It is well established that the inner core of meningococcal LPS consists of a diheptosyl-N-acetylglucosamine unit, in which the distal heptose unit Hep II can carry PEtn at the 3 or 6 po
Trang 1Identification of a novel inner-core oligosaccharide structure
1
Institute for Biological Sciences, National Research Council, Ottawa, Canada;2Institute for Molecular Medicine,
John Radcliffe Hospital, University of Oxford, UK
The structure of the lipopolysaccharide (LPS) from three
were nonreactive with mAbs that recognize common
inner-core epitopes from meningococcal LPS It is well established
that the inner core of meningococcal LPS consists of a
diheptosyl-N-acetylglucosamine unit, in which the distal
heptose unit (Hep II) can carry PEtn at the 3 or 6 position or
not at all, and the proximal heptose residue (Hep I) is
sub-stituted at the 4 position by a glucose residue Additional
substitution at the 3 position of Hep II with a glucose residue
is also a common structural feature in some strains The
structures of the O-deacylated LPSs and core
oligosaccha-rides of the three chosen strains were deduced by a
combi-nation of monosaccharide analysis, NMR spectroscopy and
MS These analyses revealed the presence of a structure not previously identified in meningococcal LPS, in which an additional b-configured glucose residue was found to sub-stitute Hep I at the 2 position This provided the structural basis for the nonreactivity of LPS with these mAbs The determination of this novel structural feature identified a further degree of variability within the inner-core oligosac-charide of meningococcal LPS which may contribute to the interaction of meningococcal strains with their host Keywords: lipopolysaccharide; mass spectrometry; Neisse-ria meningitidis; NMR; oligosaccharide
The lipopolysaccharide (LPS) of Neisseria meningitidis
contains a core oligosaccharide unit with a conserved
inner-core diheptose-N-acetylglucosamine backbone, in
can provide a point of attachment for the outer-core
oligosaccharide residues [1] Meningococcal LPS has been
classified into 12 distinct immunotypes (L1–L12), originally
defined by mAb reactivities [2], but further defined by
structural analyses The structures of LPS from
immuno-types L1/6 [3,4], L2 [5], L3 [6], L4/7 [7], L5 [8] and L9 [9]
have been elucidated The structural basis of the
immuno-typing scheme is governed by the location of a
phospho-ethanolamine (PEtn) moiety on the distal heptose residue
(Hep II) at either the 3 or 6 position, or absent, or at
both positions simultaneously [10] The length and nature
of oligosaccharide extension from the proximal heptose
residue (Hep I) and the presence or absence of a
glu-cose sugar at Hep II also dictates the immunotype The
enzyme UDP glucose 4-epimerase (GalE) is essential for
galactose into its LPS and is encoded by the gene galE [11] The absence of galactose residues from the conserved inner-core structure of meningococcal LPS has led to the utilization of mutants defective in the enzyme, resulting
in truncation of the oligosaccharide chains of the LPS at the glucose residue at Hep I, and galE mutants have been used by our group to derive mAbs to inner-core LPS epitopes [12] (unpublished data) Identified in this way, mAbs with an absolute requirement for PEtn at the 3 position [12], or the 6 position (unpublished data) of Hep II were developed mAbs were also produced that had specificities for PEtn at the 6 position of Hep II coupled with the presence of a glucose residue at the 3 position of Hep II, glucose at the 3 position of Hep II coupled with the absence of PEtn at the 6 position of Hep II and specific for
an epitope where there was no substitution with PEtn or glucose at Hep II (unpublished data) During the course of these studies, we specifically selected LPS from clinical isolates that were not reactive with these mAbs Structural analysis of meningococcal clinical isolates revealed an additional and uniquely located glucose residue in the core oligosaccharide that had not previously been identified in meningococcal LPS
Materials and methods
Growth of organism and isolation of LPS
each growth Strains 425/93 and NM115 are from the
Correspondence to A D Cox, Institute for Biological Sciences,
100 Sussex Drive, National Research Council, Ottawa,
ON K1A 0R6, Canada.
Fax: + 1 613 952 9092, Tel.: + 1 613 991 6172,
E-mail: Andrew.Cox@nrc-cnrc.gc.ca
Abbreviations: LPS, lipopolysaccharide; PEtn, phosphoethanolamine;
ESI, electrospray ionization; HSQC, heteronuclear single quantum
coherence.
