Skip to main content
NCBI home page
Immunology logoLink to Immunology
. 2000 Oct;101(2):262–270. doi: 10.1046/j.1365-2567.2000.00107.x

The lipid A region of lipopolysaccharides from Rhizobiaceae activates bone marrow granulocytes from lipopolysaccharide-hyporesponsive C3H/HeJ and C57BL/10ScCr mice

T Pedron *, R Girard *, B Jeyaretnam †, R W Carlson †, R Chaby ‡
PMCID: PMC2327079  PMID: 11012780

Abstract

We established in previous studies that the binding of Salmonella lipopolysaccharide (LPS) to constitutive receptors of low affinity triggers the expression of the inducible LPS-binding molecule CD14 in bone marrow cells (BMC) of C3H/HeOU mice, but not in BMC from C3H/HeJ mice. We show in this study that BMC from C3H/HeJ and C57BL/10ScCr mice do not express CD14 after exposure to LPSs from Salmonella enterica and Bordetella pertussis, but do express this marker when treated with several LPSs from Rhizobiaceae, or their lipid A fragments. This shows that the constitutive LPS receptor in BMC from C3H/HeJ and C57BL/10ScCr mice is fully able to trigger a complete signalling cascade. Results of cross-inhibition of the binding of radiolabelled LPS indicated that active LPSs (from R. species Sin-1 and R. galegae) and inactive LPSs (from S. enterica and B. pertussis) bind to the same site of the constitutive LPS receptor of C3H/HeJ cells. Furthermore, binding of R. species Sin-1 LPS, and signalling induced by this LPS, were both inhibited by pre-exposure of C3H/HeJ cells to B. pertussis lipid A. This correlation between binding and signalling suggests that in C3H/HeJ cells, the constitutive receptor, which recognizes a large panel of LPSs from different origins, appears selectively unable to be activated by some particular LPSs, such as those of Enterobacteria and Bordetella.

INTRODUCTION

The mechanisms by which host cells respond promptly to trace amounts of bacterial lipopolysaccharide (LPS) have been extensively investigated by a number of laboratories. However, major gaps remain in our knowledge of the different steps involved in these responses, including the LPS-recognition systems, the signal transduction elements and the transcription steps. Animals carrying a genetic defect in one of these steps should thus represent efficient tools in dissecting the mechanism involved at that level. Up to now, two mutant mouse strains have been reported to exhibit abnormal responses to LPS: the C3H/HeJ and the C57BL/10ScCr mice.

Poltorak et al.1 discovered that C3H/HeJ mice carry a mutant allele (Lpsd) of the Toll-like receptor 4 (Tlr4) gene, located on chromosome 4. This has been rapidly confirmed by other groups.2,3 As a result of the mutation, several cell types of this mouse strain, including macrophages, B cells and fibroblasts, are refractory to some purified LPSs and to their lipophilic (lipid A) fragment. For example, C3H/HeJ macrophages do not produce tumour necrosis factor-α (TNF-α), interleukin-1, interleukin-6 and nitric oxide after exposure to LPSs from Enterobacteriaceae. However, this LPS response defect can be partially corrected by administration of interferon-γ.4,5 Furthermore, other responses of C3H/HeJ macrophages to enterobacterial LPSs are preserved. This is particularly the case of NF-κB activation,6 matrix metalloproteinase-9 (gelatinase) induction,7 and Mn superoxide dismutase production.8

The second LPS-hyporesponsive mouse mutant is the C57BL/10ScCr mouse strain. No response to enterobacterial or other structurally atypical LPS has been reported in this mouse strain. In contrast to C3H/HeJ mice, in which a localized modification of TLR4 has been detected, TLR4 is completely lacking in C57BL/10ScCr mice.3,7

Another interesting tool for the dissection of cell responses to LPS is the use of structurally atypical LPSs. In this connection, it has been shown that C3H/HeJ macrophages and B lymphocytes, which cannot respond to enterobacterial LPSs, remain able to respond to some LPSs isolated from other bacteria, such as those of Pseudomonas aeruginosa,9 Porphyromonas gingivalis10,11 and Prevotella intermedia.12 This is probably due to the unique chemical features of the lipid A moieties of these LPSs, such as the presence of 15-methylhexadecanoate in P. gingivalis,10,13–15 and 3-hydroxy-15-methylhexadecanoate in P. intermedia.16

Because Gram-negative bacteria belonging to the family Rhizobiaceae infect plants instead of animals, and because the structure of their lipid A exhibits several major differences when compared with lipid A of enterobacteria, we decided to examine the influence of some LPSs isolated from this family of bacteria on cells from normal and LPS-hyporesponsive mice. Bone marrow cells (BMC) represent interesting cellular targets because we established in several previous studies that an LPS receptor of low affinity, distinct from CD14, is constitutively present on these cells,17 and that the interaction of Salmonella LPS with this receptor triggers the expression of CD14.18,19 We thus examined in this report the influence of rhizobial LPSs on this type of response of BMC.

MATERIALS AND METHODS

Materials

Culture medium (CM) was RPMI-1640 (Gibco, Grand Island, NY) containing 2 mm l-glutamine, 100 IU/ml penicillin, 100 µg/ml streptomycin, and 10% heat-inactivated (56°, 30 min) fetal calf serum (FCS; ATGC Biotechnologie, Noisy le Grand, France). The rat anti-mouse CD14 monoclonal antibody (Rm-C5-3) was from PharMingen (San Diego, CA). In fluorescence-activated cell sorter (FACS) experiments, this antibody was stained with fluorescein isothiocyanate (FITC)-labelled goat anti-rat immunoglobulin from Southern Biotechnology Associates (Birmingham, AL). Autoradiography Hyperfilm MP and all electrophoresis reagents, including molecular weight standards (rainbow markers), were from Amersham (Buckinghamshire, UK).

