- Split View
-
Views
-
Cite
Cite
Alba Silipo, Antonio Molinaro, Paola Cescutti, Emiliano Bedini, Roberto Rizzo, Michelangelo Parrilli, Rosa Lanzetta, Complete structural characterization of the lipid A fraction of a clinical strain of B. cepacia genomovar I lipopolysaccharide, Glycobiology, Volume 15, Issue 5, May 2005, Pages 561–570, https://doi.org/10.1093/glycob/cwi029
- Share Icon Share
Abstract
Burkholderia cepacia, a Gram-negative bacterium ubiquitous in the environment, is a plant pathogen causing soft rot of onions. This microorganism has recently emerged as a life-threatening multiresistant pathogen in cystic fibrosis patients. An important virulence factor of B. cepacia is the lipopolysaccharide (LPS) fraction. Clinical isolates and environmental strains possess LPS of high inflammatory nature, which induces a high level production of cytokines. For the first time, the complete structure of the lipid A components isolated from the lipopolysaccharide fraction of a clinical strain of B. cepacia is described. The structural studies carried out by selective chemical degradations, MS, and NMR spectroscopy revealed multiple species differing in the acylation and in the phosphorylation patterns. The highest mass species was identified as a penta-acylated tetrasaccharide backbone containing two phosphoryl-arabinosamine residues in addition to the archetypal glucosamine disaccharide [Arap4N-l-β-1-P-4-β-D-GlcpN-(1-6)-α-D-GlcpN-1-P-1-β-L-Arap4N]. Lipid A fatty acids substitution was also deduced, with two 3-hydroxytetradecanoic acids 14:0 (3-OH) in ester linkage, and two 3-hydroxyhexadecanoic acids 16:0 (3-OH) in amide linkage, one of which was substituted by a secondary 14:0 residue at its C-3. Other lipid A species present in the mixture and exhibiting lower molecular weight lacked one or both β-L-Arap4N residues.
Introduction
Burkholderia cepacia is a phytopathogenic Gram-negative bacterium that lives in damp or wet environments and is the causative agent of the onion rot known as slippery skin. Nowadays, the study of B. cepacia has become very important since it was recognized as an opportunistic pathogen in patients with cystic fibrosis (CF) (Govan and Deretic, 1996). These bacteria are intrinsically resistant to multiple antibiotics, and their spread among patients with CF occurs by social contact, with hospitalization as a high risk factor for acquisition. Moreover, since B. cepacia is ubiquitous, the environment is also a potential source of infection. The rapid decline of lung functions observed in CF subpopulations infected with B. cepacia complex, leading to a fatal necrotizing pneumonia accompanied with septicemia (the cepacia syndrome), is strongly feared (Blackburn et al., 2004). Recently, taxonomic studies revealed that strains identified as B. cepacia represented a complex of at least nine genomic species (genomovars) that were phenotipically very similar or even indistinguishable. Most of the genomovars of the B. cepacia complex have been assigned a binomial species name: B. cepacia (genomovar I), B. multivorans (genomovar II), B. cenocepacia (genomovar III), B. stabilis (genomovar IV), B. vietnamiensis (genomovar V), B. cepacia genomovar VI, B. ambifaria (genomovar VII), B. anthina (genomovar VIII), and B. pyrrocinia (genomovar IX) (Vandamme et al., 2003).
Clinical isolates of B. cepacia complex all possess lipopolysaccharides (LPSs) of high inflammatory potential (Zughaier et al., 1999). LPSs, also called endotoxins, are the main molecules of the outer membrane of almost all Gram-negative bacteria and play a key role in the pathogenesis and in the toxic manifestation of Gram-negative infection (Alexander and Rietschel, 2001; Mamat et al., 1998; Raetz and Whitfield, 2002). Structurally, they are amphiphilic macromolecules where a hydrophilic heteropolysaccharide (composed of a core oligosaccharide and an O-specific polysaccharide or O-chain) is covalently linked to a lipophilic moiety termed lipid A (Zähringer et al., 1994, 1999) that anchors these macromolecules to the external membrane. O-chain lacking LPSs are termed rough-type LPSs (R-LPSs) or lipooligosaccharides (LOSs). These latter may occur in both wild and laboratory strains possessing mutations in the gene clusters encoding for the O-specific polysaccharide biosynthesis or transfer to core oligosaccharide (Holst, 1999, 2002).
Lipid A is the primary immunostimulator center of LPS, promoting the activation of the innate immune system via the induction of inflammatory cytokine released by human cells (Medzhitov, 2001; Zähringer et al., 1994, 1999). The induction of uncontrolled and excessive level of these cytokines leads to the septic shock. TNF-α and IL-1, released during the first 30–90 min after the exposure to the lipid A/LPS, are the prototypic inflammatory cytokines that mediate many of the immunopathological features of LPS-induced shock. Once released, they activate a second level of inflammatory cascades, including cytokines, lipid mediators, and reactive oxygen species, as well as up-regulating cell-adhesion molecules (Alexander and Rietschel, 2001; Medzhitov, 2001). Lipid A endotoxic activity strongly varies with its primary structure, namely, fatty acids, polar heads, and carbohydrate components (Seydel et al., 2000).
Although the importance of lipid A in the immune response to Gram-negative bacteria is unquestionably recognized (Alexander and Rietschel, 2001), and in particular B. cepacia possesses LPSs of high inflammatory potential, no complete information are available on the chemical structure of lipid A from B. cepacia. Herein we report the complete structure of the lipid A fraction from the LOS isolated from a clinical strain of B. cepacia genomovar I attained by chemical, mass spectrometry (MS), and nuclear magnetic resonance (NMR) analyses.
Results
The LPS fraction obtained by extraction of dried cells (Galanos et al., 1969) showed by sodium dodecyl sulfate (SDS)–polyacrylamide gel electrohphoresis (PAGE) the run to the bottom of the gel, typical of a LOS (R-LPS). The R-LPS was further purified from other cell components by enzymatic hydrolysis with DNase, RNase, and proteinase K followed by dialysis and gel permeation chromatography. No traces of a LPS (S-LPS) were found.
