Identification and role of a 6-deoxy-4-keto-hexosamine in the lipopolysaccharide outer core of yersinia enterocolitica serotype O:3

Article abstract of DOI:10.1002/chem.200901255

The outer core (OC) region of Yersinia enterocolitica serotype 0:3 lipopolysaccharide is a hexasaccharide essential for the integrity of the outer membrane. It is involved in resistance against cationic antimicrobial peptides and plays a role in virulence

Full text of DOI:10.1002/chem.200901255

FULL PAPER
DOI: 10.1002/chem.200901255
Identification and Role of a 6-Deoxy-4-Keto-Hexosamine in the
Lipopolysaccharide Outer Core of Yersinia enterocolitica Serotype O:3
[
a, b]
[c, d]
[c]
[c, e]
Elise Pinta,
Katarzyna A. Duda,
Anna Hanuszkiewicz, Zbigniew Kaczy n´ ski,
[g] [a] [h]
[
f]
Buko Lindner, Wayne L. Miller, Heidi Hyytiꢀinen, Christian Vogel,
[
h]
[d]
[g]
Sabine Borowski, Katarzyna Kasperkiewicz, Joseph S. Lam,
[
d]
[a, i]
[c]
Joanna Radziejewska-Lebrecht, Mikael Skurnik,*
and Otto Holst
Abstract: The outer core (OC) region
(Sugp) and that WbcP is a UDP-
GlcNAc-4,6-dehydratase enzyme re-
sponsible for the biosynthesis of the
nucleotide-activated form of this rare
sugar converting UDP-2-acetamido-2-
deoxy-d-xylo-hex-4-ulopyranose (UDP-
Sugp). In an aqueous environment, the
4-keto group of this sugar was present
in the 4-dihydroxy form, due to hydra-
tion. Furthermore, evidence is provided
that the axial 4-hydroxy group of this
dihydroxy function was crucial for the
biological role of the OC, that is, in the
bacteriophage and enterocoliticin re-
ceptor structure and in the epitope of a
monoclonal antibody.
of Yersinia enterocolitica serotype O:3
lipopolysaccharide is a hexasaccharide
essential for the integrity of the outer
membrane. It is involved in resistance
against cationic antimicrobial peptides
and plays a role in virulence during
early phases of infection. We show
here that the proximal residue of the
OC hexasaccharide is a rarely encoun-
tered 4-keto-hexosamine, 2-acetamido-
deoxy-d-glucopyranose
(UDP-d-
GlcpNAc) to UDP-2-acetamido-2,6-di-
Keywords:
biosynthesis
·
hexosamines · lipopolysaccharides ·
receptors · structural biology
2,6-dideoxy-d-xylo-hex-4-ulopyranose
+
[
a] E. Pinta, Dr. H. Hyytiꢀinen, Dr. M. Skurnik
[f] Dr. B. Lindner
Infection Biology Research Program, Haartman Institute
Department of Bacteriology and Immunology
P.O.Box 21 (Haartmaninkatu 3)
Division of Immunochemistry, Research Center Borstel
Leibniz-Center for Medicine and Biosciences
23845 Borstel (Germany)
0
0014 University of Helsinki (Finland)
[
g] Dr. W. L. Miller, Dr. J. S. Lam
Department of Molecular and Cellular Biology, University of Guelph
Ontario N1G 2W1 (Canada)
Fax : (+358)9-191 26382
E-mail: mikael.skurnik@helsinki.fi
+
[
b] E. Pinta
[
h] Dr. C. Vogel, S. Borowski
Department of Medical Biochemistry and Genetics
University of Turku, 20520 Turku (Finland)
Institute of Chemistry, University of Rostock
1
8059 Rostock (Germany)
+
+
[
c] Dr. K. A. Duda, Dr. A. Hanuszkiewicz, Dr. Z. Kaczy n´ ski,
Dr. O. Holst
Division of Structural Biochemistry, Research Center Borstel
Leibniz-Center for Medicine and Biosciences
[
i] Dr. M. Skurnik
Helsinki University Central Hospital Laboratory Diagnostics
0
0029 Helsinki (Finland)
+
[
] These authors contributed equally to this work.
2
3845 Borstel (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200901255.
+
[
d] Dr. K. A. Duda, Dr. K. Kasperkiewicz, Dr. J. Radziejewska-Lebrecht
Department of Microbiology, University of Silesia
4
0-032 Katowice (Poland)
[
e] Dr. Z. Kaczy n´ ski
Department of Chemistry, University of Gda n´ sk
8
0-952 Gda n´ sk (Poland)
Chem. Eur. J. 2009, 15, 9747 – 9754
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9747

Introduction
LPS, express both OPS and OC), YeO3-R1 (OPS mutant),
and YeO3-c-trs22-R (OPS mutant expressing truncated OC)
and YeO3-c-trs24-R (OPS and OC-mutant) were attained
(the methods used are described in the Supporting Informa-
tion). Data from ESI-FTICR-MS experiments (Figure S2 in
the Supporting Information) indicated that the OC hexasac-
charide was composed of three hexose and three hexosa-
mine residues. Surprisingly though, in the LPS possessing
[1]
Lipopolysaccharide/s (LPS) are an integral part of the
outer membrane (OM) of the Gram-negative human patho-
gen Yersinia enterocolitica (Ye). Lipopolysaccharides are
structurally very diverse but are usually composed of three
[2,3]
main parts: the lipid A,
often be divided into inner (IC) and outer core (OC),
the core oligosaccharide that can
[
4–6]
[
7,8]
+
and an O-specific polysaccharide (OPS).