Mice

C57BL/10ScSn were purchased from Harlan (Gannat, France). C57BL/10ScCr mice were a gift from Dr Marina Freudenberg (Freiburg, Germany). C57BL/10ScCr, C3H/HeOU and C3H/HeJ mice were bred at the Pasteur Institute (Paris, France). Eight- to 10-week-old female mice were used in all experiments.

LPS, FITC-LPS, and 125I-LPS

The LPSs from Salmonella enterica serovar choleraesuis (serotype 62,7,14), and from Bordetella pertussis (strain 1414), and the lipid A fraction of the latter, were prepared as described previously.18,20 The four LPSs from bacteria belonging to the Rhizobiaceae (Rhizobium etli CE3, Rhizobium species Sin-1, Rhizobium galegae, and Rhizobium leguminosarum biovar trifolii 24AR) were extracted using hot phenol–water,21 and purified by gel-filtration chromatography in the presence of deoxycholate as previously described.22 Their protein contents, measured using a bicinchoninic acid assay kit from Pierce Chemical Co (Rockford, IL) and bovine serum albumin as the standard, were 1·0%, 2·0%, 1·5% and less than 0·5%, respectively. The LPSs from R. species Sin-1 and R. galegae belong to the ‘rough’ chemotype (O-chain absent), whereas the two others rhizobial LPSs do not (O-chain present).

A FITC-labelled suspension of LPS-Sc (FITC-LPS) prepared previously18 was used. Briefly, FITC (250 µl; 1 mg/ml in dimethylsulphoxide) was incubated (150 min, 20°) with a suspension of lysine-LPS (0·9 ml in 0·1 m NaHCO3; pH 9) obtained by incubation of CNBr-activated LPS (5·2 mg; 700 µl; pH 10) with lysine chloride (200 µl, 5 mg/ml in 1 m NaHCO3). After dialysis against phosphate-buffered saline (PBS), FITC-LPS (3·7 mg/ml) was stored in the dark (4°) until used.

A radiolabelled derivative of LPS-Sc was prepared as described previously.17 The CNBr-activated LPS-Sc was first coupled to tyramine, and the tyramine-labelled LPS (Tyr-LPS) was then iodinated with 125I by the chloramine-T method. After extensive dialysis, the radiolabelled LPS was separated from residual iodine by precipitation with ethanol (five volumes) at −20° for 30 min. The precipitate was recovered by centrifugation (10 min, 900 g). The pellet containing the labelled endotoxin (2·3 × 106 c.p.m./µg) was suspended in water (1 ml), and stored at −30°. Aliquots were thawed, diluted in binding buffer and sonicated before use in binding assays. This preparation of 125I-LPS binds in a specific manner to mouse BMC,17 and to cells that express CD14 constitutively (human monocytes and mouse macrophages).23 Labelling of the LPS of Rhizobium species Sin-1 was performed by the same method (specific activity: 1·9 × 106 c.p.m./µg).

FACS analysis of LPS receptors and CD14 expressed in BMC

BMC collected from mouse femurs (5 × 105 cells in 400 µl CM without FCS) were incubated (18–24 hr, 37°) with (10 ng/ml) or without LPS. When used, inhibitors were added in cell cultures (37°) 1 hr before LPS. The cultures were then maintained for 1 hr at 4°. For analysis of LPS-binding capacity, the cells were then incubated (18 hr, 4°) with FITC-LPS (0·2 µg/ml in 250 µl CM). For detection of membrane CD14, the cells were incubated first (30 min, 0°) with the rat anti-mouse CD14 monoclonal antibody (rmC5-3) and stained by reincubation (30 min, 0°) with an FITC-labelled anti-rat immunoglobulin antibody. Stained cells were layered on a 50% FCS solution, centrifuged, and the cell pellet was resuspended in 0·5 ml of staining buffer (PBS, 5% FCS and 0·02% sodium azide) containing propidium iodide (0·2 µg/ml) to stain dead cells. Fluorescent cells were detected by analysis (5000 cells per sample) on a FACS flow cytometer (FACScan, Becton-Dickinson Electronic Laboratories, Mountain View, CA) using cell quest Software. Dead cells, which incorporated propidium iodide, were gated out of analysis. Cells with a fluorescence intensity higher than the maximal level of auto-fluorescence were scored as FITC-LPS+ cells or CD14+ cells.

Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) analysis of membrane CD14

BMC were pelleted and membrane proteins were extracted with 1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propane-sulfonate (CHAPS) in 300 mm NaCl, 50 mm Tris, pH 7·5, supplemented with a cocktail of protease inhibitors (aprotinin 10 µg/ml, phenylmethylsulphonyl fluoride 1 mm, pepstatin and leupeptin at 2 µg/ml and iodoacetamide 2 mm). Solubilized proteins were analysed by SDS–PAGE in 10% polyacrylamide slab gels according to the method of Laemmli. Molecular mass markers from 14 300 to 220 000 were run in parallel. Gels were fixed in transfer buffer (20 mm Tris, 150 mm glycine, 20% methanol) and proteins were transferred onto polyvinyldifluoride (PVDF) membranes (Millipore, Bedford, MA) with a semi-dry blotting system at 45 V for 1 hr. Membranes were blocked (18 hr at 20°) with 2% bovine serum albumin (BSA) in PBS, and incubated (1 hr, 20°) with the rat anti-mouse antibody rmC5-3 (1 : 1000 in PBS containing 2% BSA). The blots were washed with 0·1% Tween-20 in PBS, and then incubated for 1 hr at 20° with a biotin-labelled goat anti-rat antibody (1 : 2500 in the same buffer). After extensive washing and incubation with peroxidase-labelled streptavidin (1 : 20 000 in 2% non-fat milk in PBS), sites with peroxidase activity were detected by chemiluminescence with the Super Signal system (Pierce, Rockford, IL) according to the guidelines of the manufacturer.

