JOURNAL OF BACTERIOLOGY, Sept. 2006, p. 6195–6206
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Vol. 188, No. 17
Separate Pathways for O Acetylation of Polymeric and
Monomeric Sialic Acids and Identification of Sialyl
O-Acetyl Esterase in Escherichia coli K1
Susan M. Steenbergen,1Young-Choon Lee,1,2Willie F. Vann,3
Justine Vionnet,3Lori F. Wright,4and Eric R. Vimr1*
Laboratory of Sialobiology and Comparative Metabolomics, Department of Pathobiology, University of Illinois at Urbana-Champaign,
Urbana, Illinois1; Department of Biotechnology, Dong-A University, Busan, South Korea2; Center for Biologics Evaluation
and Research, U.S. Food and Drug Administration, Bethesda, Maryland3; and Department of
Microbiology and Immunology, University of Rochester, Rochester, New York4
Received 4 April 2006/Accepted 15 June 2006
O acetylation at carbon positions 7 or 9 of the sialic acid residues in the polysialic acid capsule of Escherichia
coli K1 is catalyzed by a phase-variable contingency locus, neuO, carried by the K1-specific prophage, CUS-3.
Here we describe a novel method for analyzing polymeric sialic acid O acetylation that involves the release of
surface sialic acids by endo-N-acetylneuraminidase digestion, followed by fluorescent labeling and detection of
quinoxalinone derivatives by chromatography. The results indicated that NeuO is responsible for the majority
of capsule modification that takes place in vivo. However, a minor neuO-independent O acetylation pathway
was detected that is dependent on the bifunctional polypeptide encoded by neuD. This pathway involves O
acetylation of monomeric sialic acid and is regulated by another bifunctional enzyme, NeuA, which includes
N-terminal synthetase and C-terminal sialyl O-esterase domains. A homologue of the NeuA C-terminal domain
(Pm1710) in Pasteurella multocida was also shown to be an esterase, suggesting that it functions in the
catabolism of acetylated environmental sialic acids. Our combined results indicate a previously unexpected
complexity in the synthesis and catabolism of microbial sialic and polysialic acids. These findings are key to
understanding the biological functions of modified sialic acids in E. coli K1 and other species and may provide
new targets for drug or vaccine development.
Escherichia coli K1 is a versatile human and animal faculta-
tive pathogen that causes a variety of extraintestinal diseases
including sepsis, meningitis, cystitis, pyelonephritis, cellulitis,
pneumonia, and postoperative infections. The ability of E. coli
K1 to invade and traverse the mammalian epithelial cell bar-
rier also may contribute to inflammatory bowel syndromes
such as Crohn’s disease. A primary virulence determinant in
these diseases is the polysialic acid capsule or K1 antigen, a
homopolymer of 2-keto-3-deoxy-5-acetamido-7,8,9-D-glycero-
D-galacto-nonulosonic, or N-acetylneuraminic acid (Neu5Ac,
the most common sialic acid), residues connected by ?2,8-
glycoketosidic linkages (36). The kps and neu genes needed for
polysialic acid synthesis and export map to a 17-kb accretion
domain inserted near pheV (7). Mutation of neu (biosynthetic)
genes generally results in no capsular polysaccharide produced
while kps mutations usually result in intracellular accumulation
of unexported polysaccharides (46, 47). Polysialic acid is also
found on the mammalian neural cell adhesion molecule and
comprises the group B meningococcal, Pasteurella haemolytica
A2, and Moraxella nonliquefaciens capsular polysaccharides
(41). Mammalian polysialic acid regulates cell migration, axon
pathfinding and targeting, and plasticity in the embryonic and
adult nervous system (6). Molecular mimicry of this antigen by
the bacterial capsules is thought to account for the relatively
low immunogenicity of microbial polysialic acid, which has
limited the attempts to produce safe and effective capsule-
based vaccines (41). Known functions of the capsule include
inhibition of phagocytosis and other innate immune responses
to microbial infection, but despite our understanding of cap-
sule function during extraintestinal disease, we know little
about its role in colonization of the mammalian large intestine.
Increased understanding of the colonization process may sug-
gest new targets for therapeutic development.
Unlike the neural cell adhesion molecule or group B me-
ningococcal polysialic acid, the E. coli K1 capsule may exist in
an alternate form in which the individual Neu5Ac residues are
variably modified with O-acetyl esters at carbon positions 7 or
9. The O-acetyltransferase gene, neuO, responsible for these
modifications is carried on a K1-specific prophage designated
CUS-3 (13). In addition to lysogeny, neuO expression is con-
trolled by a translational switch involving slipped-strand DNA
mispairing of heptanucleotide repeats located in the 5? coding
region. This switch is designated the poly? domain, where loss
or gain of heptad repeats in any number other than a multiple
of three results in frameshift mutation and synthesis of trun-
cated (inactive) neuO gene products. The neuO contingency
locus and its mobile phage delivery vehicle account for at least
five capsule forms: (i) permanently acetylation “off” because
the cell is not a CUS-3 lysogen, (ii) stochastic variation in the
proportion of “on” and “off” forms caused by neuO frameshift-
ing, (iii) variation in the degree of acetylation, which may
depend on the length of the poly? domain, (iv) variation in the
* Corresponding author. Mailing address: Laboratory of Sialobiol-
ogy, Department of Pathobiology, University of Illinois at Urbana-
Champaign, 2522 VMBSB, 2001 South Lincoln Avenue, Urbana, IL
61802. Phone: (217) 244-7421. Fax: (217) 333-8502. E-mail: ervimr
positions of acetyl esters on individual Neu5Ac residues of the
polysialic acid chains (carbon positions 7 or 9), resulting from
nonenzymatic transesterification and (v) variation in the posi-
tioning of sialyl O-acetyl esters along the chains resulting from
incomplete acetylation. Variation in neuO and its metabolic
products thus has the capacity to alter capsule antigenicity and
physiochemical properties of the K1 cell surface, with one
locus accounting for potentially thousands of different capsular
In addition to neuO, the K1 neuD gene product annotates as
an acyltransferase (4), and the group B Streptococcus (GBS)
NeuD orthologue has been shown to be a monomeric sialic
acid O-acetyltransferase responsible for modification of the
streptococcal capsular polysaccharide (23, 24). Complementa-
tion of a GBS neuD mutant with K1 neuD?restores sialyl O
acetylation, indicating that K1 NeuD is also a monomeric O-
acetyltransferase (23). In GBS, where O-acetyl esters are
found at carbon positions 7, 8, or 9, the acetylated monomers
are activated and transferred as terminal nonreducing residues
of the capsule main chain. Therefore, neuD in E. coli K1
suggests there may be two acetylation pathways in this species,
one involving neuO for modification of polysialic acid and the
other, involving neuD, for acetylation of monomeric sialic acid.
Furthermore, neuA encodes a bifunctional enzyme including
N-terminal cytidine 5?-monophospho-N-acetylneuraminic acid
(CMP-Neu5Ac synthetase) and C-terminal esterase in both the
K1 and GBS systems (26, 52), suggesting a mechanism for
converting acetylated monomeric sialic acid to the de-O-acety-
lated forms. In this communication we describe separate path-
ways for the O acetylation of polymeric and monomeric sialic
acids and provide the first demonstration of a new class of
esterase with activity against O-acetylated sialic acids.
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. The bacterial strains and
plasmids used in this study are shown in Table 1. Plasmid pWV200b used for
overproduction of E. coli K1 NeuA was derived from pWA1 (53, 54) as a
template using PCR primers containing EcoRI and HindIII sites. Double-di-
gested PCR product was cloned into similarly digested pKK223-3, placing K1
neuA?under control of the tac promoter and the ribosome binding site of
pKK223-3. A truncated version of K1 neuA containing the N-terminal 254 amino
acid residues (neuA?) was derived by PCR from pWS313 by mutagenesis meth-
ods described earlier (42) with the reverse primer Del 254R (5?-CGGCTCGA
GCGTTATATCACTTACACTATCGAATTCCGG) to yield pWS319. This
plasmid retains the synthetase domain but lacks the C-terminal esterase domain
defined previously by Liu et al. (26). LB (Lennox formulation) was purchased
from Fisher Scientific and used as the rich medium in all experiments. Minimal
M63 salts medium (35) containing 0.4% glycerol as a carbon source was used as
the defined medium where indicated. Ampicillin or chloramphenicol was used at
100 or 20 ?g/ml, respectively, for plasmid maintenance. Cultures were grown at
37°C with vigorous aeration in a water bath equipped with a rotary shaker.
