS was prepared by suspending the LPS (15 mg/ml) in 0.2 M EDTA, pH 7.5. This suspension was mixed for 2 h at room temperature and was then dialyzed and treated as above. The dimethyl sulfoxide (Me2SO)-treated LPS was prepared by mixing the LPS in Me2SO until dissolved (15 min). This solution was dialyzed against distilled water or buffer at room temperature for 24 to 48 h. The dialyzed material was collected and stored at 40C until used. Dyed LPS was prepared as described by Jann et al. (10). Acrylamide procedures. The LPS samples were analyzed by the sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) procedure of King and Laemmli (12). Samples were routinely solubilized immediately before electrophoresis by heating a suspension of the LPS (2.5 mg/ml) in a solution containing 2% SDS, 5% 2-mercaptoethanol, 10% glycerol, 0.0625 M Tris-hydrochloride buffer (pH 6.8), and 0.0002% bromophenol blue for 1 or 2 min in a stoppered tube in a boiling-water bath. Immediately after electrophoresis, gels (0.6 by 8.5 cm) were placed in a fixing solution of 1 M KCl-10% acetic acid for 1 to 2 h (20 ml/gel). They were then soaked in a solution (20 ml/gel) of 15% trichloroacetic acid-25% isopropanol (wt/vol) until clear (1 to 2 h). The gels were further washed overnight in a solution (12 gels/liter) of 5% acetic acid-40% ethanol (vol/vol) and stained. This procedure was found to significantly enhance the details in the gels stained for polysaccharide. Protein was detected as previously described (8), and polysaccharide (periodic acid-Schiff stain) was stained by the procedure of Segrest and Jackson (34). All gels were viewed and photographed against a fluorescent lamp covered by white, translucent plastic. For elution of LPS from gels, the gels were cut into sections immediately upon completion of elec-

INFECT. IMMUN.

trophoresis. These sections were suspended in 100 ml of 0.1 M Tris-hydrochloride (pH 7.8)-0.1% SDS (Tris-SDS buffer) and chopped up in a VirTis mixer. The fine gel particles were then poured into a column and eluted at room temperature with the above Tris-SDS buffer (150- to 250-ml amounts were collected) at the rate of 1 drop/2 s. The eluate was dialyzed against distilled water at room temperature for 72 h and concentrated to 8 to 10 ml in a rotary evaporator. Five volumes of acetone were added to the concentrate, and the precipitate was collected and washed once with acetone. This material was dried under vacuum. Mouse toxicity testing. Swiss-Webster female mice, 6 to 7 weeks old (the variation in age in any given experiment was no more than 3 days), were injected intraperitoneally (0.3 ml/mouse) with LPS samples in 0.9% saline. Lethality was determined after 3 days. Chemicals. All chemicals were of reagent or analytical grade and were purchased from commercial sources. Purified Salmonella abortus-equi LPS was kindly supplied by C. Galanos (Freiburg, Germany), and Proteus mirabilis strain 1959 LPS was a gift of K. Gramska (Lodz, Poland). LPS from E. coli strain 0114:B4 (lot no. 610243) was obtained from Difco Laboratories (Detroit, Mich.). Procion red MX2B (D-4251) was a gift of the ICI America, Inc., Charlotte, N.C. Tissue culture medium RPMI 1640 was a product of Grand Island Biological Co., Berkeley, Calif. Mitogenic assay. Spleen cells (106) from SwissWebster mice were suspended in 1 ml of RPMI 1640 medium. After 92 h of incubation, a pulse of 0.5 ,uCi of [3H]thymidine (6.7 Ci/mmol) was added, and the incubation was continued for another 4 h. The cells were collected on glass fiber filters and washed once with 30 ml of distilled water. The filters were dried in scintillation vials. After the addition of 7 ml of a scintillation cocktail containing 6 g of 2,5-diphenyloxazole and 0.1 g of 1,4-bis[2]-(5-phenyloxazolyl)benzene per liter of toluene, the radioactivity was determined by using a Nuclear-Chicago Unilux III liquid scintillation spectrometer (Nuclear-Chicago Corp., Des Plaines, Ill.).

RESULTS Anticomplementary activity. C. vinosum LPS was approximately one-quarter as effective an anticomplementary substance as S. abortus-equi LPS (Fig. 1). Heating the LPS at 100°C for 5 min had no effect on the anticomplementary activity. LPS treated with EDTA or SD showed no change in anticomplementary activity, but alkaline treatment (560C for 1 h) of the LPS significantly decreased its anticomplementary activity (Fig. 1). During this study the LPS was prepared in distilled water and stored at -200C between experiments. It was observed that repeated freezing and thawing caused aggregation of the LPS and a loss of anticomplementary activity. Samples of LPS that had gone through four or

VOL. 16, 1977

LIPOPOLYSACCHARIDE OF CHROMATIUM

z 0

n U)

0

IL

w

IL.

