989
Biochem. J. (2003) 371, 989–995 (Printed in Great Britain)
Biosynthesis of 2-acetamido-2,6-dideoxy-L-hexoses in bacteria follows a
pattern distinct from those of the pathways of 6-deoxy-L-hexoses
Bernd KNEIDINGER*, Suzon LAROCQUE†, Jean-Robert BRISSON†, Nicolas CADOTTE† and Joseph S. LAM*1
*Canadian Bacterial Diseases Network, University of Guelph, Department of Microbiology, Guelph, Ontario N1G 2W1, Canada, and †Institute for Biological Sciences,
National Research Council, 100 Sussex Drive, Ottawa, Ontario K1A 0R6, Canada
6-Deoxy-L-hexoses have been shown to be synthesized from
dTDP-D-glucose or GDP-D-mannose so that the gluco/galactoconfiguration is converted into the manno/talo-configuration, and
manno/talo is switched to gluco/galacto. Our laboratory has
been investigating the biosynthesis of 2-acetamido-2,6-dideoxyL-hexoses in both Gram-positive and Gram-negative bacteria,
and in a recent paper we described the biosynthesis of the
talo (pneumosamine) and galacto (fucosamine) derivatives from
UDP-D-N-acetylglucosamine a 2-acetamido sugar [Kneidinger,
O’Riordan, Li, Brisson, Lee and Lam (2003) J. Biol. Chem. 278,
3615–3627]. In the present study, we undertake the task to test
the hypothesis that UDP-D-N-acetylglucosamine is the common
precursor for the production of 2-acetamido-2,6-dideoxy-Lhexoses in the gluco-, galacto-, manno- and talo-configurations.
We present data to reveal the steps for the biosynthesis
of the gluco (quinovosamine)- and manno (rhamnosamine)configured compounds. The corresponding enzymes WbvB,
WbvR and WbvD from Vibrio cholerae serotype O37 have been
overexpressed and purified to near homogeneity. The enzymic
reactions have been analysed by capillary electrophoresis and
NMR spectroscopy. Our data have revealed a general feature of
reaction cascades due to the three enzymes. First, UDP-D-Nacetylglucosamine is catalysed by the multi-functional enzyme
WbvB, whereby dehydration occurs at C-4, C-6 and epimerization
at C-5, C-3 to produce UDP-2-acetamido-2,6-dideoxy-L-lyxo-4hexulose. Secondly, this intermediate is converted by the C-4
reductase, WbvR, in a stereospecific reaction to yield UDP-2acetamido-L-rhamnose. Thirdly, UDP-2-acetamido-L-rhamnose
is epimerized at C-2 to UDP-2-acetamido-L-quinovose by WbvD.
Interestingly, WbvD is also an orthologue of WbjD, but not
vice versa. Incubation of purified WbvD with UDP-2-acetamido2,6-dideoxy-L-talose and analysing the reaction products by
capillary electrophoresis revealed the same product peak as when
WbjD was used. This sugar nucleotide is a specific substrate
for WbjD and is a C-4 epimer of UDP-2-acetamido-L-rhamnose.
INTRODUCTION
rhamnose [1] by RmlD or to dTDP-L-pneumose by the action
of Tll of Actinobacillus actinomycetemcomitans [2]. On the other
hand, GDP-L-fucose is synthesized by using GDP-D-mannose
as the precursor nucleotide-activated sugar, and the same seems
likely to be true for GDP-L-quinovose. In this scenario, the
first reaction step is C-4, C-6 dehydration of GDP-D-mannose
to GDP-6-dideoxy-D-lyxo-4-hexulose, catalysed by Gmd [3].
Subsequently, GDP-L-fucose is produced by the bifunctional
C-3, C-5 epimerase/C-4 reductase FX protein in humans [4].
Thus far, the biosynthesis of the nucleotide-activated form of Lquinovose has not been elucidated. This sugar has been described
as a constituent of the LPS of Legionella feeleii [5]; however,
the nucleotide sequence of the genome of this organism is not
available for the identification of the polysaccharide biosynthetic
gene locus. Our present knowledge indicates that 6-deoxy-Lhexoses are generally synthesized from dTDP-D-glucose or GDPD-mannose such that the gluco/galacto-configuration is converted
into the manno/talo-configuration, and manno/talo is switched to
gluco/galacto.
Recently we have been successful in characterizing the
pathways for the biosynthesis of UDP-2-acetamido-2,6-dideoxyL-talose (UDP-L-PneNAc) and UDP-2-acetamido-2,6-dideoxy-Lgalactose (UDP-L-FucNAc) in Pseudomonas aeruginosa and
6-Deoxy-L-hexoses and 2-acetamido-2,6-dideoxy-L-hexoses are
important constituents of the surface polysaccharide structures
of Gram-positive and Gram-negative human pathogenic
bacteria. Examples of these polysaccharide structures include
lipopolysaccharides (LPSs), capsular polysaccharides (CPs),
exopolysaccharides and other glycoconjugates. These polysaccharides are all involved in host–pathogen interactions, and a
deficiency in the production of these surface molecules invariably
results in a significant decrease in bacterial virulence.
