The HS:1 serostrain of Campylobacter jejuni has a complex
teichoic acid-like capsular polysaccharide with
nonstoichiometric fructofuranose branches and O-methyl
phosphoramidate groups
David J. McNally, Harold C. Jarrell, Jianjun Li, Nam H. Khieu, Evgeny Vinogradov,
Christine M. Szymanski and Jean-Robert Brisson
Institute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada
The Gram-negative, spiral-shaped bacterium Campylo-
bacter jejuni is one of the leading causative agents of
human enteritis and surpasses Salmonella, Shigella and
Escherichia in some regions as the primary cause of
gastrointestinal disease [1,2]. Furthermore, there is a
growing body of evidence suggesting that infection
with C. jejuni is linked to the development of Guillain-
Barre
´
and Miller-Fisher neuropathies [3–5]. Although
the bacteria now recognized as members of the genus
Campylobacter were first described at the beginning of
the 20th century, public awareness remains limited and
much of the biology of Campylobacters and the mech-
anisms by which they cause disease are still relatively
poorly understood [1].
In recent years, the health and economic burden
associated with Campylobacter infection has fueled
interest for this genus and in 2000, the genome sequence
for C. jejuni NCTC 11168 (HS:2) was reported [6].
Keywords
Campylobacter jejuni; capsular
polysaccharide; CE-ESI-MS; HR-MAS NMR;
phosphoramidate

P heteronuclear two-dimensional HR-MAS NMR at
500 MHz. In contrast, spectroscopic data acquired for hot water ⁄ phenol
purified CPS was complicated by the hydrolysis and subsequent loss of
labile groups during extraction. Collectively, the results of this study estab-
lished the importance of using sensitive isolation techniques and HR-MAS
NMR to examine CPS structures in vivo when labile groups are present.
This study uncovered how incorporation of variable O-methyl phosphor-
amidate groups on nonstoichiometric fructose branches is used in C. jejuni
HS:1 as a strategy to produce a highly complex polysaccharide from its
small CPS biosynthetic locus and a limited number of sugars.
Abbreviations
CPS, capsular polysaccharide; CE-ESI-MS, capillary electrophoresis-electrospray ionization-mass spectrometry; HMBC, heteronuclear multiple
bond coherence; HMW LPS, high-molecular-weight lipopolysaccharides; HR-MAS NMR, high resolution magic angle spinning nuclear
magnetic resonance; HSQC, heteronuclear single quantum coherence; MeOPN, O-methyl phosphoramidate CH
3
OP(O)(NH
2
)(OR).
FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS 4407
Analysis of the genome sequence revealed that this
strain possesses four gene clusters necessary for carbo-
hydrate biosynthesis [6]. Among these was the first des-
cription of a type II ⁄ III capsule locus similar to that
found in encapsulated organisms, such as those found
in Escherichia coli K1 and Neisseria meningitidis group
B [6,7]. In particular, identification of kps genes,
responsible for transferring the CPS repeat to the outer
membrane, prompted a systematic genetic analysis of
the corresponding locus and resulted in the identifica-
tion of CPS in several strains of C. jejuni [8]. The CPS

selected strains of C. jejuni were sequenced including:
serostrain HS:41 (176.83), 81–176 (HS:23 ⁄ 26), sero-
strain HS:36 (ATCC 43456), serostrain HS:23 (CCUG
10954), serostrain HS:19 (NCTC 12517) and G1
(HS:1) [14]. Comparison of the determined cps
sequences of the HS:19, HS:41 and HS:1 strains with
the genome sequenced NCTC 11168 (HS:2) strain pro-
vided evidence for multiple mechanisms of CPS vari-
ation including exchange of capsular genes and entire
clusters by horizontal transfer, gene duplication, dele-
tion, fusion and contingency gene variation [14]. The
study also demonstrated for the first time that strains
belonging to the same serogroup (e.g. 81–176
(HS:23 ⁄ 36), HS:23 and HS:36) contain capsule loci
with the same gene complement. In contrast to
C. jejuni NCTC 11168 and HS:19, the biosynthetic
region of the HS:1 strain was the smallest and was
shown to contain only 11 genes (Fig. 1a). Of import-
ance, the cps locus of the HS:1 strain was shown to
contain a tagD homologue encoding a glycerol-3-phos-
phate cytidylyltransferase necessary for the biosynthe-
sis of CDP-glycerol. Moreover, it was shown that the
(a)
(b)
Fig. 1. Predicted capsule gene schematic and determined struc-
tures for the defructosylated repeating unit (CPS-1) and complete
CPS structure (CPS-2) of the C. jejuni HS:1 serostrain. (a) Carbohy-
drate biosynthetic genes located between the genes encoding the
capsule transport system are shown from the sequenced locus of
the HS:1 strain, G1 [14]. The phase variable genes in G1 which

