Phenol-extracted Lipopolysaccharide Research Articles

Phenotypic and functional analysis of monocyte populations in cattle peripheral blood identifies a subset with high endocytic and allogeneic T-cell stimulatory capacity.

Circulating monocytes in several mammalian species can be subdivided into functionally distinct subpopulations based on differential expression of surface molecules. We confirm that bovine monocytes express CD172a and MHC class II with two distinct populations of CD14+CD16low/-CD163+ and CD14−CD16++CD163low- cells, and a more diffuse population of CD14+CD16+CD163+ cells. In contrast, ovine monocytes consisted of only a major CD14+CD16+ subset and a very low percentage of CD14−CD16++cells. The bovine subsets expressed similar levels of CD80, CD40 and CD11c molecules and mRNA encoding CD115. However, further mRNA analyses revealed that the CD14−CD16++ monocytes were CX3CR1highCCR2low whereas the major CD14+ subset was CX3CR1lowCCR2high. The former were positive for CD1b and had lower levels of CD11b and CD86 than the CD14+ monocytes. The more diffuse CD14+CD16+ population generally expressed intermediate levels of these molecules. All three populations responded to stimulation with phenol-extracted lipopolysaccharide (LPS) by producing interleukin (IL)-1β, with the CD16++ subset expressing higher levels of IL-12 and lower levels of IL-10. The CD14−CD16++ cells were more endocytic and induced greater allogeneic T cell responses compared to the other monocyte populations. Taken together the data show both similarities and differences between the classical, intermediate and non-classical definitions of monocytes as described for other mammalian species, with additional potential subpopulations. Further functional analyses of these monocyte populations may help explain inter-animal and inter-species variations to infection, inflammation and vaccination in ruminant livestock.

Peptidoglycan, not endotoxin, is the key mediator of cytokine gene expression induced in rainbow trout macrophages by crude LPS

In rainbow trout macrophages, phenol-extracted lipopolysaccharide (LPS) preparations stimulate proinflammatory cytokine gene expression but ultrapure preparations of LPS are inactive. Crude LPS preparations could potentially have a number of contaminants including peptidoglycans (PGNs), nucleic acids and lipoproteins. Thus, in the current study we individually tested potentially contaminating pathogen associated molecular patterns (PAMPs) on rainbow trout ( Oncorhynchus mykiss) macrophages to determine which ones could induce proinflammatory cytokine expression. We found that PGNs derived from Gram-negative bacteria ( Escherichia coli 0111:B4 and K12), are potent inducers of IL-1β and IL-6 gene expression and were equal to, or more potent than, crude LPS. On the other hand, PGNs of Gram-positive bacteria, DNA, RNA and lipoteichoic acid were weak stimulators, and lipid A, lipoprotein (Pam3CSK4) and ultrapure LPS were nonstimulatory. More importantly, crude LPS treated with lysozyme to degrade PGNs, exhibited greatly reduced activity in stimulating IL-1β and IL-6 gene expression, indicating that PGNs in the crude LPS are responsible for a significant amount of the proinflammatory activity. Finally, we showed that PGN treatment induces expression of COX-2 and the subsequent synthesis and release of prostaglandin E 2 (PGE 2), an important mediator of inflammatory processes. The strong stimulatory effect of E. coli PGNs by themselves on trout macrophages suggests that the recognition of Gram-negative bacteria in trout is through PGNs in the bacterial wall, and indicates that the systems responsible for bacterial recognition in invertebrates (e.g., Drosophila) may also be conserved in some vertebrates.

Evidence for different requirements in physical state for the interaction of lipopolysaccharides with the classical and alternative pathways of complement.

