Physicochemical characteristics of triacyl lipid A partial structure OM-174 in relation to biological activity


U. Seydel, Research Center Borstel, Divison of Biophysics, Parkallee 10, D-23845 Borstel, Germany. Fax: + 49 4537 188632, Tel.: + 49 4537 188232, E-mail:


The triacylated lipid A partial structure OM-174 was characterized in detail using a variety of physical and biological techniques. OM-174 aggregates adopt the micellar HI structure. The temperature (Tc) of the gel to liquid-crystalline phase transition of the hydrocarbon chains is 0 °C, from which high fluidity of the acyl chains at 37 °C can be deduced. The molecular area of a single OM-174 molecule at a surface pressure of 30 mN·m−1 is 0.78 ± 0.04 nm2. Conformational analyses, using IR spectroscopy, of the behavior of the various functional groups of OM-174 as compared with hexa-acyl lipid A suggest altered hydration of the phosphate charges and unusually strong hydration of the ester groups, which is probably related to the high accessibility of these groups to water in the micellar aggregate structure. OM-174 was shown to intercalate into a phospholipid membrane corresponding to the macrophage membrane within seconds in the presence, and within minutes to hours in the absence, of LPS-binding protein. In the Limulus amebocyte lysate assay, the triacyl lipid A is more than 105-fold less active than hexa-acyl lipid A, but only 10-fold less active in inducing IL-6 in human mononuclear cells, and equally active in inducing NO production in murine macrophages. These findings are used to explain the mechanism of the lipid A-induced cell activation.


Fourier transform infrared






critical micellar concentration


LPS-binding protein


fluorescence resonance energy transfer




N-(rhodamine B sulfonyl)-PE


Limulus amebocyte lysate


3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

Lipopolysaccharides (LPSs), particularly amphiphiles forming the lipid matrix of the outer leaflet of the outer membrane of Gram-negative bacteria [1], are also called endotoxins because of their ability to induce a variety of biological effects in mammalian cells [2]. Their lipid moiety, called lipid A, consists of a diglucosamine saccharide phosphorylated in positions 1 and 4′ and is acylated, in Enterobacteriaceae, with six or seven acyl chains in either amide or ester linkage with the diglucosamine backbone. Lipid A is known to represent the ‘endotoxic principle’ of LPS because it can induce the same spectrum of endotoxin reactions as its parent LPS [3].

The physical or physicochemical behavior of aqueous suspensions of LPS and lipid A may be important in their biological action. Physicochemical characteristics include: (a) the critical micellar concentration (CMC), which is the concentration above which, on further addition of lipids to the bulk solution, no further increase in monomer concentration takes place; (b) the shape and size of the lipid aggregates above the CMC; (c) the mobility of the hydrocarbon chains (fluidity) within the aggregates; (d) the dependence of all these parameters on pH, concentration of univalent and bivalent cations, and temperature. We have previously investigated free lipid A preparations from Escherichia coli and Salmonella minnesota under near-physiological conditions and could relate particular physicochemical characteristics to their biological activity [4–6]. In this study, the results of measurements performed with the triacyl lipid A OM-174 are presented, which differs from E. coli lipid A by the absence of the acyl chains linked directly through ester bonds to the diglucosamine backbone. This compound has been shown to be a candidate for an effective anticancer treatment [7,8]. We have found that OM-174 differs from E. coli lipid A in many physicochemical characteristics, but is still able to express cytokine-inducing activity. These findings are interpreted according to our model of cell activation, which presumes the intercalation of endotoxin monomers into the target cell membrane.

Materials and methods

Chemical extraction and lipid A isolation

OM-174 (OM PHARMA, Geneva, Switzerland) is a triacylated partial structure of natural lipid A obtained from E. coli (Fig. 1). Bacteria were treated with alkali to produce partial hydrolysis of ester-linked fatty acids in the lipid A moiety of LPS. The O-polysaccharide and core oligosaccharide regions of the E. coli LPS were then cleaved by acid treatment from the partially hydrolysed lipid A. The lipid A partial structure was purified by solvent extraction and ion-exchange and reversed-phase chromatography, resulting in an aqueous preparation of its sodium salt. This preparation was adjusted to pH 7.4 and sterilized by 0.2 µm filtration.