LPS and preliminary NMR data obtained for the par-
tially purified CPS of G1 (HS:1) and the HS:1 sero-
strain of C. jejuni [14]. For instance, we detected the
presence of at least two acid-labile groups and provi-
ded evidence showing that one of these was likely an
MeOPNCH
3
OP(O)(NH
2
)(OR) modification similar to
the one identified on the CPS structure of the genome-
sequenced strain of C. jejuni, NCTC 11168 [14,15].
In the current study, we thoroughly investigated the
chemical structure of CPS for the HS:1 serostrain
of C. jejuni. Initially, CPS was isolated from bacterial
cells using a traditional hot water ⁄ phenol method [19];
however, due to the extent of structural degradation
observed for CPS purified using this method, a gentler
procedure for isolating CPS was required to preserve
the labile constituents of HS:1 CPS. Accordingly, the
methods of Darveau and Hancock [20], Huebner et al.
[21] and Hsieh et al. [22] were combined and used to
isolate CPS from this strain of C. jejuni. High resolu-
tion NMR at 600 MHz with an ultra-sensitive cryo-
genically cooled probe was then used to elucidate the
structure of purified CPS, and HR-MAS NMR at
500 MHz was used to examine native CPS directly
on the surface of whole bacterial cells. Concurrently,
CE-ESI-MS and in-source collision-induced dissoci-
ation [23] was used to analyze the structure of purified

cells. By suspending an enzyme purified sample of
HS:1 CPS in nonbuffered D
2
O (pD 2.2), the auto-
hydrolyzed defructosylated repeating unit was obtained
(CPS-1), as well as other hydrolysis fragments. The
1
H
NMR spectrum of this auto-hydrolyzed CPS sample
showed sharp spectral lines and one major anomeric
signal for Gal H-1 (Fig. 2A). Signals originating from
the methyl group of the MeOPN modification, nor-
mally present at 3.78 p.p.m., were absent [15]. The
1
H
NMR spectrum of a hot water ⁄ phenol purified sample
of HS:1 CPS showed two broad anomeric signals for
Gal H-1 and resonances originating from the MeOPN
modification were weak and therefore difficult to
observe (Fig. 2B). In contrast, the spectrum for the
enzyme isolated CPS sample (CPS-2) showed one
signal for Gal H-1 and signals originating from the
methyl group of the MeOPN modification were sharp
and clearly discernable (Fig. 2C).
HR-MAS NMR of HS:1 cells provided valuable
insight into the nature of cell-bound CPS on the sur-
face of bacterial cells. The HR-MAS
1
H NMR spec-
trum of HS:1 cells (Fig. 2D) closely resembled the

ville, Canada) indicated that fructose also had the D
configuration (data not shown).
The chirality of naturally occurring glycerols can be
determined chemically, enzymatically or can be deduced
from the biosynthetic pathway responsible for their pro-
duction [24]. When CDP-glycerol is used as a precursor
to incorporate glycerol in the growing repeating chain
of teichoic acids, the resulting glycerol-1-phosphate
unit has the D, or R configuration [24]. Alternatively,
when glycerophosphate is biosynthetically derived from
phosphatidylglycerol, the resulting product is L-or
S-glycerol-1-phosphate [24]. Based on previous work
where we reported that the CPS biosynthetic locus of
a HS:1 strain of C. jejuni contains a tagF homologue
responsible for transferring glycerol-phosphate residues
from CDP-glycerol [14], the glycerol-1-phosphate resi-
due was concluded to have the R configuration.
High resolution NMR analysis of auto-hydrolyzed
enzyme purified CPS (CPS-1)
Due to the complexity of the NMR spectrum of the
native CPS, the backbone structure was first deter-
mined. Examination of an auto-hydrolyzed defructo-
sylated enzyme purified HS:1 CPS sample revealed a
Fig. 2. NMR analysis of purified and cell-
bound C. jejuni HS:1 CPS. (A)
1
H NMR
spectrum of an auto-hydrolyzed enzyme
purified CPS sample. (B)
1

J
P,H
¼ 10 Hz)
for cell-bound CPS. For the selective 1D
experiments, excited resonances are under-
lined.
Campylobacter jejuni HS:1 serostrain CPS D. J. McNally et al.
4410 FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS
[-4)-a-d-Galp-(1–2)-(R)-Gro-(1-P]
n
repeating unit
(CPS-1) as well as other hydrolysis products (Fig. 3).
The 1D-TOCSY of Gal H-1 revealed J-correlated
peaks for Gal H-2, H-3 and H-4 (Fig. 3A). The
1D-NOESY of Gal H-4 revealed NOEs for Gal H-3
and H-5 (Fig. 3B), and the 1D-TOCSY of Gal H-5
identified the Gal H-6 resonances (Fig. 3C). A
1D-NOESY-TOCSY experiment with selective excita-
tion of Gal H-1 ⁄ Gro H-2 was used to identify the
glycerol resonances (Fig. 3D). The Gal H-1 ⁄ Gro C-2
HMBC correlation confirmed the Gal-(1–2)-Gro link-
age. The
31
P HSQC spectrum (Fig. 3E) showed that
Gal H-4 and Gro H-1 ⁄ 1¢ were linked by a phospho-
rus atom with a chemical shift characteristic of a
monophosphate diester bond [25,26]. The
13
C HSQC
spectrum (Fig. 3G) and HMBC spectrum were used