Analysis of the constitutive LPS-binding capacity of BMC

The capacity of the cells to bind LPS was determined with the radioiodinated derivatives of LPS. Unless otherwise specified, the binding of 125I-LPS to BMC was carried out at 37° for 60 min in a binding medium consisting of RPMI-1640 containing 100 IU/ml penicillin, 100 µg/ml streptomycin, 20 mm HEPES, 2 mm l-glutamine, 50 µm 2-mercaptoethanol and 1 mm sodium pyruvate, supplemented with metabolic inhibitors,24 consisting of 1 mm ethyldiaminetetraacetic acid, 5 mm 2-deoxyglucose, 2 mm NaF and 10 mm NaN3. BMC (5 × 106 cells in 1 ml polystyrene tubes) were incubated in binding medium (total volume of 400 µl) with 125I-LPS (5 µg/ml), after a preincubation of 1 hr in the presence or absence of an unlabelled LPS used as a competitor. Unbound ligand was removed by a modification25 of the method of Tsudo et al.26 The cells were resuspended and layered on cold mixtures (0°, 200 µl) of 30% dinonyl phthalate/70% dibutyl phthalate (density of 1·025), in 1·5 ml conical microcentrifuge tubes. After centrifugation for 3 min at 10 000 g and removal of the supernatant, the tips of the tubes containing the cell pellets were cut off, and the radioactivity was measured in a gamma-counter (Kontron MR 480C). Assays were done in triplicate.

RESULTS

LPSs from Rhizobiaceae induce iLpsR/CD14 expression in mouse BMC

In previous studies18,19 we established that exposure of mouse and human BMC to small amounts of LPS from enterobacteria induces the expression of cell surface LPS-binding sites (iLpsR). The existence of LPSs with atypical structures in other Gram-negative families led us to examine whether such molecules can also induce the same effect. In this connection, LPSs from Rhizobiaceae are particularly interesting because the bacteria infect legume plants instead of animals, and because their LPSs have unique structural features. We used four rhizobial LPSs extracted from R. leguminosarum bv. trifolii 24AR, R. etli CE3, R. galegae and R. species Sin-1. The structures of their lipid A region are unique because all these lipid A contain one very long fatty acid, 27-hydroxyoctacosanoic acid. The location of this very long chain fatty acid is not known with certainty, however, a recent paper by Basu et al.27 has indicated that it may be present in the N-acyloxylacyl residue of the distal glucosamine unit. In a prior report28 on the structure of this lipid A, the workers were unable to detect any acyloxyacyl substituents by chemical analysis. Thus, these lipid A structures are unlike those of Salmonella in that they either do not have any acyloxyacyl substituents, or they have 27-hydroxyoctacosanoic acid as the sole acyloxyacyl moiety. The lipid A structures of three of these bacteria (R. etli, R. galegae and R. species Sin-1) are even more atypical, since they are devoid of phosphate, and contain a 2-aminogluconosyl (GlcN-onate) residue in their backbone (Fig. 1). We exposed BMC of LPS-responsive C3H/HeOU mice to increasing concentrations of the four rhizobial LPSs. Inducible LPS-binding sites (iLpsR) were detected by FACS with the fluorescent ligand FITC-LPS. The expression of iLpsR induced by increasing concentrations of the rhizobial LPSs was compared to that induced by a structurally ‘conventional’ LPS, isolated from an enteric bacterium (Salmonella enterica serovar choleraesuis). The results in Fig. 2(a) show that at high LPS concentrations (1 µg/ml and higher), all LPSs induce iLpsR expression. At low LPS concentrations (0·1 µg/ml or less), only the Salmonella and the R. etli CE3 LPSs were active.

Figure 1.

Figure 1

Open in a new tab

Schematic structures of the lipid A regions of the LPSs used in this study. The structures of the lipid A regions of R. species Sin-1 and R. galegae are not completely determined and are still under investigation; however, they both are devoid of phosphate and galacturonic acid, they both contain glucosamine, 2-aminogluconic acid, the very long chain fatty acyl residue 27-OHC28 : 0, and the R. galegae lipid A may have an additional glucosyl residue. It is hypothesized that their structures are very similar to that shown for R. etli CE3 with R. species Sin-1 lipid A lacking the galacturonosyl residue, and with R. galegae lipid A replacing galacturonosyl residue with glucose. The location of the 27-OHC28 : 0 fatty acyl residue found in the rhizobial lipid A molecules is not known with certainty; however, it is ester-linked,28 and a recent paper by Basu et al.27 suggests that it may be present in the N-acyloxylacyl residue of the distal glucosamine unit, as shown in this Figure.

Figure 2.

Figure 2

Open in a new tab

Influence of various LPSs on LPS receptor/CD14 expression in BMC of LPS-responsive mice. BMC (5 × 105 cells) from C3H/HeOU mice were incubated for 24 hr at 37° with various concentrations of LPSs from R. leguminosarum 24AR, R. species Sin-1, R. galegae, R. etli CE3 and S. enterica. (a) Expression of LPS receptors (iLpsR) was then detected by incubation (18 hr, 4°) with FITC-LPS (0·2 µg/ml) in medium containing 8% FCS. (b) CD14 was detected by incubation with the anti-CD14 antibody rmC5-3 (2·5 µg/ml, 30 min, 0°), and further incubation (30 min, 0°) with an FITC-labelled goat anti-rat immunoglobulin mAb. The percentage of fluorescent cells was determined by FACS analysis of the gated granulocyte population. Values represent the arithmetic mean ± SD of duplicates.

We have shown previously29 and in a recent study,30 that iLpsR expressed on BMC upon exposure to Salmonella enterica LPS is actually CD14. To determine if iLpsR induced by the rhizobial LPSs is also correlated with CD14 expression, we exposed BMC from C3H/HeOU mice to various concentrations of the different LPSs, and we analysed by FACS the reactivity of the cells with an anti-mouse CD14 antibody. The results showed complete parallelism between the dose–response curves of LPS-treated cells stained with anti-CD14 (Fig. 2b) and FITC-LPS (Fig. 2a). This indicates that the LPS-binding capacity of BMC induced either by enterobacterial or by rhizobial LPSs is correlated with the expression of CD14.