The influence of the state of aggregation of lipopolysaccharides upon their ability to interact with serum complement via either the classical or alternative pathway was studied. The anticomplement properties of two chromatographically distinct fractions of a phenol-extracted lipopolysaccharide isolated from Serratia marcescens were assessed by means of the standard sheep erythrocyte hemolytic assay and an alternative pathway-selective kinetic assay using rabbit erythrocytes. Both the high molecular weight PI fraction and the lower molecular weight PII fraction exerted anti-complement activity as determined in the sheep erythrocyte assay. Conversion of fractions PI and PII to their more soluble triethylamine salt forms resulted in a decrease in sedimentation coefficients and a corresponding loss of anticomplement activity. Further, the anticomplement activity of fractions PI and PII in the sheep erythrocyte assay was inhibited by polymyxin B, indicating a role for the lipid A region. Unlike the PII fraction, only the PI fraction can activate serum complement via the alternative pathway. This activity is not inhibited by polymyxin B, indicating that the response is not lipid A-mediated. Significantly, solubilization of the PI fraction with triethylamine had no effect on its ability to activate the alternative pathway. These studies clearly demonstrate that the interaction between lipopolysaccharides and serum complement is influenced by the state of lipopolysaccharide aggregation. However, this appears to be the case for lipopolysaccharide activation of the classical pathway but not of the alternative pathway.

Lipid A and O-chain modifications cause Rhizobium lipopolysaccharides to become hydrophobic during bacteroid development.

Modifications to the lipopolysaccharide (LPS) structure caused by three different growth conditions were investigated in the pea-nodulating strain Rhizobium leguminosarum 3841. The LPSs extracted by hot phenol-water from cultured cells fractionated into hydrophilic water and/or hydrophobic phenol phases. Most of the LPSs from cells grown under standard conditions extracted into the water phase, but a greater proportion of LPSs were extracted into the phenol phase from cells grown under acidic or reduced-oxygen conditions, or when isolated from root nodules as bacteroids. Compared with the water-extracted LPSs, the phenol-extracted LPSs contained greater degrees of glycosyl methylation and O-acetylation, increased levels of xylose, glucose and mannose and increased amounts of long-chain fatty acids attached to the lipid A moiety. The water- and phenol-phase LPSs also differed in their reactivity with monoclonal antibodies and in their polyacrylamide gel electrophoretic banding patterns. Phenol-extracted LPSs from rhizobia grown under reduced-oxygen conditions closely resembled the bulk of LPSs isolated from pea nodule bacteria (i.e. mainly bacteroids) in their chemical properties, reactivities with monoclonal antibodies and extraction behaviour. This finding suggests that, during symbiotic bacteroid development, reduced oxygen tension induces structural modifications in LPSs that cause a switch from predominantly hydrophilic to predominantly hydrophobic molecular forms. Increased hydrophobicity of LPSs was also positively correlated with an increase in the surface hydrophobicity of whole cells, as shown by the high degree of adhesion to hydrocarbons of bacterial cells isolated from nodules or from cultures grown under low-oxygen conditions. The implications of these LPS modifications are discussed for rhizobial survival and function in different soil and in planta habitats.

Tumor necrosis factor induction by an aqueous phenol-extracted lipopolysaccharide complex from Bacteroides species

The stimulation of macrophages and monocytes by lipopolysaccharide (LPS) results in the secretion of tumor necrosis factor (TNF), a cytokine which is thought to play a pivotal role in subsequent host responses. Its induction is thought to be facilitated by the binding of complexes of LPS and LPS-binding protein to CD14. The LPS of Bacteroides species was considered a weak endotoxin; however, in a recent study we have shown that the biological activity and chemical composition of the LPS from Bacteroides species are dependent on the extraction method. The present study determines the capacity of LPS extracted by aqueous phenol (the method for producing an LPS of high endotoxic activity) from four species of Bacteroides to induce TNF. Induction was investigated from human mononuclear leukocytes (MNL), THP-1 cells (with and without enhancement by vitamin D2 for CD14), and peritoneal macrophages from C3H/HeJ (LPS nonresponder) and C3H/HeN (LPS responder) mice. Escherichia coli O18K- LPS, a typical smooth LPS of heterogeneous molecular mass, was used as a control throughout. The stimulation of TNF production by E. coli LPS was between two- and fourfold more than that by Bacteroides LPS in MNL, in THP-1 cells (with enhancement for CD14), and in peritoneal macrophages from C3H/HeN mice. In THP-1 cells (without enhancement for CD14), there was no significant difference in TNF production between E. coli and Bacteroides LPSs. In peritoneal macrophages from C3H/HeJ mice, E. coli LPS stimulated no TNF production, but there was no significant difference in TNF production from peritoneal macrophages from C3H/HeJ and C3H/HeN mice by Bacteroides LPS. In all cell populations, there was a peak of TNF production after approximately 4 h of stimulation with all LPSs tested. However, other peaks of TNF production were seen in MNL and THP-1 cells (with enhancement for CD14) after stimulation with E. coli LPS only. In stimulation assays in which Bacteroides LPS was together with but in excess of E. coli LPS, it was found that TNF production from MNL and THP-1 cells (with and without enhancement for CD14) was comparable to that of Bacteroides LPS alone and not E. coli LPS alone. An anti-CD14 monoclonal antibody did not inhibit Bacteroides LPS-stimulated TNF production. However, E. coli LPS-stimulated TNF release was inhibited by an anti-CD14 monoclonal antibody, most noticeably in MNL and THP-1 cells (with enhancement for CD14).(ABSTRACT TRUNCATED AT 400 WORDS)