140 Hz, 8
transients and 256 increments. For
the selective 1D experiments, excited
resonances are underlined. Ff and Fp repre-
sent the fructofuranose and fructopyranose
monosaccharides, respectively.
Table 1. NMR proton and carbon chemical shifts d (p.p.m) for an
auto-hydrolyzed enzyme purified sample of C. jejuni HS:1 CPS
(CPS-1) and corresponding hydrolysis products. The
31
P chemical
shift for the monophosphate diester linkage was d
P
0.49 p.p.m.
Atom Type
CPS-1
d
H
d
C
A1 CH 5.20 98.9
A2 CH 3.87 70.4
A3 CH 3.98 69.9
A4 CH 4.54 75.5
A5 CH 4.17 71.5
A6 ⁄ A6¢ CH
2
3.74 ⁄ 3.74 61.6
B1 ⁄ B1¢ CH
2

it permitted the acquisition of
1
H and
13
C NMR
experiments in a relatively short period of time. The
1D-TOCSY of Gal H-1 revealed two separate reso-
nances for Gal H-2, H-3 and H4 labeled A2a,A2b,
A3a,A3b,A4a and A4b, respectively. (Fig. 4A). The
1D-NOESY of Gal H-4a showed NOEs for Gal H-2a,
Gal H-3a, Gal H-5 and Gal H-6 ⁄ 6¢, as well as for Fru
H-4 and Fru H-6 ⁄ 6¢ (Fig. 4B). Conversely, excitation
of Gal H-4b revealed NOE enhancements for Gal
H-2b, Gal H-3b, Gal H-5 and H-6 ⁄ 6¢ as well as for
Fru H-6 ⁄ 6¢ (Fig. 4C). A 1D-NOESY ⁄ TOCSY experi-
ment with selective excitation of Gal H-1 and Gro
H-1 ⁄ 1¢ permitted the assignment of Gro H-2 and Gro
H-3 ⁄ 3¢ (Fig. 4D).
The HMBC experiment revealed three-bond correla-
tions between Gal H-2, Gal H-3b and Fru C-2 indicating
that two fructose branches were present for the CPS of
C. jejuni HS:1 (data not shown). The 1D-NOESY of
Fru H-3 revealed enhancements for Fru H-1 ⁄ 1¢, H-4
and H-5 (Fig. 4E) while the 1D-TOCSY of Fru H-4
showed correlations to Fru H-3 and H-5 (Fig. 4F).
The
31
P HSQC experiment revealed monophosphate
diester linkages between Gal C-4a, C-4b and Gro C-1
with different chemical shifts at d

1D-NOESY-TOCSY of Gal H-1 (400 ms) and
Gro H-1 ⁄ 1¢ (50 ms). (E) 1D-NOESY (400 ms)
of Fru H-3. (F) 1D-TOCSY (80 ms) of Fru
H-4. (G)
31
P HSQC with
1
J
P,H
¼ 20 Hz, 8
transients and 32 increments. (H)
13
C HSQC
with
1
J
C,H
¼ 150 Hz, 80 transients and
256 increments. For the selective 1D
experiments, excited resonances are
underlined. Residue C represents Fru with
MeOPN present and residue *C, Fru with
no MeOPN.
Campylobacter jejuni HS:1 serostrain CPS D. J. McNally et al.
4412 FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS
C. jejuni HS:1 LPS (CPS) [17,31] as well as for pep-
tides and nucleic acids. Furthermore, signals belonging
to nonsubstituted b-fructofuranoside indicated that
fructose branches were variably substituted with
MeOPN groups [27–29].

ing unit, ions at m ⁄ z 639.4 and 801.6 corresponding
to (Hex)
3
+ GroP and (Hex)
4
+ GroP, respectively,
confirmed the attachment of both fructose branches on
galactose. Furthermore, ions observed at m ⁄ z 671.4,
894.6, 905.5 and 987.7 corresponding to (Hex)
3
+
(MeOPN)
2
, (Hex)
4
+ GroP + MeOPN, (Hex)
3
+
GroP + (MeOPN)
2
+ P, and (Hex)
4
+ GroP +
(MeOPN)
2
, respectively, supported that MeOPN
groups were located on both fructose branches. Of
particular importance, CE-ESI-MS ⁄ MS analysis of
m ⁄ z 732.5, corresponding to one full repeat of HS:1
CPS, showed an ion at m ⁄ z 658.2, corresponding to

