CHEMBIOCHEM
FULL PAPERS
DOI: 10.1002/cbic.201300062
Chemistry and Biology of the Potent Endotoxin from
a Burkholderia dolosa Clinical Isolate from a Cystic Fibrosis
Patient
Flaviana Di Lorenzo,[a] Luisa Sturiale,[b] Angelo Palmigiano,[b] Luigi Lembo- Fazio,[c]
Ida Paciello,[c] Carla P. Coutinho,[d] Isabel S-Correia,[d] MariaLina Bernardini,[c, e]
Rosa Lanzetta,[a] Domenico Garozzo,[b] Alba Silipo,[a] and Antonio Molinaro*[a]
This is the first report of the chemical and biological properties
of the lipooligosaccharide (LOS) endotoxin isolated from Bur-
kholderia dolosa IST4208, an isolate recovered from a cystic
fibrosis (CF) patient in a Portuguese CF center. B. dolosa is
a member of the Burkholderia cepacia complex, a group of
closely related species that are highly problematic and oppor-
tunistic pathogens in CF. B. dolosa infection leads to accelerat-
Introduction
Cystic fibrosis (CF) is a lethal autosomal recessive disorder
caused by mutations in the cystic fibrosis transmembrane reg-
ulator (CFTR) gene. These mutations alter host pulmonary de-
fenses, thereby allowing colonization by several opportunistic
bacteria. In fact, as a consequence of the genetic-based pathol-
ogy the major source of morbidity and mortality in people
with CF is chronic endobronchial infection, most commonly
with Pseudomonas aeruginosa, Staphylococcus aureus, and Bur-
kholderia cepacia complex (Bcc) bacteria.[1] Chronic colonization
by Bcc is associated with poor prognosis and decreased life
[a] Dr. F. D. Lorenzo, Prof. R. Lanzetta, Dr. A. Silipo, Prof. A. Molinaro
Dipartimento di Scienze Chimiche
Universit di Napoli Federico II, Complesso Universitario Monte S. Angelo
Via Cintia 4, 80126 Napoli (Italy)
E-mail: molinaro@unina.it
[b] Dr. L. Sturiale, Dr. A. Palmigiano, Dr. D. Garozzo
Istituto di Chimica e Tecnologia dei Polimeri–ICTP–CNR
Via P. Gaifami 18, 95126 Catania (Italy)
[c] Dr. L. L. Fazio, Dr. I. Paciello, Prof. M. Bernardini
Dipartimento di Biologia e Biotecnologie “C. Darwin”
Sapienza–Universit di Roma
Piazzale Aldo Moro, 5-00185 Roma (Italy)
[d] Dr. C. P. Coutinho, Prof. I. S-Correia
IBB—Institute for Biotechnology and Bioengineering
Centre for Biological and Chemical Engineering
Department of Bioengineering
Instituto Superior Tcnico, Technical University of Lisbon
Av. Rovisco Pais, 1049-001 Lisbon (Portugal)
[e] Prof. M. Bernardini
Istituto Pasteur-Fondazione Cenci Bolognetti
Sapienza–Universit di Roma
Piazzale Aldo Moro 5, 00185 Roma (Italy)
00185 Roma (Italy)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201300062.
 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ed loss of lung function and decreased survival. The structural
determination of its endotoxin was achieved using a combina-
tion of chemistry and spectroscopy, and has revealed a novel
endotoxin structure. The purified LOS was tested for its immu-
nostimulatory activity on human HEK 293 cells expressing TLR-
4, MD-2, and CD-14. In these assays, the LOS showed strong
proinflammatory activity.
expectancy, compared with infection of the other two major
pathogens;[2] thus it is becoming of increasing concern. Cur-
rently, Bcc comprises several related species that are phenotyp-
ically similar but genotypically distinct.[3] All Bcc species are
problematic CF pathogens because they are also multi-antibi-
otic resistant, which makes respiratory infection very difficult
to treat and impossible to eradicate. Lung transplantation is to
date the only treatment that offers improved length and quali-
ty of life to CF patients with advanced lung disease. Bcc
impact on the survival of CF patients following lung transplan-
tation is so detrimental that infection with these organisms is
considered a contraindication.[4, 5] The most prevalent clinical
species worldwide are Burkholderia cenocepacia and Burkholde-
ria multivorans. These are most often associated with cepacia
syndrome,[4] which is characterized by recurrent fever, bactere-
mia, necrotizing pneumonia, and a progressive decline that
results in premature death. However, recently, it has emerged
that high prevalence and high virulence is not confined to
these two species, but has also been described for B. cepacia[6]
and Burkholderia dolosa.[7] In particular, B. dolosa, has been
causing concern because of its virulence and transmissibility.[8]
B. dolosa is rarely associated with CF: around 3 % Bcc infections
in CF patients across the USA, and 2 % in CF patients followed
at the major Portuguese CF center in Lisbon.[9, 10]
B. dolosa was the first Bcc species to be identified as being
involved in patient-to-patient transmission: the same isolate
was found in 36 patients characterized by accelerated decline
in lung function, and, compared with patients colonized with
B. multivorans, reduced survival.[7] Furthermore, five of the pa-
tients colonized by B. dolosa with cepacia syndrome died, with
rapid decline in pulmonary function, persistent, pan-resistant
B. dolosa bacteremia, and death within four months. Under-
ChemBioChem 2013, 14, 1105 – 1115 1105
CHEMBIOCHEM
FULL PAPERS
standing how these divergent clinical outcomes arise might
allow new therapeutic strategies to emerge, to help improve
transplant outcome and to extend the life expectancy of CF
patients. Identifying differences in the biology between
B. dolosa and B. cenocepacia or B. multivorans might help to-
wards this aim. As reported over a decade ago, a number of
virulence factors can account for these divergent clinical out-
comes.[11] One of the most important virulence factors for
Gram-negative bacteria is the lipopolysaccharide (LPS) mole-
cule.[12–14]
Lipopolysaccharides are amphiphilic macromolecules that
comprise about 75 % of the outer membrane of Gram-negative
bacteria; they are indispensable for the growth and the surviv-
al of bacteria. LPSs provide a defensive barrier that helps bac-
teria to resist to antimicrobial compounds and environmental
stresses, and are involved in many aspects of host–bacterium
interactions, such as recognition, adhesion, and colonization.
They are initially cell-bound, and, upon release, they play a key
role in the pathogenesis of Gram-negative infections, causing
fever or circulatory shock. From both genetic and chemical
viewpoints, LPSs have a common structural architecture that
for smooth-type LPS (S-LPS) consists of three domains: a poly-
saccharide (O-side chain, O-specific polysaccharide, O-antigen)
covalently linked to an oligosaccharide (core) that is linked to
a glycolipid portion (lipid A), which is embedded in the outer
leaflet and anchors the macromolecule to the membrane by
electrostatic and hydrophobic interactions.[15, 17] LPSs can lack
the O-chain portion and in this case they are distinguished in
lipo-oligosaccharides (LOSs) or rough-type LPSs (R-LPSs,
Figure 1).[18] Lipid A possesses a rather conservative structure
Figure 1. General chemical structure of LPS from Gram-negative bacteria. All
forms of LPS known to date consist of a lipid A domain and a covalently
linked to a saccharide portion. The saccharide domain is, in turn, composed
of a core region and the O-specific chain.
usually consisting of a b-(1!6)-glucosamine disaccharide back-
bone phosphorylated at positions 1 and 4’, and acylated with
primary 3-hydroxy fatty acids at the positions 2 and 3 of both
glucosamine (GlcN) residues; the above-mentioned hydroxyl
groups can be further substituted by secondary acyl moiet-
ies.[18] In the core oligosaccharide, inner and outer regions are
usually distinguished: the inner core (proximal to lipid A) con-
sists typically of residues like Kdo (3-deoxy-d-manno-oct-2-ulo-
sonic acid) and heptose (Kdo is the linker between the GlcN II
of the lipid A backbone and the inner core portion); the outer
 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
core region is more variable and is usually composed of
hexose units.[15] The O-chain (the hydrophilic component of
LPS) is the most variable in the macromolecule, even within
bacteria of the same genus, and represents the antigenic de-
terminant of the LPS molecule. Structurally, the O-chain can be
a homo- or heteropolysaccharide (linear or branched) with
subunits of up to eight different sugars (Figure 1). In addition
to their essential structural function, LPSs have an important
role in eliciting the host innate immune responses. In particu-
lar, the lipid A moiety (together with other microbial glycocon-
jugates, such as peptidoglycan) is identified as a pathogen-
associated molecular pattern, and is recognized by pathogen-
recognition receptors of the host immune system.[15] Specifical-
ly, it has been reported that LPS from Bcc species is a potent
immunostimulatory agent, and has effects on mononuclear
cells and granulocytes thereby triggering the production of
pro-inflammatory cytokines.[17] Whereas activation of immune
cells is necessary for the host immune response, overproduc-
tion of cytokines can lead to severe complications; in fact Bcc
LPS is the primary contributor to the inflammatory nature of
infection in CF patients, both by promoting increased neutro-
phil recruitment and by priming neutrophil respiratory burst
responses.[19] Furthermore, LPSs extracted from different Bcc
strains have been demonstrated to exhibit diverging biological
activity.[20] Therefore, its structural characterization is essential
to understanding structure–activity relationships. Such data are
instrumental in aiding the design of antimicrobial compounds
and for the development of therapeutic strategies against the
inflammatory cascade.
In this paper, we report the complete structure of the lipo-
oligosaccharide of a clinical isolate from the Gram-negative
bacterium B. dolosa. We have carried out a complete chemical
study by means of chemical analysis, matrix-assisted laser de-
sorption/ionization mass spectrometry, and 1D and 2D NMR
spectroscopy. We have also elucidated its immunological prop-
erties.
Results and Discussion
Isolation, Purification and Biological activity of B. dolosa LOS
LOS was obtained from cells of B. dolosa IST 4208, an isolate
from a CF patient in the major Portuguese CF center. It was ex-
tracted by the hot phenol–water protocol,[21] then purified
with DNase, RNase, and proteinase K, followed by dialysis, and
gel-permeation chromatography. Purity was examined by SDS-
PAGE. The B. dolosa LPS fraction revealed that the extracted
lipopolysaccharide was a rough-type LPS, that is, an LOS.
We stimulated commercially available HEK293 cells, stably
transfected with LPS-recognizing molecular complex CD14,
MD2 and TLR4 (Invivogen) with different concentrations of
B. dolosa LOS (1, 10, and 100 ng mL1). E. coli serotype OIII:B4 is
well known to possess a fully hexa-acylated lipid A that acts as
an agonist on TLR4-MD2 receptors, so was used at the same
concentrations as a positive control. Activation of NF-kB (nu-
clear factor kappa B) and IL-8 release were the read-outs of
this experiment. The results showed that the intact LOS from
ChemBioChem 2013, 14, 1105 – 1115 1106
CHEMBIOCHEM
FULL PAPERS www.chembiochem.org