Compared to
the OC, [MÀ18] ions were identified, which were shown to
the classical Escherichia coli or Salmonella LPS organiza-
tion, the LPS of Y. enterocolitica serotype O:3 (Ye O:3)
shows some structural peculiarities. In the Ye O:3 LPS struc-
ture both the OC and OPS are linked to IC. The OC region
of Ye O:3 is a hexasaccharide unit, which plays an important
role in virulence during early phases of infection by media-
originate from the elimination of one water molecule (Figur-
e S2A–D, left panel). Such dehydration products were not
discerned from LPS of YeO3-c-trs24-R comprising only lipid
A and IC regions; thus, this reaction had to occur in the
OC.
Chemical analyses revealed the presence of sugars (in ad-
dition to fatty acids) typical for lipid A and IC, that is, 2-
amino-2-deoxy-d-glucose (d-GlcN), 3-deoxy-d-manno-oct-2-
ulosonic acid (Kdo), d-glycero-d-manno-heptose (d,d-Hep),
l-glycero-d-manno-heptose (l,d-Hep), and d-Glc. Except
for the LPS from strain YeO3-c-trs24-R, the OC specific res-
idues, that is, d-Gal and 2-amino-2-deoxy-d-galactose (d-
GalN), were also present in all LPS examined. In addition,
traces of 2-amino-2,6-dideoxy-galactose (FucN) and 2-
amino-2,6-dideoxy-glucose (QuiN) were found. At this point
it was believed that the OC contained three residues of
GalNAc, contrary to the published two GalNAc and one N-
[9]
ting resistance against bactericidal peptides.
The LPS biosynthetic gene clusters of bacteria include
genes encoding the enzymes needed for the biosynthesis of
nucleotidediphosphate-activated sugar precursors (NDP-
sugars), specific glycosyltransferases, and translocation pro-
teins. The Ye O:3 OC gene cluster contains nine genes in
the order wzx, wbcKLMNOP, gne, and wbcQ.
function of gne has been characterized and it encodes an
[10]
[
9,11,12]
The
[13]
UDP-N-acetyl-glucosamine-4-epimerase (EC 5.1.3.7). The
wbcKLMNOQ genes encode six putative glycosyltransferas-
es needed to form the six glycosidic linkages of the OC hex-
asaccharide, whereas the wzx gene encodes a flippase that
translocates the undecaprenylphosphate (Und-P)-linked OC
[11]
acetylated 6-deoxy-hexosamine (FucNAc) residues. How-
ever, the above-described elimination of one water molecule
could not be explained.
[12]
across the inner membrane.
The wbcP gene product
shows similarity to certain NDP-sugar biosynthetic enzymes
and was predicted to be involved in the biosynthesis of the
precursor of the proximal OC sugar. The wbcO gene prod-
uct is predicted to be the priming glycosyltransferase trans-
Next, it was noticed that different chemical treatments of
the LPS samples resulted in different sugar compositions. In
particular, reduction of Ye75S, YeO3-R1, and YeO3-c-trs22-
R LPS by sodium borohydride (NaBH ) or by mild hydra-
4
[14,15]
ferring the proximal sugar to the carrier lipid Und-P.
zine treatment (30 min, 378C, to O-deacylate the sample)
resulted in significantly different proportions of FucN and
QuiN, that is, in approximate molar ratios of approximately
7:3 and 1:10, respectively. All other components were pres-
ent in molar ratios similar to those of the untreated LPS.
To address this discrepancy, core oligosaccharide/s (OS)
were isolated from the LPS of Ye O:3 strains Ye75S and
YeO3-R1 either by acetate buffer hydrolysis or mild hydra-
zinolysis/hot KOH treatment, followed by purification utiliz-
ing various chromatographic steps. The compositional analy-
ses of the core OS isolated from the LPS gave different re-
sults depending on the treatment of the LPS. In addition to
d-Gal, d-GalN, l,d- and d,d-Hep, Kdo, d-Glc, and d-GlcN,
the OS from hydrazine/KOH treated YeO3-R1 LPS showed
the presence of d-QuiN in quantitative amounts, plus some
traces of d-FucN. The structure of this OS was elucidated by
NMR spectroscopy (see below), and all the sugars were
identified as pyranose residues. In contrast, the OS from
acetate buffer hydrolysis possessed only trace amounts of d-
FucpN and d-QuipN (data not shown).
The proximal sugar of the Ye O:3 OC structure was as-
signed as 2-acetamido-2,6-dideoxy-galactose (FucpNAc),
with some uncertainty in its putative phosphate substitu-
[11]
tion; therefore, the OC structure needed additional struc-
tural analyses. In this work, we present a revised structure
of the OC and show unambiguously that the first sugar is a
non-phosphorylated 2-acetamido-2,6-dideoxy-d-xylo-hex-4-
ulopyranose (Sugp). We also demonstrate that WbcP is a
UDP-N-acetyl-d-glucosamine-4,6-dehydratase that catalyzes
the conversion of UDP-GlcpNAc to UDP-Sugp, the latter
serves as the precursor of the first sugar residue in the OC
hexasaccharide biosynthesis. Furthermore, evidence is pro-
vided that the axial 4-hydroxy group of this dihydroxy func-
tion was crucial for the biological role of the OC.
Results and Discussion
The dilemma of the Ye O:3 OC composition: To establish
the nature of the proximal OC sugar, high-resolution mass
spectrometric and compositional analyses of LPS (see the
Supporting Information) isolated from strains Ye75S, YeO3-
c (two different Ye O:3 wild-type strains with regards to
Mass spectrometric analyses of the OS from acetate-hy-
drolyzed LPS showed the presence of a component with a
molecular mass of 2407.8 u, consistent with a molecule in
the lactone form comprising 1 Kdo, 4 Hep, 5 Hex, and 3
9748
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9747 – 9754

Identification and Role of a 6-Deoxy-4-Keto-Hexosamine
FULL PAPER
HexNAc residues (Figure 1A). The MS analyses of the OS
from hydrazine/KOH treated LPS showed a molecular peak
at 2985.9 u, which was consistent with the saccharide com-
posed of two phosphate groups, four HexN, two Kdo, four
Hep, five Hex, and dHexN, the latter of which was identi-
fied as QuiN (Figure 1B).
of the 4-keto form (Figure S2 in the Supporting Informa-
tion).