Lipid A was obtained by mild acid hydrolysis of the R-LPS. Compositional analysis revealed 6-substituted-GlcN and terminal-GlcN both with d configuration, 4-amino-4-deoxy-pentose with l configuration, and organic phosphate. The 4-amino-4-deoxy-pentose was identified as 4-amino-4-deoxy-l-arabinose (l-Ara4N) by comparison with the authentic l-Ara4N isolated from LPS of Burkholderia caryophylli (Molinaro et al., 2003) via acetylated O-methyl and O-octyl glycoside derivatives. Fatty acid analysis revealed the presence of (R)-3-hydroxyhexadecanoic (16:0(3-OH)) in amide linkage and (R)-3-hydroxytetradecanoic (14:0(3-OH)) and tetradecanoic acid (14:0) in ester linkage.
Analysis of lipid A
The negative-ion mode matrix assisted laser desorption ionization (MALDI) mass spectrum revealed a complex pattern of molecular ions (species 1–8, Table I and Figure 1a) representative of two distinct lipid A groups, either tetra- or pentaacylated species, according to the acylation pattern.
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1445.4 | 1 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1576.5 | 2 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1671.3 | 3 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1802.3 | 4 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1365.3 | 5 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1P |
1496.5 | 6 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 1P |
1707.3 | 7 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
1933.1 | 8 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1445.4 | 1 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1576.5 | 2 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1671.3 | 3 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1802.3 | 4 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1365.3 | 5 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1P |
1496.5 | 6 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 1P |
1707.3 | 7 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
1933.1 | 8 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1445.4 | 1 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1576.5 | 2 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1671.3 | 3 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1802.3 | 4 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1365.3 | 5 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1P |
1496.5 | 6 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 1P |
1707.3 | 7 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
1933.1 | 8 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1445.4 | 1 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1576.5 | 2 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1671.3 | 3 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1802.3 | 4 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1365.3 | 5 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1P |
1496.5 | 6 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 1P |
1707.3 | 7 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
1933.1 | 8 | Penta-acyl | 2× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
The prominent ion at m/z 1445.4 (species 1) was attributed to a tetraacylated bis-phosphorylated glucosamine disaccharide backbone possessing in ester linkage one 14:0 (3-OH) and one 14:0 chains, and in amide linkage two 16:0 (3-OH) chains, whereas species 2 at m/z 1576.5 carried an additional residue of l-Ara4N (Δm/z = +131). Species 5 and 6 were identified as the corresponding monophosphorylated species (Δm/z = −80). Species 3 and 4 were identified as pentaacylated species carrying an additional 14:0 (3-OH) fatty acid residue with respect to 1 and 2, respectively. Interestingly, MALDI mass spectrum also showed ion peaks that could be attributed to tetra- and pentaacylated species possessing two l-Ara4N residues, namely, 7 (m/z 1707.3) and 8 (m/z 1933.1). All of these peaks were also present in considerable amount as sodium adducts, especially those carrying one or two l-Ara4N residues.
The positive-ion mode MALDI mass spectrum showed pseudomolecular ions [M+Na]+ in accordance with ion peaks analysis of the previous spectrum and, in addition, two oxonium ions (Figure 1b) both arising from the glycosydic cleavage of the nonreducing unit and thus, indicative of the fatty acids distribution between the two GlcN residues. In particular, the ion m/z 933.8 could be ascribed to a triacylated GlcN II oxonium ion carrying one14:0 (3-OH), one 14:0, and a 16:0 (3-OH) residues, whereas the ion at m/z 1064.8 carried an additional Ara4N.
Analysis of de-O-acylated lipid A
An aliquot of lipid A fraction was selectively de-O-acylated by treatment with NH4OH and then analyzed via MALDI mass spectrometry (MS). This approach resulted in the location of the amide-bound acyloxyacyl moieties left unaltered by this mild hydrolysis (Silipo et al., 2002).
The negative ion mode MALDI mass spectrum (Figure 2) showed several ion peaks that could be straightforwardly attributed on the basis of the compositional analysis (Table II). The ion at m/z 1218.7 (species 9) could be ascribed to a triacylated bis-phosphorylated lipid A species with two amide-linked 16:0 (3-OH) residues, one of which was substituted by a C14:0. Species 11 at m/z 1444.4 carried an additional ester linked C14:0 (3-OH) residue (Δm/z = +226). The two main ions at m/z 1349.5 and 1575.5 (10 and 12), were identified as tri- and tetraacylated lipid A species carrying an additional l-Ara4N residue with respect to 9 and 11 ions (Δm/z = +131). These differently acylated components were also present as monophosphorylated species (Δm/z = −80). The spectrum also showed ions named 13 and 14 at m/z 1480.6 and 1706.3, respectively, present in reasonably high amount and carrying two l-Ara4N (Δm/z = +131 from 10 and 12, respectively). The positive-ion mode MALDI mass spectrum (not shown) of the product showed, beside the pseudomolecular ions in accordance with previous analysis, two oxonium ions, a major one at m/z 706.1 and a minor one at m/z 837.6. The former ion was inferred as a GlcN II oxonium ion carrying one 14:0 and one 16:0 (3-OH) residues. The latter ion at m/z 837.6 carried an additional Ara4N.
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1218.7 | 9 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 2P |
1349.5 | 10 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1444.4 | 11 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1575.5 | 12 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1480.6 | 13 | Tri-acyl | 2× 16:0 (3-OH), 1× 4:0, 2× Ara4N, 2P |
1706.3 | 14 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1218.7 | 9 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 2P |
1349.5 | 10 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1444.4 | 11 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1575.5 | 12 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1480.6 | 13 | Tri-acyl | 2× 16:0 (3-OH), 1× 4:0, 2× Ara4N, 2P |
1706.3 | 14 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1218.7 | 9 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 2P |
1349.5 | 10 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1444.4 | 11 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1575.5 | 12 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1480.6 | 13 | Tri-acyl | 2× 16:0 (3-OH), 1× 4:0, 2× Ara4N, 2P |
1706.3 | 14 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
Observed ion (m/z) . | Species . | Acyl substitution . | Proposed fatty acid, phosphate and carbohydrate composition . |
---|---|---|---|
1218.7 | 9 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 2P |
1349.5 | 10 | Tri-acyl | 2× 16:0 (3-OH), 1× 14:0, 1× Ara4N, 2P |
1444.4 | 11 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2P |
1575.5 | 12 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 1×Ara4N, 2P |
1480.6 | 13 | Tri-acyl | 2× 16:0 (3-OH), 1× 4:0, 2× Ara4N, 2P |
1706.3 | 14 | Tetra-acyl | 1× 14:0 (3-OH), 2× 16:0 (3-OH), 1× 14:0, 2× Ara4N, 2P |
Thus, by the analysis of the ammonium-treated lipid A fraction, it is evident that the secondary C14:0 fatty acid residue substitutes the amide bound 16:0 (3-OH) residue on GlcN II.