Unless indicated otherwise, all cultures were supplemented with 20 ?g/ml of
Neu5Ac and grown to early stationary phase prior to cell harvesting by low-speed
Chemical reagents. Purified bovine submaxillary gland mucin (BSM) was
kindly provided by Tony Corfield (Bristol, United Kingdom). Neu5Ac was pur-
chased from ICN (Aurora, OH). N-acetyl-9-O-acetylneuraminic acid (Neu5,9Ac2)
was a gift from Roland Schauer (Kiel, Germany), while samples of N-acetyl-
7-O-acetylneuraminic acid (Neu5,7Ac2), N-acetyl-9-O-actylneuraminic acid
(Neu5,9Ac2), N-acetyl-7,8 or 9-O-acetylneuraminic acid [Neu5,7(8),9Ac3],
and N-glycolyl-9-acetylneuraminic acid (Neu5Gc9Ac), purified from BSM as
previously described (43), were provided by Tom Warner (San Carlos, CA).
N-glycolylneurminic acid (Neu5Gc), 3-deoxy-D-manno-octulosonic acid (KDO),
para-nitrophenyl-acetate (pNP-Ac), and 1,2-diamino-4,5-methylenedioxyben-
zene (DMB) were purchased from Sigma (St. Louis, MO). All other chemicals
were purchased from Sigma or Fisher Scientific and were of the highest technical
Enzymes. Recombinant, histidine-tagged endo-N-acetylneuraminidase (endo-
N), from phage PK1E, was purified from a culture of IPTG (isopropyl-?-D-
thiogalactoside)-induced E. coli BL21(DE3) essentially as described previously
(30). CMP-Neu5Ac synthetase was purified as previously described (53, 54) from
a culture of induced cells expressing pWV200b and had a specific activity of 610.3
units/mg protein, where 1 unit activates 1 ?mol of Neu5Ac in 1 min at 37°C.
Other enzymes were used as induced soluble extracts from cells (BL21 or DH5?)
grown in overexpression broth (Zymo Research, Orange, CA) with 1 mM IPTG.
Esterase activity was assayed as described by Yu et al. (52), with the absorbance
of para-nitrophenol, resulting from hydrolysis of pNP-Ac, determined at 405 nm
after 6 min of incubation at room temperature using a Beckman DU-640 spec-
trophotometer. Complete hydrolysis of pNP-Ac was accomplished by the addi-
tion of NaOH to a final concentration of 0.1 M. Data are expressed as activity
relative to complete base hydrolysis, normalized for protein concentration as
Isolation and characterization of neuO form variants. Strains EV717 (neuO
“off”) and EV718 (neuO “on”) were isolated from EV291 essentially as described
previously (13). Briefly, EV291 was grown overnight in LB and plated for single
colonies on agar medium containing 10% (vol/vol) horse-46 anti-polysialic acid
antiserum. Colonies were examined for surrounding halos representing precip-
itation responses to polysialic acid, where acetylated capsular polysaccharide
gives no or only a weak halo response. The lengths of the poly? domains in the
two variants were determined by PCR analysis using flanking forward and flank-
ing reverse primers (13) and were found to contain 20 and 21 heptad repeats in
EV717 and EV718, respectively.
DMB analyses and TLC of ?-keto compounds. For the analysis of capsular
polysialic acid by thin-layer chromatography (TLC), cells from 10- to 30-ml
cultures were harvested by centrifugation. The pellets were resuspended in 1/10
volume of 10 mM phosphate buffer (pH 7.0) and repelleted by centrifugation.
Polysialic acid in the supernatants was either applied directly to silica gel thin-
layer plates or treated with endo-N (50 ?g) at 37°C for 2 h to release sialic acids
prior to the application. Plates were developed with n-propanol:water (7:3,
vol/vol) solvent, and sialic acids were visualized by orcinol spray as previously
described (39). For DMB analysis, oligosialic acids released by endo-N digestion
TABLE 1. Bacterial strains and plasmids used in this study
K-12/K1 hybrid (CUS-3?)
K-12/K1 hybrid (CUS-3?)
EV50 neuB neuS
Derivative of RS218 (CUS-3?)
EV5 nanA?(isogenic with EV715)
EV291 neuO “off”
EV291 neuO “on”
MC4100 K-12 (K1 null)
Cliinical K1 isolate (CUS-3?)
EV36 nanA neuD
PCR cloning vector
pGEM-T Easy neuO?
aExpresses first 254 amino acid residues of NeuA.
6196 STEENBERGEN ET AL. J. BACTERIOL.
of whole cells were hydrolyzed completely to monomers by incubation at 80°C
with an equal volume of 4 M acetic acid for 3 h. Any precipitate was removed by
centrifugation in a microcentrifuge at 16,100 ? g for 10 min at 4°C. Samples were
concentrated at least 10-fold in a Savant SpeedVac system and recentrifuged,
and 15-?l samples were subjected to DMB-labeling exactly as described in the
original procedure (18). After being labeled in the dark for 2.5 h at 50°C, samples
were centrifuged through 0.22-?m-pore-size nylon filters, and 6-?l aliquots were
immediately analyzed fluorescently (373 nm excitation and 448 nm excitation)
using a reverse phase (RP) high-performance liquid chromatography (HPLC)
system equipped with Dionex AS50 Autosampler, GP50 Gradient Pump,
RF2000 Fluorescence Detector, and Chromeleon version 6.50 software for data
management. The solvent, acetonitrile-methanol-water (8:7:84, vol/vol/vol), was
slightly modified from the 9:7:84 mixture originally described (18). One major
modification was the use of a 4.6-mm by 10-cm TSKgel Super-ODS column (2
?m, 110 Å; Tosoh Bioscience, Montgomeryville, PA), which resulted in elimi-
nation of the reagent peak or DMB-breakdown product that elutes just after
Neu5,9Ac2(18). After the completion of any given set of experiments, the
column was washed with 50% methanol before the next use. For the analysis of
intracellular sialic acids, soluble cell extracts were prepared by sonic disruption
and centrifugation to remove debris and then incubated with an equal volume of
4 M acetic acid for 30 min at 80°C before precipitated material was removed by
centrifugation and RP-HPLC analysis as described above. Data were expressed
as relative fluorescence versus time as previously described (23, 24), where
elution of Neu5Ac quinoxalinone derivatives (25) varied between 9.5 and 13.6
min. However, on a given day’s experiments the relative elution of all peaks was
consistent. Where quantitative comparisons were made, the supporting data are
posted on the Laboratory of Sialobiology website (www.cvm.uiuc.edu/path
/sialobiology). Confirmation of O acetylation was obtained by mild-base hydro-
lysis prior to DMB labeling (24).
Two molecular forms of CMP-sialic acid synthetase. Acti-
vation of sialic acid for transfer to nascent polysialic acid by
NeuS (polymerase) is catalyzed by CMP-Neu5Ac synthetase
(NeuA). Liu et al. (26) showed that the first 229 amino acid
residues of NeuA comprise the synthetase, whereas residues
228 to 419 function as an acetylhydrolase (acylesterase) with
predicted tertiary structural similarity to the ?1-subunit of bo-
vine brain platelet-activating factor acetylhydrolase (PAF-AH)
isoform I. Similarly, the GBS NeuA orthologue C terminus was
shown to be an acetyl esterase, and the authors (52) speculated
that this activity might regulate accumulation of acetylated
Neu5Ac derivatives synthesized by NeuD (23, 24). These ob-
servations suggest two molecular forms of NeuA: the long or
bifunctional form, and a short form composed of just the syn-
thetase domain. This hypothesis predicts that organisms with
neuD might acetylate monomeric sialic acids and use a long
form of NeuA to regulate intracellular O-acetylated mono-
meric sialic acid concentrations, while organisms that synthe-
size or activate sialic acids but that lack neuD might express
only the NeuA short form that lacks esterase activity. We
designated the putative sialyl O-acetyl esterase domain of the
synthetase NeuA-star (NeuA*).