(Received 21 January 2003, revised 18 February 2003,
accepted 24 February 2003)
Trang 2culture collection of E R Moxon (University of Oxford,
UK) and were grown on BHI agar plates as described [12]
Strain 1000 originates from a global collection of 34
representative serogroup B strains [13] Strain 425/93 is a
serogroup B carriage strain from the collection of P Kriz
(National Institute of Public Health, Prague, Czech
Repub-lic) isolated from the Czech Republic in 1993 [14] Strain
NM115 is a serogroup B strain from R Heyderman’s group
(Imperial College, London, UK) [15] LPS was extracted
from the fermenter-grown strains by the hot phenol/water
method as described previously and purified from the
dia-lysed aqueous phase by ultracentrifugation (45 000 r.p.m.,
was extracted from the plate-grown strains by the hot
phenol/water method and ethanol precipitation of the
each strain O-Deacylated LPS (LPS-OH) was prepared as
oligosaccharide was prepared by the following procedure
(8000 r.p.m., 20 min), and the supernatant was lyophilized,
Analytical methods
Sugars were determined as their alditol acetate derivatives
by GLC-MS as described previously [16]
Mass spectrometry
All electrospray ionization (ESI)-MS and capillary
electro-phoresis (CE)-ESI-MS analyses were carried out as
described previously [17]
NMR spectroscopy
NMR experiments were performed on Varian INOVA 500
and 400 NMR spectrometers as described previously [10]
Results
LPS was isolated from plate-grown (425/93 and NM115)
or fermenter-grown (1000 and its galE mutant) strains by
standard methods Sugar analysis of the LPS-derived
alditol acetates from the parent strains revealed glucitol,
in approximately equimolar ratios Sugar analysis of the
LPS-derived alditol acetates from the galE mutant of
O-Deacylated LPS (LPS-OH) from all strains was
prepared by hydrazinolysis, and initial analyses were
carried out by negative-ion ESI-MS and CE-ESI-MS
(Table 1) MS-MS enabled the size of the core
oligosac-charide and lipid A moiety to be deduced A consistent
variation in the size of the lipid A moiety was again
observed; this is due to a difference in the
phosphory-lation pattern of the lipid A region (unpublished data)
MS analysis indicated that there was no PEtn in the core
oligosaccharide, and this was consistent with these strains
not reacting with mAbs requiring the presence of PEtn in the core oligosaccharide (unpublished data) Sialylated glycoforms were only observed in strains NM115 and 425/93 4 Hex and sialylated 4 Hex were the major glycoforms observed for all three strains but not for the
observed for strain NM115, whereas 3 Hex glycoforms were also observed for strains 425/93 and 1000 (Table 1) Core oligosaccharide from the 1000 galE mutant strain was also prepared and examined by MS A range of molecular masses was found for the core oligosaccharide consistent with a composition of Kdo, 2Hep, GlcNAc, 2Glc with nonstoichiometric substitution with glycine and acetyl groups as observed previously for meningococcal LPS (Table 1) [18]
To elucidate the exact locations and linkage patterns of the oligosaccharide chains in the LPS, NMR studies were performed on the three LPS-OHs from the parent strains LPS-OH from strain 425/93 (Fig 1) and strain NM115
However, LPS-OH from strain 1000 initially gave a poor
addition of deuterated SDS (5 mg) and EDTA (0.5 mg)
very similar However, the anomeric signals of the lipid A amino sugars were only weakly visible for the LPS-OH from strain 1000, presumably because of extensive aggre-gation of this region of the molecule to suppress their
strains were assigned by COSY and TOCSY experiments Figure 2 shows a region of the TOCSY spectrum for the O-deacylated LPS from strain 425/93 Assignments were made by comparison with reported data for other meningococcal oligosaccharides [5–8,10], and are summar-ized in Table 2 In addition to the assignments tabulated, peaks corresponding to the axial ( 1.80 p.p.m.) and equatorial ( 2.75 p.p.m.) H-3 protons of the sialic acid
425/93 and NM115 Peaks corresponding to the acetyl groups of the N-acetylglucosamine residues and the equatorial and the axial H-3 protons of the Kdo residues were unresolved because of overlap with the N-linked fatty acid residues and the axial H-3 resonance of the sialic acid, respectively
In the representative spectrum of the LPS-OH from strain 425/93, spin systems arising from heptose residues (Hep I and Hep II) were readily identified from their
and 5.54 p.p.m (Hep II) and from the appearance of their spin systems, which pointed to manno-pyranosyl ring systems The heterogeneity observed for the ano-meric proton of Hep I was thought to be due to variation in phosphate substitution in the lipid A region