spin system a represents the form where the fructose
branch at Gal C-3 was absent.
Table 2. NMR proton and carbon chemical shifts d (p.p.m) for an
intact enzyme purified sample of C. jejuni HS:1 CPS (CPS-2). The
31
P chemical shifts for the monophosphate diester linkages of Gal
H-4a and b were d
P
0.40 p.p.m and 0.49 p.p.m., respectively. The
31
P chemical shift for the MeOPN groups was 14.67 p.p.m., and a
scalar coupling
3
J
P,H
of 11.1 Hz was observed.
Atom Type
CPS-2
d
H
d
C
A1 CH 5.40 98.8
A2a CH 4.29 68.5
A2b CH 4.28 68.3
A3a CH 4.33 69.4
A3b CH 4.40 68.7
A4a CH 4.74 77.2
A4b CH 4.69 77.3
A5 CH 4.16 72.0

C – 104.1
C3
a
CH 4.12 77.0
C4
a
CH 4.12 76.6
C5
a
CH 3.75 81.5
C6 ⁄ C6¢
a
CH
2
3.86 ⁄ 3.77 62.5
a
Chemical shift data d (p.p.m) for unsubstituted b-D-fructofurano-
side (MeOPN is absent).
D. J. McNally et al. Campylobacter jejuni HS:1 serostrain CPS
FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS 4413
HR-MAS NMR spectroscopy of cell-bound CPS
In order to characterize the heterogeneity of the CPS
in its native state, HR-MAS NMR studies were per-
formed on intact cells. As observed for the purified
CPS, two signals arising from Gal H-4, H-4a and
H-4b, were detected on the surface of HS:1 cells and
appeared to be present in equal proportions (Fig. 2).
The 1D NOESY of Gal H-1 for an enzyme purified
CPS sample showed NOEs for Gal H-2, Gro H-1 ⁄ 1¢,
Gro H-2 and Gro H-3 ⁄ 3¢ (Fig. 2E). Likewise, in the

The correlation between MeOPNatd
P
14.67 p.p.m.
with H-3 of fructofuranose was also observed. Hence,
the NOEs and
31
P HSQC indicated that structural het-
erogeneity due to different branching patterns on the
Gal residue was also present for intact cells.
Molecular dynamics simulations
Three models were constructed for the [-4)-a-d-Galp-
(1–2)-(R)-Gro-(1-P-]
n
repeating unit of HS:1 CPS
representing different substitution patterns for the fruc-
tose branches located at C-2 and C-3 of a-d-Galp
(present ⁄ absent, absent ⁄ present and present ⁄ present).
These models were then used to verify NOEs observed
during NMR analysis. A minimum energy conformer
generated using a Metropolis Monte-Carlo calculation
for HS:1 CPS with both MeOPN-substituted fructose
branches (in the same plane as the page) attached to
the repeating unit (out of plane, with P closest to the
reader) is shown in Fig. 6. Molecular dynamics simula-
tions showed that regardless of the substitution pattern
of Gal, the average interproton distance between Gal
H-1 and Gro H-2 was approximately 2.6 A
˚
± 0.2 A
˚

2
O
molecule is added to the residues unless specifically indicated.
Molecular mass (m ⁄ z)
Structure
Observed Calculated Difference
153.1 153.1 0.0 GroP
171.3 171.1 0.2 GroP +H
2
O
223.3 223.1 0.2 Hex + P ) (H
2
O)
2
254.8 254.2 0.6 Hex + MeOPN
259.0 259.1 0.1 Hex + P +H
2
O
297.3 297.2 0.1 Hex + GroP ) H
2
O
315.3 315.2 0.1 Hex + GroP
333.5 333.2 0.3 Hex + GroP +H
2
O
377.5 377.2 0.3 Hex + GroP + P ) H
2
O
385.3 385.2 0.1 (Hex)
2

O
477.3 477.3 0.0 (Hex)
2
+GroP
487.3 487.3 0.0 Hex + (GroP)
2
+H
2
O
490.8 490.4 0.4 (Hex)
2
+ Gro + MeOPN
495.5 495.4 0.1 (Hex)
2
+GroP +H
2
O
509.6 509.4 0.2 (Hex)
2
+ Gro + MeOPN
539.3 539.3 0.0 (Hex)
2
+GroP + P ) H
2
O
551.5 551.4 0.1 (Hex)
2
+GroP +Gro
557.5 557.3 0.2 (Hex)
2

3
+ MeOPN+P
667.5 667.4 0.1 (Hex)
2
+(GroP)
2
+(H
2
O)
2
671.8 671.5 0.3 (Hex)
3
+ (MeOPN)
2
701.5 701.4 0.1 (Hex)
3
+GroP+P ) H
2
O
723.3 723.5 0.2 (Hex)
2
+(GroP)
2
+Gro+H
2
O
732.3 732.5 0.2 (Hex)
3
+GroP +MeOPN
751.4 751.4 0.0 (Hex)