The family of 6-deoxy-L-hexose sugars includes Lfucose (6-deoxy-L-galactose), L-quinovose (6-deoxy-L-glucose),
L-pneumose (6-deoxy-L-talose) and L-rhamnose (6-deoxy-Lmannose). The biosynthetic pathways for these sugars, except
for L-quinovose, have been characterized in detail. dTDP-Dglucose serves as a precursor of dTDP-L-rhamnose and of dTDPL-pneumose. The first step dedicated to these pathways is C-4,
C-6 dehydration of dTDP-D-glucose to dTDP-6-dideoxy-D-xylo4-hexulose, catalysed by RmlB [1]. RmlC catalyses the D- to
L-switch in a C-3, C-5 epimerization reaction to yield dTDP6-dideoxy-L-lyxo-4-hexulose [1]. This hexulose intermediate is
then reduced in a stereospecific manner either to dTDP-L-
Key words: deoxy-hexose, lipopolysaccharide, UDP-GlcNAc,
UDP-L-RhaNAc, Vibrio cholerae.
Abbreviations used: CE, capillary electrophoresis; CP, capsular polysaccharide; HMQC, heteronuclear multiple-quantum correlation; HSQC,
heteronuclear single-quantum correlation; IPTG, isopropyl β-D-thiogalactoside; LB, Luria Bertani; LPS, lipopolysaccharide; UDP-L -FucNAc, UDP-2acetamido-2,6-dideoxy-L-galactose or UDP-N -acetyl-L-fucosamine; UDP-D-GlcNAc, UDP-2-acetamido-2-deoxy-D-glucose or UDP-N -acetyl-D-glucosamine;
UDP-L -PneNAc, UDP-2-acetamido-2,6-dideoxy-L -talose or UDP-N -acetyl-L -pneumosamine; UDP-L -QuiNAc, UDP-2-acetamido-2,6-dideoxy-L -glucose
or UDP-N -acetyl-L-quinovosamine; UDP-L-RhaNAc, UDP-2-acetamido-2,6-dideoxy-L-mannose or UDP-N -acetyl-L-rhamnosamine.
1
To whom correspondence should be addressed (e-mail [email protected]).
c 2003 Biochemical Society
990
B. Kneidinger and others
Staphylococcus aureus [6]. In that study, we found that, in
contrast with the substrates normally used to synthesize the 6deoxy-L-hexoses described above, UDP-2-acetamido-2-deoxyD-glucose (UDP-D-GlcNAc) serves as the precursor for both
the talo- and galacto-configurations. WbjB of P. aeruginosa
and its functional homologue Cap5E of S. aureus catalyse a
cascade reaction consisting of C-4, C-6 dehydratase and C-5,
C-3 epimerase activities, converting UDP-D-GlcNAc into UDP2-acetamido-2,6-dideoxy-L-lyxo-4-hexulose. This intermediate
is reduced to UDP-L-PneNAc in a stereospecific manner,
catalysed by WbjC/Cap5F. Finally, WbjD/Cap5G catalyses the
C-2 epimerization of UDP-L-PneNAc to yield a mixture of UDPL-PneNAc and UDP-L-FucNAc.
To date, the biosynthetic pathways of UDP-2-acetamido-2,
6-dideoxy-L-glucose (UDP-L-QuiNAc) and UDP-2-acetamido2,6-dideoxy-L-mannose (UDP-L-RhaNAc) have not been solved.
L-QuiNAc has been described as a constituent of the LPS
of Yersinia enterocolitica serotypes O11,23 and O11,24 [1],
Escherichia coli O98 [7], Proteus penneri 26 [8] and Proteus
vulgaris OX2 [9], as well as the capsule of Bacteroides fragilis
N.C.T.C. 9343 [10]. Another sugar, L-RhaNAc, is part of the
LPS of E. coli O3:K2ab(L):H2 [11] and B. fragilis A.T.C.C.
23745 [10]. The latter has been described to contain both LQuiNAc and L-RhaNAc in the capsule [10]. Moreover, the
capsule of Vibrio vulnificus BO62316 contains both QuiNAc
and RhaNAc, but the absolute configuration of the sugars has
not been determined, although the authors claimed that they are
most likely to be in the L-configuration [12]. Coyne and his coworkers [13] suggested that the products of three clustered genes
in B. fragilis N.C.T.C 9343, namely wcgS, wcgT and wcgU,
are involved the biosynthesis of UDP-L-QuiNAc from UDPN-acetylmannosamine (UDP-D-ManNAc); however, they do not
have biochemical evidence to substantiate the proposed pathway
involving enzymes encoded by these three genes [13].
In the present study, we report our investigation of the pathways
for the biosynthesis of UDP-L-QuiNAc and UDP-L-RhaNAc. Our
data show that, in contrast with the previously established 6deoxy-L-hexose biosynthesis pathways, the glucose derivative,
namely UDP-D-GlcNAc, serves as a precursor for 2-acetamido2,6-dideoxy-L-hexoses in the gluco-, galacto-, manno- and taloconfigurations.
EXPERIMENTAL
Materials
UDP-D-GlcNAc, NADH, NADP+ , NADPH and the antibiotics
used in this study were obtained from Sigma-Aldrich (Oakville,
Canada). Isopropyl β-D-thiogalactoside (IPTG) was purchased
from Invitrogen (Burlington, Canada) and pET28a was from
Novagen (Madison, WI, U.S.A.). Genomic DNA of V. cholerae
serotype O37 strain V52 was kindly provided by Jutta Nesper and
Joachim Reidl (University of Würzburg, Germany).