Figure 1.

Chemical structure of triacyl lipid A OM-174.

OM-174 has the diglucosamine diphosphate structure of natural E. coli lipid A. The 3-, 3′- and 4-hydroxyl groups of the diglucosamine diphosphate backbone are not substituted. The 2 amino group is substituted by a hydroxymyristate residue, and the 2′ amino group by a C14-O-C12 acyloxyacyl group.

The structure of OM-174 was determined by biochemical analysis of its constituent phosphate, glucosamine, and fatty acid groups, by MS and NMR (data not shown). MS was performed on a FAB mass spectrometer (ZAB-2SE machine; VG Elemental, Winsford, UK). The observed m/z value was 1133.55 (calculated 1133.3), with further peaks at m/z = 1053 and m/z = 951.3, representing the loss of a phosphate group and a C12 fatty acyl group, respectively, under the analytical conditions. 1H NMR (Bruker, Karlsruhe, Germany) and 13C NMR (Bruker; 90 MHz) confirmed the proposed structure.

Natural hexa-acyl lipid A was isolated from deep rough mutant Re-LPS E. coli strain F515 by mild hydrolysis with acetate buffer, purified and converted into the triethylamine salt form following standard procedures [9].

LPS from wild-type E. coli O111:B4 was obtained from Sigma (Buchs, Switzerland).

Synthetic tetra-acyl lipid A (compound 406) was a gift from S. Kusumoto (Department of Chemistry, University of Osaka, Japan).

β⇆α gel to liquid-crystalline phase transition of the acyl chains

Phase behavior was determined by Fourier-transform infrared (FT-IR) spectroscopy on a Nicolet 5-DX spectrometer (Nicolet Instruments, Offenbach, Germany) using 10 mm lipid suspensions, which were prepared by adding Hepes buffer to the appropriate amount of lipid, heating the suspension to 60 °C, vortexing it for 1 min, and recooling it to 4 °C. This procedure was repeated at least twice, and the lipid sample was cooled for 12 h before measurement. The peak position of the symmetric stretching vibration of the methylene groups νs(CH2) was taken as a measure of the acyl chain order, being ≈ 2849–2850 cm−1 in the highly ordered gel phase and 2852–2853 cm−1 in the less-ordered liquid-crystalline phase [10,11].

Conformational analysis of hydrated samples

Measurements on hydrated lipid suspensions were performed as indicated above. An aliquot (10 µL) was spread on a CaF2 crystal and allowed to stand at room temperature until all free water was evaporated. IR spectra were then recorded at room temperature and 37 °C. In the case of overlapping absorption bands, a curve-fit analysis of the original spectra was performed to obtain the single-band components. These were simulated by superpositions of Gaussian and Lorentzian band shapes, with a Gaussian fraction of 0.60 giving the best fits to the experimental band contour.

Supramolecular aggregate structure

The three-dimensional supramolecular structures of lipid/water systems were determined by small-angle X-ray diffraction with synchrotron radiation. For this, OM-174 samples in appropriate concentrations (85% water content, i.e. 15 mg in 85 µL buffer) and in the presence of 5 or 20 mm Mg2+ were temperature-cycled as indicated above and measured on the double-focusing monochromator-mirror camera X33 of the European Molecular Biology Laboratory (EMBL) outstation at the Hamburg synchrotron radiation facility HASYLAB [12]. Diffraction patterns in the range of scattering vectors 0.07 < s < 1 nm−1 (s = 2 sinΘ/λ) were recorded using a linear detector [13]. Wavelength calibration was performed with dry rat tail tendon as standard (periodicity of 65 nm) or tripalmitin (periodicity of 4.06 nm). In the diffraction patterns, the logarithm of the diffracted intensity log I was plotted against s. Spectra were evaluated as described in the literature [5,14], which allowed assignment of the spacing ratios of the main scattering maxima to defined three-dimensional structures.