O)
2
855.3 855.5 0.2 (Hex)
3
+(GroP)
2
+ P ) H
2
O
Table 3. (Continued).
Molecular mass (m ⁄ z)
Structure
Observed Calculated Difference
886.8 886.6 0.2 (Hex)
3
+(GroP)
2
+ MeOPN
894.5 894.6 0.1 (Hex)
4
+GroP +MeOPN
905.5 905.5 0.0 (Hex)
3
+GroP +(MeOPN)
2
+ P
929.8 929.6 0.2 (Hex)
3
+(GroP)
3

H-4 ⁄ Gro H-3 were on the order of 5–7 A
˚
thereby
negating the likelihood of observing these interresidue
NOEs.
Discussion
We previously demonstrated that HR-MAS NMR
can be used to rapidly compare C. jejuni CPS struc-
tures from intact cells and provided the first struc-
tural evidence that CPS is associated with Penner
serotype [14,15]. In this study, we investigated the
CPS structure for the representative HS:1 serostrain
of C. jejuni to complement data recently reported for
CPS biosynthesis in strain G1 (HS:1) [14], and to
determine the structure of labile CPS constituents
not detected by previous studies examining HMW
LPS (CPS) for the HS:1 serostrain [17,18]. Together,
different analytical methods showed that the HS:1-
type CPS of C. jejuni is complex and has a teichoic
acid-like [-4)-a-d-Galp-(1–2)-(R)-Gro-(1-P-]
n
repeating
unit with a b-d-fructofuranose branch at C-2 of Gal,
a nonstoichiometric fructose branch at C-3 of Gal
and variable MeOPN modifications on C-3 of both
fructose sugars.
By using a conventional hot water ⁄ phenol CPS
isolation method [19] and a more sensitive enzymatic
approach [20–22], we demonstrated that the method
used to isolate CPS was an important factor that influ-

nisms of infectivity and to guide the development of
effective therapeutics for this bacterium. This latter
point is illustrated by the fact that although fructose
has been reported for only a few bacterial CPSs, it was
found to be the immunodominant sugar of the cap-
sular K11 antigen of Escherichia coli O13:K11:H11
[28,29,32,33].
NMR and mass spectrometry analyses of an auto-
hydrolyzed defructosylated sample of enzyme purified
HS:1 CPS showed that it resembled teichoic acid, and
consisted of a [-4)-a-d-Galp-(1–2)-(R)-Gro-(1-P-]
n
repeating unit (CPS-1). Carbon and proton chemical
shifts were identical to those of the capsular antigen of
Neisseria meningitidis that has the same backbone [34].
Moreover, these findings supported those reported for
HMW LPS (CPS) isolated from this strain of C. jejuni
by McDonald [17], who showed it to consist of a [-4)-
a-d-Gal-(1–2)-Gro-(3-P-]
n
repeating unit. However, the
presence of fructose or MeOPN modifications was
not reported. The extraction and purification methods
used by the previous work probably resulted in the
Fig. 6. Molecular model for the C. jejuni HS:1 CPS. The [-4)-a-D-
Galp-(1–2)-(R)-Gro-(1-P-]
n
repeating unit with both MeOPN-substi-
tuted b -
D-fructofuranose branches at C-2 and C-3 of Gal. An

donor capabilities, phosphoramidates are thought to
interact nonspecifically with accessible amino acids of
proteins [38]. Furthermore, there is a growing body of
evidence suggesting that natural phosphoramidates,
such as phosphohistidine, play an important role in
two-component and phosphorelay signal transduction
pathways in bacteria that mediate responses such as
sporulation, chemotaxis, mucoidy, and flagellar move-
ment to environmental stimuli [39–45]. Accordingly, a
range of small-molecular-weight phosphoramidate
molecules have been identified that are able to elicit
similar responses from bacteria and are therefore
thought to mimic these naturally occurring phosphor-
amidate messengers [38–40,44,46]. Although a two
component system regulating growth and colonization
in response to environmental temperature was reported
for C. jejuni [47], the relationship between the biologi-
cal roles reported for phosphoramidates in other bac-
teria and the MeOPN CPS modification in C. jejuni is
not clear. Thus, the biological role of this capsular
modification in C. jejuni, much like the biosynthetic
pathway responsible for its production, is unknown at
this time.
In conclusion, in this study we determined the com-
plete structure of the CPS for the C. jejuni HS:1 sero-
strain. In doing so, we established the importance of
using mild isolation methods and noninvasive analyt-
ical techniques for examining CPS in this bacterium
due to the presence of highly labile constituents that
are easily overlooked using conventional methods.