Bacterial strains and growth conditions
E. coli strains were grown in Luria Bertani (LB) medium
(Invitrogen Inc., Burlington, ON, Canada) or on LB agar
[containing 1.5 % (w/v) Bacto Agar] at 37 ◦C. E. coli strain
Top 10 [Invitrogen Inc.; F − mcrA ∆(mrr-hsdRMS-mcrBC)
φ80lacZ∆M15 ∆lacX74 deoR recA1 araD139 ∆(araAleu)7697 galU galK rpsL (StrR ) endA1 nupG] was used for
plasmid propagation, and E. coli strain BL21(DE3) [Novagen;
B F − dcm ompT hsdS(rB − mB − ) gal λ(DE3)] was used for
c 2003 Biochemical Society
protein expression. Overexpression was induced by the addition of
IPTG at a final concentration of 1 mM. Media were supplemented
with ampicillin (100 –250 µg/ml) or kanamycin (50 µg/ml), when
necessary.
Analytical methods
BLAST [14] and ClustalW [15] were used for the analysis of
nucleotide and protein sequences. SDS/PAGE was performed
according to the method of Laemmli [16]. Gels were stained
with Coomassie Brilliant Blue R-250 (Sigma-Aldrich). Protein
concentrations were determined as described by Bradford [17].
Capillary electrophoresis (CE) analysis was performed using
a P/ACE MDQ Glycoprotein system with UV detection
(Beckman Coulter, Fullerton, CA, U.S.A.). Samples were
separated at 22 kV in 25 mM borate buffer (pH 9.5) at 25 ◦C.
Typically, reactions contained 0.5 mM UDP-D-GlcNAc, 0.1
mM NADP+ and equimolar amounts of hydride donor (NADH
or NADPH). Standard reaction buffer was 20 mM Tris/HCl,
pH 8.0, supplemented with 10 mM MgCl2 . Reactions to test the
stereospecificity of WbvD were performed with 0.5 mM UDPL-PneNAc. Enzymes (approx. 1 µg each) were added, and the
reaction mixtures were incubated at 37 ◦C for 180 min before
being analysed by CE. Reactions involving WbvD were carried
out at 25 ◦C.
For NMR analysis, all spectra were acquired using a Varian
Inova 500 MHz spectrometer equipped with a Z-gradient 3 mm
triple resonance (1 H, 13 C, 31 P) probe. The sugar nucleotide
sample was prepared in 150 µl of 90 % water/10 % 2 H2 O.
The experiments were performed at 25 ◦C with suppression of the
water resonance. The methyl resonance of acetone was used as
an internal reference at δ H 2.225 p.p.m. and δ C 31.07 p.p.m.
Standard sequences from Varian, COSY, heteronuclear
single-quantum correlation (HSQC) and 31 P heteronuclear
multiple-quantum correlation (HMQC) and selective 1D-TOCSY
and 1D-NOESY experiments were used to assign the resonances
[6,18,19].
Recombinant DNA procedures, PCR and DNA sequencing
All standard DNA recombinant procedures were performed
according to the methods described by Sambrook et al. [20] or
as recommended by the corresponding manufacturer. PCR was
carried out using a GeneAmp PCR System 2400 (Perkin Elmer,
Mississauga, Canada). DNA sequencing was performed by the
Mobix Laboratory, Hamilton, Canada. The three genes of interest
from V. cholerae were ORF7 (accession number AAM22595),
ORF8 (AAM22596) and ORF9 (AAM22597), and these have
been designated as wbvB, wbvR and wbvD respectively.
Plasmid construction
The following primers were designed for the amplification
of the putative UDP-L-QuiNAc biosynthesis genes: wbvB1F,
5 -ggactcgagcatatgttcaaggataaaactttaatg-3 ; wbv2R, 5 -gtgaattcctagcatacctgtagtg-3 ; wbvR1F, 5 -ggactcgagcatatgaaaatactaatagttggcac-3 ; wbvR2R, 5 -atgaattccttgtacctacaacagtc-3 wbv D1F,
5 -ggactcgagccatggataaattaaaagtattgac-3 ; wbvD2R, 5 -cgtggatccgtactttcaggcaaatattcg-3 . The underlined sequences correspond to the restriction sites used for cloning. PCR amplifications
were performed with Pwo polymerase (Roche, Laval, Canada).
To amplify the wbvB, wbvR and wbvD genes, the primer pairs
wbvB1F/wbvB2R, wbvR1F/wbvR2R and wbvD1F/wbvD2R
2-Acetamido-2,6-dideoxy-L-hexose biosynthesis
991
respectively were used. The wbvB and wbvR PCR products were
digested with NdeI/EcoRI and ligated into pET28a, cut with the
same enzymes. The wbvD product was cut with NcoI/BamHI
and ligated into appropriately digested pET23der [21]. The
corresponding plasmids pFuc51 (wbvB), pFuc52 (wbvR) and
pFuc53 (wbvD) were confirmed by DNA sequencing and used
for the overexpression of N-terminally histidine-tagged fusion
proteins.
Enzyme expression and purification
Expression of the enzymes was carried out at 37 ◦C in LB broth,
supplemented with 50 µg/ml kanamycin (WbvB and WbvR) or
250 µg/ml ampicillin (WbvD). Cells were grown to an absorbance
at 600 nm of 0.7 in a 330 ml culture volume; WbvB and WbvD
expression was carried out for 4 h, and WbvR was expressed for
5 h. The cells were disrupted on ice by sonication with a 0.5 inch
(12.7 mm) probe, amplitude setting 40 %, using a pulsed mode
set at 5 s on and 10 s off for 2 min (Ultrasonic Processor XL 2020;
Mandel Scientific Co. Ltd, Guelph, ON, Canada), and cell debris
and membrane fractions were removed by ultracentrifugation
at 300 000 g. Purification of the fusion proteins was achieved
using HiTrap Chelating columns (Amersham Pharmacia Biotech,
Uppsala, Sweden) as recommended by the manufacturer. Pure
WbvB and WbvD were eluted at 300 mM imidazole, whereas
pure WbvR could be obtained at 400 mM imidazole. Dithiothreitol
was added to a final concentration of 1 mM and the proteins were
stored at –20 ◦C after the addition of 40 % (v/v) glycerol. The
purity of the enzymes was checked by SDS/PAGE analysis on
a 10 % gel. The P. aeruginosa UDP-2-actamido-2,6-dideoxyL-talose 2-epimerase WbjD was overexpressed and purified as
described previously [6].