To investigate the correlation between the NeuA form and
sialic acid acetylation, we carried out a BLAST (3) search for
K1 NeuA and NeuD orthologues. As shown in Table 2, there
was general concordance between acetylation/NeuD and
NeuA*. Notable exceptions were the neuA gene products of
Pseudomonas aeruginosa and Legionella pneumophila, which
are known to acetylate the sialic acid-like molecules pseud-
aminic and legionaminic acid, respectively, and Neisseria men-
ingitidis groups C, Y, and W-135, which are known to acetylate
sialic acid but express the NeuA short form. Campylobacter
jejuni was unusual because it contains both short and long
synthetase forms. L. pneumophila expresses the NeuA short
form while the P. aeruginosa long-form C-terminal domain
was more similar to nucleotidyl transferase than esterase.
When taken together, the results shown in Table 2 suggest
that there may be two acetylation pathways in E. coli K1,
one for polysialic acid catalyzed by the neuO-encoded O-
acetyltransferase and another for acetylation of monomeric
sialic acids (Fig. 1).
Detection of O-acetylated cell surface sialic acids. Note that
the salient features of proposed sialic and polysialic acid syn-
thesis in Fig. 1 show two O acetylation pathways intercon-
nected at the substrate level by specific (NeuA*) or nonspecific
O-acetyl esterases and the possibility that NeuO could act on
the putative polysialic acid products of the NeuD-catalyzed
acetylation pathway. Testing these hypotheses required a sim-
ple method of monitoring capsule and intracellular O-acety-
lated sialic acids. Unfortunately, published methods of poly-
sialic acid isolation and characterization are laborious and
require high polysaccharide concentrations, while chemical
methods provide no information about the site(s) of O acety-
lation (28). To facilitate analytical investigation of acetylated
polymeric or monomeric sialic acids from multiple strains or
relatively small amounts of material, we took advantage of the
specificity of K1-lytic phage endo-N. Figure 2 (lane 2) shows
that digestion of polysialic acid from strain EV36 with a high
concentration of endo-N results in a mixture of monomeric
and oligomeric sialic acids. Note that the plate was overloaded
so that minor O-acetylated sialic acid could be detected if
present. A similar digestion pattern was observed when com-
mercially available polysialic acid (colominic acid), also un-
acetylated (13), was treated with endo-N (Fig. 2, lane 4). When
polysialic acid from EV36 harboring pSX785 (neuO?) was
treated with endo-N, a sialic acid species migrating faster than
Neu5Ac was detected (Fig. 2, lane 3). The relative mobility of
this species was consistent with it being Neu5,7Ac2 or
Neu5,9Ac2(32). Sensitivity of this faster-migrating species to
mild-base hydrolysis confirmed its identification as acetylated
TABLE 2. Distribution of NeuA* and NeuD in humans and in
bacteria that are known to synthesize or activate sialic,
pseudaminic, or legionaminic acid
Escherichia coli K1
aWellcome Trust Sanger Institute.
bORF, open reading frame.
cSialic acids modified with O-acetylesters. Ac, acetate.
dOpen reading frame pm1710, the first gene of the P. multocida nan operon (37).
eC-terminal domain is similar to nucleotidyl transferase instead of esterase.
VOL. 188, 2006 SIALIC ACID MODIFICATION 6197
sialic acid (not shown). The colorimetric TLC method shown
in Fig. 2 demonstrates that O acetylation neither prevents
polysialic acid hydrolysis by endo-N nor affects the composi-
tion of the limit digestion products. This conclusion suggested
a simple method for rapid and specific analysis of O-acetylated
Figure 3 shows the details of our analytical method, indicat-
ing release of oligosialic acids by endo-N treatment and sub-
sequent removal of intact cells by centrifugation. The released
material is converted to a mixture of free Neu5Ac and O-
acetylated monosaccharides (if present) by hydrolysis in dilute
acetic acid under conditions that do not destroy sialyl O-acetyl
esters. Labeling with DMB converts the sialic acids in the
hydrolysate to their quinoxalinone derivatives that have char-
acteristic retention times during RP-HPLC (Fig. 4A). Al-
though all of our analyses were carried out with pmol amounts
of sialic acids, DMB labeling is reported to detect as little as 50
to 100 fmol (18). Base treatment of BSM-derived sialic acids
resulted in loss of peaks representing Neu5,7Ac2, Neu5Ac9Gc,
Neu5,9Ac2, and Neu5,7 (8),9Ac3and the expected increased
ratio of Neu5Ac to Neu5Gc, which are both base resistant (Fig.
4B). Relative retention times of acetylated sialic acid deriva-
tives were consistent with the originally described method (18),
company literature (ProZyme Signal Product Code GKK-407),
and tandem electron spray mass spectrometry (24).
Polysialic acid O acetylation is primarily dependent on
NeuO. To detect acetylated polysialic acid independently of
FIG. 1. Biosynthesis of E. coli K1 polysialic acid and modification by O acetylation. Evidence supporting the biosynthesis and modification of
polysialic acid is provided in references 51 and 13, respectively. The monomeric modification, or NeuD-catalyzed pathway, is shown connected to
the main pathway by nonspecific (other) or specific (NeuA*) sialyl O-acetyl esterases. AcCoA, acetyl coenzyme A; Ac, acetate; ManNAc,
N-acetylmannosamine; UDP-GlcNAc, UDP-N-acetylglucosamine; PEP, phosphoenolpyruvate. Induction of the nanATEKyhcH operon caused by
derepression by NanR binding to sialic acid is indicated by the broken arrow.
FIG. 2. Sensitivity of acetylated polysialic acid to endo-N digestion.
Polysialic acid from EV36 (pGEM) or EV36 (pSX785) and commer-
cially available colominic acid (lanes 5 to 7, respectively) were digested
with recombinant endo-N and analyzed by TLC (lanes 2 to 4, respec-
tively). Lane 1 shows the migration of monomeric Neu5Ac standard
(20 ?g). Staining and migration artifacts resulted from sample over-
loading in order to detect minor O-acetylated Neu5Ac.
6198 STEENBERGEN ET AL.J. BACTERIOL.
any CUS-3 open reading frames other than neuO, strain EV36
transformed with vector or vector expressing neuO?(pSX785)
was subjected to DMB analysis. As shown in Fig. 5A, over 98%
of the sialic acid from EV36 was unacetylated, indicating that
if NeuD or any other O-acetyltransferase is active in this strain,
it can account for only about 2% of the total capsular sialic acid
residues. In contrast, polysialic acid derived from EV36 ex-
pressing pSX785 produced free Neu5Ac as well as detectable
Neu5,7Ac2and Neu5,9Ac2, which are the two known acety-
lated forms present in E. coli K1 strains lysogenized by CUS-3
(13, 18). Note the absence of triacetylated [Neu5,7 (8)9Ac3]
derivatives, which would elute after Neu5,9Ac2 (Fig. 4A), in-
dicating that NeuO most likely acetylates either the carbon-7
or carbon-9 position followed by nonenzymatic transesterifica-
tion between positions. The results of DMB capsule analysis
indicate that most of the acetylated Neu5Ac in polysialic acid
is synthesized by the NeuO pathway (Fig. 1). Nonetheless,
about 2% of the sialic acids appeared to be O acetylated by
a NeuO-independent pathway that is likely to depend on
To provide further evidence for the quantitative importance
of the NeuO-catalyzed pathway, we carried out DMB analysis
of strain EV291, a derivative of the prototypic K1 clinical
isolate RS218 known to carry the CUS-3 prophage and un-
dergo capsule form variation (13). Figure 5C shows the ex-
pected sialic acid profile of the isogenic neuO deletion mutant
EV708. This profile is nearly identical to that of EV36 shown
in Fig. 5A. Similarly, an “off” derivative (EV717) of EV291
produced mostly unacetylated polysialic acid (Fig. 5D). In con-
trast, an “on” form variant of EV291 (EV718) yielded a profile
resembling that of EV36 harboring pSX785 (Fig. 5B), except
that about 35% instead of 16% of the total sialic acid was O
acetylated (Fig. 5E). Variation in the degree of acetylation is
consistent with previously reported strain differences (19, 21,
28, 31). Note especially that in the “off” profile a small but
detectable amount of material, consistent with Neu5,7Ac2and
Neu5,9Ac2, as well as a peak representing Neu5,8Ac2, was
observed (Fig. 5D). These small amounts of NeuO-indepen-
dent acetylated sialic acids may have resulted from leakage of
intracellular monomers produced by a second O acetylation
pathway in E. coli K1 (Fig. 1). However, this hypothesis is
unlikely because, as shown below, neuA?strains do not accu-
mulate O-acetylated sialic acids. Therefore, they most likely
arose from activation and incorporation of a small proportion
of O-acetylated monomeric sialic acid that escaped NeuA*
recycling (Fig. 1). While some of the O-acetylated forms in
strain EV717 may have arisen from the stochastic proportion
of “on” forms expected to result during outgrowth of the pre-
dominant “off” form (13), this amount can be no more than 1
or 2% (compare O acetylation in Fig. 5D with that of Fig. 5C).