Molecular area of monolayers at the air/water interface

For determination of the molecular area, monolayer measurements at the air/water interface on a film balance (Langmuir trough) were performed. For this, 3 µL 1 mm OM-174 samples solubilized in chloroform/methanol (10 : 1, v/v) were spread on the aqueous subphase (100 mm NaCl in distilled water). After evaporation of the solvent, the pressure/area isotherms were recorded at a compression rate of 1.8 cm2·min−1 at room temperature. The molecular area was determined at a lateral pressure of 30 mN·m−1, which is assumed to resemble the value in biological membranes [15].

Incorporation of OM-174 into phospholipid membrane

The ability of OM-174 to incorporate itself into target cell membranes or be transported by LPS-binding protein (LBP) was determined from fluorescence resonance energy transfer (FRET) spectroscopic measurements as described elsewhere [16]. Briefly, phospholipid liposomes corresponding to the composition of the macrophage membrane (PL), i.e. phosphatidylcholine, phosphatidylserine, phosphatidylethanolamine, and sphingomyelin in a molar ratio of 1 : 0.4 : 0.7 : 0.5, were double-labeled with the fluorescent dyes N-(7-nitro-2,1,3-benzoxadiazol-4-yl)-PE (NBD-PE) and N-(Rhodamine B sulfonyl)-PE (Rh-PE), which were purchased from Molecular Probes (Eugene, OR, USA). NBD-PE and Rh-PE were dissolved in chloroform and added to the lipids in the chloroform phase to final molar proportions of [PL]/[NBD-PE]/[Rh-PE] = 100 : 1 : 1. The emission wavelength of one dye, the donor (NBD-PE), is in the range of the excitation wavelength of the second dye, the acceptor (Rh-PE). The energy transfer between these two dyes is sensitive to spatial separation. Intercalation of unlabeled molecules into the double-labeled liposomes leads to probe dilution and with that to a decrease in the efficiency of FRET: the emission intensity of the donor increases and that of the acceptor decreases (for the sake of clarity, here we only show the donor emission intensity). A preparation of 900 µL of the double-labeled PL liposomes (0.01 mm) at 37 °C was excited at 470 nm (excitation wavelength of NBD-PE), and the fluorescence emissions of NBD-PE (531 nm) and Rh-PE (593 nm) were adjusted to yield identical intensities and recorded for 50 s under continuous stirring to determine the baseline. After 50 s, unlabeled lipid A or OM-174 (100 µL, 0.1 mm) was added; after a further 50 s, the appropriate amount of LBP was added, and the emission signals were recorded for at least another 300 s.

Determination of endotoxic activity by the chromogenic Limulus test

Endotoxic activity of OM-174 as compared with control LPS E. coli O55:B5 was determined by a quantitative, kinetic assay based on the reactivity of Gram-negative endotoxin with Limulus amebocyte lysate (LAL) [17], using test kits from BioWhittaker (Verviers, Belgium; no. 50-650U).

Stimulation of NO production by murine macrophages derived from bone marrow

Bone marrow was collected from hip, femur and tibia of 6-week-old male C57/Bl6 mice, and centrifuged for 5 min at 200 g. The pellet was washed once, and the cells were resuspended in fresh medium (Dulbecco's modified Eagle's medium). Differentiation of the stem cells to macrophages was induced by cultivating 4 × 105 stem cells·mL−1 for 8 days in medium supplemented with 20% horse serum and 30% L929 supernatant-conditioned medium. L929 mouse fibroblast cell line supernatant is a medium rich in macrophage colony-stimulating factor, which was obtained by the culture of these cells to confluence in petri dishes. After 8 days incubation, most of the cells (> 95%) were macrophages attached to the petri dish. After incubation at 4 °C for 30 min, the cells were detached by pipetting, centrifuged, and washed. The concentration was adjusted to 7 × 105 macrophages per mL DH medium supplemented with 5% fetal bovine serum.