To prove this reduction–hydration hypothesis, we incubat-
1
8
ed dried LPS from strain YeO3-R1 in H2 O and immediate-
ly analyzed the product(s) by ESI-FTICR MS (Figure 2).
Figure 2. Mass spectrometric analysis of LPS from YeO3-R1 before
1
8
(
2
black line) and after (red line) exchange with H O. See text for details.
Figure 1. Charge deconvoluted ESI mass spectra of oligosaccharides iso-
lated from Y. enterocolitica LPS. A) the spectrum of the OS obtained
after acetate buffer hydrolysis. B) the spectrum of the OS obtained after
hydrazine and KOH treatment. See text for details.
The obtained mass spectra revealed peaks with shifts of 2
and 4 u in components comprising either the 4-keto or the
4
-dihydroxy form. Nonlabeled LPS containing Sugp in its 4-
Solution of the dilemma and the final core structure: The
data presented above showed that the OC region of the LPS
of Ye O:3 did not contain FucpNAc, but another sugar that
possessed the molecular mass of HexNAc and could be re-
duced by various agents to both, FucpNAc and QuipNAc.
This has led us to hypothesize that FucpNAc and QuipNAc
were derived from the reduction of 2-acetamido-2,6-di-
deoxy-d-xylo-hex-4-ulopyranose (Sugp), as reduction of d-
keto form gave a molecular mass of 3998.77 u, whereas LPS
1
8
possessing O in its 4-keto group gave rise to a mass peak
at 4000.78 u. The dihydroxy form was identified in nonla-
beled LPS by a molecular peak at 4016.76 u. The mass spec-
1
8
tra of O-labelled LPS revealed two shifts, that is, of 2 and
1
8
4 u. In the first case, Sugp possessed only one OH-group
(shift of 2 u), as shown at 4018.78 u, and in the latter case
1
8
two OH groups were present (shift of 4 u), which resulted
in a molecular peak at 4020.79 u. Such shifts were also ob-
served in all other molecular peaks consistent with LPS con-
taining the complete core OS. After having identified this,
the +1 u shift of the dHexN residue observed in GCMS
[16]
xylo-hex-4-ulopyranose yields quinovose and fucose. In an
aqueous environment, the 4-keto group undergoes hydra-
tion, resulting in a 4-dihydroxy (diol) derivative. This dihy-
droxy form possesses the same molecular mass as a
+
HexNAc. In ESI-FTICR-MS spectra, [MÀ18] ions were
analyses of LPS from YeO3-R1 after reduction with sodium
2
observed in the LPS samples, which indicated the presence
borodeuteride (NaB H ) could be interpreted as due to the
4
Chem. Eur. J. 2009, 15, 9747 – 9754
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9749

M. Skurnik et al.
incorporation of one deuterium into the 4-keto group of
Sugp. Taken together, our data provided unambiguous struc-
tural evidence for the presence of one residue of 2-acetami-
do-2,6-dideoxy-d-xylo-hex-4-ulopyranose in the OC region
of the LPS from Ye O:3. These results were substantiated by
comparison to authentic synthetic standards and also by
NMR spectroscopy (see the Supporting Information for de-
tails), by which the structure of the OS isolated after hydra-
zine/KOH treatment was determined (Figure 3A). Based on
the data from the aforementioned chemical and structural
analyses, we are able to illustrate a revised structure of the
lipid A core region of the LPS of Ye O:3 (Figure 3B).
and WbcP-E243, all exhibited enzymatic activity as demon-
strated by the consumption/decrease of the UDP-GlcpNAc
substrate peak while the emergence of a new substance elut-
ing at 13.2 min became apparent (Figure 4A). Compared to
WbcP-S212 and WbcP-E243, the wt WbcP reaction generat-
ed one additional product eluting at 16.7 min (Figure 4A).
This compound was identified as UMP by comparison to re-
actions that were spiked with UMP as a control (see Fig-
ure S4 in the Supporting Information for details). In LPS
biosynthesis UMP can be released only when the first sugar
[17]
[10]
residue is transferred from UDP-sugar to Und-P.
We
postulated that WbcP was involved in the biosynthesis of
the first sugar residue, Sugp, of
OC (see below) and that WbcP
was able to donate Sugp to the
E. coli priming glycosyltransfer-
ase present in the crude mem-
brane preparation used.
Identification of the product of
WbcP: Two homologues of
WbcP, WbpM of P. aerugino-
[
18,19]
sa,
and FlaA1 of Helico-
[20]
bacter pylori,
have been
shown to catalyze a reaction to
yield a product eluting also at
1
3.2 min in CE; and the prod-
uct was identified as UDP-
Sugp, an intermediate in the
biosynthetic pathway to gener-
Figure 3. Characterized oligosaccharide structures of LPS from Y. enterocolitica O:3. A) the structure of the
core oligosaccharide isolated after hydrazine/KOH treatment of LPS from strain YeO3-R1. B) The revised
structure of the backbone of the core-lipid A region of LPS from Y. enterocolitica O:3. All sugars are pyrano- ate UDP-N-acetyl-2,6-dideoxy-
ses. Residues A, C, D, E, G, I, J, L, N, and O are a-linked and residues B, F, H, K, K’, M, and P b-linked. Com- hexosamines (such as UDP-
ponents I, G, E are l-glycero-d-manno-, and residue J is d-glycero-d-manno-configured. All other sugars are
d-configured.