0

Iw

w

0L

LPS CONC (mg/ml) FIG. 1. Anticomplementary activity of C. vinosum and S. abortus-equi LPS. Increasing amounts of the two LPS were incubated with a constant number of units ofguinea pig complement in a volume ofl 50 p, and the hemolysis test was performed as described in the text. Symbols: C. vinosum LPS, untreated (a) and alkaline-treated (x); S. abortus-equi LPS (0). The results are an average of three experiments.

cycles of freezing and thawing had only approximately 50% of the anticomplementary activity of freshly prepared material. It has been observed that Mg2+ and Ca2+ ions inhibit the anticomplementary activity of Rform LPS (5). To test this possibility for C. vinosum LPS, either it was preincubated with Mg2+ ions and then tested for anticomplementary activity, or Mg2+ was included in the test system. In the first case, samples of LPS were incubated in 0.05 M MgCl2 for 5 min at 450C. The results showed that the Mg2+-treated LPS had significantly higher anticomplementary activity than the heated control (Fig. 2A). This experiment was repeated four times, and in each case the results were the same. In the second experiment, a quantity of C. vinosum LPS giving approximately 95% inhibition was incubated with serum in the presence of increasing amounts of MgCl2. The anticomplementary activity was lost in the presence of 0.05 ,umol of Mg2+ per ug of LPS (Fig. 2B). Calcium ion was equally effective in the suppression of anticomplementary activity. Toxicity. A comparison of the lethal endotoxic efficacy of C. vinosum LPS treated with EDTA or SD or digested with 0.25 N NaOH more

985

(56°C for 1 h) showed that only alkaline digestion resulted in any significant loss in toxicity (Table 1). After treatment with Me2SO and dialysis against distilled water, the LPS apparently remained in a disaggregated state, which, however, reaggregated upon freezing and thawing (see Discussion). Both forms appeared to be equally lethal to mice and did not differ significantly from untreated LPS (Table 1). Analytical ultracentrifugational studies. C. vinosum LPS formed a viscous, opalescent suspension in water, Tris-hydrochloride buffer (pH 8.0), or 0.2 M NaCl. This material produced a broad boundary and sedimented so rapidly in the analytical ultracentrifuge that only an approximation of its sedimentation coefficient could be obtained. However, the addition of SD or SDS to a final concentration of 0.5% to LPS in 0.1 M Tris-hydrochloride, pH 8.0, resulted in immediate clearing of the suspension. Analytical ultracentrifugation of these solutions showed that SD and SDS caused a disaggregation of the LPS into subunits that sedimented as a single sharp peak with apparent sedimentation coefficients corrected to water at 20°C (S20,., c = 1%) of 2.27 x 10-13 s (Fig. 3A) and 3.27 x 10-'3 s, respectively. Removal of the SD by dialysis against 0.1 M Tris-hydrochloride, pH 8.0, resulted in a partial reassociation ofthe subunits into aggregates with an apparent sedimentation coefficient of 3.6 x 10-13 s (Fig. 3B). C. vinosum LPS dissolved in Me2SO to give a clear yellow solution that sedimented as a single sharp peak with an apparent sedimentation coefficient of 2.58 x 10-13 s. LPS incubated with 0.25 N NaOH at 56°C for 1 h and neutralized (to give a solution of 1% LPS in 0.2 N NaCl) gave one major peak and one minor peak with apparent sedimentation coefficients of 9.76 x 10-13 and 8.86 x 10-13 s, respectively. However, if this material was dialyzed and lyophilized, it aggregated to give material that sedimented in the ultracentrifuge as a single peak with an apparent sedimentation coefficient of 79.8 x 10-13 s. The LPS that had been alkaline treated at 100°C for 1 h sedimented with apparent sedimentation coefficient of 4.88 x 10-13 s before lyophilization and 7.92 x 10-13 s after lyophilization. PAGE studies. Recent reports have shown that LPS can be separated by SDS-PAGE techniques (10, 11, 31). The results of a study of the pattern of C. vinosum LPS on polyacrylamide gels compared with LPS from some other gramnegative bacteria are shown in Fig. 4. The banding pattern of each LPS was unique. The C. vinosum LPS formed three zones: an upper diffuse, densely staining region (Fig. 4A and B); a central region composed of a series of fine

986

HURLBERT AND HURLBERT

LPS CONC (mg/ml)

INF ECT . IMMUN .

ji MOLES

Mg'*/jig L P S

FIG. 2. Effect of Mg2+ on the anticomplementary activity of C. vinosum LPS. In (A), the LPS was divided into two 0.2-ml samples (1 mg/sample). A 10-,mol amount of MgCl2 was added to one portion, and both were heated at 45°C for 5 min before testing their anticomplementary activity. Symbols: Mg2+-treated sample (0); heated control (a). In (B), constant amounts of C. vinosum LPS (150 pg) were added to a mixture of guinea pig sera and increasing amounts of MgCl2. Complement activity was carried out as described in the text. TABLE 1. Effect of alkaline digestion, SD, Me2SO, and EDTA treatment on lethal toxicity of LPS to mice Treatment