mately determine the potential of the MeOPN modifi-
cation as a useful marker and therapeutic target for
this mucosal pathogen.
Experimental procedures
Solvents and reagents
Unless otherwise stated, all solvents and reagents were
purchased from Sigma Biochemicals and Reagents (Oakville,
Canada).
Media and growth conditions
The C. jejuni HS:1 serostrain (ATCC 43429, designation
MK5-S7630) was routinely maintained on Mueller Hinton
(MH) agar (Difco, Kansas City, MO, USA) plates under
microaerophilic conditions (10% CO
2
,5%O
2
, 85% N
2
)at
37 °C. For large scale extraction of CPS, 6 L of C. jejuni
HS:1 was grown in brain heart infusion (BHI) broth
(Difco) under microaerophilic conditions at 37 °C for 24 h
with agitation at 100 r.p.m. Bacterial cells were then har-
vested by centrifugation (9000 g for 20 min) and placed in
D. J. McNally et al. Campylobacter jejuni HS:1 serostrain CPS
FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS 4417
70% (v ⁄ v) ethanol. Cells were removed from the ethanol
solution by centrifugation (9000 g for 20 min) and the bac-
terial pellet was refrigerated until extraction.
Hot water/phenol isolation of CPS

TM
ion exchange
column (Amersham Biosciences, Piscataway, NJ, USA).
Fractions containing CPS were combined, flash frozen in
an acetone ⁄ dry ice bath and lyophilized to dryness. Puri-
fied bacterial CPS was then de-salted using a SephadexÒ
superfine G-15 column (Sigma) and fractions found to
contain CPS were combined, flash frozen in an acetone ⁄
dry ice bath, lyophilized to dryness and stored at )20 °C
until further analysis.
Enzymatic isolation of CPS
An enzymatic method of isolating CPS from C. jejuni HS:1
cells was developed based on the methodologies of [20],
Huebner et al. [21] and Hsieh et al. [22]. Bacterial cells har-
vested from 6 L of BHI broth were suspended in NaCl ⁄ P
i
buffer (pH 7.4). Lysozyme was then added to a final con-
centration of 1 mgÆmL
)1
(Sigma) prior to the addition of
mutanolysin to a final concentration of 67 UÆmL
)1
(Sigma).
The bacterial cell suspension was then incubated for 24 h at
37 °C with agitation at 100 r.p.m. The mixture was then
emulsiflexed twice (21 000 psi) to lyse cells, and DNAse I
and RNAse (130 lgÆ mL
)1
DNAse I and RNAse, Sigma)
was added prior to being incubated for 4 h at 37 °C with

achieved by the addition of 0.5 mL of acetic anhydride and
heating at 85 °C for 30 min prior to being dried at room
temperature under a nitrogen stream. Alditol acetate
derived CPS sugars were then suspended in 1.5 mL of
CH
2
Cl
2
and analyzed using an Agilent 6850 series GC sys-
tem, equipped with an Agilent 19091 L-433E 50% phenyl
siloxane capillary column (30 m · 250 lm · 0.25 lm)
(170 °C to 250 °C, 2.8 °CÆmin
)1
) (Agilent Technologies,
Palo Alto, CA, USA). Alditol acetate derivatives of authen-
tic standards for common keto and aldo sugars (Sigma)
were then prepared using the same protocol outlined above.
The composition of C. jejuni HS:1 CPS was then unambig-
uously determined by comparing the retention times of
CPS alditol acetate derivatives to those of authentic stand-
ards.
Determination of absolute configuration for
enzyme purified CPS
The absolute configuration (d or l) of galactose within an
enzyme purified sample of HS:1 CPS was assigned by char-
acterization of its R-butyl glycoside using GC according to
Loentein et al. [49]. Approximately 300 lLofR-butanol
and 30 lL of acetyl chloride (Sigma) was added to 1 mg of
enzyme purified CPS. The mixture was then heated at
85 °C for 3 h prior to being dried under a nitrogen stream

Sigma) to eliminate traces of d-glucose.
HR-MAS NMR spectroscopy of cell-bound CPS
For HR-MAS analysis, C. jejuni HS:1 cells were prepared
as according to Szymanski et al. [15]. Overnight growth
from one MH agar plate was harvested and placed in 1 mL
of 10 mm potassium-buffered 98% D
2
O (pD 7.0) (Cam-
bridge Isotopes Laboratories Inc, Andover, MA, USA)
containing 10% sodium azide (w ⁄ v) for 1 h at room tem-
perature to kill cells. Cells were then pelleted by centrifuga-
tion (8900 g for 2 min), and washed once with 10 mm
potassium-buffered D
2
O. Approximately 10 lLof1%
(w ⁄ v) TSP was then added as an internal standard
(0 p.p.m) to the cell suspension prior to being loaded into a
40 lL nano-NMR tube (Varian, Palo Alto, CA, USA)
using a long tipped pipette cut diagonally approximately
1 cm from the end. HR-MAS experiments were performed
using a Varian Inova 500 MHz spectrometer equipped with
a Varian 4 mm indirect detection gradient nano-NMR
probe with a broadband decoupling coil (Varian) as previ-
ously described [2,15,51]. Spectra from 40 lL cell samples
were spun at 3 kHz and recorded in ambient temperature
(23 °C), or at 10 °C to shift the HOD signal, and all experi-
ments were performed with suppression of the HOD signal.
1
H NMR spectra of bacterial cells were acquired using the
Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence (90-(s-