Effects of natural or synthetic microbial adjuvants on induction of autoimmune thyroiditis.

Natural and synthetic adjuvants of microbial origin were compared for their capacity to potentiate the induction of experimental autoimmune thyroiditis (EAT) with the autoantigen mouse thyroglobulin (MTg). Regardless of the immunomodulator used, severe thyroiditis was observed only in EAT-susceptible strains of the k haplotype and not in EAT-resistant strains of the d haplotype. Compared to phenol-extracted lipopolysaccharide, a potent adjuvant for enhancing EAT induction, phthalyl-substituted, detoxified lipopolysaccharide, even at doses 15- to 50-fold greater, led to only low anti-mouse thyroglobulin titers and mild thyroid infiltration. The synthetic adjuvant N-acetylmuramyl-L-alanyl-D-isoglutamine (MDP) and three of its analogs, N-acetylmuramyl-L-alanyl-D-isoglutamine-L-alanyl-D-glycerol mycolate (MDP-L-Ala-Glyc-Myc), N-acetylmuramyl-L-alanyl-D-glutamyl-(decyl)methyl ester [MDP(decyl)methyl], and N-acetylmuramyl-L-alanyl-D-glutamine-alpha n-butyl ester [MDP-(Gln)-OnBu], designated murabutide, were tested in incomplete Freund adjuvant or in saline. In incomplete Freund adjuvant, MDP-L-Ala-Glyc-Myc was inefficient in inducing EAT, murabutide induced very mild involvement, and MDP and, more so, MDP(decyl)methyl were active but to a lesser degree than CFA. When saline was used, low levels of thyroid infiltration were observed in a few of the MDP-treated animals in only one experiment, whereas no lesions were observed when murabutide was used.

Differentiation between actinobacillus actinomycetemcomitans and haemophilus aphrophilus based on carbohydrates in lipopolysaccharide

In the present study, the closely related facultative, Gram-negative rods, Actinobacillus actinomycetemcomitans and Haemophilus aphrophilus, were distinguished taxonomically by means of their carbohydrate composition in phenol-extracted lipopolysaccharide. Both A. actinomycetemcomitans and H. aphrophilus lipopolysaccharide contained rhamnose, fucose, galactose, glucose, L-glycero- D-mannoheptose, galactosamine, and glucosamine. The content of galactose was approximately twice as high in lipopolysaccharide from H. aphrophilus as in lipopolysaccharide from A. actinomycetemcomitans. D-Glycero- D-mannoheptose was detected exclusively in lipopolysaccharide from A. actinomycetemcomitans where it constituted 11.8–16.7% of the sugar content. This aldoheptose may therefore serve as a marker for chemotaxonomic differentiation between A. actinomycetemcomitans and H. aphrophilus. The present study also describes fragmentation of methylheptoside derivatives of trifluoroacetic acid ( D-glycero- and L-glycero- D-mannoheptose) from A. actinomycetemcomitans as suggested by mass spectrometry.

Functional heterogeneity of the Lyb-5- B cell subpopulation: mutant xid B cells and normal Lyb-5- B cells differ in their responsiveness to phenol-extracted lipopolysaccharide.