No. of deaths/8 micea 4 mg/ 2 mg/ 1 mg/ mouse mouse mouse

8 2 0 Exptl None EDTA 8 2 0 8 2 SD 0 0 0 Alkalineb 0 6 8 1 Expt 2 None 4 7 0 Me2SO, unfrozenc 6 3 1 Me2SO, frozenc d a Determined after 3 days. b NaOH (0.25 N, 56°C, 1 h) followed by neutralization with acetic acid, dialysis, and lyophilization. c Solid NaCl was added immediately before testing to the Me2SO-treated LPS, prepared as described in the text, to bring the concentration to 0.9% saline. d Frozen overnight at -20°C. Solid NaCl was added after thawing.

bands; and a third, lower region containing a single densely staining band. At low concentrations, the fine bands in the central region are seen to extend up into the upper zone (Fig. 4B). (Alkaline, 56°C)-treated LPS gave the same banding pattern as the untreated LPS, except the fine bands were usually very faint (Fig. 4C). (Alkaline, 100°C)-treated LPS did not produce the fine bands. The C. vinosum LPS gave the same banding pattern if it was dissolved in

the SDS-2-mercaptoethanol solution without heating, after heating at 100°C for 15 min, or at 370C for 24 h. The LPS of E. coli strain B ran at the ion front, indicating that it was either highly negatively charged or very small (Fig. 4D). The LPS of both S. abortus-equi and P. mirabilis showed distinctive banding patterns with numerous bands (Fig. 4E and G). The LPS of Thiocapsa migrated as a broad, slow-moving band (Fig. 4F). Outer membrane of C. vinosum purified by differential centrifugation produced, upon dissolution and electrophoresis, the same periodic acid-Schiff-staining pattern as the purified LPS. The ordered periodicity of the fine bands ofC. vinosum LPS suggested that they are integral multiples of a basic monomeric unit. If this were the case, one would predict that if the high-molecular-weight forms of the LPS on the polyacrylamide were isolated and resolubilized, they would reproduce the original banding pattern. To test this possibility, twelve duplicate gels were divided into four regions (0 to 2, 2 to 4, 4 to 5.5, and 5.7 to 7 cm from the top), and the LPS in each of these regions was eluted and collected as described in Materials and Methods. When each of these fractions was solubilized and rerun on gels, the material from the upper two regions produced banding patterns similar to the original LPS (Fig. 5A and B).

VOL. 16, 1977

LIPOPOLYSACCHARIDE OF CHROMATIUM

I

987

I

FIG. 3. Ultracentrifuge patterns of phenol-extracted LPS after treatment with and subsequent removal of SD. (A) LPS (1%lo) in 0.5% SD in 0.1 M Tris-hydrochloride buffer (pH 8.0). (B) LPS (1%) after treatment with SD and subsequent removal of SD by dialysis against 0.1 M Tris-hydrochloride buffer (pH 8.0). The measurements were done at 244,000 x g at 48 min (1), 80 min (2), 112 min (3), and 144 min (4). The apparent sedimentation coeff;cients were 2.27 x 10-13 S (A) and 3.27 x 10-'3 S (B).

Material from the third region (4 to 5.5 cm) Contained no high-molecular-weight LPS, but did contain the central banding material as well as the fastest-moving band (Fig. 5C). Material eluted from the 5.5- to 7-cm region contained only the fastest-moving band (Fig. 5D). These data suggest that, under the experimental conditions used, the LPS disaggregates from a high-molecular-weight state to a series of lowmolecular-weight states but that reverse aggregation does not occur. However, the aggregates are moderately stable, as shown by the fact that if a single gel is sectioned into 12 equal pieces (from 0 to 7 cm) and each section is placed without further treatment on top of another gel and rerun, the material, including the fine bands, in each section migrates to its original position on the parent gel, i.e., it does not disperse throughout the gel (data not shown). The LPS of C. vinosum separates in the phenol phase during the isolation procedure (8), and it has been suggested that such distribution occurs because the LPS is bound to protein (29). To examine this possibility for C. vinosum LPS, crude LPS that had not been treated with proteolytic enzymes was run on acrylamide gels and stained with Coomassie blue. No proteinstaining material was detected, even at concentrations of LPS that overloaded the gel (500 gg). However, if the LPS gels (crude or en-

zyme-treated LPS) were first stained with periodic acid-Schiff stain and then counterstained with Coomassie blue, a faint blue zone became visible in the upper portion of the gel, but the blue stain was easily washed out if the destaining procedure was extended beyond 2 days. Whether this weakly staining material represents protein that has been unmasked by the periodic acid-Schiff stain treatment or an artifact of Coomassie blue binding with the modified LPS is not clear. Recently, Jann et al. (10) reported that LPS could be dyed with Procion dye and that the dye-LPS complex migrated in SDS-PAGE to produce a banding pattern similar to that obtained with nondyed LPS. The migration of the dyed LPS could be visualized without staining. Attempts to duplicate this procedure with C. vinosum LPS were not successful, as the dye formed an insoluble precipitate with the LPS. The reason for this was not determined. Electron microscopic characterization. Electron microscopic examination of negatively stained LPS preparations in water showed the presence of typical filamentous aggregates approximately 6.5 to 8.0 nm in width, with a few short rods and spherical forms (Fig. 6A). (Alkaline, 56°C)-treated and lyophilized, EDTA-treated, and SD-treated LPS were indistinguishable from untreated LPS.

