Glycobiology vol. 17 no. 12 pp. 1377–1387, 2007
Advance Access publication on September 20, 2007
The chitin catabolic cascade in the marine bacterium Vibrio cholerae: Characterization
of a unique chitin oligosaccharide deacetylase
Xibing Li2, Lai-Xi Wang3, Xuesong Wang2, and Saul
2Department of Biology, The Johns Hopkins University, Baltimore, MD
21218, USA; and3Institute of Human Virology, University of Maryland
Biotechnology Institute, Baltimore, MD 21201, USA
Received on June 25, 2007; revised on August 31, 2007; accepted on
September 1, 2007
Chitin, one of the most abundant organic substances in
nature, is consumed by marine bacteria, such as Vibrio
cholerae, via a multitude of tightly regulated genes (Li and
Roseman 2004, Proc Natl Acad Sci USA. 101:627–631). One
such gene, cod, is reported here. It encodes a chitin oligosac-
charide deacetylase (COD), when cells are induced by chito-
cloned (COD-6His), overproduced, and purified to appar-
ent homogeneity. COD is secreted at all stages of growth
by induced V. cholerae. The gene sequence predicts a 26 N-
terminal amino acid signal peptide not found in the isolated
protein. COD is very active with chitin oligosaccharides, is
virtually inactive with GlcNAc, and slightly active with col-
loidal ([3H]-N-acetyl)-chitin. The oligosaccharides are con-
verted almost quantitatively to products lacking one acetyl
group. The latter were characterized by mass spectrometry
(ESI-MS), and treatment with nitrous acid. COD catalyzes
the following reactions (n = 2–6): (GlcNAc)n→ GlcNAc-
GlcNH2-(GlcNAc)n−2+ Ac−. That is, COD hydrolyzes the
The gene bank sequence data show that cod is highly con-
served in Vibrios and Photobacteria. One such gene encodes
Biosci Biotech Biochem. 61:1113–1117; Ohishi et al. 2000,
J Biosci Bioeng. 90:561–563), that is specific for (GlcNAc)2,
but inactive with higher oligosaccharides. The COD enzy-
matic products, GlcNAc-GlcNH2-(GlcNAc)n, closely resem-
with Nod B: GlcNH2-(GlcNAc)3−4. The latter are key inter-
mediates in the biosynthesis of Nod factors, critically im-
portant in communications between the symbiotic nitrogen
fixing bacteria and plants. Conceivably, the COD products
remain to be defined.
Keywords: chitin oligosaccharide deacetylase/extracellular/
Chitin, a β,1-4 linked homopolymer of N-acetyl-D-glucosamine
(GlcNAc), is produced in astronomical quantities in the marine
1To whom correspondence should be addressed: Tel: +410-516-7333; Fax:
+410-516-5430; e-mail: firstname.lastname@example.org
C ?The Author 2007. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: email@example.com 1377
inundated by this highly insoluble polysaccharide, were it not
for chitinivorous bacteria that convert it to biologically useful
substances (Zobell and Rittenberg 1937; Poulicek and Jeuniaux
association with copepods is an important part of its life cycle.
It is well recognized that this association plays an important
role in V. cholerae infections in humans (Nalin 1976; Nalin
et al. 1979; Huq et al. 1983; Chakraborty et al. 1997).
We have reported that there are numerous enzymes and other
proteins in the chitin catabolic pathway and some are described
in the following references (Keyhani and Roseman 1999; Key-
hani, Li, and Roseman 2000; Park et al. 2000). Intermediates
include both fully and partially N-acetylated oligosaccharides
Asp two component sensor (Li and Roseman 2004; Meibom et
induced a disaccharide Phosphoenolpyruvate:glycose phospho-
transferase system (PTS) transport operon containing six genes
transcribed in the opposite direction. We investigated the latter,
and the gene product, and found it to be a chitin oligosaccha-
ride deacetylase (COD). The acronym COD is suggested for the
protein, and cod for the gene. COD specifically hydrolyzes only
the penultimate GlcNAc moiety at the nonreducing terminus of
Molecular cloning, expression, and purification of COD
The V. cholerae genomic sequence was used to design primers
for subcloning the cod gene. The resulting overexpression vec-
tor pET21d:VC1280 encodes eight amino acids more (LEHH-
HHHH) at the C-terminus than the wild type protein. The con-
struct was introduced into an expression Escherichia coli strain,
BL21(DE3)pLysS. We found the enzyme not only in the cells,
but also in the culture media. With prolonged induction time,
the quantity of enzyme in the culture medium increased and
accounted for the majority of expressed COD. For example,
the extracellular medium contained 83% of the total COD of an
fore, used as the enzyme source. Fractionation of the extracellu-
lar fluid with a nickel column followed by a Diethylaminoethyl-
Sepharose (DEAE) column, yielded apparently homogeneous
enzyme (Figure 2).