Analysis of dephosphorylated lipid A
An aliquot of Lipid A fraction was dephosphorylated by treatment with 48% HF and then analyzed via MALDI MS. The positive-ion MALDI mass spectrum showed several pseudomolecular ions [M+Na]+ with the same acylation pattern of the intact lipid A. Furthermore, the total absence of species containing Ara4N (Δm/z +131) clearly suggested that this monosaccharide residues were linked to lipid A backbone via phosphodiester linkage.
In addition, ion peaks were present at low molecular masses, attributable to oxonium ions deriving from the in-source cleavage of the glycoside linkage under high-power laser energy (data not shown). Two ions were identified, at m/z 852.2 and 626.1 (Δm/z +226). The ion at m/z 852.2 was consistent with a triacylated oxonium fragment, carrying one residue of GlcN II, 14:0 (3-OH), 16:0 (3-OH), and 14:0, thus demonstrating once more that the acyloxyacilamide moiety was located on GlcN II. On the other hand, the oxonium ion at m/z 626.1 kept up only the acyloxyacilamide moiety indicating that the ester linked 14:0 (3-OH) residue might lack on the GlcN II.
NMR spectroscopy of lipid A
A combination of homo- and heteronuclear NMR experiments were performed at 343 K in Me2SO-d6 on the intact lipid A aiming at the definition of the carbohydrate backbone and phosphorylation and acylation sites. Double quantum filtered (DQF)-correlation spectroscopy (COSY), total correlation spectroscopy (TOCSY), Rotating frame Overhauser enhancement spectroscopy (ROESY), 13C-1H and 31P-1H heteronuclear single quantum coherence (HSQC) spectra were used to assign 13C, 1H, and 31P resonances.
The NMR data confirmed the heterogeneity of the lipid A family from B. cepacia, revealing a carbohydrate backbone with a variety of acyl chains and phosphate substituents. The solvent Me2SO-d6, in which amide protons signals were clearly visible, was a good option to assign a complex pattern of signals deriving from a lipid A mixture of molecular species (Silipo et al., 2004). In fact, starting from the amide and the anomeric proton signals in the COSY and TOCSY spectra, all the proton resonances of the different spin systems were assigned. In particular, several spin systems attributable to α- and β-GlcN were identified (A–C and F–L, Figure 3, Table III) and for both residues the coupling constant and the chemical shift values were in agreement with the gluco-configuration of pyranose rings in a 4C1 conformation. The anomeric proton resonances of α-GlcN (GlcN I) residues A, B, and C at 5.32, 5.34, and 4.92 ppm correlated in the HSQC spectrum (Figure 4) to anomeric carbon signals resonating at 93.8 (A and B residues) and 90.0 ppm (C residue). The typical high-field carbon resonances of nitrogen bearing C-2 of the A–C residues and the low-field resonances of the correlated H-2 signals confirmed A–C spin systems as 2-deoxy-2-acylamido residues. A less intense spin system accounted for the presence of a fourth α-GlcN residue (B") further acylated at its C-3 carbon, as inferred by the occurrence of its H-3 signal at 5.12 ppm, namely, a ring proton geminal to an acyl group. The low-field 1H and 13C chemical shifts of anomeric signals of A and B residues with respect to the anomeric signals of residue C suggested phosphorylation of the anomeric position only for the first two residues, the third being present as reducing end residue.
Residue . | 1 . | 2/ NH . | 3 . | 4 . | 5ax, eq . | 6ax, eq . |
---|---|---|---|---|---|---|
A | 5.32 | 3.61/7.28 | 3.48 | 3.04 | 3.79 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.4 | 71.2 | 69.6 | 69.3 | 68.7 |
−0.66 | ||||||
B | 5.34 | 3.60/7.38 | 3.45 | 3.06 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 71.7 | 69.6 | 69.3 | 68.7 |
−2.85 | ||||||
B’ | 5.34 | 3.60/7.38 | 5.12 | 3.16 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 69.1 | 69.7 | 69.3 | 68.7 |
−2.85 | ||||||
C | 4.95 | 3.58/7.33 | 3.49 | 3.06 | 4.04 | 3.50/3.64 |