B. dolosa had a potent immunostimulatory activity as it was
able to trigger activation of NF-kB at the same level as for
E. coli LPS at all concentrations tested (Figure 2 A, left); IL-8
release matched the E. coli behavior (Figure 2 A, right).
We then assessed whether, and to what extent, B. dolosa
LOS could interfere with TLR4-mediated signaling induced by
lipid A of E. coli. Several natural and synthetic lipid A structures
have been reported to exhibit high potential as antagonists
against endotoxically active LPS. For this purpose, HEK hTLR4/
MD2/CD14 cells (HEK293 hTLR4) were pre-incubated with dif-
ferent amounts (1, 10, and 100 ng) of B. dolosa LOS for 1 h and
then re-stimulated with 10 ng of E. coli LPS for 4 h. NF-kB acti-
vation and IL-8 production were evaluated as above. We found
that cell priming (pre-incubation) with B. dolosa LOS at the
lowest concentration (1 ng mL1) was able to change both NF-
kB activation (Figure 2 B, left) and IL-8 production (Figure 2 B,
right), relative to E. coli LPS stimulation alone. At the higher
concentrations (10 and 100 ng mL1), no significant differences
were observed in IL-8 release and NF-kB activation by either of
the two LPS forms following preincubation (Figure 2 B).
Then, we tested the release of inflammatory cytokines by
bone marrow-derived macrophages (BMDMs), collected and
differentiated from wild-type C57BL/6 female mice. BMDMs
express the entire TLR repertoire in this analysis. BMDMs were
stimulated with different concentrations (1, 10, or 100 ng mL1)
of either B. dolosa LOS or E. coli OIII:B4 LPS. After 12 h of stimu-
lation, secretion of TNF-a (tumor necrosis factor alpha), IL-6,
and IL-10 was measured by an ELISA assay. We found that
100 ng mL1 B. dolosa LOS elicited significantly higher amounts
of inflammatory cytokines (e.g., TNF-a and IL-6) than E. coli LPS
(Figure 3 A). Even at 10 ng mL1, the amount of produced IL-10
with B. dolosa LOS was statistically significantly higher (p <
0.05; Figure 3 A).
Finally, we performed a competition assay in BMDMs to
assess whether B. dolosa LOS could interfere with TLR4-mediat-
ed signaling induced by LPS of E. coli. BMDMs were pre-incu-
bated for 1 hour with different amounts (1, 10, or 100 ng mL1)
of B. dolosa LOS for 1 h and then re-stimulated with
10 ng mL1 E. coli LPS for 12 h. Cytokine secretion was mea-
sured with an ELISA assay. We found that cell priming with
B. dolosa LOS did not antagonize the endotoxic activity of
E. coli LPS at all concentrations tested (Figure 3 B). Indeed, the
release of IL-6, IL-10, and TNF-a was significantly higher (p <
0.05) when BMDMs were pre-exposed to B. dolosa LOS, relative
to E. coli LPS stimulation alone (Figure 3 B).
The above immunology data confirm that the LOS from
B. dolosa is a powerful proinflammatory agent and a major vir-
ulence factor. This prompted us to assess the primary structure
of this powerful molecule.
Structural analysis of LOS from B. dolosa
Monosaccharide analysis of the LOS from B. dolosa revealed
the presence of 2,6-dideoxy-2-amino-d-glucose (d-quinovosa-

Figure 2. NF-kB activity and IL-8 production in HEK293 hTLR4 cells stimulated with B. dolosa LOS and E. coli LPS. A) NF-kB activation (left) and IL-8 secretion
(right) upon stimulation of HEK293 hTLR4 after 4 h with B. dolosa LOS or commercial hexa-acylated E. coli LPS. B) TLR4-dependent antagonist activity of
B. dolosa LOS on hexa-acylated E. coli LPS. NF-kB activation (left) and IL-8 release (right) upon stimulation of HEK293 hTLR4 after 1 h with B. dolosa LOS and
then exposed to 10 ng mL1 of E. coli LPS (4 h). Asterisk: p < 0.05 (Student t-test); NS: no stimulation.