988

INF ECT . IMMUN .

HURLBERT AND HURLBERT

i"

._-

-a

la

A

,

B

-%a

.

C

D

E

F

G

FIG. 4. Electrophoresis of LPS preparations on SDS-polyacrylamide gels containing 10% acrylamide. (A) C. vinosum, 200 jg; (B) C. vinosum, 100 jg; (C) (alkaline, 56°C)-treated C. vinosum, 200 j,g; (D) E. coli B, 200 jig; (E) S. abortus-equi, 200 jg; (F) Thiocapsa, 200 jg; (G) P. mirabilis, 200 jg.

To investigate the morphology of LPS subunits, samples of LPS were dissolved in 0.1 M Tris-hydrochloride buffer, pH 8.0-0.5% SD or SDS or in pure Me2SO. These samples were dialyzed against distilled water or 0.1 M Tris-

hydrochloride buffer, pH 8.0, for 72 h at 8°C. Samples of LPS dissolved in the Me2SO were examined directly or after dialysis. Since SD and SDS can produce artifacts, negatively stained preparations of SD or SDS-Tris-hydro-

LIPOPOLYSACCHARIDE OF CHROMATIUM

VOL. 16, 1977

989

A

B 6,_

I

C

0_

A

B

C

D

D

E

FIG. 5. Electrophoresis of LPS extracted from different regions of SDS-polyacrylamide gels. Twelve gels containing LPS were cut into sections as indicated by the arrows. The LPS was recovered from each section as described in the text, redissolved, and rerun under the original conditions. The amount of eluted LPS placed on each gel was unknown, because non-LPS material also was eluted from the gels and contributed to dry weight. (A) 0 to 2 cm; (B) 2 to 4 cm; (C) 4 to 5.5 cm; (D) 5.5 to 7 cm; (E) control LPS. The dark-staining regions at the bottoms ofthe gels are artifacts due to incomplete removal ofSDS during the washing procedure prior to staining.

chloride solutions, without LPS and with LPS but before dialysis, were prepared and compared with the SD- and SDS-treated LPS samples after dialysis. The structures observed in the samples of SD and SDS alone or with LPS before dialysis were markedly different from those observed in the dialyzed preparations. Thus, it is our belief that the material seen in the negative stains of the dialyzed samples represents LPS. When SD- or SDS-treated LPS was dialyzed

against distilled water, the solution became opalescent, and a small amount of flocculent precipitate appeared; however, Me2SO-treated LPS remained clear under these conditions. Electron microscopic examination of negative stains of the distilled water-dialyzed SD- and SDS-treated LPS showed mostly short rods, approximately 12.0 by 6.0 nm (Fig. 6B and C). The Me2SO-treated LPS appeared as small ovoid spheres, approximately 6.5 nm in diameter (Fig. 6D).

990

HURLBERT AND HURLBERT

INFECT. IMMUN.

m.,.Y,''.Li >,v8kX'

^A

>-sfl4

'^ ~

,

4¢

W

s

?,S.

S

,

-

;¢

C FIG. 6. Electron micrographs showing C. vinosum LPS, untreated (A) or treated with SDS (B), SD (C), or Me2SO (D) and dialyzed against distilled water. Bar = 0.1 um.

When the distilled water-dialyzed samples were frozen at -20°C overnight and thawed, all of them formed heavy flocculent precipitates that microscopically resembled the original

LPS (Fig. 6A). A somewhat different picture was seen when the samples were dialyzed against 0.1 M Tris-hydrochloride buffer, pH 8.0. Under these conditions, the SD- and SDS-

treated LPS solutions developed only a slight cloudiness, whereas the Me2SO-treated LPS remained clear, as before. Freezing and thawing of these samples did not result in the formation of a flocculent precipitate. Negative stains of the LPS samples dialyzed against Tris-hydrochloride had the same microscopic appearance as that of the distilled water-dialyzed LPS (Fig. 6B, C, and D). However, frozen-thawed SD- or SDS-treated LPS that had been dialyzed against Tris-hydrochloride appeared similar and seemed to be partially aggregated, as indicated by the presence of rods of various lengths (Fig. 7). Freezing and thawing did not change the microscopic appearance of the Me2SOtreated material. Mitogenic activity. Since many LPS are known to exhibit mitogenic activity (2, 19, 28), the efficacy of C. vinosum LPS in stimulating [3H]thymidine incorporation into mouse spleen lymphocytes was investigated. C. vinosum LPS was almost as effective a mitogen as E. coli LPS (Table 2). DISCUSSION In a previous paper we reported the isolation and chemical characterization of C. vinosum LPS and showed that it was different from many of the previously studied LPS in that it fractionated in the phenol phase during the hot-