days using NMR analysis at 600 MHz with an ultra-sensi-
tive, cryogenically cooled probe. Analysis of the repeating
unit and hydrolysis products over time was facilitated by
the high sensitivity of the cryoprobe as
13
C HSQC spectra
were typically acquired in approximately 1 h. For analysis
of hot water ⁄ phenol purified CPS and enzyme purified CPS
samples (CPS-2), a 3 mg sample of each was suspended in
150 lLofNH
4
HCO
3
buffered 99% D
2
O (54 mm, pD 8.6),
placed in 3 mm NMR tubes and analyzed by NMR.
For all CPS samples,
1
H NMR,
13
C HSQC, HMBC,
HMQCTOCSY, COSY, TOCSY, NOESY and selective
one-dimensional TOCSY, NOESY and NOESY-TOCSY.
NMR experiments were performed at 600 MHz (
1
H) using
a Varian 5 mm, Z-gradient triple resonance cryogenically
cooled probe (Varian). The methyl resonance of acetone
was used as an internal reference (d

used for general assignments, and selective one-dimensional
TOCSY and NOESY experiments with a Z-filter were used
for complete residue assignment and characterization of
individual spin systems [53,54].
Mass spectrometry analysis
CE-ESI-MS and CE-ESI-MS ⁄ MS analysis was performed
using a Crystal Model 310 Capillary Electrophoresis instru-
ment (ATI Unicam, Boston, MA, USA) coupled to a 4000
QTRAP mass spectrometer (Applied Biosystems ⁄ Sciex,
Concord, Canada) via a Turbo ‘V’ CE-MS probe. A sheath
solution (isopropanol ⁄ methanol, 2 : 1, v ⁄ v) was delivered at
a flow rate of 1 lLÆmin
)1
. Separations were achieved on
approximately 90 cm of bare fused-silica capillary (360 lm
outside diameter · 50 lm i.d., Polymicro Technologies,
Phoenix, AZ, USA) and 15 mm ammonium acetate ⁄ ammo-
nium hydroxide in deionized water (pH 9.0) containing 5%
(v ⁄ v) MeOH as mobile phase. A voltage of 20 kV was
D. J. McNally et al. Campylobacter jejuni HS:1 serostrain CPS
FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS 4419
typically applied during CE separation and )5 kV was used
as electrospray voltage. Mass spectra were acquired with
dwell times of 5.5 ms per step of 0.1 m ⁄ z
)1
unit in Q1 scan
mode. Tandem mass spectra were acquired in the enhanced
product ion scan (EPI) mode, using nitrogen as collision
gas. Fill time of the trap (Q3) was set to 20 ms and the
LIT scan rate was adjusted to 4000 amuÆs

Lamoureux for assistance in growing and harvesting
cells and the Natural Sciences and Engineering
Research Council of Canada (NSERC) for funding.
References
1 Ketley JM (1997) Pathogenesis of enteric infection by
Campylobacter. Microbiology 143, 5–21.
2 St Michael F, Szymanski CM, Li J, Chan KH, Khieu
NH, Larocque S, Wakarchuk WW, Brisson JR &
Monteiro MA (2002) The structures of the lipooligosac-
charide and capsule polysaccharide of Campylobacter
jejuni genome sequenced strain NCTC 11168. Eur J
Biochem 269, 5119–5136.
3 Rhodes KM & Tattersfield AE (1982) Guillain–Barre
´
syndrome associated with Campylobacter infection. Br
Med J Clin Res 285, 173–174.
4 Kaldor J & Speed BR (1984) Guillain–Barre
´
syndrome
and Campylobacter jejuni: a serological study. Br Med J
Clin Res 288, 1867–1870.
5 Jacobs BC, Endtz H, van der Meche FG, Hazenberg
MP, Achtereekte HA & van Doorn PA (1995) Serum
anti-GQ1b IgG antibodies recognize surface epitopes on
Campylobacter jejuni from patients with Miller Fisher
syndrome. Ann Neurol 37, 260–264.
6 Parkhill J, Wren BW, Mungall K, Ketley JM, Churcher
C, Basham D, Chillingworth T, Davies RM, Feltwell T,
Holroyd S, Jagels K, Karlyshev AV, Moule S, Pallen
MJ, Penn CW, Quail MA, Rajandream MA, Ruther-