Synthesis and preparation of UDP-L-RhaNAc
UDP-D-GlcNAc (5.5 µmol; 3.6 mg) was converted quantitatively
into putative UDP-L-RhaNAc by 200 µg each of WbvB and WbvR
overnight at 37 ◦C in 20 mM Tris buffer (pH 8.0) supplemented
with 10 mM MgCl2 . Equimolar amounts of NADPH were used
as the hydride donor, and the initial concentration of the cofactor
NADP+ was 10 µM. After removal of protein by ultrafiltration
(Centriplus YM3; Millipore, Bedford, MA, U.S.A.), the
nucleotide-activated sugar was purified by anion exchange on an
Econo-Pac High Q column (Bio-Rad, Mississauga, Canada) using
a linear triethylammonium bicarbonate gradient (0 –500 mM,
pH 8.0). The fractions corresponding to the putative UDP-LRhaNAc were pooled, adjusted to an acidic pH 4.5 using Dowex
50 (Bio-Rad, Mississauga, Canada) to remove CO2 , and finally
concentrated by lyophilization. 2 H2 O was added and the sample
was analysed by NMR as described above.
RESULTS AND DISCUSSION
Sequence analysis
Among bacteria containing either L-QuiNAc or L-RhaNAc in their
LPS or CP, only the gene locus of B. fragilis N.C.T.C. 9343
capsular polysaccharide B has been sequenced so far [13]. This
bacterium has no L-FucNAc or L-PneNAc in the corresponding
CP, but interestingly has two genes in the biosynthesis cluster
that are highly similar to wbjB and wbjD of P. aeruginosa
serotype O11, which are involved in the biosynthesis of
L-FucNAc. The functions of the P. aeruginosa enzymes have been
elucidated [6]. WbjB is a UDP-D-GlcNAc 4,6-dehydratase/5-
Figure 1
Clusteral organization of UDP-L-QuiNAc biosynthesis genes
Accession numbers of the protein sequences are given in parentheses. Genes are from Vibrio
cholerae serotype O37 strain V52, Bacteroides fragilis N.C.T.C. 9343, Leptospira interrogans
serovar Hardjo subtype Hardjoprajitno, Leptospira borgpetersenii serovar Hardjobovis strain
L171, Rickettsia conorii strain Malish 7 and Rickettsia prowazekii strain Madrid E. Note that all
clusters contain two of the three genes/open reading frames (ORFs) that are highly similar to
wbjB and wbjD [6]; therefore they are designated as such.
epimerase/3-epimerase, and WbjD is a UDP-2-actamido-2,6dideoxy-L-talose 2-epimerase. However, the third L-FucNAc
biosynthesis gene coding for WbjC, a UDP-2-acetamido2,6-dideoxy-L-lyxo-4-hexulose C-4 reductase, is absent from
B. fragilis N.C.T.C. 9343. This enzyme catalyses the reaction
step leading to UDP-L-PneNAc. It seems logical that the two
genes that are analogous to wbjB and wbjD are also involved
in UDP-L-QuiNAc biosynthesis. The gene downstream of the
C-2 epimerase homologue codes for an enzyme with moderate
similarity to the dTDP-dehydrorhamnose reductase RmlD, which
catalyses the final reaction step in the biosynthesis pathway
leading to nucleotide-activated L-rhamnose [1]. This reductase
might therefore catalyse the reduction of UDP-2-acetamido-2,6dideoxy-L-lyxo-4-hexulose to UDP-L-RhaNAc, a C-4 epimer of
UDP-L-PneNAc. A homologue of this reductase is present at the
O-antigen locus of Vibrio cholerae serotype O37, the surface
polysaccharide biosynthesis clusters of Leptospira interrogans
serovar Hardjo subtype Hardjoprajitno, L. borgpetersenii serovar
Hardjobovis, and the genomes of Rickettsia conorii and
R. prowazekii (Figure 1). To date, no leptospiral polysaccharide
structures have been resolved, and nothing is known about
surface polysaccharides in the obligately intracellular Rickettsiae.
Unfortunately, there is also only limited information available
on the structures of the O antigen of different serotypes of
Vibrio cholerae. It is intriguing that, in contrast with the perfect
conservation of the order of the L-FucNAc biosynthesis genes in
clusters that had been sequenced so far, the order of the three
putative L-QuiNAc/L-RhaNAc biosynthetic genes varies in the
above-mentioned bacteria (Figure 1). It should be noted that,
in all of the available sequences of the clusters containing the
wbjB/cap5E/wbvB and wbjD/cap5G/wbvD genes, these genes are
always linked to either one of the two (putative) reductases.