Binding of rhizobial LPSs to membrane CD14

Since, in the presence of serum, FITC-LPS binds essentially to the inducible CD14 expressed on stimulated BMC, it was possible to analyse the binding of an unlabelled LPS to CD14 by estimation of its ability to inhibit the binding of FITC-LPS. We used this technique to analyse the ability of the rhizobial LPSs to bind to CD14. Expression of CD14 was first induced on BMC from C3H/HeOU mice by incubation (18 hr, 37°) with Salmonella enterica LPS (10 ng/ml). Cells expressing CD14 were then exposed for 18 hr at 4° to FITC-LPS (0·2 µg/ml), in the presence of various concentrations of unlabelled rhizobial LPSs or S. enterica LPS. The results in Fig. 3 show that the rhizobial LPSs bind to CD14 at least as efficiently as the S. enterica LPS.

Figure 3.

Figure 3

Open in a new tab

Binding of S. enterica LPS to CD14 is inhibited by rhizobial LPSs. BMC from C3H/HeOU mice were incubated with S. enterica LPS (10 ng/ml, 18 hr, 37°). After equilibration for 1 hr at 4°, unlabelled LPS (0·2, 2, or 10 µg/ml) was added and, 1 hr later, the cells were stained by incubation (18 hr, 4°) with FITC-LPS (0·2 µg/ml) in medium containing 8% FCS. The mean fluorescence of the cells was determined by FACS analysis. Data represent the mean ± SD of triplicates.

Lipid A fragments induce CD14 expression in BMC from C3H/HeOU mice

To confirm by a different technique the expression of CD14 demonstrated above by FACS, we analysed the production of CD14 by Western blot. Because LPS activities are mainly due to the lipid A moiety, BMC from C3H/HeOU mice were exposed to the rhizobial LPSs and to their lipid A and polysaccharide fragments (1 µg/ml). Incubations were carried out at 37° for 18 hr in culture medium, in the absence of serum. The LPSs of S. enterica (a conventional enterobacterial LPS) and the lipid A of Bordetella pertussis (a structurally well-defined lipid A)31 were used as standards. The results in Fig. 4 show that in BMC from C3H/HeOU mice, all LPSs and lipid As, including those from Rhizobiaceae, induce the production of CD14, whereas most of the polysaccharides were inactive.

Figure 4.

Figure 4

Open in a new tab

Lipid A-induced CD14 in normal mice. BMC (5 × 106 cells/ml) from C3H/HeOu mice were incubated for 24 hr at 37° in culture medium containing 1 µg/ml of LPS, lipid A, or polysaccharide (PS) of different bacterial origin. Cell lysates (105 cells in 1% CHAPS, supplemented with a mixture of protease inhibitors) were analysed for CD14 by SDS–PAGE and Western blotting. The PVDF membrane was blocked by 2% BSA in PBS (20 hr, 20°), and incubated (1 hr, 20°) with the anti-mouse CD14 antibody.

Another method commonly used to confirm that lipid A is the biologically active part of an LPS preparation is to show that polymyxin B (PMB) blocks its biological effect. Preliminary results indicated that all the rhizobial LPSs used bind to a polymyxin–agarose column, and can be eluted with 1% deoxycholate (data not shown). We thus examined whether PMB can inhibit the CD14 expression induced by these LPSs. The results in Fig. 5 clearly show a marked dose-dependent inhibition by PMB. This can be taken as further independent evidence that the activation of BMC is induced by the lipid A region of the LPSs.

Figure 5.

Figure 5

Open in a new tab

Inhibition of LPS-induced CD14 expression by polymyxin B. LPSs from R. species Sin-1 or R. leguminosarum 24AR (5 µg/ml) were incubated (40 min, 37°) with the same volume of different concentrations of polymyxin B (0, 50, 250, 500 µg/ml) in RPMI-1640. BMC from C3H/HeOu mice (300 µl, 1·65 × 106 cells/ml) were incubated for 20 hr at 37° in the same medium with 200 µl of the LPS-PMB mixtures. Cell lysates were prepared and analysed for CD14 as described in the legend of Figure 4.

However, the results mentioned above do not mean that the polysaccharide region of some particular LPSs cannot also induce CD14 expression. Indeed, we found (Fig. 6) that the polysaccharide fragment of the LPS from R. etli CE3 is also a CD14 inducer. After incubation of BMC with 10 µg/ml of this polysaccharide, the number of LPS-binding sites detectable with FITC-LPS expressed by the cells (mean fluorescence = 23·1) was even higher than those induced by the same concentrations of the complete LPS (mean fluorescence = 18·4) or of its lipid A fragment (mean fluorescence = 10·9). This unexpected activity is probably due to the particular structure of this polysaccharide, and explains the high activity of the R. etli LPS, as compared to those of R. galegae and R. species Sin-1 (Fig. 2), in spite of the very similar structures of their hydrophobic moieties. The precise region (core or O-chain) involved in the activity of the R. etli polysaccharide remains unknown. It should be noted that the polysaccharides of R. galegae and R. species Sin-1 consist of the core region, and are inactive. But their core structures are different from that of R. etli CE3. Therefore, the possibility that the core structure of R. etli is responsible for the activity of the polysaccharide preparation cannot be excluded. However, since O-chain is, by far, predominant in this polysaccharide preparation, it is reasonable to assume that the activity of the R. etli polysaccharide is due to its O-chain region.

Figure 6.

Figure 6

Open in a new tab

Induction of iLpsR/CD14 expression by the polysaccharide and lipid A fragments of R. etli LPS. BMC from C3H/HeOu mice were incubated (20 hr, 37°) with the LPS from R. etli CE3, or its lipid A and polysaccharide fragments. Cells were either incubated with 10 µg/ml of the inducers and analysed by FACS for binding of FITC-LPS, or incubated with 1 µg/ml of the inducers and analysed by Western blot for CD14 expression. Experiments were performed as described in Fig. 2 and Fig. 4, respectively.