Aqueous preparations of OM-174 and S-form LPS from E. coli, or lipid A from E. coli diluted in 0.2% triethylamine, were diluted serially in a final volume of 100 µL in microtitre plates. A 100-µL portion of macrophage suspension was dispensed into each well. After 22 h at 37 °C under 8% CO2, cell supernatants were tested for their nitrite content, the product of reaction of NO with water, by the Griess reaction.

Stimulation of IL-6 production by human mononuclear cells from blood

For the isolation of mononuclear cells, blood was taken from healthy donors and heparinized (20 IU·mL−1). The heparinized blood was mixed with an equal volume of Hank's balanced-salt solution and centrifuged on a Ficoll density gradient for 40 min (21 °C, 500 g). The interphase layer of mononuclear cells was collected and washed three times in serum-free RPMI 1640 containing 2 mm l-glutamine, 100 U·mL−1 penicillin, and 100 µg·mL−1 streptomycin. The cells were resupended in serum-free medium and the cell number was equilibrated at 5 × 106·mL−1.

For stimulation, 200 µL heparinized whole blood or mononuclear cells (5 × 106·mL−1) per well were transferred to 96-well culture plates. Whole blood or mononuclear cells in serum-free medium were stimulated with the endotoxins. The stimuli were serially diluted in serum-free RPMI 1640 and added to the cultures at 20 µL per well. The cultures were incubated for 24 h at 37 °C and 5% CO2. Supernatants were collected after centrifugation of the culture plates for 10 min at 400 g and stored at −20 °C until determination of cytokine content.

For interleukin (IL)-6 determination, the murine cell line 7TD1 was used, the growth of which is IL-6 dependent. The culture supernatants or recombinant human IL-6 as standard (10 ng·mL−1) were diluted serially in 1 : 4 steps in microtiter plates in 50 µL 7TD1 medium (Dulbecco's modified Eagle's medium + 4.5 g·L−1 glucose, 10% fetal calf serum, 2 mm glutamine, 100 U·mL−1 penicillin, and 100 µg·mL−1 streptomycin). The cells were washed twice (10 min, 300 g) in 7TD1 medium, and adjusted to 8 × 104 cells·mL−1. An aliquot (50 µL) of this cell suspension was incubated with the supernatants or recombinant IL-6 as standard for 72 h at 37 °C in an atmosphere containing 5% CO2. For the determination of cell proliferation, the mitochondrial reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to formazan was measured. Briefly, 10 µL MTT (5 mg·mL−1 in NaCl/Pi) was added to the cells, which were incubated for 4 h with the subsequent addition of 100 µL SDS solution (5% in 50% dimethylformamide). After 2 h incubation at room temperature, the A550 was measured in an ELISA plate photometer. The absolute IL-6 content was determined with respect to the standards.


β⇆α gel to liquid-crystalline phase transition of the acyl chains

From the FT-IR spectroscopic data, for OM-174 a distinct shift of the peak position of νs(CH2) around the phase-transition temperature, Tc = 0 °C, within a narrow temperature interval (from 2850 to 2852 cm−1), and above 10 °C a gradual increase in the peak position up to 2852.5–2853.0 cm−1 can be observed (Fig. 2). For hexa-acyl lipid A from E. coli, Tc is considerably higher (45 °C). Importantly, the state of order (inversely proportional to the fluidity) at 37 °C (vertical line in Fig. 2) is low for OM-174 and high for lipid A from E. coli. The phase-transition behavior in repeated scans of the same batch, and also between different batches, of OM-174 was completely reproducible.

Figure 2.

Peak position of the symmetric stretching vibration of the methylene groups νs(CH2) in relation to temperature for OM-174 and hexa-acyl lipid A from E. coli Re-LPS.