To test the hypothesis that these three enzymes are actually
involved in the biosynthesis of UDP-L-QuiNAc and UDP-LRhaNAc, the corresponding proteins WbvB, WbvR and WbvD
from V. cholerae O37 have been overexpressed and purified to near homogeneity. In compliance with the bacterial
polysaccharide biosynthesis nomenclature scheme [22], the genes
and their corresponding proteins have been named wbv and Wbv
respectively. Wb designates proteins involved in the biosynthesis
c 2003 Biochemical Society
992
Figure 2
proteins
B. Kneidinger and others
SDS/PAGE analysis of purified UDP-L-QuiNAc biosynthesis
NHis refers to the hexa-histidine tag at the N-terminus. The Low-Range SDS/PAGE Molecular
Weight Standard kit (Bio-Rad) was used (Std).
of O antigen, and v denotes V. cholerae. B and D have been chosen
for the protein/gene designations because of the similarities to
WbjB and WbjD; finally, R stands for a protein with reductase
activity, which most probably catalyses a reaction distinct from
that of WbjC.
Figure 3
Protein purification
High-level expression of WbvB, WbvR and WbvD was easily
achieved by IPTG induction at 37 ◦C. After sonication and
ultracentrifugation, the majority of the WbvR and WbvD proteins
could be recovered from the supernatant, enabling high-yield
affinity purification using HiTrap Chelating columns. In contrast,
a much higher proportion of the overexpressed WbvB was present
as inclusion bodies than as soluble protein. Nevertheless, a yield
of approx. 300 µg of soluble WbvB was obtained, sufficient to
be used for enzyme assays. After addition of 40 % (v/v) glycerol,
the enzymes can be stored stably at –20 ◦C for several weeks.
The apparent molecular masses of the proteins, as determined
by SDS/PAGE, were in good agreement with the calculated
molecular masses, i.e. WbvB, 41.2 kDa; WbvR, 35.0 kDa;
WbvD, 44.0 kDa (Figure 2). The purified proteins were used for
functional characterization of the pathways for the biosynthesis
of nucleotide-activated L-QuiNAc and L-RhaNAc.
WvbR reduces UDP-2-acetamido-2,6-dideoxy-β-L-lyxo -4-hexulose
to UDP-β-L-RhaNAc
WbvB of V. cholerae, which is a homologue of WbjB/Cap5E,
catalysed the same reaction cascade as the previously reported
P. aeruginosa/S. aureus proteins [6]. CE analysis showed
that UDP-D-GlcNAc is converted quantitatively into a
mixture of at least two keto intermediates, namely
UDP-2-acetamido-2,6-dideoxy-α-D-xylo-4-hexulose and UDP2-acetamido-2,6-dideoxy-β-L-arabino-4-hexulose (Figure 3B).
The third intermediate and expected final WbvB reaction product
UDP-2-acetamido-2,6-dideoxy-β-L-lyxo-4-hexulose is thought to
c 2003 Biochemical Society
CE analysis of UDP-L-QuiNAc biosynthesis
(A) Standards of UDP-D-GlcNAc, NADP+ and NADPH; (B) UDP-D-GlcNAc, converted with
WbvB and NADP+ (intermediates 1 and 2 represent the 2-amino-4-keto-2,6-dideoxy-hexose
intermediates); (C) UDP-D-GlcNAc, converted with WbvB, WbvR, NADP+ and NADPH (L-RhaNAc
refers to UDP-L-RhaNAc); (D) same as (C), but spiked with UDP-D-GlcNAc; (E) UDP-D-GlcNAc,
converted with WbvB, WbvR, WbvD, NADP+ and NADPH (L-QuiNAc refers to UDP-L-QuiNAc).
a.u., arbitrary units. In contrast with UDP-D-GlcNAc and UDP-L-RhaNAc, the two sugars in trace
(E) could not be separated to baseline.
co-migrate either with the other two intermediates or with
NADP+ . Another possibility is that this sugar simply cannot be
detected due to instability or the establishment of an equilibrium
that greatly favours the other two compounds. In a previous study,
after several attempts, the purification of this third intermediate
could not be achieved; even NMR analysis of a concentrated
reaction mixture did not reveal the presence of UDP-2-acetamido2,6-dideoxy-β-L-lyxo-4-hexulose [1]. The corresponding sugar in
the dTDP-L-rhamnose biosynthesis pathway, UDP-6-dideoxy-βL-lyxo-4-hexulose, is produced by the 3,5-epimerase RmlC. The
ratio of UDP-6-dideoxy-α-D-xylo-4-hexulose to UDP-6-dideoxyβ-L-lyxo-4-hexulose has been reported to be 97:3 [23]. To date,
no NMR data have been generated for this lyxo-sugar, the reason
being that it is highly unstable, hampering attempts to obtain a
purified form of this sugar.
In CE analysis, when UDP-D-GlcNAc was incubated with
both WbvB and the putative reductase WbvR in the presence
of equimolar amounts of NADPH as a hydride donor, the
peaks corresponding to the keto intermediates disappeared, and
a compound was produced that migrated faster than UDP-DGlcNAc (Figure 3C). NADH could also be used as the reducing
agent (results not shown). To demonstrate that this sugar product
is distinct from UDP-D-GlcNAc, a WbvB/WbvR reaction mixture
was spiked with the precursor and re-analysed by CE. Two
peaks that were clearly resolved from each other were observed
2-Acetamido-2,6-dideoxy-L-hexose biosynthesis
Table 2
993
J H,H coupling constants for RhaNAc
The average error is +
− 0.3 Hz.
J (Hz)
Figure 4
NMR spectra for RhaNAc
(a) Portion of the resolution enhanced proton spectrum. One of the triethylammonium buffer
resonances is at 3.2 p.p.m. (b) 1D-TOCSY with a mixing time of 80 ms for selective excitation
of the β-RhaNAc NH resonance at 7.71 p.p.m. (c) 1D-TOCSY with a mixing time of 80 ms for
selective excitation of the α-RhaNAc NH resonance at 7.89 p.p.m. (d) 13 C HSQC spectrum. The
resonances are labelled with the atom number and ‘a’ for α-RhaNAc, ‘b’ for β-RhaNAc, and ‘c’
for ribose.