 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 1105 – 1115 1107
CHEMBIOCHEM
FULL PAPERS
Figure 3. Cytokines released from BMDMs stimulated with B. dolosa LOS and E. coli LPS. A) TNF-a, IL-6, and IL-10 released by BMDMs after stimulation with
B. dolosa LOS or E. coli LPS, as measured by ELISA at 12 h. B) TLR4-dependent antagonistic activity of B. dolosa LOS on hexa-acylated E. coli LPS in BMDMs.
TNF-a, IL-6, and IL-8 release upon stimulation of BMDMs for 1 h with B. dolosa LOS and then re-exposed to 10 ng mL1 E. coli LPS (12 h). Cytokine secretion
was measured by ELISA. Asterisk: p < 0.05 after Student t-test; NS: no stimulation.
mine, d-QuiN), 4-amino-4-deoxy-l-arabinose (l-Ara4N), d-Glc,
d-Gal, d-GlcN, l-glycero-d-manno-heptose (L,d-Hep), 3-deoxy-
d-manno-oct-2-ulopyranosonic acid (d-Kdo), and d-glycero-d-
talo-oct-2-ulopyranosonic acid (d-Ko). Methylation analysis re-
vealed the presence of terminal QuiN, terminal Glc, terminal
Gal, terminal Hep, 3-substituted Glc, 7-substituted Hep, 2,6-di-
substituted Gal, 3,4-disubstituted Hep, 3,7-disubstituted Hep,
2,3,7-trisubstituted Hep, 4,5-disubstituted Kdo, and terminal Ko
all in pyranose rings (Table S1). Fatty acids analysis showed the
presence of (R)-3-hydroxyhexadecanoic acid (C16:0 (3-OH))
with an amide linkage, and of (R)-3-hydroxytetradecanoic
(C14:0 (3-OH)) and tetradecanoic acid (C14:0) with an ester
linkage. The overall chemical composition matched those of
archetypal Burkholderia LPSs/LOSs.[19] In order to determine the
primary structure of the oligosaccharide of B. dolosa LOS, the
lipid A part was cleaved off by mild hydrolysis with acetate
buffer, and the oligosaccharide fraction was purified by gel
permeation chromatography and characterized by NMR spec-
troscopy. A combination of homo- and heteronuclear 2D NMR
experiments (DQF-COSY, TOCSY, ROESY, NOESY, 1H,13C HSQC,
1 13
H, C HSQC-TOCSY, and 1H,13C HMBC, Figures 4, 5, S1–S6) was
performed, in order to assign all the spin systems and to estab-
lish the monosaccharide sequence. In particular, starting from
anomeric proton signals, TOCSY and DQF-COSY spectra al-
lowed the correct identification and attribution of all ring pro-
tons and the subsequent assignment of each carbon atom
from analysis of the 1H,13C HSQC spectrum. The anomeric con-
figuration of each monosaccharide was attributed on the basis
of the 3JH-1,H-2 coupling constant values obtained by DQF-COSY,
whereas the relative configuration of sugar residues was as-
signed on the basis of the 3JH,H ring coupling constants; finally,
 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
the intra-residue pattern of dipolar correlations gave further
confirmation of the anomeric configurations. Both inter-residu-
al NOE contacts (obtained by NOESY experiments) and long-
range HMBC correlations were useful to identify the sequence
of monosaccharides in the oligosaccharide. In the anomeric
region of the 1H NMR spectrum (Figure 4), 11 anomeric signals
were identified (A–M, Table 1); furthermore, the signals at 1.85/
2.05 ppm were attributed to the H-3 methylene protons of the
Kdo residue. The relative intensities and the shifts of anomeric
signals suggests marked heterogeneity, typical of non-stoichio-
metric carbohydrate substitutions. In accordance with chemical
analysis, spin systems B, C, D, E, G were all identified as Hep
residues, as indicated by their 3JH-1,H-2 and 2JH-2,H-3 coupling con-
stants (below 3 Hz), and by the intra-residue NOE of H-1 with
H-2. Moreover, the 13C chemical shift value of C6 of these hep-
tose residues (all below 72 ppm) allowed us to identify them
as l,d-Hep residues. Residues A and H (Table 1) were identified
as galacto-configured monosaccharides because of their low
3
JH-3,H-4 and 3JH-4,H-5 values (~ 3 Hz and 1 Hz respectively). In par-
ticular, residue A was identified as a-Gal, based on the H-1 and
C1 chemical shifts (5.35 and 97.3 ppm, Figure S4, Table 1), the
3
JH-1,H-2 value, and the intra-residual NOE contact of H-1 with H-
2 (all in agreement with an a-anomeric configuration and a 4C1
ring conformation). The anomeric 1JC1,H-1 3JH-1,H-2 values (164 and
8.0 Hz, respectively) of residue H (H-1 at 4.56 ppm) indicated
a b-anomeric configuration; this was further corroborated by
intra-residue NOE connectivity between H-1 and H-3/H-5 sig-
nals in the NOESY spectrum. Spin systems F, F’, L, and M
(Table 1) were identified as Glc residues by their characteristi-
cally large 3JH,H ring coupling constants (~ 9 Hz). The strong
intra-residue NOE contacts of H-1 L and M with H-3 and H-5 of
ChemBioChem 2013, 14, 1105 – 1115 1108
CHEMBIOCHEM
FULL PAPERS www.chembiochem.org

trum showed the presence of

a correlation of H-2 I, at
3.69 ppm, with a nitrogen-bear-
ing carbon signal at 54.1 ppm.
The downfield shift of the H-2 I
proton resonance was diagnostic
of N-acetylation at this position;
this was confirmed by the di-
polar correlation of H-2 I with
the methyl protons of the acetyl
group at 1.96 ppm. The spin
system of Kdo K was assigned
starting from the diastereotopic
H-3 methylene proton signals,
resonating in a shielded region
at 1.85 and 2.28 ppm (H-3ax and
H-3eq, respectively). The Kdo resi-
due was present in multiple
forms, because of its free reduc-
ing end; nevertheless, the sig-
nals from the a-reducing unit
were clearly assignable, and the
a-anomeric orientation at C-2
Figure 4. 1H NMR spectrum of core oligosaccharide, with detail of anomeric region. Anomeric signals of spin sys- was assigned by the chemical
tems are designated as in Table 1. shift values of H-3 and by the
values of the 3JH-7,H-8a and 3JH-7,H-8b
coupling constants.[22, 23] Ko resi-
3
both residues L and M, together with the JH-1,H-2 coupling con- due (J) was detected by the presence of the characteristic
stants (7 Hz) were diagnostic of a b-configuration, whereas the inter-residue NOE contact in 2D NOESY spectra between H-3eq
intra-residue NOE contact of H-1 with H-2, and the 3JH-1,H-2 cou- of Kdo (K) and H-6 of Ko (J); this is also diagnostic for a a-d-
pling constants (3 Hz) were indicative of an a-anomeric config- Ko-(2!4)-a-d-Kdo linkage.[22, 24, 25] The downfield shift of the
uration for residues F and F’. Residue I was recognized as a b- carbon resonances identified the glycosylated positions: O-2
QuiN: the gluco configuration was indicated by the 3JH,H ring and O-6 of residue A; O-2, O-3, and O-7 of B; O-3 and O-4 of
coupling constants (~ 9 Hz). The intra-residue NOE contacts of C; O-3 and O-7 of D; O-3 of F’; and O-4 of Kdo (K). Residues E,
H-1 with H-3 and H-5, and the 3JH-1,H-2 coupling constant were F, G, H, I, L, M, and J were non-reducing terminal sugars, in
indicative of a b-anomeric configuration. The 1H,13C HSQC spec- agreement with the methylation data. The oligosaccharide se-

Figure 5. Detail of overlapped TOCSY (gray) and NOESY (black) spectra of core oligosaccharide. The relevant inter-residual NOE contacts are reported.