4

991

LIPOPOLYSACCHARIDE OF CHROMATIUM

VOL. 16, 1977

0

i,

TABLE 2. Stimulation of [3H]thymidine incorporation into mouse spleen cultures by C. vinosum and E. coli LPS LPS

LPS (pAg)

Control 10 E. colia 10 C. vinosum a Strain 0114:B4.

[3H]thymidine incorporation (cpm/culture) 388 12,536

7,555

Stimulation index

32 19

phenol extraction procedure (9). However, it was clear from the analytical data that the basic structure of C. vinosum LPS is similar to that of other gram-negative bacteria, but its sugar and fatty acid composition is unique. Further, we showed that C. vinosum LPS is toxic to mice. The present paper shows that C. vinosum LPS is similar in its biological activities and physicochemical characteristics to enteric LPS. The anticomplementary activity of the C. vinosum LPS is within the range found for enteric LPS (4, 5) and suggests that it has a common molecular structure with enteric LPS. The fact that mild alkaline treatment destroys its anticomplementary activity, as it does for the enterics (5, 23), further supports this view. Since this activity has been shown to lie with the lipid A portion of enteric LPS (5), it is assumed that it is the lipid A fraction of C. vinosum LPS that is responsible for its anticomplementary activity. However, proof of this must await further investigation. The stimulation of anticomplementary activity induced by pretreatment of the LPS with Mg2+ is the opposite of that observed for Salmonella minnesota LPS (5), but this may be the result of changes in the configuration of the LPS at low Mg2+ concentrations, which better expose the anticomplementary groups of the LPS. However, at higher Mg2+ (or Ca2+) concentrations the anticomplementary activity is suppressed, as was observed for the R-form LPS of S. minnesota

(5). As with other LPS preparations (4), C. vinosum LPS is an effective anticomplementary agent in a high-molecular-weight soluble state, but it loses activity when it loses solubility upon repeated freezing and thawing. Several investigations on the effect of alkaline digestion of LPS in aqueous solution have shown that such treatment frequently detoxifies it and destroys its anticomplementary activity (5, 22, 23, 38). In addition, it has been FIG. 7. Electron micrograph showing C. vinosum found that alkaline treatment of LPS usually LPS treated with SD and dialyzed against 0.1 M reduces its molecular weight (23). These studies have shown that the rate and extent of these Tris-hydrochloride buffer, pH 7.8. Bar 0.1 ,um. =

992

HURLBERT AND HURLBERT

effects are influenced by the concentration of the alkali, the temperature, and the solvent used. The LPS of C. vinosum appears to be typical, in that mild alkaline digestion completely detoxifies it and greatly diminishes its anticomplementary activity. The effect of different alkali treatments on the molecular weight of C. vinosum LPS is marked. At 56°C, the LPS is clearly disaggregated to lower-molecular-weight forms, which can be partially reaggregated by lyophilization. The aggregating effect of lyophilization on dispersed LPS has previously been noted (4). However, if the LPS is alkaline digested at 100°C under the same conditions, it is converted to a molecularweight species of one-half the sedimentation value of the 56°C-treated LPS, and lyophilization has much less effect on its size. If it is assumed that the alkaline treatment acts primarily to hydrolyze the ester-linked fatty acids of the lipid A portion of the LPS (5, 23), then it follows that these molecules play a crucial role in the formation of highly aggregated LPS. Treatment of the C. vinosum LPS with detergents or Me2SO results in a reversible dissociation into subunits similar to that observed for E. coli LPS (29) and for other gram-negative bacteria (37). From the size of the subunits seen in the electron microscope, Me2SO appears to be the most effective dispersing agent; however, the sedimentation coefficient data suggest that all the dispersing agents used produce approximately the same size of subunit. Since it is not possible to examine LPS microscopically in the presence of SD or SDS, this could not be verified. The size of the subunits (-6.5 by 12.0 nm) that are seen in the SD- and SDS-treated and dialyzed preparations suggests that they are dimers of the Me2SO subunits. It is not clear why there is more aggregation in SD- or SDStreated and dialyzed LPS than with Me2SOtreated LPS. It may be that since Me2SO is a smaller molecule than either SD or SDS, it dialyzes away too rapidly to allow reaggregation. The reason for the range in apparent sedimentation coefficients of the LPS in the disaggregating agents is probably the result of different degrees of association between the LPS and the agents employed. Olins and Warner (26) reported that SDS binds tightly to the LPS of Azotobacter vinelandii. The extent of reaggregation upon dialysis of the disaggregated LPS is markedly dependent upon both the dialysis solution employed and the subsequent treatment of the material, i.e., freezing and thawing or lyophilization. The increase in apparent sed-