of the molecule (unpublished data) The a-configurations were evident for the heptosyl residues from the occur-rence of intraresidue NOEs between the H-1 and H-2 resonances only The remaining resolved residues in the a-anomeric region at 5.07 and 5.39 p.p.m and a minor signal at 5.50 p.p.m were determined to be gluco-pyranose amino sugars, from the appearance of their spin systems and the fact that the H-2 resonances of 3.89
Trang 3C14:
Trang 4single quantum coherence (HSQC) experiment with 13C
diagnostic of amino-substituted carbons The signals at 5.39 and 5.50 p.p.m were attributable to the a-glucos-amine residue of lipid A The heterogeneity of this residue was probably due to variation in phosphate substitution patterns in the lipid A region of the molecule (unpublished data) There were no other residues in the a-anomeric region, and of particular interest was the absence of an a-glucose residue that is often found substituting the 3 position of the Hep II residue in mAb B5 nonreactive strains such as immunotype strains L2 [5] and L5 [8] and clinical strain BZ157 [10] The remainder
of the anomeric resonances in the low-field region (4.45– 6.00 p.p.m.) of the spectrum were all attributable to b-linked residues by virtue of their chemical shifts and, in
coupling constants Three of these resonances at 4.72 (GlcNAc), 4.57 (Glc I) and 4.50 (Glc II) p.p.m were assigned to the gluco-configuration from the appearance
of their spin systems The resonance at 4.72 p.p.m was attributed to an amino sugar because its H-2 resonance
in the low-field region at 4.55 (Gal II) and 4.45 (Gal I) p.p.m were assigned to galacto-pyranosyl residues from the appearance of their characteristic spin systems
to the H-4 resonance in a TOCSY experiment
The sequence of glycosyl residues of the LPS-OH from
NOE measurements between anomeric and aglyconic protons on adjacent glycosyl residues (Table 2) Thus a lacto-N-neotetraose oligosaccharide unit attached to the proximal heptose residue was readily identified, as was
N-acetylglucosamine These are common structural motifs which have been identified in most N meningitidis immunotypes Intriguingly, a novel NOE contact was observed between the anomeric resonance of the Glc II residue at 4.50 p.p.m and the H-2 resonance of the Hep I residue at 4.20 p.p.m (Fig 3A) Similarly there was a NOE connectivity between the H-1 resonance of the Hep I residue at 5.45 p.p.m and the H-1 resonance
of the Glc II resonance at 4.50 p.p.m (Fig 3B) Taken together these NOE data suggested that the Hep I residue was also substituted at the 2 position by the Glc II residue This behaviour has been observed previ-ously for a b-Glc residue replacing a heptose residue at the 2 position [19], because of the proximity of the heptose H-1 proton enabling a NOE effect between the anomeric protons This structural arrangement has not been previously observed in meningococcal LPS Con-firmatory data for this novel linkage were obtained from
resonance for the H-2 proton of Hep I at 4.20 p.p.m
substitution at the 2 position of this Hep I residue
69–71 p.p.m for the H-2 resonance of the Hep I residue from the meningococcal immunotype strains LPS [6,7] Additional evidence for a novel substitution pattern
at Hep I was provided by an alteration in the inter-NOE
Trang 5contacts observed from the H-1 resonance of the Glc I
residue When Glc I substitutes Hep I at the 4 position,
NOE contacts are usually observed between the anomeric
proton resonance of the Glc I residue and the H-4 and
H-6 proton resonances of the Hep I residue [10] In the
present structure, a NOE contact was only observed to
the H-4 resonance, suggesting that a change in
confor-mation has occurred at Hep I caused by the presence of
the Glc II residue at the 2 position It was also possible
to discriminate between the 4 position of Hep I and the 2
position of Hep II, as the location of the Glc I residue,
because of the absence of characteristic H-1 to H-1 NOE
contacts normally observed for substitution of a heptose
residue at the 2 position [19] Almost identical results
were obtained from the LPS-OH of strain 1000 and strain NM115 Chemical shifts for the Gal II residue of strain 1000 were different because in this strain, Gal II is
a terminal residue whereas in strain 425/93 and NM115
it is substituted by sialic acid at the 3 position, as evidenced by the differences in the chemical shifts for the
strain 1000 and NM115 are summarized in Table 2, confirming that each strain had the same LPS structure, the major 4 Hex glycoform of which is depicted in Fig 5