Supramolecular aggregate structure

From the X-ray diffraction spectra (Fig. 3), exhibiting a periodicity of ≈ 8.78 nm and three further reflections at 8.78/√3, 8.78/√4, and 8.78/√7, it can be concluded that OM-174 aggregates assume, under near-physiological conditions, the micellar HI structure. In this structure, the lipid molecules are packed with their backbone on a cylindrical surface with the acyl chains directed inward. The cylinders themselves are hexagonally packed. From the occurrence of the √7 reflection, the existence of a cubic phase can be excluded. Theoretically, the inverted hexagonal structure HII would also match the indexing. However, from considerations of the molecular geometry as well as from diffraction patterns of the HII-structure observed for lipid A [4,5], which display a completely different intensity ratio of the reflections, this possibility can be excluded. Measurements at different temperatures showed slight changes of the spectra in the range 20–70 °C, but no change in the basic structure, and complete reversibility of the original spectrum after recooling to 20 °C (data not shown). From these measurements, it can be deduced that the geometrical shape of the OM-174 molecule is conical–convex, i.e. the cross-section of the hydrophilic moiety is higher than that of the hydrophobic one.

Figure 3.

X-ray small-angle diffraction patterns of OM-174 at a water content of 85%, 5 mm Mg2+, and 20 °C. The spacings of the different diffraction maxima (d = 1/s) fit in the relationships 5.10 nm = 8.78 nm/√3, 4.41 nm = 8.78 nm/√4, and 3.30 nm = 8.78 nm/√7 typical for a hexagonal HI phase

Intramolecular and intermolecular conformations

A detailed FT-IR spectroscopic analysis was performed to study the behavior of various functional groups in the molecule.

The IR spectra in the wavenumber range 1800–900 cm−1(Fig. 4A,B) for OM-174 and lipid A, respectively, clearly indicate characteristic differences: the intensity ratio of the ester carbonyl stretching band (at 1715–1740 cm−1) to that of the amide bands (I at 1650 and II at 1545 cm−1) is considerably lower for OM-174 than for lipid A and corresponds to the smaller number of ester bonds of the former. The different intensity ratios of the vibrational bands resulting from the diglucosamine backbone in the range 1150–1000 cm−1, however, indicate quite different conformations of the respective functional groups. The ester carbonyl stretching vibrational band in the range 1760–1700 cm−1 is split into three single components [18]. The observation that in OM-174 preparations the component at the lowest wavenumber (1710–1717 cm−1) has the highest intensity indicates an unusually high degree of hydration of the ester groups (Fig. 5A). This behavior differs considerably from that of lipid A, for which the band component at ≈ 1733 cm−1 is dominating, corresponding to a less disturbed ester carbonyl vibration (Fig. 5B). Similarly, the antisymmetric stretching vibration of the negatively charged groups νas(PO2) are split into several band components [18,19]. The components between 1250 and 1270 cm−1, which correspond to the phosphate vibration weakly disturbed by water adsorption, are much less extensively expressed than for lipid A but are still present, whereas the component between 1270 and 1300 cm−1 is relatively strong (Fig. 6A). The comparison with the phosphate conformations of natural (Fig. 6B) and synthetic lipid A structures (data not shown) shows that the conformation of the bisphosphoryl backbone of OM-174 is different from that of fully or only partially acylated structures such as the tetra-acyl lipid A precursor IVa (synthetic compound ‘406’), to which, however, the greatest similarity could be observed.

Figure 4.

Infrared spectra of hydrated samples of OM-174 (A) and lipid A from E. coli Re-LPS (B) in the wavenumber range 1800–900 cm−1.

Figure 5.

Infrared spectra of hydrated samples of OM-174 (A) and lipid A from E. coli Re-LPS (B) in the wavenumber range 1760–1700 cm−1 of the ester carbonyl stretching vibration ν(C = O). The single-band components were obtained by curve-fit analysis (see Materials and methods)

Figure 6.

Infrared spectra in the wavenumber range 1320–1180 cm−1 of the antisymmetric stretching vibration of the negatively charged phosphate νas(PO2) for hydrated OM-174 (A) and lipid A from E. coli Re-LPS (B).