α-GlcN I | 90.0 | 53.4 | 70.3 | 69.6 | 69.5 | 68.7 |
D | 5.38 | 3.56 | 3.87 | 3.40 | 3.54/4.04 | |
β-Ara4N | 94.2 | 70.2 | 70.8 | 50.2 | 58.6 | |
−2.85 | ||||||
E | 5.29 | 3.55 | 3.82 | 3.38 | 3.57/3.97 | |
β-Ara4N | 93.8 | 70.2 | 69.3 | 50.1 | 58.8 | |
−0.66 | ||||||
F | 4.84 | 3.68/7.68 | 4.93 | 4.02 | 3.46 | 3.66/3.48 |
β-GlcN II | 99.5 | 52.6 | 73.7 | 72.5 | 74.2 | 61.8 |
−0.26 | ||||||
G | 4.72 | 3.66/7.68 | 4.93 | 4.03 | 3.49 | 3.70/3.50 |
β-GlcN II | 100.4 | 53.2 | 73.7 | 73.9 | 74.6 | 62.0 |
−0.26 | ||||||
H | 4.68 | 3.63/7.59 | 3.60 | 4.07 | 3.04 | 3.51/3.60 |
β-GlcN II | 100.7 | 53.4 | 67.9 | 70.1 | 73.7 | 61.1 |
−0.66 | ||||||
I | 4.80 | 3.68/7.51 | 3.54 | 3.45 | 3.42 | 3.60/3.48 |
β-GlcN II | 99.7 | 52.6 | 70.2 | 71.6 | 75.6 | 61.9 |
L | 4.78 | 3.66/7.51 | 3.53 | 3.53 | 3.20 | 3.60/3.50 |
β-GlcN II | 97.7 | 52.6 | 70.2 | 70.3 | 74.9 | 61.9 |
Fatty acids | αa/αb | β | γ | (CH2)n | CH3 | |
14:0 (3-OH) | 2.30/2.22 | 3.82 | 1.31 | 1.25 | 0.86 | |
O-linked | 42.1 | 65.9 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OH) | 2.21/2.14 | 3.78 | 1.33 | 1.25 | 0.86 | |
N-linked | 42.6 | 66.5 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OR) | 2.35/2.25 | 5.00 | 1.45 | 1.22 | 0.86 | |
N-linked | 39.6 | 69.8 | 32.3 | 28.2 | 12.9 |
Residue . | 1 . | 2/ NH . | 3 . | 4 . | 5ax, eq . | 6ax, eq . |
---|---|---|---|---|---|---|
A | 5.32 | 3.61/7.28 | 3.48 | 3.04 | 3.79 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.4 | 71.2 | 69.6 | 69.3 | 68.7 |
−0.66 | ||||||
B | 5.34 | 3.60/7.38 | 3.45 | 3.06 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 71.7 | 69.6 | 69.3 | 68.7 |
−2.85 | ||||||
B’ | 5.34 | 3.60/7.38 | 5.12 | 3.16 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 69.1 | 69.7 | 69.3 | 68.7 |
−2.85 | ||||||
C | 4.95 | 3.58/7.33 | 3.49 | 3.06 | 4.04 | 3.50/3.64 |
α-GlcN I | 90.0 | 53.4 | 70.3 | 69.6 | 69.5 | 68.7 |
D | 5.38 | 3.56 | 3.87 | 3.40 | 3.54/4.04 | |
β-Ara4N | 94.2 | 70.2 | 70.8 | 50.2 | 58.6 | |
−2.85 | ||||||
E | 5.29 | 3.55 | 3.82 | 3.38 | 3.57/3.97 | |
β-Ara4N | 93.8 | 70.2 | 69.3 | 50.1 | 58.8 | |
−0.66 | ||||||
F | 4.84 | 3.68/7.68 | 4.93 | 4.02 | 3.46 | 3.66/3.48 |
β-GlcN II | 99.5 | 52.6 | 73.7 | 72.5 | 74.2 | 61.8 |
−0.26 | ||||||
G | 4.72 | 3.66/7.68 | 4.93 | 4.03 | 3.49 | 3.70/3.50 |
β-GlcN II | 100.4 | 53.2 | 73.7 | 73.9 | 74.6 | 62.0 |
−0.26 | ||||||
H | 4.68 | 3.63/7.59 | 3.60 | 4.07 | 3.04 | 3.51/3.60 |
β-GlcN II | 100.7 | 53.4 | 67.9 | 70.1 | 73.7 | 61.1 |
−0.66 | ||||||
I | 4.80 | 3.68/7.51 | 3.54 | 3.45 | 3.42 | 3.60/3.48 |
β-GlcN II | 99.7 | 52.6 | 70.2 | 71.6 | 75.6 | 61.9 |
L | 4.78 | 3.66/7.51 | 3.53 | 3.53 | 3.20 | 3.60/3.50 |
β-GlcN II | 97.7 | 52.6 | 70.2 | 70.3 | 74.9 | 61.9 |
Fatty acids | αa/αb | β | γ | (CH2)n | CH3 | |
14:0 (3-OH) | 2.30/2.22 | 3.82 | 1.31 | 1.25 | 0.86 | |
O-linked | 42.1 | 65.9 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OH) | 2.21/2.14 | 3.78 | 1.33 | 1.25 | 0.86 | |
N-linked | 42.6 | 66.5 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OR) | 2.35/2.25 | 5.00 | 1.45 | 1.22 | 0.86 | |
N-linked | 39.6 | 69.8 | 32.3 | 28.2 | 12.9 |
Residue . | 1 . | 2/ NH . | 3 . | 4 . | 5ax, eq . | 6ax, eq . |
---|---|---|---|---|---|---|
A | 5.32 | 3.61/7.28 | 3.48 | 3.04 | 3.79 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.4 | 71.2 | 69.6 | 69.3 | 68.7 |
−0.66 | ||||||
B | 5.34 | 3.60/7.38 | 3.45 | 3.06 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 71.7 | 69.6 | 69.3 | 68.7 |
−2.85 | ||||||
B’ | 5.34 | 3.60/7.38 | 5.12 | 3.16 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 69.1 | 69.7 | 69.3 | 68.7 |
−2.85 | ||||||
C | 4.95 | 3.58/7.33 | 3.49 | 3.06 | 4.04 | 3.50/3.64 |
α-GlcN I | 90.0 | 53.4 | 70.3 | 69.6 | 69.5 | 68.7 |
D | 5.38 | 3.56 | 3.87 | 3.40 | 3.54/4.04 | |
β-Ara4N | 94.2 | 70.2 | 70.8 | 50.2 | 58.6 | |
−2.85 | ||||||
E | 5.29 | 3.55 | 3.82 | 3.38 | 3.57/3.97 | |
β-Ara4N | 93.8 | 70.2 | 69.3 | 50.1 | 58.8 | |
−0.66 | ||||||
F | 4.84 | 3.68/7.68 | 4.93 | 4.02 | 3.46 | 3.66/3.48 |
β-GlcN II | 99.5 | 52.6 | 73.7 | 72.5 | 74.2 | 61.8 |