To confirm the novel linkage pattern at Hep I, methy-lation analysis was carried out on the core oligosaccharide from the galE mutant of strain 1000 As expected, 1,2,3,4,5-O-acetyl-6,7-di-O-methylheptitol was identified (data not shown), consistent with the presence of the 2,3,4-trisubsti-tuted Hep I residue
Discussion
Strains 425/93 and NM115 were initially identified in a collection of meningococcal serogroup B isolates by virtue
of their lack of reactivity with mAbs specific for defined meningococcal LPS inner-core epitopes (unpublished data) Strain 425/93 came from a collection of carriage isolates [14] whereas the NM115 strain was isolated from patients with meningococcal sepsis [15] Strain 1000 was initially identified
in a collection of meningococcal clinical isolates by virtue of its lack of reactivity with mAbs identified to require PEtn at the distal heptose residue (Hep II) in the inner core ([12]; unpublished data) All previous structural studies on meningococcal LPS had revealed the same substitution pattern at the proximal heptose residue (Hep I), wherein Hep II substituted Hep I at the 3 position, and the first glucose residue (Glc I) of the oligosaccharide chain substituted Hep I at the 4 position Structural analysis of LPS from meningococcal strains 1000, 425/93 and NM115 revealed the same arrangement However, an additional glucose residue (Glc II) was also identified and found to substitute Hep I at the 2 position This organization at
Fig 2 b-Anomeric region of the TOCSY spectrum from the
O-deac-ylated LPS of N meningitidis strain 425/93 The spectrum was
recor-ded in D 2 O at 25 °C.
Fig 1 Anomeric region of the 1D1H-NMR
spectrum of the O-deacylated LPS from
N meningitidis strain 425/93 The spectrum
was recorded in D 2 O at 25 °C.
Trang 6Hep I has not been observed previously in meningococcal
LPS The observed lack of reactivity with several inner-core
mAbs would suggest that this novel substitution pattern
either alters the core conformation or masks
inner-core epitopes Molecular modelling studies are underway to
determine if the 2,3,4-trisubstituted Hep I residue adopts a
unique conformation because of the constraints of
substi-tution at the three ring positions A unique conformation
would also be consistent with the observed differences in the
NOE contacts for the Glc I to Hep I linkage In all three
strains examined, in which the Hep I residue of the LPS
bears the additional Glc II residue at the 2 position, only an
interresidue NOE to the H-4 proton of Hep I is observed
from Glc I The well-established NOE contacts between the
anomeric proton resonance of the Glc I residue and the H4
and H6 proton resonances of the Hep I residue [10],
observed in previous studies of meningococcal LPS, were
not observed, consistent with an altered pattern of
substi-tution at Hep I, suggesting that a change in conformation
has occurred caused by the presence of the Glc II residue at
the 2 position of Hep I Experiments will be initiated to
attempt to identify the gene encoding the
glucosyltrans-ferase responsible for the addition of this b-glucose residue
to the 2 position of Hep I Trisubstitution of the Hep I
residue of LPS has been observed in other bacterial species
An additional glucose residue has been identified previously
at the 6 position of the Hep I residue of LPS from Vibrio
residues have been identified as substituents at the 2 position
of Hep I with a b-configured glucuronic acid residue in the LPS of Vibrio parahaemolyticus O12 [22], and an a-configured galactose residue in the LPS from
the LPS of V parahaemolyticus O12 is identical with that identified here except that in the meningococcal strains investigated here it is a glucose residue at the 2 position It is intriguing that only a small number of the meningococcal strains so far examined elaborate this LPS structure, and it will be interesting to see how common this substitution pattern is and, perhaps more crucially, how many men-ingococcal strains possess the genetic machinery required to elaborate this novel structure These analyses have therefore revealed further potential for variation in the inner-core LPS of meningococcal strains The potential to vary the degree of substitution at the Hep I residue provides
conformation of its LPS epitopes and possibly affect its interaction with the host
Table 2. 1H-NMR chemical shifts (recorded at 25 °C, in D 2 O relative to HOD at 4.78 p.p.m) and NOE data for the LPS-OH from strains (i) 1000, (ii) 425/93, and (iii) NM115 ND, not determined.