Demonstration of polysaccharide capsule in Campylo-
bacter jejuni using electron microscopy. Infect Immun
69, 5921–5924.
14 Karlyshev AV, Champion OL, Churcher C, Brisson
JR, Jarrell HC, Gilbert M, Brochu D, St Michael F,
Li J, Wakarchuk WW, Goodhead I, Sanders M,
Stevens K, White B, Parkhill J, Wren BW &
Szymanski CM (2005) Analysis of Campylobacter
jejuni capsular loci reveals multiple mechanisms for
the generation of structural diversity and the
ability to form complex heptoses. Mol Microbiol 55,
90–103.
Campylobacter jejuni HS:1 serostrain CPS D. J. McNally et al.
4420 FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS
15 Szymanski CM, St Michael F, Jarrell HC, Li J, Gilbert
M, Larocque S, Vinogradov E & Brisson JR (2003)
Detection of conserved N-linked glycans and phase-vari-
able lipooligosaccharides and capsules from campylo-
bacter cells by mass spectrometry and high resolution
magic angle spinning NMR spectroscopy. J Biol Chem
278, 24509–24520.
16 Roberts IS (1996) The biochemistry and genetics of cap-
sular polysaccharide production in bacteria. Annu Rev
Microbiol 50, 285–315.
17 McDonald AG (1993) Lipopolysaccharides from Cam-
pylobacter. PhD Thesis, York University, North York,
Canada.
18 Moran, AP, Penner, JL & Aspinall, GO (2000)
Campylobacter lipopolysaccharides. In Campylobacter
(Nachamkin I & Blaser MJ, eds), pp. 241–257.

(1956) Principles of phosphorus chemistry. II. Nuclear
magnetic resonance measurements. J Am Chem Soc 78,
5715–5726.
26 Nielsen ML, Pustinger JV & Strobel J (1964) Phospho-
rus-31 nuclear magnetic resonance chemical shifts of
phosphorus compounds. J Chem Eng Data 9, 167–170.
27 Bock K & Pedersen C (1983) Carbon-13 nuclear mag-
netic resonance spectroscopy of monosaccharides. Adv
Carbohydr Chem Biochem 41, 27–66.
28 Hofmann P, Jann B & Jann K (1985) Structure of the
fructose-containing K52 capsular polysaccharide of uro-
pathogenic Escherichia coli O4:K52:H Eur J Biochem
147, 601–609.
29 Rodriguez ML, Jann B & Jann K (1990) Structure and
serological properties of the capsular K11 antigen of
Escherichia coli O13:K11:H11. Carbohydr Res 196, 101–
109.
30 Guijarro JI, Gonzalez-Pastor JE, Baleux F, San Millan
JL, Castilla MA, Rico M, Moreno F & Delepierre M
(1995) Chemical structure and translation inhibition
studies of the antibiotic microcin C7. J Biol Chem 270,
23520–23532.
31 Aspinall GO, McDonald AG, Raju TS, Pang H, Moran
AP & Penner JL (1993) Chemical structures of the core
regions of Campylobacter jejuni serotypes O:1, O:4,
O:23, and O:36 lipopolysaccharides. Eur J Biochem 216,
880.
32 Volpi N (2003) Milligram-scale preparation and purifi-
cation of oligosaccharides of defined length possessing
the structure of chondroitin from defructosylated capsu-

proteins by low molecular weight phospho-donors. Proc
Natl Acad Sci USA 89, 718–722.
41 Welch M, Oosawa K, Aizawa S & Eisenbach M (1993)
Phosphorylation-dependent binding of a signal molecule
D. J. McNally et al. Campylobacter jejuni HS:1 serostrain CPS
FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS 4421
to the flagellar switch of bacteria. Proc Natl Acad Sci
USA 90, 8787–8791.
42 Amster-Choder O & Wright A (1997) BglG, the
response regulator of the Escherichia coli bgl operon, is
phosphorylated on a histidine residue. J Bacteriol 179,
5621–5624.
43 McEvoy MM & Dahlquist FW (1997) Phosphohisti-
dines in bacterial signaling. Curr Opin Struct Biol 7,
793–797.
44 Mayover TL, Halkides CJ & Stewart RC (1999) Kinetic
characterization of CheY phosphorylation reactions:
comparison of P-CheA and small-molecule phospho-
donors. Biochemistry 38 , 2259–2271.
45 Hoch JA (2000) Two-component and phosphorelay sig-
nal transduction. Curr Opin Microbiol 3, 165–170.
46 Silversmith RE, Appleby JL & Bourret RB (1997) Cata-
lytic mechanism of phosphorylation and dephosphoryla-
tion of CheY: kinetic characterization of imidazole
phosphates as phosphodonors and the role of acid cata-
lysis. Biochemistry 36, 14965–14974.
47 Bras A, Chatterjee S, Wren BW, Newell DG & Ketley
JM (1999) A novel Campylobacter jejuni two-component
regulatory system important for temperature-dependent
growth and colonization. J Bacteriol 181, 3298–3302.

55 Humphrey W, Dalke A & Schulten K (1996) VMD: vis-
ual molecular dynamics. J Mol Graph 14, 33–38.
Campylobacter jejuni HS:1 serostrain CPS D. J. McNally et al.
4422 FEBS Journal 272 (2005) 4407–4422 ª 2005 FEBS