(Figure 3D). This reduction product was purified by anionexchange chromatography and analysed by NMR spectroscopy.
In contrast with UDP-L-PneNAc, which proved to be stable,
the purified UDP-β-L-RhaNAc readily decomposed to the
monosaccharide and UDP.
The 1 H spectrum of the UDP-RhaNAc sample is presented
in Figure 4(a). The sample had degraded to UMP, UDP,
α-RhaNAc and β-RhaNAc. Resonances from the triethylammonium buffer were also observed at 1.28 and 3.20 p.p.m.
In a 31 P-HMQC experiment, no 31 P correlations could be
found to the proton anomeric resonances of RhaNAc, indicating
hydrolysis of the sugar. Proton assignments were made using
the COSY experiment. Assignments for 13 C were made from
an HSQC experiment (Figure 4d and Table 1). From the 1D
selective TOCSY experiment on the NH sugar resonances and
Table 1
NMR chemical shifts for RhaNAc
Chemical shifts were measured at 500 MHz (1 H) in 2 H2 O at 25 ◦C ( +
− 0.2 p.p.m. error for
δ C and +
− 0.02 p.p.m. for δ H ). The internal acetone CH3 resonance was set at δ H 2.225 p.p.m.
and δ C 31.07 p.p.m.. NAc CH3 (δ H , δ C ) cross-peaks were at (2.10 p.p.m., 23.0 p.p.m.) and
(2.06 p.p.m., 22.7 p.p.m.).
Chemical shift (p.p.m.)
H-1
C-1
H-2
C-2
H-3
C-3
H-4
C-4
H-5
C-5
H-6
C-6*
N-H
α-RhaNAc
5.03
93.5
4.26
54.3
3.97
69.1
3.42
73.1
3.87
68.8
1.28
7.89
β-RhaNAc
4.98
93.4
4.41
55.0
3.76
72.4
3.30
72.8
3.43
73.1
1.28
7.71
Unit
* Not determined due to interference from the T1 ridge of the triethylammonium buffer
resonance at 1.28 p.p.m. in the HSQC spectrum.
Unit
J 1,2
J 2,3
J 3,4
J 4,5
J 5,6
J 2,NH
α-RhaNAc
β-RhaNAc
1.5
1.7
4.7
4.5
9.8
9.9
9.6
9.8
6.5
6.2
9.3
9.8
the proton spectrum, accurate proton chemical shifts and proton
coupling constants (Table 2) could be obtained for the H-1 to
H-5 resonances. Proton coupling constants were typical of a
manno-pyranose configuration, as compared with the data in
a study by Perry et al. [24]. The C-6 CH3 RhaNAc proton
chemical shifts were determined from a COSY experiment, since
their resonances overlapped with the triethylammonium buffer
resonance at 1.28 p.p.m. From the 1D-NOESY experiment of the
anomeric resonance at 5.03 p.p.m., the only major NOE was on
H-2, indicating an α-pyranose ring configuration for that residue.
Consequently the other residue was assigned as the β-pyranose
form. The 1 H, 13 C and 31 P chemical shifts for UDP and UMP were
similar to those reported previously [18].
WbvD catalyses epimerization of UDP-β-L-RhaNAc
to UDP-β-L-QuiNAc
In our previous study [6], WbjD was shown to epimerize UDPβ-L-PneNAc to UDP-β-L-FucNAc in an equilibrium reaction.
To investigate whether WbvD of V. cholerae catalyses the equivalent reaction step in UDP-β-L-QuiNAc biosynthesis, the
WbvB/WbvR reaction mixture was incubated with WbvD, and
the formation of a new sugar peak distinct from the UDP-βL-RhaNAc peak was observed (Figure 3E). However, the new
product could not be baseline-separated from UDP-β-L-RhaNAc.
In contrast, the use of WbjD in this reaction did not produce the
same peaks in CE analysis. This proved that WbjD did not catalyse
this conversion, suggesting stereospecificity with respect to
C-4, despite the striking similarity between the WbjD and WbvD
proteins (62 % identity and 78 % similarity; Figure 5).
Both WbvD and WbjD are capable of epimerizing UDP-β-L-PneNAc
to UDP-β-L-FucNAc
The stereospecificity of the V. cholerae enzyme WbvD was
investigated using purified UDP-β-L-PneNAc [6]. WbjD from
P. aeruginosa epimerizes this sugar to a mixture of UDP-β-LPneNAc and UDP-β-L-FucNAc (Figure 6B). The V. cholerae
enzyme WbvD acted on both C-4 epimers, catalysing the same
reaction, and produced the same product peak as the reaction when
WbjD was used as enzyme (Figure 6C). The ‘reverse’ experiment
using WbjD to test its reactivity on UDP-β-L-RhaNAc could
not be performed due to the intrinsic instability of this sugarnucleotide. As there is no selective pressure for catalysing the
conversion of UDP-β-L-PneNAc as well as UDP-β-L-RhaNAc in
an organism that does not contain L-QuiNAc, it is not surprising
that the Pseudomonas enzyme is stereospecific.