Trang 7We are grateful to P Kriz(National Institute of Public Health, Prague)
and M Maiden (Oxford) who kindly provided strain 425/93 from a
collection of carriage strains from the Czech Republic We gratefully
acknowledge the contribution of O Harrison, C Ison and R
Heyder-man (Imperial College, London) who provided strain NM115 We
thank J Li for CE-MS-MS studies and D W Hood for valuable
discussions.
References
1 Kahler, C.M & Stephens, D.S (1998) Genetic basis for bio-synthesis, structure, and function of meningococcal lipooligo-saccharide (endotoxin) Crit Rev Microbiol 24, 281–334.
2 Scholten, R.J., Kuipers, B., Valkenburg, H.A., Dankert, J., Zollinger, W.D & Poolman, J.T (1994) Lipo-oligosaccharide immunotyping of Neisseria meningitidis by a whole-cell ELISA with monoclonal antibodies J Med Microbiol 41, 236–243.
3 Di Fabio, J.L., Michon, F., Brisson, J & Jennings, H.J (1990) Structure of L1 and L6 core oligosaccharide epitopes of Neisseria meningitidis Can J Chem 68, 1029–1034.
4 Wakarchuk, W.W., Gilbert, M., Martin, A., Wu, Y., Brisson, J.R., Thibault, P & Richards, J.C (1998) Structure of an alpha-2,6-sialylated lipooligosaccharide from Neisseria meningitidis immunotype L1 Eur J Biochem 254, 626–633.
5 Gamian, A., Beurret, M., Michon, F., Brisson, J.R & Jennings, H.J (1992) Structure of the L2 lipopolysaccharide core oligo-saccharides of Neisseria meningitidis J Biol Chem 267, 922–925.
Fig 3 Portion of (A) the b-anomeric region and (B) the a-anomeric
region of the NOESY spectrum from the O-deacylated LPS of
N meningitidis strain 425/93 The spectra were recorded in D 2 O at
25 °C In (B) for clarity the spectrum is shown at low intensity At
higher intensity levels a similar series of cross-peaks were observed
from the Hep I anomeric resonance at 5.42 p.p.m including NOE
contacts to Glc I H1 and Kdo H5.
Fig 4 2D1H-13C-HSQC spectrum of a region of the O-deacylated LPS from N meningitidis strain 425/93 illustratingthe substitution pattern of the Hep residues The spectrum was recorded in D 2 O at
25 °C.
Fig 5 Structure of the major 4 Hex LPS glycoform from N menin-gitidis strains 1000, 425/93 and NM115.
Trang 86 Pavliak, V., Brisson, J.R., Michon, F., Uhrin, D & Jennings, H.J.
(1993) Structure of the sialylated L3 lipopolysaccharide of
Neis-seria meningitidis J Biol Chem 268, 14146–14152.
7 Kogan, G., Uhrin, D., Brisson, J.R & Jennings, H.J (1997)
Structural basis of the Neisseria meningitidis immunotypes
including the L4 and L7 immunotypes Carbohydr Res 298,
191–199.
8 Michon, F., Beurret, M., Gamian, A., Brisson, J.R & Jennings,
H.J (1990) Structure of the L5 lipopolysaccharide core
oligo-saccharides of Neisseria meningitidis J Biol Chem 265, 7243–
7247.
9 Jennings, H.J., Johnson, K.G & Kenne, L (1983) The structure of
an R.-type oligosaccharide core obtained from some
lipopoly-saccharides of Neisseria meningitidis Carbohydr Res 121, 233–
241.
10 Cox, A.D., Li, J., Brisson, J.-R., Moxon, E.R & Richards, J.C.
(2002) Structural analysis of the lipopolysaccharide from Neisseria
meningitidis strain BZ157 galE: localisation of two
phosphoetha-nolamine residues in the inner core oligosaccharide Carbohydr.
Res 337, 1435–1444.
11 Jennings, M.P., van der Ley, P., Wilks, K.E., Maskell, D.J.,
Poolman, J.T & Moxon, E.R (1993) Cloning and molecular
analysis of the galE gene of Neisseria meningitidis and its role in
lipopolysaccharide biosynthesis Mol Microbiol 10, 361–369.