These considerations suggest that there is a minor or second-
ary pathway for the O acetylation of monomeric sialic acid and
that about 2% of these forms are ultimately incorporated into
polysialic acid independently of NeuO.
A second O acetylation pathway for monomeric sialic acid.
Lewis et al. (24) reported that a mutation in the GBS neuA
orthologue resulted in accumulation of Neu5,7Ac2, Neu5,8Ac2,
and Neu5,9Ac2, suggesting that NeuA normally out-competes
NeuD or other O-acetyltransferases for Neu5Ac substrate. Al-
ternatively, loss of the NeuA* esterase activity in this mutant
might have accounted for the increase in acetylated forms (52).
FIG. 3. DMB-labeling and analysis of capsular polysialic acid. En-
capsulated E. coli K1 (smooth rectangle with wavy lines) expressing
acetylated (Ac?) or unacetylated (Ac?) polysialic acid (wavy lines) is
treated in step 1 with recombinant endo-N (lawn mower icon), pro-
ducing oligosialic acids and intact cells with shorn capsules; treated
cells are removed by centrifugation. After acetic acid hydrolysis to
produce monomeric sialic acids (Sia’s) in step 2, the reducing ends
were labeled with DMB (step 3). The resulting fluorescent quinoxali-
none derivatives are analyzed by RP-HPLC as described in the text
FIG. 4. DMB analysis of sialic acid standards from BSM. (A) BSM
subjected to steps 2 to 4 shown in Fig. 3. (B) After treatment of BSM
in step 2, base hydrolysis of O-acetyl esters was carried out prior to
completion of steps 3 and 4. Note the relative increase of the Neu5Ac
peak to Neu5Gc after base treatment, indicating the conversion of
O-acetylated sialic acids to Neu5Ac. The identities of the unshaded
peaks in this and succeeding figures are unknown.
VOL. 188, 2006SIALIC ACID MODIFICATION 6199
If a similar pathway is operable in E. coli K1, loss of NeuA in
a mutant also lacking sialic acid aldolase (NanA) should result
in accumulation of acetylated sialic acids (Fig. 1). Elimination
of NanA would be necessary to detect this phenotype because
in a neuA mutant the increased pool of Neu5Ac binds NanR
repressor, leading to induction of the nanATEKyhcH catabolic
operon and destruction of intracellular sialic acids (22, 29, 30).
When a whole-cell extract (sonicate) of the nanA neuA dou-
ble mutant EV715 was subjected to DMB analysis, four peaks
other than free Neu5Ac were observed (Fig. 6A). On the basis
of BSM-derived standards (Fig. 4A) and analogy to the results
of tandem electrospray mass spectrometry (24), the three
peaks eluting with retention times greater than the Neu5Ac
peak represent Neu5,7Ac2, Neu5,8Ac2, and Neu5,9Ac2, re-
spectively. Confirmation that these peaks were derived from
O-acetylated sialic acids was shown by sensitivity to mild-base
treatment carried out prior to DMB-labeling (Fig. 6B). Note
the increase in the ratio of peak c (Neu5Ac) to peak a in Fig.
6B, indicating conversion of O-acetylated sialic acids to
Neu5Ac. Peak a, with the earliest retention time, eluted even
before Neu5Gc (Fig. 4A), suggesting that it might be the 8-car-
bon ?-keto sugar acid KDO (34). Spiking a sample of the
EV715 extract with Neu5Gc prior to DMB analysis resulted in
the addition of one new peak to the profile (Fig. 6C), showing
that peak a was not Neu5Gc. Chromatography of derivatized
KDO, Neu5Gc, or KDO plus Neu5Gc unambiguously con-
firmed that peak a was derived from KDO (not shown). We
assume that KDO in the intracellular extracts results from
partial release of free KDO during acid treatment from lipo-
polysaccharide present as contaminating membrane material.
However, some kps genes (kpsF and kpsU) have been shown to
FIG. 5. NeuO is responsible for the majority of O-acetylated poly-
sialic acid in E. coli K1. The indicated strains were subjected to DMB
analysis as shown in Fig. 3. The percentages (relative to Neu5Ac) of
O-acetylated sialic acids in panels A to E are 1.9, 16.4, 3.1, 4.0 and 35.2,
respectively. The quantitative data supporting the relative percentages
are presented in a file, designated DMB labeling, that is freely avail-
able at www.cvm.uiuc.edu/path/sialobiology.
FIG. 6. Intracellular accumulation of O-acetylated sialic acids in E.
coli K1 neuA mutants. Intracellular sialic acids from the unencapsu-
lated E. coli K1 mutant EV715 (nanA neuA) and its isogenic nanA?
derivative EV716 were subjected to DMB analysis (panels A and D,
respectively). Sialic acids from strain EV715 were treated with base to
demonstrate the presence of O-acetyl esters (B) or supplemented with
20 pmol of Neu5Gc prior to the analysis (C). Peaks a to f represent the
relative elution of quinoxalinone derivatives of KDO, Neu5Gc,
Neu5Ac, Neu5,7Ac2, Neu5,8Ac2and Neu5,9Ac2, respectively.
6200STEENBERGEN ET AL. J. BACTERIOL.
function in the KDO biosynthetic pathway, warranting future
studies of KDO metabolism in E. coli K1 strains. Because free
sialic acid is not detectable by colorimetric or amperometric
methods in wild-type E. coli K1 (29), we concluded that loss of
NeuA allows accumulation of sufficient free Neu5Ac in a nanA
mutant background to produce a detectable pool of acetylated
forms, which by analogy to GBS may result from the action of
NeuD (Fig. 1). Note that the combined amount of diacetylated
forms is at least twice that of the intracellular Neu5Ac con-
centration (Fig. 6A). Although the concentration of all sialic
acid derivatives was reduced in the nanA?neuA mutant,
EV716, acetylated forms were still detectable (Fig. 6D). We
assume that the known relative resistance of acetylated sialic
acids to sialate aldolase encoded by nanA accounts for this
Although LB contains a low concentration of free Neu5Ac
(45), all of the acetylated forms synthesized by EV715 and
EV716 (Fig. 6) were dependent on the mutant background.
This conclusion was substantiated by analysis of an extract
from EV78, an E. coli K-12 derivative of strain MC4100 that
lacks the kps/neu accretion domain. Note the accumulation of
Neu5Ac by EV78 but the absence of diacetylated forms (Fig.
7A). As expected, when EV78 was grown in nonsupplemented
minimal medium, no free Neu5Ac peak was observed (Fig.