Lipid As from rhizobial LPSs induce CD14 expression in BMC from C3H/HeJ and C57BL/10ScCr mice

Two mutant mouse strains, C3H/HeJ and C57BL/10ScCr, have been reported to exhibit, at the cellular level, very low responses to LPSs of different bacteria. It was therefore important to determine whether BMC from these mutants also exhibit lower responses to the rhizobial LPSs. In a first experiment, we analysed by Western blot the CD14 produced by BMC from C3H/HeJ mice exposed to three rhizobial LPS and lipid As (R. galegae, R. species Sin-1 and R. leguminosarum 24AR), as compared to that produced in response to S. enterica LPS and to B. pertussis lipid A. The results are presented in Fig. 7(a). As expected, the cells did not respond to S. enterica LPS and B. pertussis lipid A. In contrast, the three rhizobial LPSs and lipid As induced CD14 production in the BMC of this hyporesponsive mouse strain. Very similar results were obtained with these compounds when the responses of BMC from the mutant C57BL/10ScCr strain and the normal histocompatible C57BL/10ScSn strain were compared (Fig. 7b). Again, the B. pertussis lipid A induced CD14 production in the normal strain, but was inactive in the mutant strain, whereas the three rhizobial lipid As induced CD14 in BMC from both strains.

Figure 7.

Figure 7

Open in a new tab

Induced expression of CD14 in BMC. BMC from C3H/HeJ mice (a), or from C57BL/10ScSn or C57BL/10ScCr mice (b) were incubated with 1 µg/ml LPS from different bacteria (a), or with 1 µg/ml of the lipid A fragment isolated from different LPSs (a,b). The cells were analysed for CD14 production by Western blot, as described in the legend of Fig. 4.

Cross-reactivity of Bordetella, Salmonella and Rhizobium LPSs with BMC binding sites

We demonstrated in previous studies that compounds structurally unrelated to LPS can induce CD14 expression in BMC from C3H/HeOU and C3H/HeJ mice. For example, the glycoside ring of staurosporine can do so, and the mechanism involved is clearly different from that induced by LPS.29 The fact that the effects of LPS and staurosporine in LPS-responsive BMC are additive suggests that the receptors involved are different. Cholera toxin, and cAMP also induce CD14 expression in C3H/HeOU and C3H/HeJ mice after interaction with the ganglioside GM1, or with a purinoreceptor, respectively.30 Therefore, the ability of rhizobial LPSs to induce CD14 expression in BMC from C3H/HeOU and C3H/HeJ mice can be due either to the interaction with an LPS receptor, or to the interaction with another receptor. To examine this possibility, we analysed the cross-reactivities between rhizobial and non-rhizobial LPSs with LPS-binding sites of BMC. In a first experiment (Fig. 8a), we found that the binding of a radiolabelled LPS of S. enterica to BMC of C3H/HeJ mice was inhibited by unlabelled LPSs from B. pertussis, R. species Sin-1 and R. galegae. In contrast, the binding of the radiolabelled LPS was not inhibited by cholera toxin, another CD14 inducer. This result suggests that the four LPSs cross-react with the same binding site(s) of BMC, whereas cholera toxin does not. This conclusion was confirmed by a second experiment in which a radiolabelled preparation of the R. species Sin-1 LPS was used. We found that the binding of this rhizobial LPS was inhibited by unlabelled LPSs from B. pertussis and R. galegae (Fig. 8b). We also found that the polysaccharide fragments of S. enterica and R. species Sin-1 do not inhibit the binding of the two radiolabelled LPSs used (data not shown), thus suggesting that cross-reactivities are mediated by the lipid A regions of the LPSs.

Figure 8.

Figure 8

Open in a new tab

Inhibition of the binding of radiolabelled LPSs to BMC of C3H/HeJ mice. BMC from C3H/HeJ mice (5 × 106 cells) were incubated for 1 hr at 37° with 125I-labelled LPS of S. enterica (a) or R. species Sin-1 (b), in the binding medium, in the presence of unlabelled inhibitors [50 µg/ml or 100 µg/ml in (a), 100 µg/ml in (b)]. Inhibitors were added 60 min before the radiolabelled LPS. Bound ligand was measured in the pellet after centrifugation of the cells through phthalate. In both (a) and (b), inhibition values obtained with 100 µg/ml of unlabelled R. galegae LPS (34 ± 2% and 18 ± 2% inhibition of the binding of radiolabelled S. enterica and R. species Sin-1, respectively), were taken as the controls corresponding to 100% inhibition of specific and cross-reactive binding sites (0% of binding). Binding of the radiolabelled LPS from S. enterica (a) or R. species Sin-1 (b) are expressed as percentages of the corresponding control, and represent means ± SD of triplicate determinations.

Lipid A from B. pertussis inhibits CD14 expression induced by a rhizobial lipid A in C3H/HeJ cells

To determine if the induction of CD14 by rhizobial lipid As is mediated by the constitutive LPS receptor, an attractive possibility should be to inhibit this response by occupation of the receptor with a ligand which does not induce the same effect. The B. pertussis LPS was a good candidate since we show above that it binds to the constitutive receptor which recognizes the R. species Sin-1 LPS in C3H/HeJ cells, but does not induce CD14 expression in these cells. Therefore, we examined whether the B. pertussis lipid A can inhibit the activation induced by the R. species Sin-1 lipid A. The results in Fig. 9(a) show that the production of CD14 induced by R. species Sin-1 lipid A in BMC from C3H/HeJ mice was markedly inhibited if B. pertussis lipid A was added in the culture medium 60 min before. We also found (Fig. 9b) that B. pertussis lipid A does not inhibit the production of CD14 induced by cholera toxin, which is mediated by an irrelevant receptor (the ganglioside GM1). This indicates that inhibition by B. pertussis lipid A is not due to some general suppressive effect. Therefore, the results in Fig. 9 strongly suggest that activation of C3H/HeJ cells by the lipid A region of rhizobial LPS is triggered by the constitutive receptor which recognizes B. pertussis lipid A.

