Structural determination of the O-chain polysaccharide
from
Agrobacterium tumefaciens
, strain DSM 30205
Cristina De Castro
1
, Olga De Castro
2
, Antonio Molinaro
1
and Michelangelo Parrilli
1
1
Dipartimento di Chimica Organica and Biochimica, Universita
`
di Napoli; Complesso Universitario Monte Sant’ Angelo, Napoli, Italy;
2
Dipartimento di Biologia Vegetale, Universita
`
di Napoli, Italy
Agrobacterium tumefaciens is a Gram-negative, phytopatho-
genic bacterium and is characterized by an unique mode of
action on dicotyledonous plants: it is able to genetically
modify the host, and because of this feature, it is used as a
tool for transgenic plants. Many experiments have demon-
strated that lipopolysaccharides (LPSs) play an important
role for the disease development, as they are involved in the
adhesion process of the bacterium on the plant cell wall.
Despite the wealth of information on the role of LPS on
phytopathogenesis, the present paper appears as the first
report on the molecular primary structure of the O-chain

accepted mechanism, A. tumefaciens is attracted to wound
sites of the root surfaces by chemotaxis, and the presence of
phenolic compounds, such as acetosyringone, in synergy
with a certain class of monosaccharides (
D
-glucose,
D
-galactose,
L
-arabinose) triggers the activation of the
virulence genes [2]. In order to transfer its T-DNA into
the plant cell, the bacterium has to be adsorbed on the
wounded area; this event is modulated by the components
of the external membrane of the bacterium, both the
proteins and the lipopolysaccharides (LPS) [3]. In the latter
case, the interaction is based on the recognition of a portion
of the lipopolysaccharide, defined with the term epitope, by
particular receptor proteins [4] situated on the plant cell
wall. In fact, it is possible to saturate these receptors with an
LPS solution leading to the protection of the plant from the
bacterial action. Further studies showed that the epitope
recognized by the plant is located on the O-antigenic part of
LPS, that is the O-chain as demonstrated by the reduced
virulence of bacterial mutants of the O-antigenic part [4,5].
Despite the wealth of information regarding the biologi-
cal role of the LPS components, there are no data available
on their chemical structure so far. However, the information
we do have gives us some insight into the pathogenesis
mechanism.
MATERIAL AND METHODS

Eur. J. Biochem. 269, 2885–2888 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02955.x
with 0.85% NaCl, ethanol, acetone and diethyl ether.
Typically, 10 L of culture yielded to 0.5 g of dry cells.
Isolation and purification of the LPS fraction
Dried cells were extracted according to the phenol/water
method [6]. Both phases were separately dialyzed against
distilled water, freeze-dried and screened by 12% SDS/
PAGE [7] on a miniprotean gel system from Bio-Rad; the
samples (4 lg) where run at constant voltage (150 V) and
stained according to the procedure of Kittelberger [8].
Lipopolysaccharide material was found exclusively in the
water phase.
LPS fraction was further purified from proteic material,
and low molecular mass glucan, on Sephacryl HR 400
(Pharmacia, 1.5 · 90 cm, eluent NH
4
HCO
3
50 m
M
, flow
0.4 mLÆmin
)1
), eluate was monitored with a R.I. refrac-
tometer (R410 Waters) and the collected peaks screened
again on SDS/PAGE, leading to 27 mg (5.8% yield respect
dry cells) of LPS fraction.
Chemical composition analysis
Monosaccharides were analysed as acetylated methyl gly-
coside derivatives and lipids as methyl esters, according the

peak) and 24.038 (major peak); 2-(+)-octyl-
D
-arabinoside:
23.387 (minor peak), 24.229 (major peak) and 24.833 (minor
peak); -(+)-octyl-
L
-arabinoside: 23.736 (minor peak),
24.197 (major peak) and 24.467 (minor peak).
GC-MS analysis conditions for both fatty acids, methyl
and octyl glycoside derivatives were the same and were run
on a Hewlett-Packard 5970 instrument, using a SPB-5
capillary column (Supelco; 30 m · 0.25 inside diameter;
flow rate 0.8 mLÆmin
)1
; He as the carrier gas), with the
temperature program: 150 °C for 5 min, 150 to 300 °Cat
5.0 °CÆmin
)1
and 300 °Cfor15min.Massspectrawere
recorded using a ionization energy of 70 eV and a ionizing
current of 0.2 mA.
Glycosyl-linkage analysis of LPS, was performed accord-
ing to the procedure of Sandford [10]. The permethylated
lipopolysaccharide was recovered in the organic layer of the
water/chloroform extraction and converted into its partially
methylated alditol acetates [11], which were analyzed by
GC-MS, with the following temperature program: 80 °C
2 min, 80 to 240 °Cat4°CÆmin
)1
and 240 °Cfor15min.