−0.26 | ||||||
G | 4.72 | 3.66/7.68 | 4.93 | 4.03 | 3.49 | 3.70/3.50 |
β-GlcN II | 100.4 | 53.2 | 73.7 | 73.9 | 74.6 | 62.0 |
−0.26 | ||||||
H | 4.68 | 3.63/7.59 | 3.60 | 4.07 | 3.04 | 3.51/3.60 |
β-GlcN II | 100.7 | 53.4 | 67.9 | 70.1 | 73.7 | 61.1 |
−0.66 | ||||||
I | 4.80 | 3.68/7.51 | 3.54 | 3.45 | 3.42 | 3.60/3.48 |
β-GlcN II | 99.7 | 52.6 | 70.2 | 71.6 | 75.6 | 61.9 |
L | 4.78 | 3.66/7.51 | 3.53 | 3.53 | 3.20 | 3.60/3.50 |
β-GlcN II | 97.7 | 52.6 | 70.2 | 70.3 | 74.9 | 61.9 |
Fatty acids | αa/αb | β | γ | (CH2)n | CH3 | |
14:0 (3-OH) | 2.30/2.22 | 3.82 | 1.31 | 1.25 | 0.86 | |
O-linked | 42.1 | 65.9 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OH) | 2.21/2.14 | 3.78 | 1.33 | 1.25 | 0.86 | |
N-linked | 42.6 | 66.5 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OR) | 2.35/2.25 | 5.00 | 1.45 | 1.22 | 0.86 | |
N-linked | 39.6 | 69.8 | 32.3 | 28.2 | 12.9 |
Residue . | 1 . | 2/ NH . | 3 . | 4 . | 5ax, eq . | 6ax, eq . |
---|---|---|---|---|---|---|
A | 5.32 | 3.61/7.28 | 3.48 | 3.04 | 3.79 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.4 | 71.2 | 69.6 | 69.3 | 68.7 |
−0.66 | ||||||
B | 5.34 | 3.60/7.38 | 3.45 | 3.06 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 71.7 | 69.6 | 69.3 | 68.7 |
−2.85 | ||||||
B’ | 5.34 | 3.60/7.38 | 5.12 | 3.16 | 3.80 | 3.50/3.66 |
α-GlcN I | 93.8 | 53.5 | 69.1 | 69.7 | 69.3 | 68.7 |
−2.85 | ||||||
C | 4.95 | 3.58/7.33 | 3.49 | 3.06 | 4.04 | 3.50/3.64 |
α-GlcN I | 90.0 | 53.4 | 70.3 | 69.6 | 69.5 | 68.7 |
D | 5.38 | 3.56 | 3.87 | 3.40 | 3.54/4.04 | |
β-Ara4N | 94.2 | 70.2 | 70.8 | 50.2 | 58.6 | |
−2.85 | ||||||
E | 5.29 | 3.55 | 3.82 | 3.38 | 3.57/3.97 | |
β-Ara4N | 93.8 | 70.2 | 69.3 | 50.1 | 58.8 | |
−0.66 | ||||||
F | 4.84 | 3.68/7.68 | 4.93 | 4.02 | 3.46 | 3.66/3.48 |
β-GlcN II | 99.5 | 52.6 | 73.7 | 72.5 | 74.2 | 61.8 |
−0.26 | ||||||
G | 4.72 | 3.66/7.68 | 4.93 | 4.03 | 3.49 | 3.70/3.50 |
β-GlcN II | 100.4 | 53.2 | 73.7 | 73.9 | 74.6 | 62.0 |
−0.26 | ||||||
H | 4.68 | 3.63/7.59 | 3.60 | 4.07 | 3.04 | 3.51/3.60 |
β-GlcN II | 100.7 | 53.4 | 67.9 | 70.1 | 73.7 | 61.1 |
−0.66 | ||||||
I | 4.80 | 3.68/7.51 | 3.54 | 3.45 | 3.42 | 3.60/3.48 |
β-GlcN II | 99.7 | 52.6 | 70.2 | 71.6 | 75.6 | 61.9 |
L | 4.78 | 3.66/7.51 | 3.53 | 3.53 | 3.20 | 3.60/3.50 |
β-GlcN II | 97.7 | 52.6 | 70.2 | 70.3 | 74.9 | 61.9 |
Fatty acids | αa/αb | β | γ | (CH2)n | CH3 | |
14:0 (3-OH) | 2.30/2.22 | 3.82 | 1.31 | 1.25 | 0.86 | |
O-linked | 42.1 | 65.9 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OH) | 2.21/2.14 | 3.78 | 1.33 | 1.25 | 0.86 | |
N-linked | 42.6 | 66.5 | 36.6 | 28.2 | 12.9 | |
16:0 (3-OR) | 2.35/2.25 | 5.00 | 1.45 | 1.22 | 0.86 | |
N-linked | 39.6 | 69.8 | 32.3 | 28.2 | 12.9 |
Five spin systems could be recognized as β-GlcN (GlcN II) residues (F–L, Table III) and fully assigned with the TOCSY and COSY spectra. The anomeric proton and carbon resonances, together with the intraresidual nuclear Overhauser effect (NOE), found in the ROESY spectrum, between H-1 and H-3 and H-5 of each residue, supported the β-configuration of these residues. The C-2 resonances of F–L possessed the typical chemical shift of nitrogen-bearing carbons, whereas the low-field chemical shift of their correlated proton signals indicated N-acylation at these positions. On the other hand, the H-3 signals of F and G were shifted down-field at 4.93 ppm, indicating O-acylation at these positions for only these two residues, whereas H-3 signals of H-L spin systems were found in the typical ring proton region of the spectrum.
The β-(1→6) disaccharide backbone was eventually recognized by the ROESY spectrum in which a strong interresidual NOE contact of the anomeric proton signals of the β-GlcNs with H-6 signals of the α-GlcNs A–C (3.50–3.66 ppm) was present. This was in agreement with methylation data and HSQC spectrum in which all C-6 signals (68.7 ppm) of spin systems A–C showed downfield displacement due to glycosylation (see previous discussion and Table III).