12 Plested, J.S., Makepeace, K., Jennings, M.P., Gidney, M.A.,
Lacelle, S., Brisson, J., Cox, A.D., Martin, A., Bird, A.G.,
Tang, C.M., Mackinnon, F.G., Richards, J.C & Moxon, E.R.
(1999) Conservation and accessibility of an inner core
lipopoly-saccharide epitope of Neisseria meningitidis Infect Immun 67,
5417–5426.
13 Seiler, A., Reinhardt, R., Sakari, J., Caugant, D.A & Achtman,
M (1996) Allelic polymorphisms and site-specific recombination
in the opc locus of Neisseria meningitidis Mol Microbiol 19, 841–
856.
14 Jolley, K.A., Kalmusova, J., Feil, E.J., Gupta, S., Musilek, M.,
Kriz, P & Maiden, M.C.J (2000) Carried meningococci in the
Czech Republic: a diverse recombining population J Clin.
Microbiol 38, 4492–4498.
15 Harrison, O.B., Robertson, B.D., Faust, S.N., Jepson, M.A.,
Goldin, R.D., Levin, M & Heyderman, R.S (2002) Analysis of
pathogen–host cell interactions in purpura fulminans: expression of
capsule, type IV pili, and porA by Neisseria meningitidis in vivo.
Infect Immun 70, 5193–5201.
16 Lysenko, E., Richards, J.C., Cox, A.D., Stewart, A., Martin, A., Kapoor, M & Weiser, J.N (2000) The position of phosphoryl-choline on the lipopolysaccharide of Haemophilus influenzae affects binding and sensitivity to C-reactive protein-mediated killing Mol Microbiol 35, 234–245.
17 Mackinnon, F.G., Cox, A.D., Plested, J.S., Tang, C.T., Make-peace, K., Coull, P.A., Wright, J.C., Chalmers, R., Hood, D.W., Richards, J.C & Moxon, E.R (2002) Identification of a gene (lpt-3) required for the addition of phosphoethanolamine to the lipopolysaccharide inner core of Neisseria meningitidis and its role
in mediating susceptibility to bactericidal killing and opsonopha-gocytosis Mol Microbiol 43, 931–943.
18 Cox, A.D., Li, J & Richards, J.C (2002) Identification and localisation of glycine in the inner core lipopolysaccharide of Neisseria meningitidis Eur J Biochem 269, 4169–4175.
19 Risberg, A., Masoud, H., Martin, A., Richards, J.C., Moxon, E.R & Schweda, E.K.H (1999) Structural analysis of the lipo-polysaccharide oligosaccharide epitopes expressed by a capsule-deficient strain of Haemophilus influenzae Rd Eur J Biochem.
261, 171–180.
20 Cox, A.D., Brisson, J.-R., Thibault, P., Perry, M.B (1997) Structural analysis of the lipopolysaccharide from Vibrio cholerae serotype O22 Carbohydr Res 303, 191–208.
21 Brisson, J.-R., Crawford, E., Uhrin, D., Khieu, N.H., Perry, M.B., Severn, W.B & Richards, J.C (2002) The core oligosaccharide component from Mannheimia (Pasteurella) haemolytica serotype A1 lipopolysaccharide contains 1-glycero- D -manno- and D -gly-cero- D -manno-heptoses Analysis of the structure and conforma-tion by high-resoluconforma-tion NMR spectroscopy Can J Chem 80, 949–963.
22 Kondo, S., Za¨hringer, U., Seydel, U., Sinnwell, V., Hisatsune, K.
& Rietschel, E.T (1991) Chemical structure of the carbohydrate backbone of Vibrio haemolyticus serotype O12 lipopolysaccharide Eur J Biochem 200, 689–698.
23 Aspinall, G.O., Monteiro, M.A., Pang, H., Kurjanczyk, L.A & Penner, J.L (1995) Lipo-oligosaccharide of Campylobacter lari strain PC 637 Structure of the liberated oligosaccharide and an associated extracellular polysaccharide Carbohydr Res 279, 227– 244.
24 Aspinall, G.O., Monteiro, M.A & Pang, H (1995) Lipo-oligo-saccharide of Campylobacter lari type strain ATCC 35221 Structure of the liberated oligosaccharide and an associated extracellular polysaccharide Carbohydr Res 279, 245–264.