phase-sensitive gradient-HSQC) were measured using
standard Bruker software.
For homonuclear experiments, typically 256 FIDs of
1024 complex data points were collected, with 40 scans per
FID. In all cases, the spectral width was set to 10 p.p.m. and
the frequency carrier was placed at the residual water peak.
A mixing time of 200 ms was used in the NOESY
experiment. For the HSQC spectrum, 256 FIDS of 1024
complex points were acquired with 50 scans per FID, the
GARP sequence was used for
13
C decoupling during
acquisition. Processing and plotting were performed with
a standard Bruker
XWINNMR
1.3 program.
RESULTS AND DISCUSSION
A. tumefaciens, strain DSM 30205 (type strain referenced
also as B6), possesses an S-type LPS as shown by the typical
ladder appearance located in the upper part of the gel
electrophoresis (Fig. 1).
The aqueous phase of the phenol/water treatment was
purified by GPC in order to remove other contaminants as
low molecular mass glucans and nucleic material.
The purified fraction was subjected to compositional
analysis and revealed the presence of 3-hydroxymyristic acid
together with minor amounts of palmitic, 3-hydroxy-
palmitic, 2-hydroxy-palmitic and stearic acids.
Monosaccharide composition revealed the presence of
Kdo and mannose in traces and the absence of heptose

analysis directly on the O-chain moiety, that provided
spectra with a resolution better than that of LPS spectra.
The separation of the O-chain and of lipid A moieties was
achieved selecting very mild conditions (sodium acetate at
pH 4.50 with 0.1% SDS at 100 °Cfor2h)inorderto
hydrolyse the Kdo linkage without effecting the acid-labile
furanosidic unit.
Combining the information from the analysis of the
COSY and NOESY spectra (Fig. 3) and HSQC, the
complete assignment of the
1
Hand
13
C signals was achieved
(Table 1).
Starting from the anomeric proton signals, it was possible
to identify all the protons of each residue through the
interproton scalar connectivity measured by a COSY
spectrum.
The broad singlet at 5.22 p.p.m. was assigned to the
anomeric proton A-1 of the arabinofuranose unit on the
basis of its correlations with the carbon signals at
110.7 p.p.m. [12], in addition, the low field chemical shift
of the A-3 carbon signal at 84.7 p.p.m., confirmed the
glycosylation of this position.
Fig. 2. 125 MHz carbon spectrum of lipopolysaccharide fraction from
A. tumefaciens B6 DSM 30205. Residue A: (3)-a-
D
-Araf-(1fi.Residue
B: (3)-a-

3.1 J
4,5
3.6 J
5,5¢
12.3 J
4,5¢
5,6
B 4.97 d 3.88 dd 3.95 3.97 4.27 q 1.21 d
100.2 68.3 78.4 73.1 68.3 16.5
3)-a-L-Fuc-(1fi J
1,2
4.0 J
2,3
9,9
aa
J
5,6
6.6 J
5,6
6.6
a
Overlapping signals.
Fig. 3. Section of NOESY (black) and COSY (grey) spectra of O-chain
moiety. Residue A: (3)-a-
D
-Araf-(1fi.ResidueB:(3)-a-
L
-Fucp-(1fi.
Ó FEBS 2002 O-chain structure from A. tumefaciens (Eur. J. Biochem. 269) 2887
The analysis of the spin system of B unit showed intense

whereas proton B-1 had a strong dipolar coupling with
proton A-3 and only a very weak one with proton A-4. The
two residues showed also some intraresidue diagnostic
NOEs, in particular the correlation between B-3 and B-5
suggested 1,3 diaxial orientation of these protons, and the
B-4/B-5 correation was expected due to the galacto-confi-
guration of this residue.
In conclusion, the spectroscopical information agreed
with the chemical analysis composition performed on the
purified S-type LPS. The O-chain structure is built of the
following repeating disaccharide unit:
3)-a-d-Araf-(1!3)-a-l-Fucp-(1!
This structure is the first reported for the Agrobacterium
genus, and in contrast to its apparent simplicity, it presents
some peculiarities.
As mentioned in the introduction, this bacterium
requires external factors to trigger its own virulence.
Such factors are provided from the wounded plant cell
wall and are phenolic compounds in synergy with
particular monosaccharides as
D
-galactose,
D
-fucose and
L
-arabinose [2]. On the other hand, the absolute config-
urations of the O-chain constituent residues is
D
for
arabinose and

progress to estimate the in vitro biological activity of the
O-chain.
ACKNOWLEDGEMENTS
The authors thank the ÔCentro di Metodologie Chimico-FisicheÕ of the
University Federico II of Naples for NMR facilities, the ÔProgetto
Giovani Ricercatori 2000Õ and L. R. 41/94 prot. CCAMAA370B2000
for financial support.
REFERENCES
1. Sigee, D.C. (1993) Bacterial Plant Pathology: Cell and Molecular
Aspects. Cambridge University Press, Cambridge.
2. Cangelosi, G.A., Ankenbauer, R.G. & Nester, E.W. (1990) Sugars
induce the Agrobacterium virulence genes through a periplasmic
binding protein and a transmembrane signal protein. Proc. Natl
Acad. Sci. USA 87, 6708–6712.
3. Pueppke, S.G. & Benny, U.K. (1984) Adsorption of tumorigenic
Agrobacterium tumefaciens cells to susceptible potato tuber tissues.
Can. J. Microbiol. 30, 1030.
4. Matthysse, A.G. (1986) Initial Interactions of Agrobacterium
Tumefaciens with plant host cells. CRC Crit. Rev. Microbiol. 13,
281.
5. New, P.B., Scott, J.J., Ireland, C.R., Farrand, S.K., Lippincott,
B.B. & Lippincott, J.A. (1983) Plasmid pSa causes loss of LPS-
mediated adherence in Agrobacterium. J. Gen. Microbiol. 129,
3657–3660.
6. Westphal, O. & Jann, K. (1965) Bacterial lipopolysaccharides
extraction with phenol-water and further applications of the
procedure. Methods Carbohydr. Chem. 5, 83–91.
7. Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature 97, 620–628.
8. Kittelberger, R. & Hilbink, F. (1993) Sensitive silver-staining