In the present study, responses stimulated by phenol-extracted lipopolysaccharide (LPS(phenol)) and butanol-extracted LPS (LPS(butanol)) were used to assess the possibility that xid B cells might not be identical to the Lyb-5- B cells present in normal mice. It was found that xid B cells responded well only to LPS(butanol) whereas normal B cells responded well to both LPS(butanol) and LPS(phenol). Thus, LPS(butanol) appeared to be a TI-1 antigen and LPS(phenol) appeared to be a TI-2 antigen. In contrast to classical TI-2 responses, however, responses stimulated by LPS(phenol) did not exhibit a stringent requirement for accessory cells. Furthermore, if LPS(phenol) were a classical TI-2 antigen, it should only activate Lyb-5+ B cells. To determine if the responsiveness of normal B cells to LPS(phenol) were due, at least in part, to the stimulation of normal Lyb-5- B cells, the responsiveness of normal neonatal B cells and normal adult B cells that had been pretreated with anti-Lyb-5.1 + C was assessed. It was found that both normal neonatal B cells and normal adult Lyb-5- B cells did respond well to LPS(phenol). Thus, even though LPS(phenol) does not stimulate xid B cells, these data demonstrate that LPS(phenol) is different from other TI-2 antigens. More importantly, these data also demonstrate that xid B cells and normal Lyb-5- B cells are not identical. It is hypothesized that the normal Lyb-5- B cell subpopulation is heterogeneous, consisting of an Lyb-5(1)- and an Lyb-5(2)-B cell subset with the xid mutation blocking the differentiation of Lyb-5(1)-B cells into Lyb-5(2)-B cells.

A pomona serogroup-specific, agglutinating antigen in leptospira, identified by monoclonal antibodies

Monoclonal antibodies were produced by hybridoma cell lines derived by fusion of mouse NS-1 myeloma cells with splenocytes from mice immunized with Leptospira interrogans serovar pomona. One hybridoma (A3) produced an IgG2a antibody which agglutinated all leptospires of the Pomona serogroup but not leptospires representative of serovars of any other serogroup. The antibody precipitated in immunodiffusions with either alkali- or phenol-extracted lipopolysaccharide and also with the TM antigen of Yanagawa et al., indicating a common determinant on both antigens. A3 antibody opsonized viable leptospires for phagocytosis by mouse macrophages in vitro.

Complement proteins and macrophages. II. The secretion of factor B by lipopolysaccharide-stimulated macrophages.

The secretion of factor B by mouse peritoneal macrophages was found to be enhanced following in vivo or in vitro stimulation with lipopolysaccharide (LPS). The intravenous administration of LPS to mice of various strains caused an increased release of factor B but not the release of acid phosphatase by the peritoneal macrophages obtained from the stimulated mice. In vitro stimulation of cultured macrophages with LPS resulted in an enhanced secretion of both factor B and acid phosphatase. The dose-dependent augmentation of factor B secretion by LPS was found in the macrophages from LPS-responsive C3H/HeN mice, whereas the macrophages from LPS-unresponsive C3H/HeJ mice did not respond to either phenol-extracted LPS or butanol-extracted LPS. The ability of LPS to cause the enhancement of factor B secretion by macrophages was abolished by alkali or acid treatment of LPS, indicating that its lipid A part was responsible for the observed effect.

Gentamicin-lipopolysaccharide interactions inPseudomonas aeruginosa

Gentamicin was found to bind toPseudomonas aeruginosa lipopolysaccharide. The interaction was partially ionic in nature and probably involved the core region of the lipopolysaccharide molecule. A comparison of phenol-extracted lipopolysaccharide derived fromP. aeruginosa ATCC 9027 and a lipopolysaccharide “core”-defective strain indicated that changes in this region of the molecule were responsible for gentamicin-biding differences. Lysozyme and magnesium ion were both capable of disrupting this interaction. The nature of the interaction appeared to be related to the amount of organic phosphate associated with the lipopolysaccharide and may play a role in the barrier function of the cell envelope.

Lymphoid procoagulant activity and mitogenesis in the C3H/HeJ mouse: discordant response to lipopolysaccharide stimulation.