In the NMR spectra, two other carbohydrate spin systems were found with anomeric proton signals at 5.38 (D) and 5.29 ppm (E). The examination of the COSY and TOCSY spectra allowed the assignment of H-1 to H-5 proton resonances of both residues, whereas no H-6 signals were found, in accordance with the presence of two pentopyranose residues. Chemical shifts, coupling constant values and the NOE contact found in the ROESY spectrum between H-1 of both residues and H-3 and H-5ax were typical of β-configurations. Moreover, both the 3JH,H ring values, in particular 3JH3,H4 and 3JH4,H5 values of 3 and 1 Hz, respectively, and the chemical shifts were diagnostic of two arabino-configurations in a 1C4 conformation. The carbon resonances, assigned from inspection of the HSQC spectrum, were typical of unsubstituted monosaccharide residues evidently present in dissimilar chemical/magnetic environment, and in particular, the high-field carbon resonance of C-4 of both D and E identified them as 4-deoxy-4-amino-arabinose. These results were in agreement with compositional analysis in which Ara4N was found.
The 31P-NMR spectrum showed three different groups of signals in different regions of the spectrum, at −0.26 ppm, −0.66 ppm, and −2.85 ppm. A 31P-1H HSQC (Figure 5) experiment allowed the recognition and location of these three phosphate groups. The phosphorous resonance at −0.26 ppm correlated to two proton signals at 4.02 and 4.03 ppm identified as H-4 of two residues of β-GlcN II, namely F and G. At higher field, the phosphorous signal at −0.66 ppm connected to a proton at 5.32 ppm attributed to H-1 of α-GlcN I A. The same 31P signal was also accountable for a double correlation with two very different proton signals at 5.29 ppm (H-1 E) and 4.07 ppm (H-4 H). Even though the resolution of the spectra was not high enough to allow a discrimination between H-1 E and H-1 A, the double correlation in the 31P-1H HSQC spectrum was plainly attributed to H-1 E of the Ara4N residue attached to C-4 of GlcN II of lipid A via phosphodiester bond (see also MS data). Another signal responsible for a phosphodiester bond and very high in intensity was found at −2.85 ppm in the same spectrum between two anomeric protons at 5.34 and 5.38 ppm, α-GlcN B and β-Ara4N D spin systems. Thus, the other Ara4N residue was linked to C-1 of α-GlcN B via a phosphate group.
On the basis of the NMR data, the phosphate substitution pattern can be resumed in this way. As for GlcN I of lipid A, spin system A possesses a phosphate group at anomeric position, whereas spin system B possesses a phosphoryl-arabinosamine group (D residue) linked in a phosphodiester bond, spin system C does not possess phosphate at all. As for GlcN II, spin systems F and G possess a phosphate group at C-4 position, spin system H possesses a phosphoryl-arabinosamine group (E residue) linked at C-4 in a phosphodiester bond, and I and L do not link any phosphate group.
A further classification can be drawn according to the acylation pattern. GlcN I residues are mostly not O-acylated at position 3 (species A–C); only residue B" shows this feature. GlcN II residues F and G are both O-acylated at position 3, whereas the other three residues, H, I, and L, are not.
All the NMR data, in particular the presence of multiple spin systems accounting for only two monosaccharide residues (GlcN I and II), are explainable with the coexistence of various lipid A disaccharide backbones differing for acylation, phosphorylation, and further carbohydrate substituents. The lipid A fraction of B. cepacia can be defined at this stage as a heterogeneous mixture of species all sharing the same bis-N-acylated disaccharide backbone [β-D-GlcpN-(1Æ6)-α-d-GlcpN].
The NMR spectra also showed spin systems attributable to the acyl chains. In the HSQC spectrum, two cross-peaks between the signals at 65.9 and 3.82 and at 66.5 and 3.78 ppm were attributed to H3/C3 of 14:0(3-OH) and 16:0(3-OH). In the COSY and TOCSY spectra these protons at 3.82 and 3.78 ppm correlated with the two diastereotopic α-methylene protons and with the γ-methylene protons of the acyl chains (Table III). A cross-peak between a proton at 5.00 ppm and a carbon at 69.8 ppm was identified as the H-3 proton signal of 16:0(3-O-acyl) moiety of 16:0(3-O-(14:0)). This proton signal correlated with the α-methylene and the γ-methylene signal of the acyl chains, at 2.35/2.25 and ppm 1.45 ppm, respectively.
Concluding structural remarks
According to NMR and MS data on intact and selectively degraded lipid A, the ion peaks of MALDI mass spectrum of native lipid A (Figure 1a) could be now accurately assigned. The lipid A family is constituted by a heterogeneous mixture of tetra- and pentaacylated species differing in the phosphorylation pattern (Figure 6). The highest mass ion peak, 8 at m/z 1933.1 ([M+Na+] at m/z 1955.1) is consistent with a pentaacylated disaccharide backbone substituted by two phosphoryl-arabinosamine residues [β-l-Arap4N-1-P→4-β-d-GlcpN-(1→6)-α-d-GlcpN-1→P-1-β-l-Arap4N], carrying in ester linkage two 14:0(3-OH) chains and in amide linkage two 16:0(3-OH) chains, one of which, on the GlcN II, is further substituted by a secondary fatty acid, a 14:0 residue. Species 4 (m/z 1802.3) lacks Ara4N on GlcN II, whereas species 3 (m/z 1671.3) lacks both Ara4N residues. The ion 7 consists in the same saccharide backbone of 8 but lacks a 14:0(3-OH) fatty acid. The two prominent tetraacylated species 2 (m/z 1576.5) and 1 (m/z 1445.4) lack one and two residues of Ara4N, respectively. Monophosphorylated species (5 and 6) lacking the anomeric phosphate group (Δm/z 80) in respect to 1 and 2 are also present.
Discussion
Microorganisms from the B. cepacia complex are important pathogens in CF and are associated with increased rates of sepsis and death. They are made up of nine closely related species known as genomovars. A wealth of data indicate that the LPS fraction is one of major virulence factors (Zughaier et al., 1999), but only few are related to the structure of LPS components, in particular to the inner core–lipid A region. B. cepacia LPS possesses in the inner core the peculiar trisaccharide Ara4N → d-glycero- d-talo-oct-2-ulosonic acid (Ko) →3-deoxy-d-manno-oct-2-ulosonic acid (Kdo) (Gronow et al., 2003; Isshiki et al., 1998, 2003). Ko monosaccharide has been found, as component of bacterial LPS in only a small number of bacteria: Acinetobacter (Vinogradov et al., 1997), Yersinia pestis (Vinogradov et al., 2002), and B. caryophylli (Molinaro, unpublished data). However, the enzymatic steps for the biosynthesis and incorporation of Ko are not established yet.