Figure 9.

Figure 9

Open in a new tab

CD14 induction by a rhizobial lipid A is inhibited by B. pertussis lipid A in C3H/HeJ cells. (a) BMC from C3H/HeJ mice were preincubated (1 hr, 37°) in the presence or absence of B. pertussis lipid A (10 µg/ml), and reincubated (20 hr, 37°) with or without R. species Sin-1 lipid A (1 µg/ml). The cells were collected, lysed and analysed by Western blot for the detection of CD14 as described in the legend of Fig. 4. (b) BMC from C3H/HeJ mice were preincubated (1 hr, 37°) with or without 20 µg/ml of B. pertussis lipid A, and reincubated (20 hr, 37°) with or without cholera toxin (0,1 µg/ml) or R. species Sin-1 lipid A (10 µg/ml).The expression of membrane-bound CD14 was detected by FACS as described in the legend of Fig. 2(b).

DISCUSSION

In this study, we have shown that several LPSs isolated from Rhizobiaceae can activate mouse bone marrow cells by inducing the expression of CD14. This was rather surprising because the lipid A structures of these rhizobial LPSs are markedly different from those of enterobacterial LPSs: they all contain a very long-chain fatty acid (27-hydroxyoctacosanoic acid), and three of the rhizobial LPSs used are devoid of phosphate, and contain a 2-aminogluconosyl residue in their backbone. The reported absence of acyloxyacyl substituents28 and the recent possibility that the 27-hydroxyoctacosanoic acid may be the sole acyloxyacyl substituent27 are particularly interesting observations since it has been claimed that the presence of acyloxyacyl residues at the 3′ position of lipid A may play an important role in the biological effects of LPS. It has been shown for instance that the LPS from the mutant BMS67C12 Escherichia coli K-12, which is devoid of myristoyl fatty acid at the 3′ R-3-hydroxymyristate position, exhibits a markedly reduced ability to induce the production of TNF-α and macrophage inflammatory protein-1α (MIP-1α) by human monocytes.32 A constructed Salmonella mutant with the same defect in lipid A biosynthesis was also less lethal in mice, and the killed bacteria induced less pro-inflammatory cytokines and inducible nitric oxide synthase in a macrophage cell line.33 However, other examples show that the absence of 3′-acyloxyacyl residues may not be so critical for LPS activities. This is particularly the case of the Bordetella pertussis LPS. The lipid A region of this LPS contains a single hydroxylated fatty acid at the 3′ position of lipid A, instead of the usual acyloxyacyl group found in enterobacterial LPSs.31 In spite of this structural difference, the B. pertussis LPS retains the capacity to induce TNF-α production in macrophages,34 and has many of the typical in vivo activities of other LPSs.35 We also show in the present study that the B. pertussis LPS is an efficient inducer of CD14 in BMC of normal mice. The observation that B. pertussis lacks acyloxyacyl residues at the 3′ position of lipid A and can still induce CD14 expression in BMC clearly indicates that this structural feature is not important for this activity. It should be noted that for three rhizobial LPSs, the concentration required to induce CD14 expression (1–5 µg/ml) is much higher than that required by the Salmonella enterica LPS (10 ng/ml). This may be due to the location of the 27-hydroxyoctacosanoic acid residue (i.e. whether or not it is present as an acyloxyacyl residue), or to one of the three other marked structural differences mentioned above. Further structural analysis is in progress to determine the exact location of the very long chain fatty acid in the rhizobial lipid A structures.

The second finding of this study is that the rhizobial LPSs are active in BMC from the two ‘LPS-hyporesponsive’ mouse strains C3H/HeJ and C57BL/10ScCr. The amount of CD14 induced by these structurally atypical LPSs in the BMC of the mutant strains is however, much lower than that induced in normal BMC (data not shown). These effects of rhizobial LPSs are not analogous to those of other CD14 inducers such as staurosporine29 and cholera toxin30 since we showed that the former can be inhibited by B. pertussis lipid A. Furthermore, we found that in BMC from C3H/HeJ mice, the LPSs of S. enterica, B. pertussis, R. galegae and R. species Sin-1 cross-react with the same constitutive receptor. It should be recalled here that we demonstrated in a previous study17 that the constitutive receptor of BMC combines an LPS binding capacity with low affinity and the functional ability to activate the cells. This already established result, and the two observations of the present study mentioned above, when taken together, indicate that the effects induced by the rhizobial LPSs in the mutant mouse strains are mediated by the constitutive LPS receptor. Therefore, the second conclusion of our study is that, in spite of the defect in the Lps gene, the signalling moiety of the constitutive LPS receptor in BMC from C3H/HeJ and C57BL/10ScCr mice is fully able to trigger a complete signalling cascade leading to CD14 expression. This also means that this signalling moiety is not TLR4, since the gene coding for this molecule is completely absent in C57BL/10ScCr mice.1,3 The possibility that this signalling moiety could be TLR2 cannot be excluded and is rather attractive.

The observation that the constitutive LPS receptor of C3H/HeJ cells can trigger a complete signalling cascade, which is however, inhibited by B. pertussis lipid A, strongly suggests a third important conclusion: in C3H/HeJ cells the constitutive LPS receptor can distinguish between lipid A structures. The lipid A fragments of B. pertussis and rhizobial LPSs are antagonists and agonists of this constitutive receptor, respectively. The observation that B. pertussis LPS inhibits both the binding of rhizobial LPS (Fig. 8b) and the signalling induced by this LPS (Fig. 9a) indicates that binding to the constitutive receptor is required for signalling. On the other hand, the observation that the receptor binds to the three families of LPSs (Salmonella, Bordetella and Rhizobium) whereas only the latter induces a response in C3H/HeJ cells, indicates that binding is not sufficient for activation. Because binding is necessary but not sufficient, it follows that binding to the receptor, and activation of the receptor, are dissociated events. This type of dissociation is not unique: Cunningham et al.36 reported that a mutant endotoxin lacking a single secondary acyl chain binds normally to CD14, but fails to induce TNF-α.