INFECT. IMMUN.

imentation coefficient of SD-treated LPS dialyzed against Tris-hydrochloride buffer is much less than that observed by Ribi et al. (29) for similarly treated E. coli LPS but is similar to the response of Rhodopseudomanas capsulata LPS (40). Because of the heterogeneity of LPS in other characteristics, it is not surprising to observe a range in reaggregation response in different LPS treated in a similar fashion. The microscopic appearance of untreated as well as SD- and SDS-treated and dialyzed C. vinosum LPS is similar to that reported for E. coli LPS (29). The effect of freezing and thawing and lyophilization on inducing aggregation of water-dialyzed LPS is probably the result of the LPS being forced out of solution and concentrated during ice formation. Olins and Warner (26) observed a similar response of Azotobacter LPS, and Ribi et al. (29) reported that SDtreated Salmonella enteritidis LPS subunits are reaggregated during reextraction with phenol. Galanos and Luderitz (4) reported that lyophilization converted a 10S species of S. abortus-equi LPS to a 110S form. The reason that LPS can be induced to form these large aggregates is unknown, but it has been suggested by Milner et al. (17, 18) that they are stable continuous or fringe micelles. The stability of these aggregates in aqueous environments and the requirement for detergents or similar substances to disperse them, as well as the observation that (alkaline, 100°C)-treated samples do not reaggregate, are all consistent with the assumption that hydrophobic regions are involved. The complex banding pattern of several LPS in SDS-PAGE has not been previously reported, but there have been numerous reports of heterogeneity within the LPS isolated from one organism (10, 24, 35). The LPS of Serratia marcescens has been fractionated on the basis of size (27) and charge (25), and the fractions have been shown to differ in a number of physical and biological effects (27). Five fractions of Brucella melitensis LPS, differing in their sugar compositions, have been obtained by diethylaminoethyl chromatography (14). LPS fractions with different sizes have been isolated from Enterobacter (13) and E. coli (20). In one report (1), E. coli LPS was fractionated by sucrose density gradient into large and small particles, with a continuous distribution of size in between. The nature of this heterogeneity is unclear, but the fact that we have obtained the same banding pattern on SDS-PAGE of C. vinosum outer membranes purified by differential centrifugation in distilled water makes it clear that the multiple bands are not a product

VOL. 16, 1977

LIPOPOLYSACCHARIDE OF CHROMATIUM

993

of the phenol extraction procedure. Rather, the have to be undertaken to determine if this is ordered pattern of the fine bands appears to be the case for C. vinosum LPS. due to a self-associating polymer series. Such ACKNOWLEDGMENTS series have been shown to be characteristic for We thank the staff of the Washington State University a number of polymeric proteins that, under the Microscope Center and S. Gurusiddaish of the appropriate experimental conditions, arrange Electron Lab for their technical assistance. themselves in a series of integral multiples of Analytical This investigation was supported in part by funds prothe basic monomeric unit (33, 39). The patterns vided for biological and medical research by the State of of the successive bands in such a series are Washington Initiative Measure no. 171. related by the geometric constant g = l+lp,u/DI,u LITERATURE CITED (,u = electrophoretic mobility), and from this it 1. Beer, H., T. Staehelin, H. Douglas, and A. I. Braude. follows that a plot of the log of the mobility 1965. Relationship between particle size and biologiversus the band number produces a straight cal activity of E. coli boivin endotoxin. J. Clin. Inline (33, 39). In the case of C. vinosum LPS, the vest. 44:592-602. geometric constant, calculated from the gel 2. Chiller, J. M., B. J. Skidmore, D. C. Morrison, and W. 0. Weigle. 1973. Relationship of the structure of bacshown in Fig. 5E, was found to be 0.9601, with a terial lipopolysaccharides to its function in mitogenestandard deviation of +0.0096, and a plot of the sis and adjuvanticity. Proc. Natl. Acad. Sci. U.S.A. log mobility versus band number produces a 70:2129-2133. straight line (Fig. 8). Therefore, the data are 3. Cluff, L. E. 1971. Effects of lipopolysaccharides (endotoxins) on susceptibility to infections, p. 399-413. In consistent with the conclusion that C. vinosum S. Kadis, G. Weinbaum, and S. J. Ajl (ed.), Microbial LPS, as well as LPS from other bacteria, is toxins: a comprehensive treatise, vol. 5, Bacterial capable of forming a homologous series of oligoendotoxins. Academic Press Inc., New York. mers in detergent gels. The structure of the 4. Galanos, C., and 0. Luderitz. 1976. The role of the physical state of lipopolysaccharides in the interacvarious oligomers is unknown, but the fact that tion with complement: high molecular weight as pre(alkaline, 100°C) digestion destroys the ability requisite for the expression of anti-complementary of the LPS to form this series suggests that activity. Eur. J. Biochem. 65:403-408. fatty acids are crucial to the formation of these 5. Galanos, C., E. T. Rietschel, 0. Luderitz, and 0. Westphal. 1971. Interaction of lipopolysaccharides and aggregation states. If further investigation of lipid A with complement. Eur. J. Biochem. 19:143this phenomenon verifies that it follows the 152. same rules as protein systems, it may be possi- 6. Gewurz, H., R. Snyderman, S. E. Mergenhagen, and H. S. Shin. 1971. Effects of endotoxin lipopolysacchable to use this procedure as a simple and rapid rides on the complement system, p. 127-149. In S. means of estimating LPS molecular weight Kadis, G. Weinbaum, and S. J. Ajl (ed.), Microbial