Both B. fragilis A.T.C.C. 23745 and V. vulnificus BO62316 have
been described to contain L-FucNAc, L-QuiNAc and L-RhaNAc in
their CP [10,11]. Unfortunately, at present no genetic information
on the corresponding polysaccharide biosynthesis loci in these
organisms is available. We predict that there are at least four
genes in these clusters dedicated to the biosynthesis of the
c 2003 Biochemical Society
994
Figure 5
B. Kneidinger and others
Sequence alignment of V. cholerae WbvD and P. aeruginosa WbjD
Alignment analysis was performed by using Clustal W [15]. Black boxes (∗ ) denote identical amino acid residues, and grey boxes (·) denote amino acid residues that are similar in terms of their
polar or hydrophobic properties.
Figure 6
CE analysis of WbvD stereospecificity
(A) UDP-L-PneNAc; (B) UDP-L-PneNAc converted with WbjD; (C) UDP-L-PneNAc converted
with WbvD.
nucleotide-activated precursors of these three sugars, namely
wbvB/wbjB, wbvR and wbjC as well as either one or both of
the wbvD and wbjD genes.
Conclusion
In the present study we provide evidence that UDP-D-GlcNAc
is the precursor substrate for the biosynthesis of 2-acetamido2,6-dideoxy-L-hexoses in the gluco-, galacto-, manno- and
talo-configurations (Scheme 1). WbvB of V. cholerae is highly
c 2003 Biochemical Society
Scheme 1 Biosynthesis pathways leading to UDP-2-acetamido-2,6dideoxy-L-hexoses
I, UDP-D-GlcNAc; II, UDP-2-acetamido-2,6-dideoxy-L-lyxo -4-hexulose; III, UDP-L-PneNAc;
IV, UDP-L-FucNAc; V, UDP-L-RhaNAc; VI, UDP-L-QuiNAc.
2-Acetamido-2,6-dideoxy-L-hexose biosynthesis
similar to WbjB (in P. aeruginosa) and its orthologue Cap5E
(in S. aureus). Consistent with previous observations on the
enzymic activities of WbjB and Cap5E [6], WbvB is also
a multifunctional enzyme possessing UDP-D-GlcNAc 4,6dehydratase/5-epimerase/3-epimerase activities. Thus WbvB is
also a key enzyme for the biosynthesis of all four configurations
of dideoxy-L-hexoses. The pathways diverge at the reduction
step, where the L-lyxo-hexulose intermediate is stereospecifically
reduced to either UDP-β-L-PneNAc or UDP-β-L-RhaNAc. The
identity among the amino acid sequences of the corresponding
reductases, WbjC/Cap5F or WbvR, is fairly low. However,
they all share conserved motifs containing the Rossmann fold
and the catalytic triad Ser-Tyr-Lys. Thus it is evident that the
stereospecificity of catalysis by these enzymes in sugar nucleotide
biosynthesis is not due to variations found in a single residue
or domain. WbjD/Cap5G and WbvD, which share high amino
acid sequence identity, catalyse the final steps of the pathways,
namely the epimerization at C-2 to yield UDP-β-L-FucNAc
or UDP-β-L-QuiNAc. It is intriguing to observe that, despite
their highly similar amino acid sequences, the enzymes are not
interchangeable and show unique substrate specificity in each
case. WbjD, which is active on UDP-β-L-PneNAc, is incapable
of catalysing its C-4-epimer UDP-β-L-RhaNAc, whereas its
homologue WbvD is capable of catalysing the C-2-epimerization
of both UDP-β-L-RhaNAc and UDP-β-L-PneNAc. In this case,
a single residue difference seems to be responsible for the
stereospecificity of the enzyme–substrate reaction.
We thank Jutta Nesper and Joachim Reidl (University of Würzburg, Germany) for kindly
providing us with genomic DNA of Vibrio cholerae serotype O37 strain V52. This work
was supported by operating grants to J.S.L. from the Canadian Institute for Health
Research (CHIR; #MOP-14687) and the Natural Science and Engineering Research
Council-Collaborative Health Research Projects program (#251007-02). An equipment
grant (#MMA-41558) to J.S.L. for the purchase of the CE instrument was obtained from
CIHR. J.S.L. is a Zellers Senior Scientist and the recipient of a Marsha Morton scholarship
from the Canadian Cystic Fibrosis Foundation.
REFERENCES
1 Graninger, M., Kneidinger, B., Bruno, K., Scheberl, A. and Messner, P. (2002) Homologs
of the Rml enzymes from Salmonella enterica are responsible for dTDP-beta-L-rhamnose
biosynthesis in the gram-positive thermophile Aneurinibacillus thermoaerophilus DSM
10155. Appl. Environ. Microbiol. 68, 3708–3715
2 Nakano, Y., Suzuki, N., Yoshida, Y., Nezu, T., Yamashita, Y. and Koga, T. (2000) Thymidine
diphosphate-6-deoxy-L-lyxo -4-hexulose reductase synthesizing dTDP-6-deoxy-L-talose
from Actinobacillus actinomycetemcomitans . J. Biol. Chem. 275, 6806–6812
3 Sturla, L., Bisso, A., Zanardi, D., Benatti, U., De Flora, A. and Tonetti, M. (1997)
Expression, purification and characterization of GDP-D-mannose 4,6-dehydratase from
Escherichia coli . FEBS Lett. 412, 126–130
4 Tonetti, M., Sturla, L., Bisso, A., Benatti, U. and De Flora, A. (1996) Synthesis of
GDP-L-fucose by the human FX protein. J. Biol. Chem. 271, 27274–27279
5 Sonesson, A., Jantzen, E., Tangen, T. and Zahringer, U. (1994) Chemical characterization
of lipopolysaccharides from Legionella feeleii, Legionella hackeliae and Legionella
jordanis . Microbiology 140, 2663–2671
6 Kneidinger, B., O’Riordan, K., Li, J., Brisson, J. R., Lee, J. C. and Lam, J. S. (2003) Three
highly conserved proteins catalyze the conversion of UDP-N -acetyl-d-glucosamine to
precursors for the biosynthesis of O antigen in Pseudomonas aeruginosa O11 and
capsule in Staphylococcus aureus Type 5. Implications for the
UDP-N -acetyl-L-fucosamine biosynthetic pathway. J. Biol. Chem. 278, 3615–3627
995
7 Marsden, B. J., Bundle, D. R. and Perry, M. B. (1994) Serological and structural
relationships between Escherichia coli O:98 and Yersinia enterocolitica O:11,23 and