QuipNAc and UDP-FucpNAc
of P. aeruginosa). To verify the
structure of the WbcP reaction
product, we applied a specific
Biosynthesis of Sugp: The investigation of the biosynthesis
of the proximal sugar, Sugp, of Ye O:3 OC was initiated
based on in silico analysis of the nine genes in the biosyn-
combined reaction in which the reaction product was fed to
WbgX, an aminotransferase from Plesiomonas shigel-
[21]
loides. The substrate of WbgX is UDP-Sugp and it produ-
[9,12]
[22]
thesis cluster, which suggested a role for wbcP.
The
ces UDP-2-acetamido-4-amino-2,4,6-trideoxy-d-galactose.
amino acid sequence of WbcP is 48% identical to that of
To confirm the identity of the intermediate, we used a bio-
chemical assay in which purified WbgX was added to the
enzyme reactions carried out with wt WbcP, WbcP-S212, or
WbcP-E243 (Figure 4B). The CE analyses showed that the
product peak of the WbcP enzymatic activity, eluting at
13.2 min, decreased, which demonstrated that WbgX could
use the product of WbcP as a substrate. This indicated that
the final WbcP product was UDP-Sugp, the intermediate in
the UDP-QuipNAc and UDP-FucpNAc synthesis. The
UMP peak in wt WbcP reactions was unaffected as expect-
ed. These results suggested that WbcP is a UDP-GlcNAc-
4,6-dehydratase (EC number is pending, a request has been
made to IUBMB).
[18]
WbpM (see Figure S3 in the Supporting Information),
a
Pseudomonas aeruginosa enzyme involved in the biosynthe-
sis of UDP-FucpNAc. The N-terminal third of WbpM func-
tions as a membrane anchor, whereas the C-terminal two
[18]
thirds encompass the catalytic constituent.
To facilitate
purification of WbcP, we designed two N-terminally truncat-
ed versions, WbcP-E243 and WbcP-S212, and the latter was
predicted to allow for better exposure of the N-terminal
His-tag (see the Supporting Information for details). WbcP-
S212 was purified to near homogeneity (> 95% purity),
whereas full length WbcP and WbcP-E243 were only ob-
tained as sonicated cell extracts.
The catalytic activity of WbcP was assessed by CE using
various NDP sugars as substrates. Wild-type (wt) WbcP that
was partially purified as a constituent of the cell extract, and
the two truncated versions of purified proteins, WbcP-S212
The UDP-Sugp reductase activity is missing in Ye O:3
and expression of the UDP-Sugp reductase activity in Y. en-
terocolitica disturbs OC biosynthesis. Our chemical data pro-
vided the evidence that in Ye O:3, the Sugp and not its re-
9750
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9747 – 9754

Identification and Role of a 6-Deoxy-4-Keto-Hexosamine
FULL PAPER
corresponding activities of these two enzymes are missing
from Ye O:3.
The DOC-PAGE analysis revealed that LPS was only par-
tially substituted by OC. This could be due to several differ-
ent reasons: 1) depleting the pool of UDP-Sugp by WbpK
Figure 5. Effect of WbpK and WbpV on OC synthesis in YeO3-R1.
Silver-stained DOC-PAGE analysis of isolated LPS from Ye O:3 strains
[
28]
YeO3-R1, YeO3-c-OCR (a strain with the OC gene cluster deleted),
YeO3-R1/pEPwbpVRe, YeO3-R1/pEPwbpK, and YeO3-R1/pEPwbpVc.
Note that the plasmid pEPwbpVRe carrying the wbpV gene in reverse
orientation has no effect on LPS pattern. The IC arrow indicates a LPS
band that contains lipid A plus IC and the IC+OC arrow a LPS band
that contains lipid A plus IC and OC.
Figure 4. CE analyses of WbcP reactions. A) WbcP reaction uses UDP-
GlcNAc (UG) as a substrate and peaks at 13.2 and 16.7 min are formed.
Pellet (P) or supernatant (S) extracts from E. coli BL21 ACHUTNGRENNUG( DE3) cells carry-
ing plasmids pET-28a (vector only, Ve), pETWbcP (pWbcP), or
pETWbcP-S212 (pS212). The identity of peaks is indicated by arrows.
Panel B, WbcP-WbgX coupled assay. CE analysis of different cell extract
reactions after addition of purified WbgX. The open arrow points to the
peak of WbgX product.
or WbpV would decrease the OC biosynthesis, 2) the pri-
ming glycosyltransferase WbcO would transfer QuipNAc or
FucpNAc less efficiently to Und-P than 4-keto-Sugp, 3) nei-
ther Und-P-P-QuipNAc nor Und-P-P-FucpNAc could be
used as a substrate by WbcQ (UDP-GalpNAc:Und-P-P-
[24]
duced derivative is incorporated in the OC hexasaccharide.
This could either be due to a tight coupling of the biosynthe-
sis of the UDP-Sugp to the priming glycosyltransferase reac-
tion; thus sheltering the substrate from the reductases, or a
lack of the reductase activity from Ye O:3. We were able to
address this question indirectly by expressing the wbpK and
wbpV genes in trans in the Ye O:3 strain YeO3-R1. WbpK
and WbpV are putative 4,6-reductases of P. aeruginosa asso-
ciated with the conversion of the keto intermediate, UDP-
Sugp, to UDP-FucpNAc and UDP-QuipNAc, respective-
ly. If the activities of WbcP and the primase were tightly
coupled we should not see any change in OC expression.
On the other hand, if the expression of wbpK and wbpV in
trans caused disturbance in the OC biosynthesis, it would in-
dicate that the reductase activity was abrogated in Ye O:3.