To date, complete structural data on lipid A of B. cepacia are lacking, whereas, in our opinion, these are mandatory for the comprehension of the unique biological activity of this molecule. Previously, a partial structural elucidation of the de-O-acylated LPS from the deep-rough mutant Ko2b from B. cepacia type strain ATCC 25416 was accomplished by MS and gas chromatography (GC)-MS analyses (Gronow et al., 2003). The LPS, once more, possesses the inner core trisaccharide Ara4N→Ko→Kdo linked to lipid A moiety that contains two Ara4N, two GlcN, two phosphate, and two C16:0 (3-OH) residues, all in stoichiometric amounts. However, no information are given on the precise chemical structure of lipid A, namely, the phosphate substitution, Ara4N residues location, GlcN absolute configuration, site of linkages and acylation, and fatty acids location and their absolute configuration.
The exact chemical structure of lipid A from B. cepacia genomovar I LPS has been determined within this frame. In free-living culture, the clinical isolated B. cepacia genomovar I expresses only R-form LPS, namely, a LOS, which is made of a heterogeneous lipid A consisting of several chemical species whose carbohydrate backbone is differently acylated and phosphorylated. In addition, it possesses a further carbohydrate substitution, consisting of one or two Ara4N residues. From a chemiotaxonomy point of view, the fatty acids substitution is perfectly corresponding to the one found in the lipid A of B. caryophylli (Molinaro et al., 2002), the only Burkholderia lipid A elucidated thus far.
During animal infections, lipid A activates the innate immune system through interaction with Toll-like receptor (TLR-4) on macrophage and endothelial cell surfaces (Medzhitov et al., 2001).
An important component of host response is the production of cationic peptides, which act as antimicrobial molecules by linking to lipid A. An advantageous reply of the bacterium is a covalent modification of lipid A consisting in the addition of ethanolamine or Ara4N, thus reducing the net charge and ion attraction for the peptides. Likewise, the lipid A backbone substitution by such chemical appendages is a chemical strategy of Gram-negative bacteria during Gram-positive attack to avoid the binding to lipid A of cationic bacterial peptides, as polymyxin (Zhou et al., 2001).
The biosynthesis of the attachment of l-Arap4N to lipid A in LPSs of S. enterica and E. coli has been elucidated only recently (Raetz and Whitfield, 2002; Trent et al., 2001a,b). An inner membrane enzyme (ArnT) was identified to be expressed in polymyxin-resistant mutants that adds one or two l-Arap4N residue to lipid A or its precursors. In E. coli, biosynthesis of the core region begins with the attachment of two Kdo residues to precursor IVA, the tetraacylated and bis-phosphorylated GlcN-disaccharide, which is performed by one Kdo-transferase WaaA (KdtA) and which results in Kdo2-lipid IVA. Completion of lipid A employs two additional acylation steps of Kdo2-lipid IVA, and then the Kdo residue that is attached to lipid A is substituted at O-5 by l,d-Hepp, followed by further steps of core biosynthesis (Raetz and Whitfield, 2002). Most interestingly, ArnT adds two l-Arap4N residues to Kdo2-lipid IVA and to Kdo2-lipid A (Re-LPS), whereas to lipid IVA substrate it is able to transfer only one Ara4N residue to anomeric phosphate of GlcN I (Trent et al., 2001a,b). Both Escherica coli and Salmonella enterica lipid As show substitution by l-Arap4N residues or by other chemical covalent appendages as ethanolamine only under particular stress conditions, for example, E. coli when grown in ammonium metavanadate and S. enterica in mutants resistant to polymyxin (Trent et al., 2001a,b; Zhou et al., 2001). Conversely, B. cepacia expresses a very intricate lipid A when isolated from patients affected by CF; consequently, the esterification of phosphate by l-Arap4N can be deemed critical for bacterial survival in pulmonary airways.
Interestingly, the other frequent cationic binding group (e.g., ethanolamine) has not been found in lipid A from B. cepacia, thus pointing to a unique role for l-Arap4N. Hence, the transferase enabling the attachment of l-Arap4N residues to lipid A and whichever earlier implicated enzymatic step may be a potential novel target for the future developments of drugs against B. cepacia.
Materials and methods
Bacterial growth and LPS extraction
B. cepacia mucoidstrain BTS7 (Lagatolla et al., 2002) was isolated from a CF patient attending the Regional Centre for Cystic Fibrosis, Children Hospital Burlo Garofolo, Trieste. The bacteria were cultivated in liquid Luria Bertani medium enriched with 0.2% glucose for 3 days at 30°C with shaking. The broth culture was centrifuged; the cells were washed with 0.5% NaCl and freeze-dried (yield: 11 g dried cells).
LOS was extracted by the phenol/chloroform/petroleum ether procedure as described (Galanos et al., 1969) and revealed by SDS–PAGE 12%. The gel was stained with silver nitrate for detection of LOSs (Kittelberger and Hilbink, 1993). To get rid of all the cell contaminants, the LPS fraction was further subjected to enzymatic hydrolysis with RNase, DNase, and proteinase K followed by a size-exclusion chromatography on Sephacryl S-300 in 50 mM NH4CO3 (yield 170 mg, 1.5% of dried cells).
Preparation of Lipid A, de-O-acylated Lipid A, and dephosphorylated Lipid A
Free lipid A was obtained by treatment of LOS (10 mg) with 0.1 m sodium acetate buffer (pH 4.4) containing 1% SDS (100°C, 3 h). The solution was then lyophilized, treated with 2 M HCl:EtOH (1:100, v/v)to remove the SDS, evaporated, dissolved in water, and ultracentrifuged (100.000 ×g, 4°C, 90 min). The obtained precipitate (free lipid A) was washed with water (yield: 4.8 mg, 48% of the LOS). Thin-layer chromatography (TLC) of lipid A was carried out on Silica Gel 60 TLC plates (20 × 20 cm, 0.25 μm thickness), eluted with CHCl3/MeOH/H2O (100:75:15, by volume) and CHCl3/MeOH/H2O/triethylamine (30:12:2:0.1). Compounds were visualized by spraying the plate with 10% ethanolic H2SO4 and charring.