7B). However, when the medium was supplemented with ex-
ogenous Neu5Ac, accumulation of intracellular sialic acid was
detected (Fig. 7C). Note that the failure of supplemented
EV78 to produce diacetylated sialic acids means that E. coli
lacking the kps/neu genes does not express Neu5Ac O-acetyl-
transferase. Furthermore, the failure to accumulate diacety-
lated forms by the nanA neuB neuS triple mutant, EV239
(compare Fig. 7D and E) suggests efficient de-O acetylation by
NeuA*. Although as expected EV36 (wild type) does not ac-
cumulate detectable free sialic acid, the peak eluting with the
longest retention time represents DMB-labeled pyruvate, an-
other ?-keto acid that most likely results from sialic acid ac-
cumulation and subsequent NanA cleavage during polysialic
acid biosynthesis (Fig. 7F). Finally, to determine if activation
of sialic acids protects acetylated forms from NeuA*, we ana-
lyzed an extract from strain EV136 that accumulates CMP-
Neu5Ac due to a NeuS polymerase defect (29, 38, 51). Al-
though the free Neu5Ac peak resulting from acid hydrolysis of
CMP-Neu5Ac in this extract was apparent, acetylated forms
were not detected, indicating that if acetylated sialic acids are
activated by the synthetase, NeuA* deacetylates them prior to
DMB labeling (Fig. 7G). We conclude from the results shown
in Fig. 5 to 7 that the monomeric O-acetyltransferase, prob-
ably NeuD (23), produces diacetylated sialic acids that are
normally deacetylated by NeuA* or nonspecific esterase
(Fig. 1). As shown in Fig. 5, only a small percentage of the
acetylated monomers are ever incorporated into capsular
polysialic acid (Fig. 5). This is a major distinction with the
FIG. 7. Endogenous NeuA* prevents intracellular accumulation of
O-acetylated sialic acids in vivo. Wild type or the indicated mutants
were grown under the specified conditions and intracellular ?-keto
acids analyzed by DMB labeling and RP-HPLC. (A) Growth in non-
supplemented LB shows presence of contaminating Neu5Ac but lack
of exogenous O-acetylated sialic acids. (B) Growth in nonsupple-
mented minimal medium shows expected absence of Neu5Ac in a K-12
strain lacking kps and neu genes. (C) Growth in supplemented
minimal medium showing uptake and accumulation of Neu5Ac. (D)
Growth of the triple mutant in nonsupplemented minimal medium.
(E) Growth of the triple mutant in supplemented minimal medium.
(F) Growth of the wild type in nonsupplemented LB. (G) Growth of
the polymerase mutant in nonsupplemented LB.
VOL. 188, 2006SIALIC ACID MODIFICATION 6201
NeuD catalyzed pathway in GBS, where 50 to 60% of O-
acetylated monomeric sialic acids are added to the strepto-
coccal polysaccharide (23, 24).
O-acetyl esterase (NeuA*) activity of NeuA. To determine if
NeuA has sialyl O-acetyl esterase activity, recombinant syn-
thetase was added to an extract of EV715. As shown in Fig. 8A,
all three diacetylated peaks were sensitive to the NeuA* activ-
ity of the purified enzyme. Heating enzyme at 90°C for 5 min
prior to addition to the EV715 extract eliminated esterase
activity, indicating the absence of nonspecific de-O acetylation
(Fig. 8B). Using half the amount of enzyme as the experiment
shown in Fig. 8A resulted in incomplete deacetylation (Fig.
8C). That contamination of the enzyme preparation by non-
specific esterase is an unlikely explanation for these results was
apparent from the extended (5 h) incubation of the EV715
extract at 37°C. Though this treatment resulted in some loss of
Neu5,7Ac2, it had only a minor effect on Neu5,8(9)Ac2(Fig.
8D), while all three acetylated forms were eliminated after
incubation with purified NeuA for 30 min (Fig. 8A).
NeuA* has been shown to hydrolyze pNP-Ac between pH
7.5 and 9.0 (26, 52). As shown in Table 3, the normalized
recombinant NeuA activity against this model substrate was as
great as the hydrolysis produced by alkali treatment, indicating
that NeuA* rapidly and completely deacetylated pNP-Ac. Sim-
ilarly, an extract containing overproduced K1 NeuA, but not
one containing the overproduced NeuA short form from group
B meningococci, expressed NeuA* activity that was at least
nine times higher than background (Table 3). This background
activity was likely due to nonspecific esterase because the same
amount of background was observed in an extract prepared
from EV715, which lacks NeuA because of mutation (Table 3).
In contrast, an extract containing overproduced Pm1710 from
Pasteurella multocida (37) had elevated esterase activity while
an extract containing the overproduced short form of P. mul-
tocida NeuA (Pm0187) did not (Table 3). When taken to-
gether, the results shown in Fig. 8 and Table 3 indicate that
NeuA*, in addition to its expected activity against pNP-Ac (26,
52), is a sialyl O-acetyl esterase. This conclusion is supported
by the identity of conserved blocks I, II, III, and V found in
GDSL esterases of the SGNH-hydrolase family (2) whose
members include conserved catalytic SGNH residues at the
indicated positions of each polypeptide shown in Fig. 9,
strongly suggesting that Pm1710 also is a sialyl O-acetyl ester-
ase (37). Further confirmation that NeuA* is an acetyl esterase
was obtained by showing that an extract containing overpro-
duced, truncated K1 NeuA composed of just the first 254
residues (synthetase domain) lacked elevated activity against
pNP-Ac (Table 3).
Contribution of NeuD to O acetylation of polysialic acid. To
determine if the small amount of NeuO-independent O-acety-
lated sialic acids represents monomers that escaped NeuA*
recycling and were subsequently incorporated into polysialic
acid, we used exogenous Neu5Ac to rescue the nanA neuD
double mutant RS2887 (12). In addition to its activity as a
monomeric sialic acid O-acetyltransferase (23), NeuD is re-
quired for sialic acid synthesis (4, 12). The synthetic defect of
a neuD mutant can be rescued by adding Neu5Ac to the growth
medium as long as the cell lacks NanA to prevent the lyase
from destroying transported monosaccharide. When polysialic
acid from sialic acid-rescued RS2887 was analyzed by DMB
labeling, there were no detectable O-acetylated forms (not
shown), indirectly suggesting that NeuD is necessary for the
FIG. 8. Detection of NeuA* in vitro. Extracts of EV715 were
treated or not with purified recombinant K1 NeuA under the specified
conditions prior to DMB analysis. (A) NeuA added for 30 min at 37°C
before labeling. (B) Extract treated as in panel A with heat-inactivated
(90°C for 5 min) NeuA. (C) Extract treated as described in panel A but
with half the amount of NeuA. (D) Extract held on ice for 5 h com-
pared to a sample of the sample extract incubated at 37°C for 5 h prior
TABLE 3. Relative esterase activities of various NeuA forms and
Pm1710 against pNP-Ac
Plasmid (gene)NeuA form Sourcea
E. coli K1
N. meningitidis group B
E. coli K1
aSpecies from which the gene was cloned or strain used as source of extract.
bNormalized to 50 ?g of protein unless indicated otherwise.
cFirst gene of the P. multocida nan operon; lacks synthetase domain (37).
dTruncated K1 neuA lacking NeuA* (Table 1).
ePurified K1 NeuA from overexpressing pWV200b (Table 1).
fRelative activity produced by 10 ?g of purified K1 NeuA.
6202 STEENBERGEN ET AL.J. BACTERIOL.
minor O acetylation detected in wild-type strains lacking neuO
(Fig. 5). We conclude that NeuD is a functional monomeric
O-acetyltransferase in E. coli K1 but that NeuA* recycles most
monomers to Neu5Ac before incorporation into polymer (Fig.
1). The exact function of this minor pathway is unknown,
although we assume that NeuD is responsible for the O acet-
ylation of monomeric sialic acids in species such as GBS, E. coli
O104, and others (23). However, our results unambiguously
demonstrate the quantitative importance of the NeuO-cata-
lyzed pathway, suggesting that NeuD, at least in E. coli K1,
plays only a minor role in capsule modification.
Identification and function of NeuA*. Our results indicate
that sialyl O-acetyl esterase activity is catalyzed by the long
form of NeuA, a bifunctional CMP-Neu5Ac synthetase in E.
coli K1 and other bacteria that decorates their surfaces with
sialic acids (Table 2). On the basis of three-dimensional mo-
lecular modeling and biochemical activity, Liu et al. (26) con-
cluded that the K1 NeuA C terminus (residues 229 to 419) was
a PAF-AH. These authors speculated that PAF cleavage may
function in E. coli K1 invasion of the blood-brain barrier after
proteolytic release of PAF-AH from the N-terminal synthetase
domain (26). We think this suggestion is unlikely for at least six
reasons. First, NeuA is a cytoplasmic enzyme, and even if it
were proteolytically processed in vivo, how it gains access to
PAF would still remain unclear. Even if PAF were somehow
accessible as a substrate, it is unclear how hydrolysis would
affect endothelial invasion. E. coli K1 invasion of the blood-
brain barrier has been studied extensively in tissue culture, and
the model seems unconvincingly complicated since mutants
with in vitro invasion defects still cause meningitis in the ro-
dent disease model and have only modest invasion defects in
vivo (20). Second, our demonstration that O-acetylated sialic
acids are NeuA* substrates provides the first direct evidence
for this activity, suggesting that PAF may not be a physiological
substrate. Third, the failure to accumulate O-acetylated sialic
acids in a neuA?strain points to a function of NeuA* in
regulating the intracellular monomeric O-acetylated sialic acid
concentration. Fourth, the obligate animal commensal and fac-
ultative pathogen P. multocida expresses CMP-sialic acid syn-
thetase encoded by the short form of neuA (pm0187). P. mul-
tocida also expresses a homologue of neuA* (pm1710), which
maps as the first gene of a functional nan operon for catabo-
lism of environmental sialic acids (37). The concentration of
O-acetylated sialic acids in nonhuman mammalian serum is
more than twice that of Neu5Ac (18), suggesting that the
pm1710 gene product may function to remove O-acetyl esters
after the cell transports acetylated sialic acids from the host. O
acetylation is known to inhibit sialate aldolase (NanA), further
suggesting that the NeuA* physiological substrates, indepen-
dent of species, are O-acetylated sialic acids. Fifth, the active
residues in blocks I to III and V of NeuA* are more similar to
esterases such as Pm1710 and TesA than to PAF-AH (Fig. 9).