(33). The preliminary data indicate that the LPS of C. vinosum is almost as active a mitogen as E. coli LPS. Other LPS have been shown to be B-cell mitogens (19, 28), but further studies will 1*0 z

0-9

4c

ID 0*8-

t-

0-

F

w

-I-

0 6-

0

,-j

o-4F 0

- -I

_

I

4

8

I

I

12

16

20

24

BAND NUMBER

FIG. 8. Plot of mobility offine bands against band number. Data taken from Fig. 5E.

toxins: a comprehensive treatise, vol. 5, Bacterial endotoxins. Academic Press Inc., New York. 7. Hinshaw, L. B. 1971. Release of vasoactive agents and the vascular effects of endotoxin, p. 209-275. In S. Kadis, G. Weinbaum, and S. J. Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 5, Bacterial endotoxins. Academic Press Inc., New York. 8. Hurlbert, R. E., J. R. Golecki, and G. Drews. 1974. Isolation and characterization of Chromatium vinosum membranes. Arch. Microbiol. 101:169-185. 9. Hurlbert, R. E., J. Weckesser, H. Mayer, and J. Fromme. 1976. Isolation and characterization of the lipopolysaccharide of Chromatium vinosum. Eur. J. Biochem. 68:365-371. 10. Jann, B., K. Reske, and K. Jann. 1975. Heterogeneity of lipopolysaccharides. Analysis of polysaccharide chain lengths by sodium dodecylsulfate-polyacrylamide gel electrophoresis. Eur. J. Biochem. 60:239-246. 11. Johnson, R. G., M. B. Perry, I. J. McDonald, and R. R. B. Russell. 1975. Cellular and free lipopolysaccharides of some species of Neisseria. Can. J. Microbiol. 21:1969-1980. 12. King, J., and U. K. Laemmli. 1971. Polypeptides of the tail fibres of bacteriophage T4. J. Mol. Biol. 62:465477. 13. Koeltzow, D. E., and H. E. Conrad. 1971. Structural heterogeneity in the lipopolysaccharide of Aerobacter aerogenes NCTC 243. Biochemistry 10:214-224. 14. Lacave, C., J. Asselineau, A. Serre, and J. Roux. 1969. Comparaison de la composition chimique d'une fraction lipopolysaccharidique et d'une fraction polysaccharidique isolees de Brucella melitensis. Eur. J. Biochem. 9:189-198.

994

HURLBERT AND HURLBERT

15. Luderitz, O., C. Galanos, V. Lehmann, M. Nurminen, E. T. Rietschel, G. Rosenfelder, M. Siman, and 0. Westphal. 1973. Lipid A: chemical structure and biological activity. J. Infect. Dis. 128:517-629. 16. Luderitz, O., 0. Westphal, A. M. Staub, and H. Nikaido. 1971. Isolation and chemical and immunological characterization of bacterial lipopolysaccharides, p. 145-233. In G. Weinbaum, S. Kadis, and S. J. Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 4, Bacterial endotoxins. Academic Press Inc., New York. 17. Milner, K. C., R. L. Anacker, K. Fukushi, W. T. Haskins, M. Landy, B. Malmgren, and E. Ribi. 1963. Symposium on relationship of structure of microorganisms to their immunological properties. III. Structure and biological properties of surface antigens from gram-negative bacteria. Bacteriol. Rev. 27:352368. 18. Milner, K. C., J. A. Rudbach, and E. Ribi. 1971. General characteristics, p. 1-65. In G. Weinbaum, S. Kadis, and S. J. Ajl (ed.), Microbial toxins; a comprehensive treatise, vol. 4, Bacterial endotoxins. Academic Press Inc., New York. 19. Moller, G., 0. Sjoberg, and J. Anderson. 1973. Immunogenicity, tolerogenicity and mitogenicity of lipopolysaccharides. J. Infect. Dis. 128:552-556. 20. Morrison, D. C., and L. Leive. 1975. Fractions of lipopolysaccharide from Escherichia coli 0111:B4 prepared by two extraction procedures. J. Biol. Chem. 250:2911-2919. 21. Nagler, A. L., and S. M. Levenson. 1971. Experimental hemorrhagic and endotoxic shock, p. 341-398. In S. Kadis, G. Weinbaum, and S. J. Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 5, Bacterial endotoxins. Academic Press Inc., New York. 22. Neter, E., 0. Westphal, 0. Luderitz, E. A. Gorzynski, and E. Eichenberger. 1956. Studies ofenterobacterial lipopolysaccharides. Effects of heat and chemicals on erythrocyte-modifying, antigenic, toxic and pyrogenic properties. J. Immunol. 76:377-385. 23. Niwa, M., K. C. Milner, E. Ribi, and J. A. Rudbach. 1969. Alteration of physical, chemical, and biological properties of endotoxin by treatment with mild alkali. J. Bacteriol. 97:1069-1077. 24. Nowotny, A. 1971. Chemical and biological heterogeneity of endotoxins, p. 309-329. In G. Weinbaum, S. Kadis, and S. J. Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 4, Bacterial endotoxins. Academic Press Inc., New York. 25. Nowotny, A., K. R. Cundy, N. L. Neale, A. M. Nowotny, S. P. Thomas, and D. J. Tripodi. 1966. Relation of structure to function in bacterial 0-antigens. IV. Fractionation of the components. Ann. N.Y. Acad. Sci. 133:586-603. 26. Olins, A. L., and R. C. Warner. 1967. Physiochemical studies on lipopolysaccharide from the cell wall of Azotobacter vinelandii. J. Biol. Chem. 242:4994-5001. 27. Oroszlan, S. I., and P. T. Mora. 1963. Dissociation and reconstitution of an endotoxin. Biochem. Biophys. Res. Commun. 12:345-349.