Abstract Lymphoid cells from the C3H/HeJ mouse respond abnormally to bacterial lipopolysaccharide (LPS) in a number of in vitro functions. Abnormalities in both macrophage and lymphocyte populations have been suggested. In the present study, splenic mononuclear cells were isolated from C3H/HeJ, C3H/St, A/J, and BALB/c mice, and analyzed for mitogenesis and induction of procoagulant activity (PCA) by LPS prepared by either phenol or butanol extraction methods. All strains responded vigorously with resultant increased 3H thymidine uptake to Con A and PHA, and all but the C3H/HeJ responded well to butanol or phenol extracted LPS. However, the C3H/HeJ was unresponsive to phenol-extracted LPS and exhibited a poor but significant response to butanol-extracted LPS. In contrast, lymphoid cells from all strains, including the C3H/HeJ, exhibited a marked and equal increase of PCA in response to both phenol- and butanol-extracted LPS by using both splenic and peripheral blood lymphoid cells. The results obtained indicate that although the C3H/HeJ mouse may be unresponsive to a number of LPS-stimulated functions, there is no associated defect in PCA generation, a lymphocyte collaboration-dependent event.

Immunochemical and biological properties of the outer membrane-associated lipopolysaccharide and protein of Rochalimaea quintana.

The outer membrane complex of Rochalimaea quintana was isolated by use of ethylene-eiaminetetraacetate and was compared biochemically and biologically both with lipopolysaccharide (LPS) isolated by phenol-water extraction of whole organisms and with lipids isolated by chloroform-methanol extractions of the phenol-water insoluble residues. The outer membrane consisted of protein and LPS components, as distinguished by precipitin tests with sera from patients with trench fever or tests with hyperimmune animal sera. The outer membrane protein component, but not LPS, also reacted with sera from infections with Rickettsia tsutsugamushi. The LPS contained 2-keto-3-deoxy-octonate and heptose. The outer membrane and phenol-extracted LPS were reactive in the chick embryo lethality test, limulus assay, and complement activation. Outer-membrane activity was confined to the LPS component. The lipid extracts were reactive in chick embryo lethality and limulus assays, but did not activate complement.

BCG-induced enhancement of endotoxin sensitivity in C3H/HeJ mice. I. In vivo studies.

C3H/HeJ mice exhibit a marked insensitivity to bacterial lipopolysaccharide (LPS) in vivo. Pretreatment of these mice with viable BCG organisms 11 days before LPS administration renders them sensitive to the lethal effects of a highly purified, phenol-extracted LPS. Other in vivo responses to LPS are increased in BCG-infected C3H/HeJ mice in parallel with enhanced lethality. These include 1) the elevation of serum interferon, 2) the production of the acute phase reactant, serum amyloid A (SAA), and 3) hypoglycemia. However, BCG infection has only a minimal effect on anti-LPS antibody production. BCG-infected C3H/HeJ mice approach the LPS sensitivity of normal C3H/HeN mice, but the enhanced LPS sensitivity is transient and decreases over a 2-month period. The ability of BCG to induce LPS sensitivity in C3H/HeJ mice demonstrates that LPS unresponsiveness is not due to an absolute defect in this strain, but rather, a partially reversible state of hyporesponsiveness. In addition, these findings, in conjunction with other observations, suggest that the enhancement of LPS sensitivity induced by BCG infection is mediated primarily through an effect on T cells and/or macrophages rather than B lymphocytes.

Influence of lipopolysaccharide on graft versus host reactivity of lipopolysaccharide-unresponsive C3H/HeJ mice.