Finally, other neuroinvasive strains like N. meningitidis and
Haemophilus influenzae express the short form of neuA, indi-
cating that NeuA* is not essential for cell invasion. Evidence
for the in vivo function of NeuA* in vivo may be obtained by
constructing an E. coli K1 mutant that lacks star (O-acetyl
esterase) activity while retaining synthetase.
Human serum contains over 30 times less O-acetylated sialic
acid than Neu5Ac (18), which may explain why an obligate
human commensal like H. influenzae lacks a pm1710 homo-
logue despite the functional expression of other nan catabolic
genes (44). Note that H. influenzae neuA encodes the short
FIG. 9. Sequence alignment of four conserved blocks in the SGNH-hydrolase family. Blocks I, II, III, and V found in enzymes of the
SGNH-hydrolase family (2) are shown boxed, including the NeuA* domains of GBS and K1 NeuA; the esterase encoded by P. multocida open
reading frame pm1710; E. coli thioesterase I (TesA), accession no. AAC37396; and human PAF-AH, accession no. AAH07863. Other accession
numbers are given in Table 2. Single-letter amino acid designations are used, and the numbering in parentheses indicates the relative positions of
each block to the N terminus. Note that block I is always found at the N terminus of esterases, consistent with its relative positions in Pm1710,
TesA, and PAF-AH. Note that in the two CMP-sialic acid synthetases, the esterase domains are located at the C termini of these bifunctional
enzymes. Conserved residues are in boldface, where asterisks indicate catalytic residues.
VOL. 188, 2006SIALIC ACID MODIFICATION 6203
form of the synthetase (Table 2), suggesting that it does not
modify its surface with O-acetylated sialic acids. Because most
E. coli strains do not synthesize or activate sialic acids, and
therefore lack neuA, the NeuA* physiological substrates in E.
coli K1 may be the monomeric O-acetylated sialic acids pro-
ase), NeuB (synthase), and NeuD (Fig. 1), instead of environ-
mentally derived sialic acids. The niche of the human and
animal large intestine occupied by E. coli K1 and other E. coli
is a complex and poorly understood microbial environment
that includes a diverse set of organisms expressing sialidases,
sialyl O-acetyl esterases, and other catabolic enzymes directed
against mucins and mucin-derived monosaccharides (8–10),
suggesting that the E. coli nan system may primarily scavenge
Neu5Ac instead of O-acetylated forms. However, over half of
all mucin sialic acid residues are O acetylated (8–10), suggest-
ing that E. coli K1 NeuA* may confer a selective advantage
over other E. coli lacking the esterase. For example, in a study
of healthy pregnant females, E. coli K1 was the most prevalent
aerobic species in 38% of the participants (31), indicating that
it may express specialized colonization factors such as efficient
deacetylation of environmental sialyl O-acetyl monomers. The
temperature dependence of capsule synthesis and nicotin-
amide auxotrophy of many K1 strains also may indicate ongo-
ing adaptation to an animal environment (48), in which case
NeuA* could play an important role in colonization and per-
Contribution of the monomeric O acetylation pathway to
capsule modification. Although NeuA*-sensitive O-acetylated
sialic acids accumulate in a neuA mutant, most of the acety-
lated sialic acids in the capsule are derived from the NeuO-
catalyzed pathway. However, 2 to 4% of residues in polysialic
acid are O acetylated by the NeuO-independent pathway, in-
dicating that all E. coli K1 stains, regardless of CUS-3 status,
contain a small amount of modified residues. The apparent
synthesis and reconversion of most O-acetylated monomeric
sialic acids to Neu5Ac would seem to be a wasteful or nones-
sential process (Fig. 1). However, E. coli serotype O104 (Table
2) synthesizes O-acetylated monomeric sialic acid as a struc-
tural component of lipopolysaccharide (15). NeuA* offers a
mechanism to regulate the degree of acetylation by controlling
the concentration of O-acetylated sialic acids available for sur-
face modification. In contrast, no O acetylation of sialic acid
was reported for the E. coli O145 serotype (14), suggesting that
there may be differences in the relative activity of NeuA* or
activity of NeuD in some strains. In GBS the relative concen-
tration of O-acetylated monomeric sialic acids is low unless the
neuA orthologue is inactivated (24). Although Lewis et al. (24)
ascribed this phenotype to a competition between the syn-
thetase and O-acetyltransferase encoded by neuD (23), our
results support NeuA* as the molecular explanation for the
accumulation of O-acetyl forms in an E. coli K1 neuA mutant.
Indeed, it is unclear how the levels of GBS polysaccharide O
acetylation reported could occur in the presence of NeuA*,
unless the activity of this enzyme in GBS is less than in E. coli
K1. In vitro comparisons of NeuA* proteins from E. coli K1
and GBS against the model esterase substrate pNP-Ac indicate
that the E. coli esterase is relatively more active (52).
Similarly, the groups C, Y, and W-135 meningococcus neuA
homologues lack the genetic information to code for NeuA*,
and the degree of O acetylation in these strains is 90%, 47%,
and 66%, respectively (21), which is much higher than in E. coli
K1 strains lacking neuO. Meningococcal strains also lack NeuD
(Table 2), supporting the negative correlation between the
degree of acetylation and NeuA*. In contrast to these NeuA*-
negative meningococcal strains, the degree of capsule acetyla-
tion in E. coli K1 CUS-3?strains may approach 100% (28)
despite the simultaneous occurrence of sialyl O-acetyl esterase.
Our current results show that capsule O acetylation is largely
dependent on neuO. That NeuO acetylates polymeric instead
of monomeric sialic acid (19) presumably accounts for the lack
of an effect of NeuA* on the degree of polysialic acid O
acetylation, since it is as yet unclear whether NeuA* recognizes
O-acetylated polysialic acid. However, the degree of E. coli K1
polysialic acid acetylation is strain specific (13, 19, 28). Sialic
acids from stains lacking neuO either due to mutation or nat-
urally because they are not CUS-3 lysogens contain few O-
acetyl esters. In contrast, CUS-3 lysogens expressing neuO in
the “on” form vary over a wide range in the degree of acety-
lation (28), which may not be solely dependent on the stochas-
tic proportion of cells in the “off” form (13). How or even
whether the degree of acetylation is affected during host col-
onization relative to that observed in cells grown in vitro, as in
the current study, provides fertile new ground for investigating
the biological functions of variable capsule modification in the
context of the host-microbe interaction.