O:11,24 lipopolysaccharide O-antigens. Biochem. Cell Biol. 72, 163–168
8 Shashkov, A. S., Arbatsky, N. P., Widmalm, G., Knirel, Y. A., Zych, K. and Sidorczyk, Z.
(1998) Structure and cross-reactivity of the O-specific polysaccharide of Proteus penneri
strain 26, another neutral Proteus O-antigen containing 2-acetamido-2,6dideoxy-L-glucose (N -acetyl-L-quinovosamine). Eur. J. Biochem. 253, 730–733
9 Cedzynski, M., Knirel Yu, A., Amano, K. I., Swierzko, A. S., Paramonov, N. A.,
Senchenkova, S. N. and Kaca, W. (1997) Structural and serological studies of the
O-antigen of the bacterium Proteus vulgaris OX2 (serogroup O2) used in the Weil-Felix
test. Biokhimiya (Moscow) 62, 15–20
10 Kasper, D. L., Weintraub, A., Lindberg, A. A. and Lonngren, J. (1983) Capsular
polysaccharides and lipopolysaccharides from two Bacteroides fragilis reference strains:
chemical and immunochemical characterization. J. Bacteriol. 153, 991–997
11 Jann, B. and Jann, K. (1968) 2-Amino-2,6-dideoxy-L-mannose (L-rhamnosamine)
isolated from the lipopolysaccharide of Escherichia coli 03:K2ab(L):H2. Eur. J. Biochem.
5, 173–177
12 Reddy, G. P., Hayat, U., Bush, C. A. and Morris, Jr, J. G. (1993) Capsular polysaccharide
structure of a clinical isolate of Vibrio vulnificus strain BO62316 determined by
heteronuclear NMR spectroscopy and high-performance anion-exchange
chromatography. Anal. Biochem. 214, 106–115
13 Coyne, M. J., Kalka-Moll, W., Tzianabos, A. O., Kasper, D. L. and Comstock, L. E. (2000)
Bacteroides fragilis NCTC9343 produces at least three distinct capsular polysaccharides:
cloning, characterization, and reassignment of polysaccharide B and C biosynthesis loci.
Infect. Immun. 68, 6176–6181
14 Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W. and Lipman,
D. J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search
programs. Nucleic Acids Res. 25, 3389–3402
15 Thompson, J. D., Higgins, D. G. and Gibson, T. J. (1994) CLUSTAL W: improving the
sensitivity of progressive multiple sequence alignment through sequence weighting,
position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22,
4673–4680
16 Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of
bacteriophage T4. Nature (London) 227, 680–685
17 Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72,
248–254
18 Koplin, R., Brisson, J. R. and Whitfield, C. (1997) UDP-galactofuranose precursor
required for formation of the lipopolysaccharide O antigen of Klebsiella pneumoniae
serotype O1 is synthesized by the product of the rfbDKP O1 gene. J. Biol. Chem. 272,
4121–4128
19 Brisson, J. R., Sue, S. C., Wu, W. G., McManus, G., Nghia, P. T. and Uhrin, D. (2002)
NMR of carbohydrates: 1D homonuclear selective methods. In NMR Spectroscopy of
Glycoconjugates (Jimenez-Barbero, J. and Peters, T., eds), pp. 59–93, Wiley-VCH,
Weinhem
20 Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989) Molecular Cloning: A Laboratory
Manual, 2nd edn, CSH Laboratory Press, Cold Spring Harbor, NY
21 Newton, D. T. and Mangroo, D. (1999) Mapping the active site of the Haemophilus
influenzae methionyl-tRNA formyltransferase: residues important for catalysis and tRNA
binding. Biochem. J. 339, 63–69
22 Reeves, P. R., Hobbs, M., Valvano, M. A., Skurnik, M., Whitfield, C., Coplin, D., Kido, N.,
Klena, J., Maskell, D., Raetz, C. R. and Rick, P. D. (1996) Bacterial polysaccharide
synthesis and gene nomenclature. Trends Microbiol. 4, 495–503
23 Graninger, M., Nidetzky, B., Heinrichs, D. E., Whitfield, C. and Messner, P. (1999)
Characterization of dTDP-4-dehydrorhamnose 3,5-epimerase and
dTDP-4-dehydrorhamnose reductase, required for dTDP-L-rhamnose biosynthesis in
Salmonella enterica serovar Typhimurium LT2. J. Biol. Chem. 274, 25069–
25077
24 Perry, M. B., MacLean, L. L. and Brisson, J. R. (1993) The characterization of the
O-antigen of Escherichia coli 064:K99 lipopolysaccharide. Carbohydr. Res. 248,
277–284
Received 10 January 2003/6 February 2003; accepted 7 February 2003
Published as BJ Immediate Publication 7 February 2003, DOI 10.1042/BJ20030099
c 2003 Biochemical Society