Our results supported the latter scenario; the DOC-
PAGE analysis of LPS showed that expression of both
WbpK and WbpV in Ye O:3 affected the LPS banding pro-
file and that the LPS in the transformed strains was only
partially substituted with OC (Figure 5). Data from GC and
GCMS analyses of the LPS showed the presence of both IC
and OC components in both strains (data not shown). In ad-
dition, unlike wt LPS, the 6-deoxy-hexosamines QuipNAc
and FucpNAc were present in LPS of YeO3-R1/pEPwbpVc
and YeO3-R1/pEPwbpK, respectively. These results demon-
strated that 1) WbpV and WbpK function as 4,6-reductases
in UDP-QuipNAc and UDP-FucpNAc biosynthesis, respec-
tively, 2) they were both active in vivo in Ye O:3, and 3) the
Sugp a1-3 GalpNAc-transferase), which resulted in a LPS
substituted only by QuipNAc or FucpNAc, 4) WbcQ could
use UndPP-QuipNAc and/or UndPP-FucpNAc as sub-
strates, albeit at lower efficiency, and this would result in re-
duced amounts of OC substitutions, or 5) either the OC-flip-
pase, Wzx, or OC-ligase could transfer or ligate the Quip-
NAc or FucpNAc-containing OC at lower efficiency. To ad-
dress these alternatives, LPS of the strains were analyzed by
ESI-FTICR MS and also by Capillary Skimmer Dissociation
(CSD) ESI-FTICR MS (see Table S3 in the Supporting In-
formation). The ESI-FTICR MS measurements confirmed
that both YeO3-R1/pEPwbpVc and YeO3-R1/pEPwbpK
LPS contained 6d-HexNAc in the OC structure instead of
the Sugp present in YeO3-R1. Furthermore, methylation
analysis of these LPS preparations confirmed the presence
of 3-substituted QuipNAc and FucpNAc (data not shown),
respectively, thus disproving alternative (3). These results,
therefore, suggest that one of the alternatives (2), (4), and
(5) is correct; however, at present we cannot differentiate
between them.
[
23]
The biological significance of the keto–diol group of Sugp:
As described above, the mild hydrazine treatment of YeO3-
R1 LPS quantitatively reduces the Sugp of the OC to Quip-
NAc. As this treatment does not affect the N-acylated resi-
dues, the only additional difference in the hydrazine-treated
LPS is loss of the two O-linked fatty acids. Thus, to evaluate
the biological significance of the Sugp in the Ye O:3 LPS
Chem. Eur. J. 2009, 15, 9747 – 9754
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9751

M. Skurnik et al.
structure we tested the neutralization potential of natural
and hydrazine-treated LPS to the lytic bacteriophage fR1-
the reducing effect of hydrazine on Sugp. If Sugp would be
present in LPS of YeO3-R1/pEPwbpK, traces of QuiN in
addition to FucN should be detectable after hydrazine treat-
ment. This was indeed the case and we could roughly esti-
mate that less than 5–10% of the LPS had contained the
Sugp (data not shown).
3
tor.
7 and enterocoliticin that both utilize the OC as a recep-
[
25,26]
Our results showed that the hydrazine-treated LPS
À1
at 500 mgmL was incapable of neutralizing either fR1-37
or enterocoliticin. In contrast, the use of 5 mgmL of the wt
À1
LPS was able to completely neutralize both. Furthermore,
À1
as little as 0.5 mgmL of wt LPS was effective in completely
inhibiting the activity of 100 units of enterocoliticin (see
Table S4 in the Supporting Information). In addition, mAb
Conclusions
[27,28]
2
5
2
B5, specific for Ye O:3 OC,
ng of Ye O:3 LPS in a dot-blotting assay. In contrast, mAb
B5 did not react with up to 2 mg of the hydrazine-treated
readily interacted with
We have provided conclusive evidence that the OC structure
in Ye O:3 contains Sugp as the reducing end (proximal)
sugar instead of FucNAc as was reported earlier.
[11]
We
LPS (see Figure S5 in the Supporting Information). These
data collectively indicated that the Sugp of the Ye O:3 OC
is biologically important; it either forms part of the receptor
structure of both bacteriophage fR1-37 and enterocoliticin
and the mAb 2B5 epitope or it allows the OC to take the
correct 3D structure that QuipNAc would prevent.
showed that WbcP is a UDP-GlcNAc-4,6-dehydratase that
catalyzes the conversion of UDP-d-GlcNAc to UDP-Sugp.
The 4,6-dehydratase reaction mechanism is likely to proceed
analogous to that described in detail for a Streptomyces 4,6-
dehydratase that uses dTDP-glucose as a substrate; the
enzyme uses NAD and active-site tyrosine and threonine to
[29]
To investigate if the presence of QuipNAc or FucpNAc in
the OC structure has any consequences in situ on the bacte-
rial surface, Ye O:3 strains transformed with wbpV (YeO3-
R1/pEPwbpVc) and wbpK (YeO3-R1/pEPwbpK), respec-
tively, were tested for the sensitivity to fR1-37 and entero-
coliticin. When compared to YeO3-R1, both recombinant
strains showed altered but different sensitivities. YeO3-R1/
pEPwbpK was fully susceptible and YeO3-R1/pEPwbpVc
was almost fully resistant to fR1-37 infection, whereas both
strains were highly resistant to enterocoliticin (data not
shown). These sensitivity changes could be a direct effect of
the substitution of Sugp by QuipNAc or FucpNAc in the
OC structure or due to decreased expression of the modified
OC as demonstrated by DOC-PAGE analysis (Figure 5). Al-
ternatively, minute amounts of Sugp containing OC could
be expressed by the recombinant strains that could be suffi-
cient to allow phage infection but not for receptor recogni-
tion by enterocoliticin.