In conclusion, because of their unique structural features, the lipid A regions of rhizobial LPSs are less active than conventional lipid As in BMC from normal mice, but are more active than the latter in BMC from the two mutant mouse strains. An LPS receptor constitutively present on this cell type is bound by various LPSs from different bacterial species, but triggers a signal, in a TLR4-independent way, only in response to some particular LPSs such as those from Rhizobiaceae.

Acknowledgments

This work was supported by the Centre National de la Recherche Scientifique (grant 1961), and by grants from the Pasteur Institute (grant 3540), and the Fondation pour la Recherche Médicale. We thank Marina Freudenberg (Max-Planck Institut fur Immunbiologie, Freiburg, Germany) for providing male and female C57BL/10ScCr mice for breeding.

REFERENCES

  • 1.Poltorak A, He X, Smirnova I, et al. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science. 1998;282:2085–8. doi: 10.1126/science.282.5396.2085. [DOI] [PubMed] [Google Scholar]
  • 2.Hoshino K, Takeuchi O, Kawai T, Sanjo H, Ogawa T, Takeda Y, Takeda K, Akira S. Cutting edge: Toll-like receptor 4 (TLR4) – deficient mice are hyporesponsive to lipopolysaccharide: evidence for TLR4 as the Lps gene product. J Immunol. 1999;162:3749–52. [PubMed] [Google Scholar]
  • 3.Qureshi ST, Larivière L, Leveque G, Clermont S, Moore KJ, Gros P, Malo D. Endotoxin-tolerant mice have mutations in Toll-like receptor 4 (Tlr4) J Exp Med. 1999;189:615–25. doi: 10.1084/jem.189.4.615. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Akagawa KS, Kamoshita K, Onodera S, Tokunaga T. Restoration of lipopolysaccharide-mediated cytotoxic macrophage induction in C3H/HeJ mice by interferon-gamma or a calcium ionophore. Jap J Cancer Res. 1987;78:279–87. [PubMed] [Google Scholar]
  • 5.Beutler B, Krochin N, Milsark IW, Luedke C, Cerami A. Control of cachectin (tumor necrosis factor) synthesis: mechanisms of endotoxin resistance. Science. 1986;232:977–80. doi: 10.1126/science.3754653. [DOI] [PubMed] [Google Scholar]
  • 6.Ding A, Hwang S, Lander HM, Xie QW. Macrophages derived from C3H/HeJ (Lpsd) mice respond to bacterial lipopolysaccharide by activating NF-κB. J Leuk Biol. 1995;57:174–9. doi: 10.1002/jlb.57.1.174. [DOI] [PubMed] [Google Scholar]
  • 7.Jin F, Nathan CF, Ding A. Paradoxical preservation of a lipopolysaccharide response in C3H/HeJ macrophages: induction of matrix metalloproteinase-9. J Immunol. 1999;162:3596–600. [PubMed] [Google Scholar]
  • 8.Gibbs LS, Del Vecchio PJ, Shaffer JB. Mn and Cu/Zn SOD expression in cells from LPS-sensitive and LPS-resistant mice. Free Radic Biol Med. 1992;12:107–11. doi: 10.1016/0891-5849(92)90003-y. [DOI] [PubMed] [Google Scholar]
  • 9.Pier GB, Markham RB, Eardley D. Correlation of the biologic responses of C3H/HeJ mice to endotoxin with the chemical and structural properties of the lipopolysaccharides from Pseudomonas aeruginosa and Escherichia coli. J Immunol. 1981;127:184–91. [PubMed] [Google Scholar]
  • 10.Tanamoto K, Azumi S, Haishima Y, Kumada H, Umemoto T. The lipid A moiety of Porphyromonas gingivalis lipopolysaccharide specifically mediates the activation of C3H/HeJ mice. J Immunol. 1997;158:4430–6. [PubMed] [Google Scholar]
  • 11.Ogawa T, Shimauchi H, Uchida H, Mori Y. Stimulation of splenocytes in C3H/HeJ mice with Porphyromonas gingivalis lipid A in comparison with enterobacterial lipid A. Immunobiology. 1996;196:399–414. doi: 10.1016/S0171-2985(96)80062-3. [DOI] [PubMed] [Google Scholar]
  • 12.Kirikae T, Nitta T, Kirikae F, Suda Y, Kusumoto S, Qureshi N, Nakano M. Lipopolysaccharides (LPS) of oral black-pigmented bacteria induce tumor necrosis factor production by LPS-refractory C3H/HeJ macrophages in a way different from that of Salmonella LPS. Infect Immun. 1999;67:1736–42. doi: 10.1128/iai.67.4.1736-1742.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Kumada H, Haishima Y, Umemoto T, Tanamoto K. Structural study on the free lipid A isolated from lipopolysaccharide of Porphyromonas gingivalis. J Bacteriol. 1995;177:2098–106. doi: 10.1128/jb.177.8.2098-2106.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Ogawa T. Chemical structure of lipid A from Porphyromonas (Bacteroides) gingivalis lipopolysaccharide. FEBS Lett. 1993;332:197–201. doi: 10.1016/0014-5793(93)80512-s. [DOI] [PubMed] [Google Scholar]
  • 15.Ogawa T. Immunobiological properties of chemically defined lipid A from lipopolysaccharide of Porphyromonas (Bacteroides) gingivalis. Eur J Biochem. 1994;219:737–42. doi: 10.1111/j.1432-1033.1994.tb18552.x. [DOI] [PubMed] [Google Scholar]
  • 16.Nichols FC. Distribution of 3-hydroxy iC17: 0 in subgingival plaque and gingival tissue samples: relationship to adult periodontitis. Infect Immun. 1994;62:3753–60. doi: 10.1128/iai.62.9.3753-3760.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Girard R, Pedron T, Chaby R. Functional lipopolysaccharide receptors of low affinity are constitutively expressed on mouse bone marrow cells. Immunology. 1997;91:391–8. doi: 10.1046/j.1365-2567.1997.00275.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Girard R, Pedron T, Chaby R. Endotoxin-induced expression of endotoxin binding sites on murine bone marrow cells. J Immunol. 1993;150:4504–13. [PubMed] [Google Scholar]