Function of the NeuD-catalyzed pathway. Our results indi-
cate that the NeuD-catalyzed addition of O-acetyl esters to
Neu5Ac is a continuous process in E. coli K1 but that NeuA*
limits accumulation of O-acetylated forms or their incorpora-
tion into capsule. However, while neuD?(pRS361) comple-
mented a ?neuD mutant (RS2887) for sialic acid synthesis, the
plasmid did not result in detectable O-acetylated sialic acids
when expressed in strain EV78 grown in sialic acid-supple-
mented medium (not shown). Thus, while NeuD is the most
likely candidate for the E. coli K1 sialyl O-acetyltransferase,
our results do not provide direct evidence for this conclusion
despite the recent observation that K1 NeuD appears to acet-
ylate Neu5Ac in GBS (23). Note that Lewis et al. (23) did not
identify acetyl-coenzyme A as the presumed two-carbon donor
nor demonstrate biochemically that either K1 or GBS NeuD
functions as an acetylase in vitro. Both of these biochemical
conditions have been demonstrated for NeuO (13). It may be
that NeuD requires interaction with other kps or neu gene
products in order to be an active O-acetyltransferase, and a
direct interaction between NeuD and NeuB has been demon-
strated in vivo (11). However, in unpublished experiments,
plasmids expressing neuBAC or neuDBAC in EV78 did not
produce detectable intracellular Neu5Ac, suggesting that stoi-
chiometry or interactions with still other gene products not
present in the K-12 background may be critical. Evidence that
O-acetylated monomers are not necessarily protected from
NeuA* upon activation by the synthetase domain was evident
from the phenotype of EV136 shown in Fig. 7G. This mutant
has a defect in neuS and accumulates high intracellular con-
centrations of CMP-Neu5Ac (29, 38) but no detectable O-
acetylated forms, suggesting that K1 NeuA either does not
activate acetylated monomers or, if so, then they, too, must be
NeuA* substrates. However, because GBS NeuA clearly acti-
vates O-acetylated Neu5Ac, the low incorporation of NeuO-
6204 STEENBERGEN ET AL. J. BACTERIOL.
independent acetylated sialic acids in polysialic acid presum-
ably reflects the efficiency of NeuA* and, perhaps, the
efficiency of the polymerization process (39). It will be inter-
esting to determine whether NeuA* is active against O-acety-
lated polysialic acid. Our current results suggest that the small
but constant amount of NeuO-independent acetylation of
polysialic acid may be a target for potential vaccination against
all K1 strains.
Finally, the NeuD-catalyzed O acetylation pathway may be
irrelevant to the biology of E. coli K1. This possibility follows
from the bifunctional nature of neuD, where in addition to its
acetylase activity, NeuD is required for sialic acid synthesis
(12). Dissection of these two functions has been accomplished
for GBS NeuD by site-directed mutagenesis, resulting in loss of
the acetylase activity but retention of the sialic acid biosyn-
thetic phenotype (23). Therefore, it may be that only the bio-
synthetic function of NeuD is important in E. coli K1. The
situation where only one of two biochemical functions is im-
portant to polysialic acid biosynthesis might be analogous to
KpsF, which is enzymatically active in KDO-precursor biosyn-
thesis (27) but also includes a cystathionine-?-synthase or
Bateman domain (5) that is now understood to bind adenosine
ligands and function in a variety of cellular regulatory phenom-
ena (33). E. coli kpsF mutants are defective in export of poly-
sialic acid that accumulates intracellularly (7). Thus, the enzy-
matic functions of NeuD and KpsF may be of little or no
relevance to polysialic acid synthesis or its in vivo functions. In
either case our results indicate that the metabolism of micro-
bial monomeric and polymeric sialic acids is far more complex
than previously thought. The absence of at least some of these
metabolic processes in mammals suggests new targets for po-
tential drug or vaccine development (47–49).
This research was supported by NIH grant R01 AI042015 (E.R.V.).
Y.-C. Lee was supported in part by the LG Yeonam Foundation,
South Korea, during his sabbatical with the Laboratory of Sialobiology.
We thank Kerry Helms for expert photographic assistance. We are
grateful to Mark and Theresa Kuhlenschmidt for technical training in
the use of the Dionex HPLC components used for analysis of DMB-
labeled ?-keto compounds, and Richard Silver for stimulating discus-
sions. We gratefully acknowledge the kind gifts of BSM or O-acety-
lated standards from T. Corfield, T. Warner, and R. Schauer and to U.
Rutishauser for pEndo-N.
1. Achtman, M., A. Mercer, B. Kusecek, A. Pohl, M. Heuzenroeder, W. Aaronson,
A. Sutton, and R. P. Silver. 1983. Six widespread bacterial clones among
Escherichia coli K1 isolates. Infect. Immun. 39:315–335.
2. Akoh, C. C., G. C. Lee, Y. C. Liaw, T. H. Huang, and J. F. Shaw. 2004. GDSL
family of serine esterases/lipases. Prog. Lipid Res. 43:534–552.
3. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller,
and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation
of protein database search programs. Nucleic Acids Res. 25:3398–3402.
4. Annunziato, P. W., L. F. Wright, W. F. Vann, and R. P. Silver. 1995. Nucle-
otide sequence and genetic analysis of the neuD and neuB genes in region 2
of the polysialic acid gene cluster of Escherichia coli K1. J. Bacteriol. 177:
5. Bateman, A. 1997. The structure of a domain common to archaebacteria and
the homocystinuria disease protein. Trends Biochem. Sci. 22:12–13.
6. Bruse ´s, J. L., and U. Rutishauser. 2001. Roles, regulation, and mechanism of
polysialic acid function during neural development. Biochimie 83:635–643.
7. Cieslewicz, M., and E. R. Vimr. 1997. Reduced polysialic acid expression in
Escherichia coli K1 mutants with chromosomal defects in kpsF. Mol. Micro-
8. Corfield, A. P., S. A. Wagner, L. J. O’Donnell, P. Durdey, R. A. Mountford,
and J. R. Clamp. 1993. The roles of enteric bacterial sialidase, sialate O-
acetyl esterase and glycosulfatase in the degradation of human colonic mu-
cin. Glycoconj. J. 10:72–81.
9. Corfield, A. P., R. Wiggins, C. Edwards, N. Myerscough, B. F. Warren, P.
Soothill, M. R. Millar, and P. Horner. 2003. A sweet coating—how bacteria
deal with sugars. Adv. Exp. Med. Biol. 535:3–15.
10. Corfield, T. 1992. Bacterial sialidases—roles in pathogenicity and nutrition.
11. Daines, D. A., and R. P. Silver. 2000. Evidence for multimerization of Neu
proteins involved in polysialic acid synthesis in Escherichia coli K1 using
improved LexA-based vectors. J. Bacteriol. 182:5267–5270.
12. Daines, D. A., L. F. Wright, D. O. Chaffin, C. E. Rubens, and R. P. Silver.
2000. NeuD plays a role in the synthesis of sialic acid in Escherichia coli K1.
FEMS Microbiol. Lett. 189:281–284.
13. Deszo, E. L., S. M. Steenbergen, D. I. Freedberg, and E. R. Vimr. 2005.
Escherichia coli K1 polysialic acid O-acetylation transferase gene, neuO, and
the mechanism of capsule form variation involving a mobile contingency
locus. Proc. Natl. Acad. Sci. USA 102:5564–5569.
14. Feng, L., S. N. Senchenkova, J. Tao, A. S. Shashkov, B. Liu, S. D. Shevelev,
P. R. Reeves, J. Xu, Y. A. Knirel, and L. Wang. 2005. Structural and genetic
characterization of enterohemorrhagic Escherichia coli O145 O antigen and
development of an O145 serogroup-specific PCR assay. J. Bacteriol. 187:
15. Gamian, A., E. Romanowska, J. Ulrich, and J. Defaye. 1992. The structure of
the sialic acid-containing Escherichia coli O104 O-specific polysaccharide
and its linkage to the core region in lipopolysaccharide. Carbohydr. Res.
16. Gonzalez, M. D., C. A. Lichtensteiger, and E. R. Vimr. 2001. Adaptation of
signature-tagged mutagenesis to Escherichia coli K1 and the infant-rat model
of invasive disease. FEMS Microbiol. Lett. 198:125–128.
17. Gonzalez, M. D., C. A. Lichtensteiger, R. Caughlan, and E. R. Vimr. 2002.
Conserved filamentous prophage in Escherichia coli O18:K1:H7 and Yersinia
pestisi biovar Orientalis. J. Bacteriol. 184:6050–6055.
18. Hara, S., M. Yamaguchi, Y. Takemori, K. Furuhata, H. Ogura, and M.
Nakamura. 1989. Determination of mono-O-acetylated N-acetylneuraminic
acids in human and rat sera by fluorometric high-performance liquid chro-
matography. Anal. Biochem. 179:162–166.
19. Higa, H. H., and A. Varki. 1988. Acetyl-coenzyme A: polysialic acid O-
acetyltransferase from K1-positive Escherichia coli. J. Biol. Chem. 279:42765–
20. Huang, S. H., Y.-H. Chen, Q. Fu, M. Stins, Y. Wang, C. Wass, and K. S. Kim.
1999. Identification and characterization of an Escherichia coli invasion gene
locus, ibeB, required for penetration of brain microvascular endothelial cells.