To elucidate this, we tested the ability of mAb 2B5 to
react with LPS from the recombinant strains by Western im-
munoblotting. Indeed, mAb 2B5 recognized the OC-con-
taining LPS bands from both recombinant strains (data not
shown). This suggested that either the first sugar of the OC
structure is not crucial for the antibody epitope or the Sugp
residue was present as a minor fraction in the OC band. The
latter interpretation was more likely to be correct since puri-
fied hydrazine-treated LPS (with Sugp reduced to Quip-
NAc) was not recognized by mAb 2B5 (see Figure S5 in the
Supporting Information). In chemical analyses, QuipNAc
and FucpNAc were detectable as constituents of the OC
structures of strains YeO3-R1/pEPwbpVc and YeO3-R1/
pEPwbpK, respectively. Furthermore, as no peaks due to
Sugp in the diol form were detected in ESI-FTICR MS
either under soft ionization conditions or with nonspecific
fragmentation by CSD it was likely that the majority of the
OC in both strains contained either QuipNAc or FucpNAc
and that only amounts below the detection level of the Sugp
were present if at all. To address this, we again made use of
accomplish the reaction. Sugp has never been identified
as a constituent of Yersinia cell envelope structures before;
however, we recently sequenced a LPS gene cluster of Yersi-
[30]
nia sp. strain ꢂ125 KOH2 isolated from iceberg lettuce.
Interestingly, the gene cluster, similar to that of the OC
gene cluster of Ye O:3, encodes highly identical homologues
[30]
of WbcO and WbcP. Thus, it is plausible that the LPS of
this strain carries Sugp. Work is underway to resolve the
structure of its LPS.
However, Sugp was identified as a component of the cap-
sular polysaccharide from Streptococcus pneumoniae type
[
31]
5,
Vibrio ordalii O:2,
and in three different OPS, that is, from the LPS of
[32]
Flavobacterium columnare ATCC
[33]
[34]
43622
and Pseudoalteromonas rubra ATCC 29570T.
Additionally, in the mutant strain Rhizobium etli CE 166,
Sugp was identified as the sugar linking the OPS to the core
region, which in the wt strain was brought about by d-Quip-
[35]
NAc. It was suggested that due to transposon insertion in
the lpsQ gene (a homologue of wbpV) the strain lacked the
UDP-Sugp reductase activity, which in the wild-type strain
was suggested to convert UDP-Sugp to UDP-d-QuipNAc.
The data presented in this study support this conclusion.
In all these cases, the identification of Sugp has been
rather straightforward, in particular by NMR spectroscopy
and reduction experiments. This was not the case in the
analysis of the Ye O:3 core region, since the characteristic
1
3
signal of the diol group at about 94 ppm in the C NMR
spectrum could not be identified. Then, early MS data let us
think of the presence of three HexN residues instead of two,
which had been published earlier. These analytical difficul-
ties had obviously been the reason for proposing FucpN as
[11]
the sugar linking the OC to the IC of the Ye O:3 LPS.
While the genetics associated with the biosynthesis of
UDP-QuipNAc and UDP-FucpNAc have been described
for many bacteria, neither the genetics nor in-depth bio-
chemical characterization of the biosynthesis of UDP-Sugp
biosynthesis in bacteria has been described until this study.
As the metabolic steps leading to the biosynthesis of UDP-
9752
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9747 – 9754

Identification and Role of a 6-Deoxy-4-Keto-Hexosamine
FULL PAPER
QuipNAc and UDP-FucpNAc required two enzymes, the
UDP-GlcNAc-4,6-dehydratase followed by the 4,6-reduc-
tase, it is apparent that the latter enzymatic step is missing
from those organisms that carry Sugp, as was demonstrated
in this work for Ye O:3.
with UDP-2-acetamido-2-deoxy-d-galactopyranose (UDP-GalpNAc) or
UDP-GlcpNAc in an appropriate buffer. Reaction mixtures were ana-
lyzed by CE (see the Supporting Information for details). WbcP-S212
was purified from the crude extract supernatant by nickel-chelation affin-
ity chromatography followed by desalting with a PD-10 column.
Enzyme assays: The purified WbcP-S212 or the membrane fraction con-
taining wt WbcP was incubated with UDP-GlcpNAc in suitable buffer at
Interestingly, the presence of the Sugp instead of Quip-
NAc in R. etli mutant strain CE166 suggests that the glyco-
syltransferase for UDP-QuipNAc is able to use UDP-Sugp
as a substrate, but less efficiently since the level of OPS de-
tected in CE166 was lower than that of the wild-type strain
3
78C for 3 h. The reaction products were analyzed by CE. When necessa-
ry the enzyme reactions were spiked with UMP and/or UDP before the
samples were analyzed by CE. To confirm the identity of the WbcP reac-
tion product, sequential reactions were used in which the WbcP reaction
product was subjected to the enzymatic activity of WbgX of P. shigel-
loides. To reaction mixtures obtained with the wt WbcP or WbcP-S212 as
described above, the WbgX was added in a suitable reaction mixture.
The WbgX reaction was allowed to proceed at 378C for 1 h after which
the products were analyzed by CE (see the Supporting Information).
[35]
CE3. The situation in our UDP-Sugp-4,6-reductase over-
expressing recombinant strains was consistent with the
above observation concerning the glycosyltransferase for
UDP-QuipNAc in R. etli CE166, that is, QuipNAc could be
introduced to the OC of Ye O:3, but the overall expression
of the OC was decreased relative to the wild-type LPS of
the wild-type Ye O:3 strain.
Bacteriophage, enterocoliticin, and monoclonal antibody assays: Serial
dilutions of lyophilized LPS diluted into aqua were incubated either with
[
12,26]
[25,39]
bacteriophage fR1–37
or bacteriocin enterocoliticin.