Molecular area

To study the influence of the reduced number of acyl chains in OM-174 as compared with lipid A on the area requirement of a single molecule, monolayer measurements at the air/water interface were performed on a film balance (Fig. 7). From these measurements, a value of 0.78 ± 0.04 nm2 for a single OM-174 molecule at 30 mN·m−1 was determined, which is 63% of the value found for lipid A from E. coli (1.23 ± 0.05 nm2). Literature data [20] from crystallographic investigations give a value of 0.185 nm2 for the minimal area requirement of a highly ordered acyl chain. This is in agreement with the experimental value for fully (hexa)acylated lipid A, whereas the above value for OM-174 (0.78 ± 0.04 nm2) is considerably higher than 3 × 0.185 nm2 = 0.555 nm2. From this deviation, a strong influence of the kind and distribution of the acyl chain binding to the polar backbone on the area requirement can be followed.

Figure 7.

Pressure/area isotherms of an OM-174 monolayer as compared with lipid A from E. coli Re-LPS at room temperature. The molecular space requirements of the two lipids were determined at 30 mN·m−1, which is assumed to correspond to the lateral pressure in lipid bilayers [15]. The standard deviation of the molecular areas was calculated from the evaluation of five different batches each.

It can be seen from Fig. 7 that the general course of the pressure/area isotherm for OM-174 is different from that of lipid A. The latter exhibits a steep increase in lateral pressure (from 20 to 40 mN·m−1) with a small reduction in the film area, whereas OM-174 shows a gradual increase in pressure within a larger area reduction. These observations clearly demonstrate that OM-174 monolayers remain in the unordered liquid expanded state up to lateral pressures of 40 mN·m−1 at room temperature, which is in agreement with a very low value for Tc (0 °C, see above). In contrast, the lipid A monolayer undergoes a transition from the unordered liquid expanded to the ordered liquid condensed state at ≈ 10 mN·m−1 (at 20 °C) corresponding to the high Tc value of 45 °C.

Incorporation into phospholipid liposomes

FRET spectroscopy was used to investigate whether OM-174 on its own or in the presence of LBP is able to be incorporated into a phospholipid membrane resembling the composition of the macrophage cytoplasmic membrane (PLΜΦ). The results are shown in Fig. 8. Clearly, the increase in NBD fluorescence intensity for lipid A and OM-174 after addition of LBP indicates rapid intercalation into the PLΜΦ membrane on the timescale of seconds for OM-174 and minutes for lipid A. In the absence of LBP, lipid A is not incorporated at all (data not shown), whereas OM-174 is, although at a much lower rate than in the presence of LBP. Intercalation of OM-174 continues for up to at least 1 h of incubation in the absence of LBP (data not shown). In control experiments, the effect of LBP on PLΜΦ alone, without addition of lipids, was measured and showed a slight increase in NBD fluorescence intensity at the time of addition, but no further increase at later times.

Figure 8.

NBD (donor)-fluorescence intensity in relation to time upon addition of lipid A from E. coli Re-LPS and OM-174 to a phospolipid mixture resembling the composition of the macrophage membrane (PL) and subsequent addition of LBP as compared with a control [NaCl/Pi (PBS)]. Furthermore, OM-174 was also monitored in the absence of LBP.

In vitro LAL activity

Table 1 shows a comparison of the endotoxin activity of the OM-174 preparation with that of a standard E. coli S-form LPS. The results show that the triacylated lipid structure is more than 105 times less reactive in this assay than LPS.

Table 1. LAL activity of OM-174 compared with that of control LPS. EU, endotoxin units.
Product at 1 mg·mL−1LAL activity
LPS E. coli O55:B5107
OM-17436.5 ± 22.1

Stimulation of NO production by murine macrophages derived from bone marrow

OM-174 is a strong inducer of the NO response by murine macrophages (Fig. 9). At concentrations above 1 µg·mL−1, it was a stronger inducer than S-form LPS from O111:B4 or native E. coli lipid A. Below 1 µg·mL−1, the activities of the three substances were similar, although the activity of OM-174 decreased more rapidly with decreasing concentration than that of LPS or native lipid A.

Figure 9.