 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 1105 – 1115 1109
CHEMBIOCHEM
FULL PAPERS www.chembiochem.org

ed at O-2 and O-6; thus, the NOE correlation of H-6 A (4.00/

Table 1. 1H and 13C NMR chemical shifts of the oligosaccharide derived
from mild hydrolysis of the LOS from Burkholderia dolosa. 3.57 ppm) with H-1 F (4.95 ppm) of a-Glc gave evidence of the
substitution of residue A at O-6 by the a-Glc F. An alternative
Chemical shift d (1H/13C) spin system was identified for residue F, namely residue F’; this
Unit 1 2 3 4 5 6 7 8
was recognized as a 3-a-substituted Glc that was, in turn, non-
A 5.35 3.56 3.73 3.99 3.64 4.00/ stoichiometrically substituted at position 3 by b-Gal H, as
3.57
shown by the NOE contact of H-3 F’ (3.85 ppm) with H-1 of
2,6-a-Gal 97.3 74.4 71.3 70.3 71.2 64.6
B 5.20 4.08 3.91 3.55 3.70 3.64 3.63/3.55 residue H (4.56 ppm). Residue A was, in turn, also substituted
2,3,7-a-Hep 99.1 77.2 79.9 68.9 70.7 71.9 70.1 at position O-2 by the a-Hep residue D (4.97 ppm), as showed
C 5.19 3.95 3.93 3.97 3.81 3.91 3.92/3.49 by the NOE correlation of H-2 of A (3.56 ppm) with H-1 of D.
3,4-a-Hep 99.1 65.7 71.6 71.6 72.1 68.7 63.0
Unit D was, in turn, glycosylated at O-3 by residue E as attest-
D 4.97 3.88 4.03 3.70 3.53 3.95 3.63/3.55
3,7-a-Hep 97.4 70.7 76.5 71.5 72.2 70.4 70.1 ed by the NOE contact of H-3 D (4.03 ppm) with H-1 E
E 4.95 4.00 4.11 3.64 3.53 4.07 3.63 (4.95 ppm). Finally, residue D was shown to be glycosylated, in
t-a-Hep 102.3 69.8 69.3 71.1 71.1 67.9 62.9 a non-stoichiometrical fashion, at O-7 (3.63/3.55 ppm) by resi-
F 4.95 3.71 3.70 3.43 3.58 3.74
due I, as evidenced by the NOE correlation between H-1 I
t-a-Glc 97.4 68.5 70.6 71.5 71.2 60.3
F’ 4.90 3.67 3.85 3.44 3.44 3.76 (4.50 ppm) and H-7 (3.63/3.55 ppm) of residue D. Thus, in sum-
3-a-Glc 97.6 71.1 82.5 72.9 68.0 60.3 mary, all of the data allowed us to establish the overall oligo-
G 4.82 3.88 3.76 3.64 3.53 3.91 3.62 saccharide structure reported below.
t-a-Hep 100.6 69.9 70.4 71.1 71.1 68.7 62.9
H 4.56 3.49 3.60 3.82 3.64 3.67
t-b-Gal 103.3 71.1 72.0 68.5 75.2 60.5
I 4.50 3.69 3.73 3.24 3.40 1.21
t-b-QuiNAc 100.6 54.1 74.7 75.6 75.4 16.2
L 4.46 3.22 3.58 3.56 3.35 3.64
t-b-Glc 102.0 73.4 75.7 68.9 76.3 60.8
M 4.30 3.26 3.58 3.40 3.35 3.64
t-b-Glc 101.3 72.2 75.8 71.6 76.4 60.8
K – – 1.85/ 4.02 4.16 3.71 3.78 3.87/
2.28 3.58
4,5-a-Kdo 174.1 n.d. 33.4 71.9 69.1 71.6 69.3 62.7
J – – 3.70 3.93 n.d 3.56 3.99 3.86/
3.59
t-a-Ko n.d. n.d. 70.8 71.6 n.d 71.8 71.9 62.8 The HMBC spectrum confirmed the assigned structure, as it
n.d.: not determined. contained all the required long-range correlations to demon-
strate the proximity of the residues.

quence was established on the basis of the intra-residue NOE

contacts identified in the ROESY (not shown) and NOESY spec-
tra (Figure 5), and in the long-range scalar correlations in the
HMBC spectrum (not shown). The linkage of the heptose C (H-
1 at 5.19 ppm) to O-5 of Kdo K was proven by the NOE con-
nectivity between H-1 of the Hep C and H-5 of K (4.16 ppm).
Given the l,d relative configuration of Hep C, the presence of
further NOE contacts between H-1 of Hep C and H-6 and H-7
of K ultimately proved the d-configuration for the Kdo resi-
due.[24] Residue C was in turn substituted at O-3 and O-4; the
NOE contacts (Figure 5) of H-4 C (3.97 ppm) and H-3 C
(3.93 ppm) with H-1 of residue L (4.46 ppm) evidenced that
the O-4 of a-Hep C was glycosylated by residue L, the b-Glc.
Moreover, residue C was also substituted at O-3 by residue B
of a-Hep, according to the NOE of H-3 and H-2 C with H-1 B.
Residue B was identified as a 2,3,7-trisubstituted a-Hep. The
NOE contacts of H-1 M (4.30 ppm) with H-1 and H-2 B (5.20
and 4.08 ppm, respectively) evidenced that the O-2 of residue
B was glycosylated by the b-Glc residue M. Furthermore, the
NOE correlation of H-3 B (3.91 ppm) with H-1 A (5.35 ppm)
proved the substitution of residue B at O-3 with the a-Gal resi-
due A. Residue B was also glycosylated at O-7 by the terminal
a-Hep G, as attested by the NOE contact between H-1 G
(4.82 ppm) and H-7 B (3.63/3.55 ppm). Residue A was substitut-
Structural characterization by MALDI mass spectrometry of
the intact LOS and lipid A
Intact LOS was analyzed by MALDI-MS to gain further informa-
tion on both lipid A and the core region. This approach is very
useful as it allows the study of integral LOS molecules[26, 27]
without any chemical manipulation, thus preventing the loss
(either by alkaline or acid treatment) of labile groups (e.g.,
phosphate, acetyl), typically present on LPSs. LOS was analyzed
by using a sample preparation procedure specifically optimized
for these amphiphilic molecules.[26, 27] The negative-ion MALDI
mass spectrum of intact LOS showed, in addition to the molec-
ular ions related to the native LOS mixture (m/z 3400–4000),
fragments between m/z 1400 and 2300 (Figure 6 A). These ori-
ginated from the b-elimination cleavage of the labile glycoside
bond between Kdo and the lipid A moiety, which yielded
either oligosaccharide (OS) ions (B-type ions) or lipid A ions.
This low-mass region (Figure 6 B) presented a very heterogene-
ous OS mixture, including the peak at m/z 2225.6 (OS1), which
matched a dodecasaccharide of five Hep residues, five Hex res-
idues, one Kdo residue, and one Ko residue—the core oligo-
saccharide structure determined (above) by NMR spectroscopy.
Minor species were also present: m/z 2063.9 (OS2), 2033.5
(OS3), 1871.6 (OS4), and 2164.0 (OS5); these differed from OS1