The influ-
ences of the LPS to the infectivity of fR1–37 and the activity of entero-
coliticin were measured as described in the Supporting Information. Dot
Intriguingly, we demonstrated that the axial 4-hydroxy
group of this dihydroxy function is crucial for the biological
role of the Ye O:3 OC, that is, in the bacteriophage and en-
terocoliticin receptor structure and in the epitope of a mon-
oclonal antibody. We also postulate that the dihydroxy func-
tion is crucial for the pathogenic potential of Ye O:3; how-
ever, at present we cannot test this experimentally until we
have successfully engineered a strain expressing LPS that
would be fully substituted by QuipNAc- or FucpNAc-con-
taining OC (Figure 5). Such a strain should have the OC
biosynthetic machinery, for example, WbcQ, the OC-flip-
pase and the OC-ligase, adapted to the QuipNAc or Fucp-
NAc-containing OC. More work is underway towards attain-
ing this goal.
blotting with LPS preparations were performed by using a monoclonal
[27,28]
antibody (mAb) 2B5 specific for OC
tion (see the Supporting Information).
with chemiluminescence detec-
Construction of plasmids pEPWbpVc, pEPWbpVRe, and pEPWbpK:
[
40]
The wbpV and wbpK genes were cloned as PCR products to pTM100.
For wbpV and wbpK, chromosomal DNA of P. aeruginosa serotype O6
[
22]
strain (ATCC 33354) and plasmid pET-28a-wbpK,
respectively, were
used as templates. Plasmid pEPWbpVc carries the wbpV gene down-
stream of the tetracycline-resistance gene promoter of pTM100 in correct
orientation, pEWbpVRe, in reverse orientation, and plasmid pEPWbpK,
the wbpK gene, in correct orientation. All three plasmids were mobilized
into YeO3-R1 (see the Supporting Information for details).
Acknowledgements
This work was supported by funding from the Academy of Finland (proj-
ects 104361 and 50441 to M.S.), from the Magnus Ehnrooth Foundation
Experimental Section
(
to E.P.), from the Canadian Institutes of Health Research (CIHR, #
MOP-14687 to J.S.L.), and from the Deutsche Forschungsgemeinschaft
LI-448/1–1 to B.L.). W.L.M. was a recipient of a Doctoral Research
Bacterial strains, culture conditions, and LPS isolation: The Ye O:3
strains Ye75S, 6471/76-c (YeO3-c), YeO3-R1, YeO3-c-trs22-R, and
YeO3-c-trs24-R used for the structural studies have been described earli-
(
[
9,36–38]
Award from CIHR and J.S.L. holds a Canada Research Chair in Cystic
Fibrosis and Microbial Glycobiology. We thank A. Liljegren and L. Kalin
(
tance, H. Moll (RCB) for help with GCMS, H. Kꢀßner (RCB) for re-
cording NMR spectra and Dr. U. Zꢀhringer (RCB) for valuable discus-
sions. Parts of this work were presented at the 14th European Carbohy-
drate Symposium (September 2–7th, 2007, Lꢃbeck, Germany) and at the
Meeting of the International Endotoxin and Innate Immunity Society
er.
Culture conditions and LPS isolation procedures as well as the
chemical and structural analytical methods are described in detail in the
UH), J. Karhu (UT), and B. Kunz (RCB) for excellent technical assis-
Supporting Information.
1
8
18
Experiment with H
2 2
O: The LPS from YeO3-R1 was dissolved in H O
(
Medical Isotopes, INC) in a tightly closed vial and evaporated under ni-
trogen by utilizing two thin needles driven in through the teflon septum
of the vial cap. This was done twice. Then the sample was dissolved in a
1
8
2
mixture of 2-propanol, triethylamine, and H O, as described for mass
(
July 30th–August 2nd, 2008, Edinburgh, UK).
spectrometry in the Supporting Information, and immediately investigat-
ed by ESI-FTICR MS.
Cloning of wbcP and truncated versions into expression vector: Full
length and different N-terminally truncated derivatives of Ye O:3 wbcP
were cloned as PCR products into an expression vector pET-28a (Nova-
gen). Chromosomal DNA from Ye O:3 strain 6471/76-c was used as a
template in PCR. A detailed protocol is provided in the Supporting In-
formation. The obtained plasmids were named pETWbcP (expressing
[
1] O. Holst, S. Mꢃller-Loennies in Comprehensive Glycoscience, Vol. 1
Eds.: J. P. Kamerling, G.-J. Boons, Y. C. Lee, A. Suzuki, N. Tanigu-
chi, A. G. J. Voragen), Elsevier, Oxford, 2007, pp. 123–175.
2] U. Zꢀhringer, B. Lindner, E. T. Rietschel, Adv. Carbohydr. Chem.
Biochem. 1994, 50, 211–276.
3] U. Zꢀhringer, B. Lindner, E. T. Rietschel in Endotoxin in Health and
Disease (Eds.: H. Brade, S. M. Opal, S. N. Vogel, D. C. Morrison),
Marcel Dekker, New York, 1999, pp. 93–113.
(
[
[
full length WbcP), pETWbcP-S212 (expressing
D AHCTUNGTERNNUNG( 1–211)-truncated
WbcP), and pETWbcP-E243 (expressing D(1–242)-truncated WbcP).
AHCTUNGTRENNUNG
Expression and analysis of WbcP, WbcP-S212, and WbcP-E243: For effi-
cient protein expression, plasmids pETWbcP, pETWbcP-S212, and
pETWbcP-E243 were transformed into E. coli BL21 ACHUTNGRENNUG( DE3) (Novagen)
cells. In crude extract assays sonicated cell membranes containing WbcP
and supernatants containing WbcP-S212 or WbcP-E243 were incubated
[4] O. Holst, FEMS Microbiol. Lett. 2007, 271, 3–11.
[5] O. Holst, Trends Glycosci. Glycotechnol. 2002, 14, 87–103.
[6] O. Holst in Endotoxin in Health and Disease (Eds.: H. Brade, S. M.
Opal, S. N. Vogel, D. C. Morrison), Marcel Dekker, NewYork, 1999,
pp. 115–154.
Chem. Eur. J. 2009, 15, 9747 – 9754
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9753

M. Skurnik et al.
[
7] Y. A. Knirel, N. K. Kochetkov, Biochemistry (Moscow) 1994, 59,
325–1383.
8] P.-E. Jansson in Endotoxin in Health and Disease (Eds.: H. Brade,
[25] E. Strauch, H. Kaspar, C. Schaudinn, C. Damasko, A. Konietzny, P.