  • 19.Pedron T, Girard R, Turco SJ, Chaby R. Phosphatidylinositol-anchored molecules and inducible lipopolysaccharide binding sites of human and mouse bone marrow cells. J Biol Chem. 1994;269:2426–32. [PubMed] [Google Scholar]
  • 20.Tahri-Jouti MA, Mondange M, Ledur A, Auzanneau FI, Charon D, Girard R, Chaby R. Specific binding of lipopolysaccharide to mouse macrophages. II- Involvement of distinct lipid A substructures. Mol Immunol. 1990;27:763–70. doi: 10.1016/0161-5890(90)90085-e. [DOI] [PubMed] [Google Scholar]
  • 21.Westphal O, Jann K. Bacterial lipopolysaccharides. Meth Carbohydr Chem. 1965;5:83–91. [Google Scholar]
  • 22.Reuhs BL, Kim JS, Badgett A, Carlson RW. Production of cell-associated polysaccharides of Rhizobium fredii USAD205 is modulated by apigenin and host root extract. Mol Plant-Microbe Interact. 1994;7:240–4. doi: 10.1094/mpmi-7-0240. [DOI] [PubMed] [Google Scholar]
  • 23.Girard R, Pedron T, Kosma P, Chaby R. A synthetic analog of the 3-deoxy-d-manno-2-octulosonic acid disaccharide moiety of rough-type endotoxins does not bind to mouse peritoneal macrophages and human monocytes. Infect Immun. 1993;61:3616–24. doi: 10.1128/iai.61.9.3616-3624.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Schmid SL, Carter LL. ATP is required for receptor-mediated endocytosis in intact cells. J Cell Biol. 1990;111:2307–18. doi: 10.1083/jcb.111.6.2307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gessani S, Testa U, Varano B, et al. Enhanced production of LPS-induced cytokines during differentiation of human monocytes to macrophages. Role of LPS receptors. J Immunol. 1993;151:3758–66. [PubMed] [Google Scholar]
  • 26.Tsudo M, Kozak RW, Goldman CK, Waldmann TA. Demon-stration of a non-Tac peptide that binds interleukin 2: a potential participant in a multichain interleukin 2 receptor complex. Proc Natl Acad Sci USA. 1986;83:9694–8. doi: 10.1073/pnas.83.24.9694. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Basu SS, White KA, Que NLS, Raetz CRH. A deacylase in Rhizobium leguminosarum membranes that cleaves the 3-O-linked β-hydroxymyristoyl moiety of lipid A precursors. J Biol Chem. 1999;274:11150–8. doi: 10.1074/jbc.274.16.11150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Bhat UR, Forsberg LS, Carlson RW. Structure of lipid A component of Rhizobium leguminosarum bv. phaseoli lipopolysaccharide. Unique nonphosphorylated lipid A. J Biol Chem. 1994;269:14402–10. [PubMed] [Google Scholar]
  • 29.Pedron T, Girard R, Inoue K, Charon D, Chaby R. Lipopolysaccharide and the glycoside ring of staurosporine induce CD14 expression on bone marrow granulocytes by different mechanisms. Molec Pharmacol. 1997;52:692–700. doi: 10.1124/mol.52.4.692. [DOI] [PubMed] [Google Scholar]
  • 30.Pedron T, Girard R, Chaby R. Exogenous cyclic AMP, cholera toxin, and endotoxin induce expression of the lipopolysaccharide receptor CD14 in murine bone marrow cells: role of purinoreceptors. Clin Diagnost Lab Immunol. 1999;6:885–90. doi: 10.1128/cdli.6.6.885-890.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Caroff M, Deprun C, Richards JC, Karibian D. Structural characterization of the lipid A of Bordetella pertussis strain 1414 endotoxi. J Bacteriol. 1994;176:5156–9. doi: 10.1128/jb.176.16.5156-5159.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Tian L, White JE, Lin HY, et al. Induction of Mn SOD in human monocytes without inflammatory cytokine production by a mutant endotoxin. Am J Physiol. 1998;275:C740–C747. doi: 10.1152/ajpcell.1998.275.3.C740. [DOI] [PubMed] [Google Scholar]
  • 33.Khan SA, Everest P, Servos S, et al. A lethal role for lipid A in Salmonella infections. Mol Microbiol. 1998;29:571–9. doi: 10.1046/j.1365-2958.1998.00952.x. [DOI] [PubMed] [Google Scholar]
  • 34.Lasfargues A, Chaby R. Endotoxin-induced tumor necrosis factor (TNF): selective triggering of TNF and interleukin-1 production by distinct glucosamine-derived lipids. Cell Immunol. 1988;115:165–78. doi: 10.1016/0008-8749(88)90171-2. [DOI] [PubMed] [Google Scholar]
  • 35.Chaby R, Caroff M. Lipopolysaccharides of Bordetella pertussis endotoxin. In: Wardlaw AC, Parton R, editors. Pathogenesis and Immunity in Pertussis. Chichester: John Wiley and Sons; 1988. pp. 247–71. [Google Scholar]
  • 36.Cunningham MD, Bajorath J, Somerville JE, Darveau RP. Escherichia coli and Porphyromonas gingivalis interactions with CD14: implications for myeloid and non-myeloid cell activation. Clin Infect Dis. 1999;28:497–504. doi: 10.1086/515158. [DOI] [PubMed] [Google Scholar]

Articles from Immunology are provided here courtesy of British Society for Immunology

RESOURCES