Lipid A was partially de-O-acylated by treatment with 12% NH4OH at 25°C for 18 h (Silipo et al., 2002). Lipid A was dephosphorylated by treatment with 48% aqueous HF (4°C, 48 h).
General and analytical procedures
Monosaccharides were identified as acetylated O-methyl glycosides derivatives. After methanolysis (2 M HCl/MeOH, 85°C, 24 h) and acetylation with acetic anhydride in pyridine (85°C, 30 min) the sample was analyzed by GC-MS. The absolute configuration of the monosaccharides was obtained according to the published method (Leontein and Lönngren, 1978).
Methylation analysis was carried out on dephosphorylated product: briefly, the sample (1 mg) was kept at 4°C, 48 h, in HF 48% (200 μl) and then evaporated under a stream of nitrogen. Methylation was performed with methyl iodide as published (Ciucanu and Kerek, 1984). The hydrolysis of the methylated sugar backbone was performed with 4 M trifluoracetic acid (100°C, 4 h), and the partially methylated product, after reduction with NaBH4, was converted into alditol acetates with acetic anhydride in pyridine at 80°C for 30 min and analyzed by gas-liquid chromatography (GLC)-MS as described.
Total fatty acid content was obtained by acid hydrolysis lipid A. Briefly, lipid A was first treated with 4 M HCl (4 h, 100°C) and then with 5 M NaOH (30 min, 100°C). Fatty acids were then extracted in CHCl3, methylated with diazomethane, and analyzed by GLC-MS. The ester-bound fatty acids were selectively released by base-catalyzed hydrolysis with NaOH 0.5M/MeOH (1:1 v/v, 85°C, 2 h), then the product was acidified, extracted in CHCl3, methylated with diazomethane, and analyzed by GLC-MS. The absolute configuration of fatty acids was determined as described (Rietschel, 1976).
All GLC analyses were performed on a Hewlett-Packard 5890, SPB-5 capillary column (0.25 mm × 30 m, Supelco, Bellefonte, PA). For sugar methylation analysis and O-methyl glycosides derivatives the temperature program was: 150°C for 2 min, then 2°C min−1 to 200°C for 0 min, then 10°C min−1 to 260°C for 11 min, then 8°C min−1 to 300°C for 20 min; for absolute configuration analysis, it was: 150°C for 8 min, then 2°C min−1 to 200°C for 0 min, then 6°C min−1 to 260°C for 5 min. For fatty acids analysis the temperature program was 80°C for 2 min, then 8°C min−1 to 300°C for 15 min.
MS
MALDI time-of flight (TOF) MS of native and TLC-separated fractions of lipid A, partially de-O-acylated and dephosphorylated lipid A was performed with a Voyager DE-PRO MALDI-TOF mass spectrometer (Applied Biosystems, Foster City, CA), in the positive- and negative-ion mode at an acceleration voltage of 24 kV. In general, compounds were dissolved in CHCl3/MeOH 4:1 at a concentration of 1 μg/μl. One microliter of the sample was then mixed (directly on the metallic sample surface) with 1 μl of a 20 mg/ml solution of 2,5-dihydroxybenzoic acid (Aldrich, Steinheim, Germany) in acetonitrile/0.1 M trifluoroacetic acid 7:3. The mass spectra shown are the average of at least 50 single scans. An external mass scale calibration was performed with similar compounds of known chemical structure (i.e., lipid A from E. coli).
NMR spectroscopy
1H- and 13C-NMR spectra of lipid A were recorded on Varian INOVA 500 equipped with a reverse probe at 343 K in DMSO-d6. 13C and 1H chemical shifts are expressed in δ relative to dimethyl sulfoxide (δH 2.49, δC 39.7). 31P-NMR spectra were recorded on a Bruker DRX 400 spectrometer equipped with a reverse probe at 343 K in DMSO-d6. Aqueous 85% phosphoric acid was used as external reference (0.00 ppm) for 31P-NMR spectroscopy.
ROESY was measured using data sets (t1 × t2) of 2048 × 1024 points, and 16 scans were acquired. A mixing time of 300 ms was employed. DQF phase-sensitive COSY experiment was performed with 0.258 s acquisition time using data sets of 4096 × 1024 points, and 64 scans were acquired. The TOCSY experiment was performed with a spinlock time of 120 ms, using data sets (t1 × t2) of 4096 × 1024 points, and 16 scans were acquired. In all homonuclear experiments the data matrix was zero-filled in the F1 dimension to give a matrix of 4096 × 2048 points and was resolution-enhanced in both dimensions by a shifted sine-bell function before Fourier transformation. Coupling constants were determined on a first-order basis from 2D phase-sensitive DQF-COSY (Rance et al., 1983). The HSQC experiment spectrum was measured in the 1H-detected mode via single quantum coherence with proton decoupling in the 13C domain, using data sets of 2048 × 512 points, and 64 scans were acquired for each t1 value. The experiments were carried out in the phase-sensitive mode according to the method of States et al. (1982) and the data matrix was extended to 2048 × 1024 points using forward linear prediction extrapolation (Stern et al., 1996).
Dr. Y. Herasimenka is gratefully acknowledged for the cultivation of the B. cepacia cells. NMR experiments were carried out on a 500 MHz spectrometer of Consortium INCA (L488/92, Cluster 11) at Centro Interdipartimentale Metodologie Chimico Fisiche, Università di Napoli Federico II.
References
Author notes
2Dipartimento di Chimica Organica e Biochimica, Università degli Studi di Napoli Federico II, Via Cintia 4, I-80126 Napoli, Italy, and 3Dipartimento di Biochimica, Biofisica e Chimica delle Macromolecole Università degli Studi di Trieste, Via L. Giorgieri 1, I-34127 Trieste, Italy