Dersch, M. Skurnik, B. Appel, Adv. Exp. Med. Biol. 2003, 529, 249–
251.
1
[
S. M. Opal, S. N. Vogel, D. C. Morrison), Marcel Dekker, New York,
[26] S. Kiljunen, K. Hakala, E. Pinta, S. Huttunen, P. Pluta, A. Gador, H.
Lçnnberg, M. Skurnik, Microbiology 2005, 151, 4093–4102.
[27] K. Pekkola-Heino, M. K. Viljanen, T. H. Stꢆhlberg, K. Granfors, A.
Toivanen, APMIS 1987, 95, 27–34.
1
999, pp. 155–178.
[
9] M. Skurnik, R. Venho, J.-A. Bengoechea, I. Moriyꢄn, Mol. Micro-
biol. 1999, 31, 1443–1462.
[
[
10] C. R. Raetz, C. Whitfield, Annu. Rev. Biochem. 2002, 71, 635–700.
11] J. Radziejewska-Lebrecht, M. Skurnik, A. S. Shashkov, L. Brade, A.
Rozalski, B. Bartodziejska, H. Mayer, Acta Biochim. Pol. 1998, 45,
[28] M. Biedzka-Sarek, R. Venho, M. Skurnik, Infect. Immun. 2005, 73,
2232–2244.
[29] S. T. Allard, W. W. Cleland, H. M. Holden, J. Biol. Chem. 2004, 279,
2211–2220.
[30] T. Niskanen, R. Laukkanen, A. Murros, J. Bjçrkroth, M. Skurnik, H.
Korkeala, M. Fredriksson-Ahomaa, Int. J. Food Microbiol. 2009,
129, 150–156.
[31] P. E. Jansson, B. Lindberg, U. Lindquist, Carbohydr. Res. 1985, 140,
101–110.
1
011–1019.
12] M. Skurnik, R. Venho, P. Toivanen, A. Al-Hendy, Mol. Microbiol.
995, 17, 575–594.
[
[
1
13] J. A. Bengoechea, E. Pinta, T. Salminen, C. Oertelt, O. Holst, J.
Radziejewska-Lebrecht, Z. Piotrowska-Seget, R. Venho, M. Skur-
nik, J. Bacteriol. 2002, 184, 4277–4287.
[
14] M. S. Anderson, S. S. Eveland, N. P. Price, FEMS Microbiol. Lett.
[32] I. Sadovskaya, J. R. Brisson, N. H. Khieu, L. M. Mutharia, E.
Altman, Eur. J. Biochem. 1998, 253, 319–327.
2
000, 191, 169–175.
[
[
15] N. P. Price, F. A. Momany, Glycobiology 2005, 15, 29R–42R.
16] R. Madsen in Oxidation and Reduction (Eds.: B. Fraser-Reid, K.
Tatsuta, J. Thiem, Springer, Heidelberg, Germany, 2001, pp. 195–
[33] L. L. MacLean, M. B. Perry, E. M. Crump, W. W. Kay, Eur. J. Bio-
chem. 2003, 270, 3440–3446.
[34] M. Kilcoyne, A. S. Shashkov, Y. A. Knirel, R. P. Gorshkova, E. L.
Nazarenko, E. P. Ivanova, N. M. Gorshkova, S. N. Senchenkova,
A. V. Savage, Carbohydr. Res. 2005, 340, 2369–2375.
[35] L. S. Forsberg, K. D. Noel, J. Box, R. W. Carlson, J. Biol. Chem.
2003, 278, 51347–51359.
2
29.
[
17] S. Borowski, D. Michalik, H. Reinke, C. Vogel, A. Hanuszkiewicz,
K. A. Duda, O. Holst, Carbohydr. Res. 2008, 343, 1004–1011.
[
[
18] C. Creuzenet, J. S. Lam, Mol. Microbiol. 2001, 41, 1295–1310.
19] C. Creuzenet, R. V. Urbanic, J. S. Lam, J. Biol. Chem. 2002, 277,
[36] K. Wartenberg, J. Lysy, W. Knapp, Zentralbl. Bakteriol. Abt. 1 Orig.
A 1975, 230, 361–366.
2
6769–26778.
[
20] N. Ishiyama, C. Creuzenet, W. L. Miller, M. Demendi, E. M. Ander-
[37] M. Skurnik, J. Appl. Bact. 1984, 56, 355–363.
[38] A. Al-Hendy, P. Toivanen, M. Skurnik, Infect. Immun. 1992, 60,
870–875.
[39] E. Strauch, H. Kaspar, C. Schaudinn, P. Dersch, K. Madela, C. Ge-
winner, S. Hertwig, J. Wecke, B. Appel, Appl. Environ. Microbiol.
2001, 67, 5634–5642.
son, G. Harauz, J. S. Lam, A. M. Berghuis, J. Biol. Chem. 2006, 281,
2
4489–24495.
21] J. G. Shepherd, L. Wang, P. R. Reeves, Infect. Immun. 2000, 68,
056–6061.
[
[
6
22] W. L. Miller in Investigation of enzymes involved in sugar-nucleotide
biosynthesis pathways, PhD Thesis University of Guelph (Guelph),
[40] T. Michiels, G. R. Cornelis, J. Bacteriol. 1991, 173, 1677–1685.
2
006.
23] M. Bꢅlanger, L. L. Burrows, J. S. Lam, Microbiology 1999, 145,
505–3521.
24] E. Pinta, K. A. Duda, A. Hanuszkiewicz, J. A. Bengoechea, H. Hyy-
tiꢀinen, B. Lindner, J. Radziejewska-Lebrecht, O. Holst, M. Skurnik,
unpublished results.
[
[
3
Received: May 12, 2009
Published online: August 20, 2009
9754
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9747 – 9754