INEECT . IMMUN . 28. Peavy, D. L., J. W. Shands, W. H. Adler, and R. T. Smith. 1973. Selective effects of bacterial endotoxins on various subpopulations of lymphoreticular cells. J. Infect. Dis. 128:591-599. 29. Ribi, E., R. L. Anacker, R. Brown, W. T. Haskins, B. Malmgren, K. C. Milner, and J. A. Rudbach. 1966. Reaction of endotoxin and surfactants. I. Physical and biological properties of endotoxin treated with sodium deoxycholate. J. Bacteriol. 92:1493-1509. 30. Roantree, R. J. 1971. The relationship of lipopolysaccharide structure to bacterial virulence, p. 1-37. In S. Kadis, G. Weinbaum, and S. J. Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 5, Bacterial endotoxins. Academic Press Inc., New York. 31. Russell, R. R. B., and K. G. Johnson. 1975. SDS-polyacrylamide gel electrophoresis of lipopolysaccharides. Can. J. Microbiol. 21:2013-2018. 32. Russell, R. R. B., K. G. Johnson, and I. J. McDonald. 1975. Envelope proteins in Neisseria. Can. J. Microbiol. 21:1519-1534. 33. Scurzi, W., and W. N. Fishbein. 1973. The geometric mobility sequence in polyacrylamide gel electrophoresis of polymeric proteins: its significance for polymer geometry and electrophoretic theory. Trans. N.Y. Acad. Sci. 35:396-416. 34. Segrest, J. P., and R. L. Jackson. 1972. Molecular weight determination of glycoproteins by polyacrylamide gel electrophoresis in sodium dodecylsulfate. Methods Enzymol. 28:54-63. 35. Shands, J. W. 1971. The physical structure of bacterial lipopolysaccharides, p. 127-144. In G. Weinbaum, S. Kadis, and S. J Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 4, Bacterial endotoxins. Academic Press Inc., New York. 36. Snell, E. S. 1971. Endotoxin and the pathogenesis of fever, p. 277-340. In S. Kadis, G. Weinbaum, and S. J. Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 5, Bacterial endotoxins. Academic Press Inc., New York. 37. Sultzer, B. M. 1971. Chemical modification of endotoxin and inactivation of its biological properties, p. 91-126. In S. Kadis, G. Weinbaum, and S. J. Ajl (ed.), Microbial toxins: a comprehensive treatise, vol. 5, Bacterial endotoxins. Academic Press Inc., New York. 38. Tripodi, D., and A. Nowotny. 1966. Relation of structure to function in bacterial 0-antigens. V. Nature of active sites in endotoxic lipopolysaccharides ofSerratia marcescens. Ann. N.Y. Acad. Sci. 133:604-621. 39. Ugel, A. R. 1971. Fractionation and characterization of an oligomeric series of bovine keratohyalin by polyacrylamide gel electrophoresis. Anal. Biochem. 43:410-426. 40. Weckesser, J., G. Drews, and R. Ladwig. 1972. Localization and biological and physicochemical properties of the cell wall lipopolysaccharide of Rhodopseudomonas capsulata. J. Bacteriol. 110:346-353. 41. Westphal, O., 0. Luderitz, and F. Bister. 1952. Uber die Extraktion von Bakterien mit Phenol/Wasser. Z. Naturforsch. Teil B 7:148-155.