Infect. Immun. 67:2103–2109.
21. Jones, C., and X. Lemercinier. 2002. Use and validation of NMR assays for
the identity and O-acetyl content of capsular polysaccharides from Neisseria
meningitidis used in vaccine manufacture. J. Pharm. Biomed. Anal. 30:1233–
22. Kalivoda, K. A., S. M. Steenbergen, E. R. Vimr, and J. Plumbridge. 2003.
Regulation of sialic acid catabolism by the DNA binding protein NanR in
Escherichia coli. J. Bacteriol. 185:4806–4815.
23. Lewis, A. L., M. E. Hensler, A. Varki, and V. Nizet. 2006. The group B
streptococcal sialic acid O-acetyltransferase is encoded by neuD, a conserved
component of bacterial sialic acid biosynthetic gene clusters. J. Biol. Chem.
24. Lewis, A. L., V. Nizet, and A. Varki. 2004. Discovery and characterization of
sialic acid O-acetylation in group B Streptococcus. Proc. Natl. Acad. Sci. USA
25. Lin, S.-L., S. Inoue, and Y. Inoue. 2000. Acid-base properties of the reaction
product of sialic acid with fluorogenic reagent, 1,2-diamino-4,5-methy-
lenedioxybenzene (DMB). Carbohydr. Res. 329:447–451.
26. Liu, G., C. Jin, and C. Jin. 2004. CMP-N-acetylneuraminic acid synthetase
from Escherichia coli K1 is a bifunctional enzyme: identification of minimal
catalytic domain for synthetase activity and novel functional domain for
platelet-activating factor acetylhydrolase activity. J. Biol. Chem. 279:17738–
27. Meredith, T. C., and R. W. Woodard. 2006. Characterization of Escherichia
coli D-arabinose 5-phosphate isomerase encoded by kpsF: implications for
group 2 capsule biosynthesis. Biochem. J. 395:427–432.
28. Ørskov, F., I. Ørskov, A. Sutton, R. Schneerson, W. Lin, W. Egan, G. E. Hoff,
and J. B. Robbins. 1979. Form variation in Escherichia coli K1: determined
by O-acetylation of the capsular polysaccharide. J. Exp. Med. 149:669–685.
29. Ringenberg, M. C., C. A. Lichtensteiger, and E. R. Vimr. 2001. Redirection
of sialic acid metabolism in genetically engineered Escherichia coli. Glyco-
30. Ringenberg, M. A., S. M. Steenbergen, and E. R. Vimr. 2003. The first
committed step in the biosynthesis of sialic acid by Escherichia coli K1 does
not involve a phosphorylated N-acetylmannosamine intermediate. Mol. Mi-
31. Sarff, L. D., G. H. McCracken, M. S. Schiffer, M. P. Glode, I. Ørskov, and F.
Ørskov. 1975. Epidemiology of Escherichia coli K1 in healthy and diseased
newborns Lancet 1:1099–1104.
VOL. 188, 2006SIALIC ACID MODIFICATION6205
32. Schauer, R. 1978. Characterization of sialic acids. Methods Enzymol. 50: Download full-text
33. Scott, J. W., S. A. Hawley, K. A. Green, M. Anis, G. Stewart, G. A. Scullion,
D. G. Norman, and D. G. Hardie. 2004. CBS domains form energy-sensing
modules whose binding of adenosine ligands is disrupted by disease muta-
tions. J. Clin. Investig. 113:274–284.
34. Seveno, M., M. Bardor, T. Paccalet, V. Gomord, P. Lerouge, and L. Faye.
2004. Glycoprotein sialylation in plants? Nat. Biotechnol. 22:1470–1471.
35. Silhavy, T. J., M. L. Berman, and L. W. Enquist. 1984. Experiments with
gene fusions. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,
36. Silver, R. P., and E. R. Vimr. 1990. Polysialic acid capsule of Escherichia coli
K1, p. 39–60. In B. H. Iglewski and V. L. Clark (ed.), Molecular basis of
microbial pathogenesis, vol. IX. Academic Press, San Diego, Calif.
37. Steenbergen, S. M., C. A. Lichtensteiger, R. Caughlan, J. Garfinkle, T. E.
Fuller, and E. R. Vimr. 2005. Sialic acid metabolism and systemic pasteurel-
losis. Infect. Immun. 73:1284–1294.
38. Steenbergen, S. M., and E. R. Vimr. 1990. Mechanism of polysialic acid chain
elongation in Escherichia coli K1. Mol. Microbiol. 4:603–611.
39. Steenbergen, S. M., and E. R. Vimr. 2003. Functional relationships of the
sialytransferases involved in expression of the polysialic acid capsule of
Escherichia coli K1 and K92 and Neisseria meningitidis groups B or C. J. Biol.
40. Steenbergen, S. M., T. J. Wrona, and E. R. Vimr. 1992. Functional analysis
of the sialyltransferase complexes in Escherichia coli K1 and K92. J. Bacte-
41. Stein, D. M., J. Robbins, M. A. Miller, F.-Y. C. Lin, and R. Schneerson. 2006.
Are antibodies to the capsular polysaccharide of Neisseria meningitidis group
B and Escherichia coli K1 associated with immunopathology? Vaccine 24:
42. Stoughton, D. M., G. Zapata, R. Picone, and W. F. Vann. 1999. Identification
of Arg-12 in the active site of Escherichia coli K1 CMP-sialic acid synthetase.
Biochem. J. 343:397–402.
43. Varki, A., and S. Diaz. 1984. The release and purification of sialic acids from
glycoconjugates: methods to minimize the loss and migration of O-acetyl
groups. Anal. Biochem. 137:236–247.
44. Vimr, E., C. Lichtensteiger, and S. Steenbergen. 2000. Sialic acid metabo-
lism’s dual function in Haemophilus influenzae. Mol. Microbiol. 36:1113–
45. Vimr, E. R. 1992. Selective synthesis and labeling of the polysialic acid
capsule in Escherichia coli K1 strains with mutations in nanA and neuB. J.
46. Vimr, E. R., W. Aaronson, and R. P. Silver. 1989. Genetic analysis of chro-
mosomal mutations in the polysialic acid gene cluster of Escherichia coli K1.
J. Bacteriol. 172:1106–1117.
47. Vimr, E. R., K. A. Kalivoda, E. L. Deszo, and S. M. Steenbergen. 2004.
Diversity of microbial sialic acid metabolism. Microbiol. Mol. Biol. Rev.
48. Vimr, E. R., and S. M. Steenbergen. 2006. Mobile contingency locus controlling
Escherichia coli K1 polysialic acid capsule acetylation. Mol. Microbiol. 60:828–
49. Vimr, E. R., and S. M. Steenbergen. 2006. Targeting microbial sialic acid
metabolism for new drug development, p. 125–150. In C. A. Bewley (ed.),
Protein-carbohydrate interactions in infectious diseases: chemistry and biol-
ogy. Royal Society of Chemistry, Cambridge, United Kingdom.
50. Vimr, E. R., and F. A. Troy. 1985. Identification of an inducible catabolic
system for sialic acids (nan) in Escherichia coli. J. Bacteriol. 164:845–853.
51. Vimr, E. R., and F. A. Troy. 1985. Regulation of sialic acid metabolism in
Escherichia coli: role of N-acylneuraminate pyruvate-lyase. J. Bacteriol. 164:
52. Yu, H., W. Ryan, H. Yu, and X. Chen. 2006. Characterization of a bifunc-
tional cytidine 5?-monophosphate N-acetylneuraminic acid synthetase
cloned from Streptococcus agalactiae. Biotechnol. Lett. 28:107–113.
53. Zapata, G., P. P. Roller, J. Crowley, and W. F. Vann. 1993. The role of
cysteine residues 129 and 329 in Escherichia coli K1 CMP-NeuAc synthase.
Biochem. J. 295:485–491.
54. Zapata, G., W. F. Vann, W. Aaronson, M. S. Lewis, and M. Moos. 1989.
Sequence of the cloned Escherichia coli K1 CMP-N-acetylneuraminic acid
synthetase gene. J. Biol. Chem. 264:14769–14774.
6206 STEENBERGEN ET AL.J. BACTERIOL.