 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 1105 – 1115 1110
CHEMBIOCHEM
FULL PAPERS www.chembiochem.org

moiety after acetate buffer hy-

drolysis. This treatment left the
lipid A fraction partially dephos-
phorylated (not shown). In par-
ticular, the positive-ion MS/MS
spectrum of the pseudomolecu-
lar ion [M+Na] + at m/z 1387.89
(corresponding to the tetra-acy-
lated monophosphorylated lipid
A) presented a very interesting
fragmentation pattern, with a
significant fragmentation due to
the rupture of the glycosidic
linkage between the two GlcN
units (Figure S7). It gave rise to
a tri-acylated B-type ion at m/
z 954.6 carrying one 14:0 (3-OH),
one 14:0, and one 16:0 (3-OH)
residue at the nonreducing GlcN
unit, and a very intense diacylat-
ed B-type ion (base-peak) at m/z
726.4 that arose from further
loss of the 14:0 moiety (Fig-
ure S7). The assignment of the
main LOS molecular ions (Fig-
ure 6 C) resulted from the combi-
nation of the lipid A moieties
and the core oligosaccharide
species. LOS composed of tetra-
acylated lipid A and OS species
were found at m/z 3668.8 (L1 +
Figure 6. MALDI TOF MS analysis of intact LOS from B. dolosa; A) Negative-ion mass spectrum showing both LOS OS1), 3799.6 (L2 + OS1), 3930.8
molecular ions and their ion fragments, attributable to the core oligosaccharides and to the reported lipid A struc- (L3 + OS1), 3637.9 (L2 + OS2),
ture(s). B) Detail of the lower-mass region (m/z 1250–2450), including lipid A and core oligosaccharide (OS) ion
3476.4 (L1 + OS3), 3607.5 (L2 +
fragments and their assignments. C) Detail of the higher-mass region (m/z 3000–4280) showing the composition
of each LOS species; this emerges from the heterogeneity of the core fraction and lipid A components, which is OS3), 3738.6 (L3 + OS3 and L2 +
mainly due to non-stoichiometrically linked terminal sugar moieties. OS4) and 3445.9 (L2 + OS5). In
summary, MS analysis of the
intact LOS completed and con-
by the lack of one terminal Hex and/or Hep and one Ara4N firmed the previous structural hypotheses and allowed the full
residues (Figure 6 B). Each OS peak possessed a minor but con- assignment of the LOS structure from the cystic fibrosis patho-
firmed twin-ion peak (Dm/z 44), which matches neutral loss of gen B. dolosa (Scheme 1).
a CO2 molecule from the reducing terminal Kdo. It is worth
noting that the oligosaccharide ion with the terminal QuiNAc
Discussion
residue was not detected in the MALDI analysis. The same
mass region (m/z 1400–2300) also included ion peaks derived Chronic infection with Bcc is a major cause of premature death
from a mixture of tetra-acylated lipid A, distinguishable by the in patients with cystic fibrosis, due to its inherent resistance to
absence or presence of one/two Ara4N residues, (Figure 6 a). most used antibiotics, to person-to-person spread of highly
Species L1 (m/z 1444.8) matched a tetra-acylated bis-phos- transmissible strains, and to its persistence, despite a specific
phorylated disaccharide backbone carrying one 14:0 (3-OH) and pronounced antibody response.[28] Infections of B. dolosa,
acyl chain in the ester linkage and two 16:0 (3:OH) acyl chains a Bcc species, broke out among individuals with cystic fibrosis
in amide linkage; one of these, on GlcN II, was further substi- in Boston;[7, 29] it chronically colonized CF patients and spread
tuted by a secondary 14:0 fatty acid. Species L2 at m/z 1575.6 from person to person, thus causing concerns in several CF
(Dm/z 131 from L1) and L3 at m/z 1706.1 (Dm/z 131 from L2) centers.[30] In the major Portuguese CF treatment center only
were tetra-acylated lipid A, and carried one and two Ara4N res- one case of B. dolosa infection was detected during 18 years of
idues, respectively. The identity of either a primary ester-linked epidemiological surveillance of respiratory infections involving
14:0 (3-OH) or a secondary amide-linked 14:0 fatty acid on the Bcc.[10] The infected CF patient was chronically colonized for
GlcN II was established by the MS/MS analysis of the lipid A 5.5 years with B. dolosa, until death following severe pulmona-