Stimulation of NO production in murine macrophages from bone marrow after incubation with different concentrations of OM-174, LPS O111:B4 S-form from E. coli, or lipid A from E. coli Re-LPS in the range 0.0005–500 µg·mL−1. The results are expressed as µm nitrite. Basal (unstimulated) values were subtracted in each case. Values represent the mean ± SD from three wells. The data presented are representative of four independent experiments.

Stimulation of IL-6 production by human mononuclear cells

The results of IL-6 production in human mononuclear cells induced by OM-174, lipid A from E. coli, and the synthetic tetra-acyl lipid A 406 are shown in Fig. 10. Clearly, the biological response of the triacyl sample is approximately one order of magnitude lower than that of the hexa-acyl sample. It is surprising, however, that the tetra-acyl lipid A is much less active than OM-174.

Figure 10.

Stimulation of IL-6 production in human mononuclear cells after incubation with different concentrations of OM-174, lipid A from E. coli Re-LPS and the synthetic tetra-acyl lipid A partial structure 406 in the range 1–10 000 ng·mL−1. The data are representative of three independent experiments.


OM-174 is a triacylated partial structure of the hexa-acylated natural lipid A. It has been reported to induce regression of tumors in rats bearing established colon tumors [7,8]. Furthermore, OM-174 was found not to be directly toxic to tumor cells, but the observed effects involved the host-mediated antitumor reaction.

Despite the fact that the backbone, a 1,4′-bisphosphorylated β(1→6)-linked diglucosamine, is identical for OM-174 and lipid A, their overall physicochemical behavior, as well as biological reactivity, is very different, as indicated by the Tc, the fluidity at the physiological temperature (37 °C), the supramolecular conformation of aggregates, and the molecular area within a monolayer. These parameters are, of course, not independent. For example, a reduction in the number of acyl chains leads to a decrease in Tc from 43 °C for the hexa-acyl lipid A to 30 °C for a penta-acyl structure and 15 °C for a tetra-acyl partial structure [18] down to 0 °C for the triacyl OM-174. Concomitantly, the supramolecular structures at 37 °C changes from an inverted cubic (hexa-acyl lipid A [5]), through a lamellar (tetra-acyl lipid A, unpublished data) to a micellar HI (OM-174) structure. The decrease in the number of acyl chains and the concomitant change in the supramolecular conformation are accompanied by the ability to bind increasing amounts of water in the interface region. Furthermore, the CMC, which can only be approximated for hexa-acyl lipid A to be lower than 0.1 µm[21], can be estimated for OM-174 to lie around 0.1 mm (unpublished results). The different physicochemical parameters, thus, vary systematically when going from the hexa-acyl to the triacyl compound.

The expression of biological activity of a given endotoxin is a very complex function of its physicochemical characteristics. It was found that a prerequisite for the induction of cytokine production in monocytes/macrophages is the tendency of the endotoxin molecules to adopt, at least partially, a non-lamellar inverted supramolecular structure implying a conical–concave shape for the individual molecules [22–24]. Endotoxins fulfilling this basic condition can be further distinguished by different acyl chain fluidities at 37 °C (corresponding to different Tc values), i.e. samples with the greatest fluidity at 37 °C exhibit highest biological activity and vice versa [22,23]. Great fluidity at 37 °C per se, however, is not necessarily related to biological activity; for example, lipid A from Rhodobacter capsulatus is largely inactive but has very fluid acyl chains [25]. Whether endotoxin activity results more from the interaction of aggregates or from that of monomers with target cell membranes is a matter of controversy [26–29]. Irrespective of this, we have shown that the molecular shape of the lipid A moiety is a determinant of biological activity. Thus, the individual lipid A molecules in non-lamellar aggregate structures have a non-cylindrical conical–concave shape and may cause strong disturbance in the host cell membrane, for instance in the direct vicinity of a signal-transducing protein. It is known that the lipid A molecules either intercalate by themselves (hydrophobic interaction; unpublished results), or with the help of serum proteins such as LBP [16], into the target cell membranes. For both processes, the number of endotoxin monomers at a given concentration is important for the ability to intercalate into cell membranes. The ‘residence time’ of monomers within micellar structures is of the order of 10−4 s, whereas for lamellar structures this value is ≈ 10+4 s [20]; no data are available for inverted structures. Consequently, in the concentration range used in the biological test systems, all OM-174 aggregates are rapidly dissolved, and the free monomers may readily intercalate into the cell membranes of monocytes/macrophages. This is demonstrated in Fig. 8, which shows that OM-174 is incorporated in the presence of LBP, but also intercalates on its own, into a cell membrane corresponding to the composition of the macrophage membrane on a relatively short timescale. Lipid A and LPS do not intercalate in the absence of LBP at comparable or even higher concentrations [16]. This is apparently related to the fact that samples with lamellar or inverted structures are stable, at least on a timescale of days, their disaggregation and subsequent transport into membranes being catalysed by proteins such as LBP and CD14 [30–33].