It was initially reported that lipopolysaccharide (LPS)-unresponsive C3H/HeJ mice are refractory to LPS at the B-lymphocyte level, but more recently it has been shown that other cells are similarly unaffected. The current study was undertaken to study an in vivo LPS-modulated disease process involving macrophage-T cell interactions. Adult CBA/J and C3H/HeJ mice were used as spleen donors, and graft versus host reactions were induced in BALB/c neonates. Prior LPS treatment of CBA/J adults decreased the ability of their spleen cells to cause fatal graft versus host disease in BALB/c neonates, whereas no difference was found between injection of spleen cells from normal or LPS-treated C3H/HeJ mice. Similar results were obtained with these cell types when the mouse spleen mixed leukocyte culture system was used. In a carbon clearance assay for stimulation of the reticuloendothelial system with LPS, it was found that the rate of phagocytosis was significantly increased in BALB/c and CBA/J mice 72 h after inoculation of LPS. No stimulation was seen in rate of carbon uptake in the C3H/HeJ animals after treatment with phenol-extracted LPS or with butanol-extracted LPS. An LPS-induced protective serum factor was produced only in the LPS-responsive CBA/J mice and was specific for the syngeneic cells.

Physicochemical and in situ observations on the adhesion of gliding bacteria to surfaces

The ability of Flexibacter BH3 to adhere to solid surfaces and to overcome the horizontal drag involved in gliding across the surfaces was considered in terms of the Stefan adhesion principle. The extracellular slime produced by Flexibacter BH3 was suitable as a Stefan adhesive because it exhibited viscous properties characteristic of a linear colloid, increasing the adhesiveness of the bacterium but allowing translational motion across the surface. The water-soluble slime was a glycoprotein, containing glucose, fucose, galactose and some uronic acid. Vesicles and tubules on the outer surface of Flexibacter BH3 possessed trilaminar membranes, contained 2-keto-3-deoxyoctonate (KDO), and showed identity with phenol-extracted lipopolysaccharide (LPS) in gel-diffusion tests. Sections of Flexibacter BH3 gliding on a gold film overlaying an agar medium reveraled a highly convuluted cell envelope outer membrane, portions of which closely conformed to the microcontours of the gold surface. Possible mechanisms of gliding are discussed in relation to this close association with solid surface features, to the finding that flexibility and spiral motion are not essential for gliding, and to evidence revealing the extrusion of slime in advance of “pathfinder” bacteria.

Biological and physicochemical properties of the lipopolysaccharide of Chromatium vinosum.

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 sulfate-treated 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 release of alkaline phosphatase and of lipopolysaccharide during the growth of rough and smooth strains of Salmonella typhimurium.

A rough strain of Salmonella typhimurium, which has a defect in the "inner core" region of its lipopolysaccharide, was shown to release more alkaline phosphatase into the medium during growth than did an isogenic smooth strain of the same organism. The pattern of enzyme release varied during growth in that early logarithmic phase cells retained alkaline phosphatase in their periplasmic space and at their cell surface while cells in later growth phases shed as much as 6% of their content of this enzyme into the medium. The reaction product localization technique used in this study enables us to assess enzyme production and localization in individual cells of a culture. The release of the enzyme was paralleled by a release of lipopolysaccharide from the cells. The released alkaline phosphatase shows a strong tendency to associate with purified phenol-extracted lipopolysaccharide.

Characteristics of complement-dependent release of phospholipid from Escherichia coli.

The release of (32)P-labeled bacterial phospholipid from a smooth Escherichia coli by serum components depends on complement activated by antibody. Phospholipid release in excess antibody tends to be proportional to the concentration of complement as does the release of other cellular constituents. Phospholipids are not simply stripped off during cell lysis. Whereas 94% of the total phospholipid freed from E. coli by mechanical lysis sediments at centrifugal forces sufficient to sediment molecules of 10(6) molecular weight, similar centrifugation sediments only 50% of the phospholipid released by antibody-complement. In fact, after mechanical lysis more than 50% of the phospholipid sediments at velocities sufficient to bring down cell envelopes. Although the bulk of the bacterial phospholipid is located in the cell envelopes, isolated (32)P-labeled cell envelopes and phenol-extracted lipopolysaccharide fails to release phospholipids in the presence of antibody-complement. Moreover, ethylenediaminetetraacetic acid, which like antibody-complement causes loss of cellular selective permeability and prepares E. coli cell walls for the action of lysozyme, releases only small amounts of phospholipid from E. coli and these are sedimentable. The most likely mechanism of phospholipid release caused by antibody-complement appears to be the activation directly or indirectly of an enzyme which is present only in the intact cells.