 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2013, 14, 1105 – 1115 1111
CHEMBIOCHEM
FULL PAPERS
Scheme 1. Structure of LOS from B. dolosa. Dotted lines indicate non-stoichiometric sub-
stitution. The anomeric configuration of lipid A residues was assigned based on all Bur-
kholderia LPS lipid As elucidated to date.
ry deterioration. The LOS chemical structure described in this
work was from the first B. dolosa isolate that was retrieved
from the patient. In another epidemiologic study dealing with
clinical implications of chronic infection with B. dolosa in cystic
fibrosis patients, it was also reported that acquisition of this
organism was associated with accelerated decline in lung func-
tion and reduced survival,[31, 32] differently to other Bcc infec-
tions. These findings provide additional evidence that B. cepa-
cia complex species differ in their virulence and clinical impli-
cations and emphasize the importance of examining at chemi-
cal level the virulence factors of all strains.[31] The hallmark of
Bcc infection in CF is the inflammation that depends on the
peculiar biochemical properties of the lipopolysaccharide,
which is able to induce host defense responses and tissue
damage. Thus, the full primary structure of LPS from all Bcc
species and their LPS structure–function relationships are clini-
cally relevant.
 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
Within this frame we report for the first time the
complete structure and biological activity of LOS
from a B. dolosa clinical isolate retrieved from a CF
patient who died due to progressive loss of lung
function. In this case, the composition of the
B. dolosa lipid A species displayed a relatively homo-
geneous tetra-acylated lipid A bearing Ara4N resi-
due(s).[20] The substitution of Ara4N in LPS molecules
is essential for bacterial survival, as it prevents the in-
teractions with antibiotic compounds and host anti-
microbial peptides.[33] These residues are positively
charged under physiological conditions and so they
reduce the net charge on the bacterial membrane
surface. By preventing ionic attraction, these residues
prevent membrane permeability by positively
charged antimicrobial peptides.
The core oligosaccharide from B. dolosa has a
novel structure. From the chemical point of view, the
B. dolosa LOS also has no negative charge at the
outer core, which contains neither phosphate groups
nor uronic acids. This confers on the LPS an overall
neutral characteristic that renders it resistant to cat-
ionic antimicrobial peptides. The Burkholderia genus
has a highly conserved LPS inner core structure, with
several characteristics replicated in the current study.
The main feature is the presence of the [3,4-a-L,d-
Hep-(1!5)-a-d-Kdo-(4!2)-a-d-Ko] trisaccharide in
the inner core. Notably, within this genus, the Ko resi-
due is frequently not at the terminal, but rather
bears an Ara4N residue. The presence of Ko mono-
saccharide is limited to a few bacterial LPS molecules,
such as those of Acinetobacter,[34] Yersinia,[35] and Ser-
ratia.[36] Usually, as in the present case, the a-l,d-Hep
of the above-mentioned trisaccharide bears another
Hep residue linked at O-3 and glycosylated at O-4 by
a b-d-Glc residue. In this B. dolosa strain, the latter
Hep residue of inner core is glycosylated at O-2, O-3,
and O-7, as previously noted for B. multivorans;[19]
however, the heptose carries an additional heptose
at O-7, as found in B. cenocepacia ET-12 LOS,[36] and
at O-2 there is a b-d-Glc, and at O-3 there is a a-d-Gal. This
latter monosaccharide was, in turn, substituted at position O-2
by an a-Hep residue and at O-6 by a a-d-Glc which can be, in
turn, non-stoichiometrically glycosylated by a b-d-Gal residue.
Finally, the latter Hep was glycosylated at position O-3 by an-
other Hep. Interestingly, this a-Hep-(1!3)-a-Hep disaccharide
is infrequently found as a component of the outer core of
LOSs,[13] and its presence might involve the existence of an ad-
ditional different heptosyl transferase. Moreover, in the outer
core region, we found the presence of a non-stoichiometric
residue, the b-d-QuiNAc linked to the a-Hep-(1!3)-a-Hep dis-
accharide. In summary, B. dolosa LOS possesses a peculiar
structure, the biosynthesis of which requires further study.
From a biological point of view, we observed that B. dolosa
LOS acts as potent inflammatory agent through the TLR-4, MD-
2, CD-14 pathway, thereby inducing high levels of NF-kB and
IL-8 release from HEK293 cells stably transfected for LPS com-
ChemBioChem 2013, 14, 1105 – 1115 1112
CHEMBIOCHEM
FULL PAPERS
plex recognition (HEK293 hTLR4). These data were corroborat-
ed by the results obtained from analysis of inflammatory cyto-
kines released in BMDMs after stimulation with intact LOS from
B. dolosa. The high levels of TNF-a, IL-10, and IL-6 secretion
confirm the virulence and support the concerns about this
cystic fibrosis pathogen. Lipid A contributes to the majority of
the endotoxic activity of LPS. A number of factors influence
the lipid A biological activity, including the number and the
distribution of acyl chains, the phosphorylation pattern, and
the presence of charged groups on the polar heads. The corre-
lation between increasing acylation of lipid A and elevated cy-
tokine induction has previously been reported in Pseudomo-
nas[38] and Burkholderia.[38] Interestingly, lipid A from B. dolosa
did not show high levels of acylation, but the entire LOS still
remains a potent immunostimulator with HEK293 cells. Taking
into account these findings, it might be speculated that the
novel and heterogeneous structure of the core oligosaccharide
portion plays a key function in the molecular mechanism of
the interaction with the TLR-4/MD-2 receptors, thus demon-
strating a notable role in the pathogenicity of B. dolosa. Fur-
ther studies will be performed to test this hypothesis.
In conclusion, in this work we have chemically and biologi-
cally elucidated for the first time the LOS endotoxin from
B . dolosa. This study improves the understanding of the endo-
toxin structure–function relationship, which is of pivotal impor-
tance in the comprehension of the overall process of patho-
genesis of such important microorganisms.
Experimental Section
Growth of B. dolosa cells for LOS extraction: B. dolosa IST4208
isolate, from which LOS was extracted, was obtained from the res-
piratory secretions of a CF patient under surveillance at the Hospi-
tal de Santa Maria CF Center, in Lisbon. Isolate B. dolosa IST4208
was obtained in 2005 from a CF patient who passed away. Consent
was obtained from the patient, and the isolate was provided by
the hospital for this particular study; the patient’s anonymity was
preserved. This isolate was identified based on the polymorphisms
of the recA gene,[40] followed by multilocus sequence typing
(MLST) analysis. The MLST sequences of the allelic type of seven
loci were assigned and submitted to the B. cepacia complex MLST
database (http://pubmlst.org/bcc/) under the designation ST-668;
the obtained profile confirmed the identification as B. dolosa. Cells
of B. dolosa IST4208 were grown overnight in lysogeny broth (LB;
Difco, Sparks, MD), then inoculated into LB (initial OD640 nm 0.05)
and grown in shake flasks with orbital agitation (250 rpm, 37 8C)
until mid-exponential phase. The cell culture was diluted (standar-
dized OD640 nm 0.2  0.02), and a sample (100 mL) was plated onto
LB agar (Difco) plates. These plates were incubated (37 8C, 24 h) to
allow cells to grow on the agar surface, to mimic bacterial growth