Biological activity of OM-174 on murine and human monocytes/macrophages should also be discussed in this context. Triacyl lipid A is a strong inducer of NO (Fig. 9) and a moderate inducer of IL-6 production (Fig. 10). These results may be understood in view of the ability of OM-174 to intercalate into target cell membranes as described above and by its preference for non-lamellar aggregate structures, although micellar rather than inverted, leading to a conical–convex shape of the individual molecules. These conical molecules may exert mechanical stress on the putative signaling protein, as proposed earlier for conical–concave shaped hexa-acyl lipid A [16].

Also the lack of LAL activity of OM-174 (Table 1) may be seen in the light of its unusual physicochemical characteristics. It has been shown that LPS monomers activate the enzymatic LAL reaction cascade to a much smaller extent [27] than aggregates at comparable concentrations. In the LAL activity measurements, for the reasons discussed above, OM-174 is present mainly in the monomeric state, whereas LPS is present in the aggregated state.

Our findings seem to suggest that widely accepted general ideas on the molecular requirements for endotoxic molecules to express biological activity should be revised. It was thought that full biological activity is expressed by a lipid A molecule consisting of a hexa-acylated, bisphosphorylated β1,6-linked d-GlcN disaccharide [2,34]. This was based on observations in various different test systems, in particular cytokine-inducing capacity, LAL activity, pyrogenicity, and lethal toxicity in animal models. All these activities are usually observed for endotoxically active compounds, but not for inactive compounds, except for LAL activity, which is sometimes found for otherwise inactive substances [35]. Here, we describe an endotoxin partial structure with high cytokine-inducing capacity but almost no LAL activity.

In this context, the results of Funatogawa et al. [36] should be mentioned. These authors found cytokine-inducing capacity for a synthetic triacyl monosaccharide lipid A partial structure (GLA-60), although the applied lipid concentration was two orders of magnitude higher (10 µg·mL−1) than that of hexa-acyl lipid A (100 ng·mL−1). This has never previously been observed for a monosaccharide lipid A partial structure (see review [37]). From these results and the findings presented here for OM-174, it should be emphasized, that in vivo studies on pyrogenicity and lethal toxicity should also be included to provide further valuable information on structural requirements for the different aspects of endotoxicity.

In summary, we have characterized a triacyl lipid A partial structure, OM-174, which has the ability to induce cytokine responses but has very low LAL activity. These observations have been interpreted in the light of the molecular and supramolecular characteristics that we have determined, to give insights into the nature of the biologically active unit of lipid A and endotoxin. The findings of biological activity for the triacyl lipid A partial structure OM-174 in terms of its capacity to induce NO and IL-6 production in monocytes/macrophages, in relation to its physicochemical characteristics and, in particular to respective data for hexa-acyl lipid A, emphasize the necessity for comprehensive analysis, as presented here, for an understanding of endotoxin reactions.


We thank G. von Busse for performing IR spectroscopy, D. Koch for monolayer measurements, and S. Kusumoto for the gift of synthetic tetra-acyl lipid A 406. This work was supported by the Deutsche Forschungsgemeinschaft (SFB 367, project B8 and BR 1070/2-1).