in CF lungs. Cells were scraped, collected, autoclaved, and lyophi-
lized.
Extraction, purification, and SDS-PAGE analysis of LOS: Dried
cells were extracted by the phenol/water method.[21] After exten-
sive dialyses against distilled water, the extracted phases were sub-
jected to enzymatic digestion to remove nucleic acids and protein
contaminants. Then both the water and phenol fractions were ana-
lyzed by 13.5 % SDS-PAGE; the LPS fraction was found exclusively
 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
in the water phase, as suggested by the presence of the typical
ladder pattern of its migration in the gel.
Isolation of oligosaccharide OS: Isolation of OS was obtained by
mild acid hydrolysis. Purified LOS (~ 10 mg) was dissolved in ace-
tate buffer (1 mL, pH 4.4). SDS (1 mg mL1) was added, and hydrol-
ysis proceeded (100 8C, 3 h). After lipid A removal by centrifugation
(10 000 g, 30 min), the water-soluble product was purified by gel
filtration chromatography on a Bio-Gel P-6 column (Bio-Rad, Her-
cules, CA).
Isolation of lipid A: Free lipid A was obtained by hydrolysis
(100 8C, 3 h) of LOS in sodium acetate buffer (100 mm, pH 4.4). The
solution was extracted three times with CHCl3/MeOH/H2O
(100:100:30, v/v/v) and centrifuged (5000 g, 4 8C, 15 min). The
organic phase contained lipid A, and the water phase contained
the core oligosaccharide.
Chemical analysis: Determination of the sugar residues and of
their absolute configurations by GC-MS analysis was carried out as
described elsewhere.[41, 42] Monosaccharides were identified as
acetylated O-methyl glycoside derivatives. After methanolysis (2 m
HCl/MeOH, 85 8C, 24 h) and acetylation with acetic anhydride in
pyridine (85 8C, 30 min), the sample was analyzed by GC-MS. The
absolute configurations of the sugar residues were determined by
GC-MS analysis of the acetylated (+)-O-2-octyl glycoside derivatives
and comparison with authentic standards. Linkage analysis was
carried out by methylation of the complete saccharide portion as
described previously:[43] the sample was methylated with iodome-
thane, hydrolyzed (100 8C, 2 h) with trifluoroacetic acid (2 m), car-
bonyl reduced with NaBD4, acetylated with acetic anhydride and
pyridine, and analyzed by GC-MS.
NMR analysis: 1D and 2D 1H NMR spectra were recorded in D2O
(300 K, pD 7) with a 600 DRX spectrometer (Bruker) equipped with
a cryoprobe. The spectrometer was internally calibrated with ace-
tone (dH = 2.225 ppm; dC = 31.45 ppm). Rotating frame Overhauser
enhancement spectroscopy (ROESY) and nuclear Overhauser en-
hancement spectroscopy (NOESY) experiments were performed
using data sets (t1  t2) of 4096  256 points with mixing times be-
tween 100 and 400 ms. Double quantum-filtered phase-sensitive
COSY experiments were executed with sets of 4096  512 points.
Total correlation spectroscopy (TOCSY) experiments were per-
formed with spinlock times of 100 ms with data sets (t1  t2) of
4096  256 points. The data matrix in all the homonuclear experi-
ments was zero-filled in both dimensions to give a matrix of 4 K 
2 K points, and was resolution-enhanced in both dimensions using
a cosine-bell function before Fourier transformation. Coupling con-
stants were determined by 2D phase-sensitive DQF-COSY.[44, 45] Het-
eronuclear single quantum coherence (HSQC) and heteronuclear
multiple bond correlation (HMBC) experiments were performed in
1
H-detection mode by single-quantum coherence with proton de-
coupling in the 13C domain using data sets of 2048  256 points.
Experiments were carried out in the phase-sensitive mode.[46] A
60 ms delay was used for the evolution of long-range correlations
in the HMBC experiment. The data matrix in all the heteronuclear
experiments was extended to 2048  1024 points by using forward
linear prediction extrapolation.
MALDI TOF MS: MALDI-TOF mass spectra of intact LOS were re-
corded in linear mode and negative ion polarity on a Voyager STR
(PerSeptive Biosystems, Framingham, MA) equipped with delayed-
extraction technology. Ions formed by a pulsed UV laser (nitrogen
laser, l = 337 nm) were accelerated by 24 kV. Reflectron MALDI TOF
MS and MALDI TOF/TOF MS2 of lipid A were performed with
a 4800 Proteomic analyzer (Applied Biosystems) with a Nd:YAG
ChemBioChem 2013, 14, 1105 – 1115 1113
CHEMBIOCHEM
FULL PAPERS
laser (355 nm). External calibration was performed: mass accuracy
was better than 75 ppm. Mass resolution of the spectra obtained
in reflector mode was approximately 13 000. R-type LPS and lipid A
MALDI preparations were performed as previously reported.[27, 47]
HEK293 hTLR4/CD14/MD2 cell culture, transfection, and stimula-
tion: Stably transfected HEK cell line 293-hTLR4-MD2-CD14 (Invivo-
Gen, San Diego, CA) was cultured in DMEM (Lonza, Basel, Switzer-
land) with 10 % FBS (Euroclone, Pero, Italy). Blasticidin-S
(10 mg mL1; InvivoGen) and HygroGold (50 mg mL1; InvivoGen)
were added to cell cultures according to the manufacturer’s in-
structions. The cells were transiently transfected by using PolyFect
Transfection Reagent (Qiagen) according to the manufacturer’s in-
structions. For NF-kB studies, the cells were seeded into 96-well
plates at a concentration of 3  105 cells per mL, then transfected
overnight with a reaction mixture of PolyFect transfection reagent
(1 mL, Qiagen), and plasmids pGL3.ELAM.tk (150 ng; Promega),
pRLTK (15 ng; Promega).[47] pGL3.ELAM.tk encodes Firefly luciferase
under control of the NF-kB promoter; pRLTK encodes Renilla luci-
ferase, and was used as a control. HEK 293 hTLR4/MD2/CD14 cells
were exposed to different concentrations of B. dolosa LOS or E. coli
LPS (1, 10, or 100 ng mL1; Ultra-pure LPS-EB, InvivoGen), and stim-
ulation was for 4 h.
In the competition assays HEK 293 hTLR4/MD2/CD14 cells were
primed with B. dolosa LOS (1, 10, or 100 ng mL1) for 1 h and then
re-stimulated with of E. coli LPS (10 ng mL1). After 4 h of stimula-
tion, NF-kB-dependent luciferase activity was measured by a Dual-
Luciferase Reporter Assay System (Promega, Milan, Italy), as previ-
ously reported.[48] IL-8 production was quantified by a DuoSet
ELISA assay (R&D Systems, Minneapolis, MN) according to the man-
ufacturer’s instructions.
Bone marrow-derived macrophages (BMDMs) culture and stimu-
lation assay: Bone marrow-derived macrophages (BMDMs) were
derived from bone marrow cells collected from five-week-old wild-
type C57BL/6 female mice (Charles River Laboratories; all animals
were maintained in a specific pathogen-free animal facility at the
Sapienza University of Rome (Italy)). BMDMs were differentiated
over 7 days in RPMI 1640 medium (Lonza), supplemented with
heat-inactivated FBS (10 %, Euroclone), l-glutamine (2 mm), sodium
pyruvate (5 %), non-essential amino acids (NEAA 100  ; 100 mm,
Euroclone), b-mercaptoethanol (0.5 %, Gibco), and macrophage
colony-stimulating factor (30 ng mL1; Miltenyi Biotec, Calderara di
Reno, Italy). At 7 days, BMDMs were characterized by immunostain-
ing with FITC-F4/80 and Mac-1/CD11b (clone M1/70) monoclonal
antibodies (Abcam, catalogue number ab105155 and BD Pharmin-
gen, catalogue number 557397, respectively) by flow cytometric
analysis in a FACSCalibur cytometer (BD Biosciences).
For stimulation assays, BMDMs were seeded in 24-well plates (5 
105 cells per well), exposed to different concentrations of B. dolosa
LOS or E. coli LPS (1, 10, or 100 ng mL1) and incubated for 12 h.
Cell suspension supernatants were recovered and processed to
measure TNF-a (R&D Systems, catalogue number DY410), IL-10
(R&D Systems, catalogue number DY417), and IL-6 (R&D Systems,
catalogue number DY406) production by DuoSet ELISA assay ac-
cording to manufacturer’s instructions.
In the competition assays, BMDMs were primed with B. dolosa LOS
(1, 10, or 100 ng mL1) for 1 h and then re-stimulated with of E. coli
LPS (10 ng mL1). After 12 h of re-stimulation, TNF-a, IL-10, and IL-6
production was quantified by ELISA assay (DuoSet) according to
manufacturer’s instructions.
 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
Statistical analysis: Data are reported as mean  S.D, and the
numbers of independent experiments are indicated in each figure
legend. Statistical calculations and tests were performed by using
the Student’s t-test. A p value < 0.05 was considered statistically
significant.
Acknowledgements
A.M., A.S., M-L.B., I.S-C, acknowledge COST Action BM1003 “Mi-
crobial cell surface determinants of virulence as targets for new
therapeutics in cystic fibrosis.” This work has been partially sup-
ported to A.M. by Italian Cystic Research Foundation, grant FFC
11#2010 with the contribution of Pastificio Giovanni Rana s.p.a.
The contribution of P. Azevedo L. Lito and J. Melo-Cristino, Hospi-
tal de Santa Maria, and A. S. Moreira, IBB/CEBQ, Instituto Superi-
or Tcnico, to the epidemiology and identification of the CF iso-
late of B. dolosa, is gratefully acknowledged. This work was par-
tially supported by “Fundażo para a CiÞncia e a Tecnologia”,
FCT, Portugal, through a post-doctoral fellowship (SFRH/BPD/
81220/2011) awarded to C.P.C. R.L. acknowledges the CREME proj-
ect.
Keywords: Burkholderia cepacia complex (Bcc) · Burkholderia
dolosa · lipooligosaccharide · mass spectrometry · NMR
spectroscopy
[1] E. Mahenthiralingam, T. A. Urban, J. B. Goldberg, Nat. Rev. Microbiol.
2005, 3, 144 – 156.
[2] C. R. Hansen, T. Pressler, K. G. Nielsen, P. Ø. Jensen, T. Bjarnsholt, N.
Hoiby, J. Cystic Fibrosis 2010, 9, 51 – 58.
[3] E. Yabuuchi, Y. Kosako, H. Oyaizu, I. Yano, H. Hotta, Y. Hashimoto, T.
Ezaki, M. Arakawa, Microbiol. Immunol. 1992, 36, 1251 – 1275. Erratum:
Microbiol. Immunol. 1993, 37, 335.
[4] A. De Soyza. , P. A. Corris, J. Heart Lung Transplant. 2003, 22, 954 – 958.
[5] D. Hadjiliadis, M. P. Steele, C. Chaparro, L. G. Singer, T. K. Waddell, M. A.
Hutcheon, R. D. Davis, D. E. Tullis, S. M. Palmer, S. Keshavjee, J. Heart
Lung Transplant. 2007, 26, 834 – 838.
[6] M. V. Cunha, A. Pinto-de-Oliveira, L. Meirinhos-Soares, M. J. Salgado, J.
Melo-Cristino, S. Correia, C. Barreto, I S-Correira, J. Clin. Microbiol. 2007,
45, 1628 – 1633.
[7] L. A. Kalish, D. A. Waltz, M. Dovey, G. Potter-Bynoe, A. J. McAdam, J. J.
LiPuma, C. Gerard, D. Goldman, Am. J. Respir. Crit. Care Med. 2006, 173,
421 – 425.
[8] J. J. LiPuma, T. Spilker, L. H. Gill, P. W. Campbell, III, L. Liu, E. Mahenthi-
nalingam, Am. J. Respir. Crit. Care Med. 2001, 164, 92 – 96.
[9] J. J. LiPuma, Clin. Microbiol. Rev. 2010, 23, 299 – 323.
[10] C. P. Coutinho, S. C. dos Santos, A. Madeira, N. P. Mira, A. S. Moreira, I.
S-Correia, Front Cell Inf. Microbiol. 2011, 1, DOI:10.3389/
fcimb.2011.00012.
[11] E. Mahenthiralingam, T. Coenye, J. W. Chung, D. P. Speert, J. R. W. Govan,
P. Taylor, P. Vandamme, J. Clin. Microbiol. 2000, 38, 910 – 913.
[12] G. Seltmann, O. Holst, The Bacterial Cell Wall, Springer, Heidelberg,
2001.
[13] O. Holst in Endotoxin in Health and Disease (Eds.: H. Brade, S. M. Opal,
S. N. Vogel, D. C. Morrison,), Dekker, New York, 1999, pp. 115 – 154.
[14] C. R. H. Raetz, C. Whitfield, Annu. Rev. Biochem. 2002, 71, 635 – 700.
[15] C. Alexander, E. T. Rietschel, J. Endotoxin Res. 2001, 7, 167 – 202.
[16] A. De Soyza, F. Di Lorenzo, A. Silipo, R. Lanzetta, A. Molinaro, Carbohydr.
Chem. 2012, 38, 13 – 39.
[17] P. Vandamme, D. Henry, T. Coenye, S. Nzula, M. Vancanneyt, J. J. LiPuma,
D. P. Speert, J. R. W. Govan, E. Mahenthiralingam, FEMS Immunol. Med.
Microbiol. 2002, 33, 143 – 149.
[18] J. E. Hughes, J. Stewart, G. R. Barclay, J. R. W. Govan, Infect. Immunol.
1997, 65, 4281 – 4287.
ChemBioChem 2013, 14, 1105 – 1115 1114
CHEMBIOCHEM
FULL PAPERS
[19] J. E. Hughes, J. Stewart, G. R. Barclay, J. R. W. Govan, Infect. Immunol.
1997, 65, 4281 – 4287.
[20] A. De Soyza, A. Silipo, R. Lanzetta, J. R. Govan, A. Molinaro, Innate
Immun. 2008, 14, 127 – 144.
[21] O. Westphal, K. Jann, Methods Carbohydr. Chem. 1965, 5, 83 – 91.
[22] G. I. Birnbaum, R. Roy, J.-R. Brisson, H. Jennings, J. Carbohydr. Chem.
1987, 6, 17 – 39.
[23] O. Holst, J. E. Thomas-Oates, H. Brade, Eur. J. Biochem. 1994, 222, 183 –
194.
[24] K. Bock, E. Vinogradov, O. Holst, H. Brade, Eur. J. Biochem. 1994, 225,
1029 – 1039.
[25] Y. Isshiki, K. Kawahara, U. Zhringer, Carbohydr. Res. 1998, 313, 21 – 27.
[26] L. Sturiale, D. Garozzo, A. Silipo, R. Lanzetta, M. Parrilli, A. Molinaro,
Rapid Commun. Mass Spectrom. 2005, 19, 1829 – 1834.
[27] L. Sturiale, A. Palmigiano, A. Silipo, Y. A. Knirel, A. P. Anisimov, R. Lanzet-
ta, M. Parrilli, A. Molinaro, D. Garozzo, J. Mass Spectrom. 2011, 46, 1135 –
1142.
[28] E. Evans, I. R. Poxton, J. R. W. Govan, J. Med. Microbiol. 1999, 48, 825 –
832.
[29] R. Biddick, T. Spilker, A. Martin, J. J. LiPuma, FEMS Microbiol. Lett. 2003,
228, 57 – 62.
[30] J. J. Lipuma, K. A. Marks-Austin, D. S. Holsclaw, Jr., G. B. Winnie, P. H. Gilli-
gan, T. L. Stull, Pediatr. Infect. Dis. J. 1994, 13, 716 – 719.
[31] E. Caraher, C. Duff, T. Mullen, S. McKeon, P. Murphy, M. Callaghan, S.
McClean, J. Cystic Fibrosis 2007, 6, 49 – 56.
[32] R. M. Aris, J. C. Routh, J. J. LiPuma, D. G. Heath, P. H. Gilligan, Am. J.
Respir. Crit. Care Med. 2001, 164, 2102 – 2106.
[33] M. A. Hamad, F. Di Lorenzo, A. Molinaro, M. A. Valvano, Microbiology
2012, 85, 962 – 974.
[34] E. V. Vinogradov, S. Mller-Loennies, B. O. Petersen, S. Meshkov, J. E.
Thomas-Oates, O. Holst, H. Brade, Eur. J. Biochem. 1997, 247, 82 – 90.
 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
[35] E. V. Vinogradov, B. Lindner, N. A. Kocharova, S. N. Senchenkova, A. S.
Shashkov, Y. A. Knirel, O. Holst, T. A. Gremyakova, R. Z. Shaikhutdinova,
A. P. Anisimov, Carbohydr. Res. 2002, 337, 775 – 777.
[36] E. V. Vinogradov, B. Lindner, G. Seltmann, J. Radziejewska-Lebrecht, O.
Holst, Chem. Eur. J. 2006, 12, 6692 – 6700.
[37] A. Silipo, A. Molinaro, T. Ieran, A. De Soyza, L. Sturiale, D. Garozzo, C.
Aldridge, P. A. Corris, C. M. A. Khan, R. Lanzetta, M. Parrilli, Chemistry
2007, 13, 3501 – 3511.
[38] R. K. Ernst, E. C. Yi, L. Guo, K. B. Lim, J. L. Burns, M. Hackett, S. I. Miller,
Science 1999, 286, 1561 – 1565.
[39] A. Silipo, A. Molinaro, P. Cescutti, E. Bedini, R. Rizzo, M. Parrilli, R. Lanzet-
ta, Glycobiology 2005, 15, 561 – 570.
[40] K. Vermis, T. Coenye, J. J. LiPuma, E. Mahenthiralingam, H. J. Nelis, P.
Vandamme, Int. J. Syst. Evol. Microbiol. 2004, 54, 689 – 691.
[41] K. Leontein, J. Lçnngren, Methods Carbohydr. Chem. 1978, 62, 359 – 362.
[42] C. De Castro, M. Parrilli, O. Holst, A. Molinaro, Methods Enzymol. 2010,
480, 89 – 115.
[43] S. Hakomori, J. Biochem. 1964, 55, 205 – 208.
[44] U. Piantini, O. W. Sorensen, R. R. Ernst, J. Am. Chem. Soc. 1982, 104,
6800 – 6801.
[45] M. Rance, O. W. Sørensen, G. Bodenhausen, G. Wagner, R. R. Ernst, K.
Wthrich, Biochem. Biophys. Res. Commun. 1983, 117, 479 – 485.
[46] D. J. States, R. A. Haberkorn, D. J. Ruben, J. Magn. Reson. 1982, 48, 286 –
292.
[47] A. Silipo, L. Sturiale, D. Garozzo, C. ee Castro, R. Lanzetta, M. Parrilli,
W. D. Grant, A. Molinaro, Eur. J. Org. Chem. 2004, 2263 – 2271.
[48] G. Nigro, L. L. Fazio, M. C. Martino, G. Rossi, I. Tattoli, V. Liparoti, C. De C-
astro, A. Molinaro, D. J. Philpott, M. L. Bernardini, Cell. Microbiol. 2008,
10, 682 – 695.
Received: February 4, 2013
Published online on June 3, 2013
ChemBioChem 2013, 14, 1105 – 1115 1115