The purification procedure gave a 30% recovery and 7.5-fold
purification. The apparent molecular weight on sodium dode-
cyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) is
X Li et al.
Fig. 1. V. cholerae chitooligosaccharide utilizing gene cluster—The cod (VC1280) and neighboring genes are shown. The neighboring genes are transcribed in the
opposite direction. The acronyms cel and Cel are used in the annotations for the first four of these genes, VC1281 to VC1284. The annotations were based on the
concept that these were cryptic genes involved in the uptake and utilization of cellobiose. This concept is incorrect as explained in the Discussion. The corrected
annotations are shown in the lower two lines of the figure, namely that these genes are induced by chitobiose, and are involved in the utilization of
Fig. 2. SDS-PAGE of COD—Lane 1, protein markers. Lane 2, crude extracts
of uninduced cells. Lane 3, crude extracts of induced cells. Lane 4, the culture
medium of induced cells. Lane 5, COD eluted from Ni2+column. Lane 6,
COD after purification on DEAE column (SDS gel deliberately over-loaded
with protein). The identity of COD was confirmed by N-terminal sequence
(see Table I), and mass spectrometry (see Results).
Subcellular distribution of COD in V. cholerae VCXB21
The enzyme expressed in E. coli was found both in the cells
and the extracellular media. Therefore, it became important to
V. cholerae cells grown in 50% ASW-HEPES minimal medium
were induced with 1 mM (GlcNH2)2, and the cultures separated
into the following fractions: extracellular fluid (culture media),
periplasmic, inner membrane, outer membrane, and cytoplas-
mic. Each fraction was concentrated and the quantity of COD
present in each was determined by immunoelectrophoresis.
COD was detected only in the extracellular fluid (growth
medium), and accounted for about 10% of the total protein in
the extracellular space.
To confirm and extend this result, V. cholerae cells were
and cell fractions isolated as indicated above. Again COD was
found only in the extracellular medium, and only from induced
cells. It was not detectable at the early log stage, around 3795
molecules/cell at mid log stage, 3930 molecules/cell at late log
stage and 710 molecules/cell at stationary stage. These results
indicate that the enzyme is actively secreted and is not being
released into the medium by some other mechanism such as
cell lysis. The decrease in enzyme in the stationary phase prob-
ably reflects proteolysis. A decreased secretion was expected
as the inducer, (GlcNH2)2, was depleted from the medium,
but this could not account for the change from 3795 to 710
molecules/cell, since such a change would require a fivefold
increase in cell number, and this would have been detected.
Sequence analysis of COD
The cod gene, annotated as VC1280 from the Institute for
Genomic Research (TIGR) microbial V. cholerae genome
database, contains 431 amino acids before processing. A 26
amino acids signal peptide was predicted by program Signal
P 3.0 server (Technical University of Denmark) giving a 405
amino acid mature protein. This is exactly the same as de-
termined by N-terminal protein sequencing shown in Table I
for the recombinant protein. With our 6 His-tag protein, the
enzyme contains 413 amino acids and the calculated molec-
ular weight is 46,086 Da, in good agreement with the appar-
ent molecular weight on SDS-PAGE. This value was further
demonstrated by mass spectrometry of pure recombinant COD
Chitin oligosaccharide deacetylase (COD)
Table I. N-terminal amino acid sequences of chitooligosaccharide deacetylase
aDeduced from gene sequence.
enzyme from E. coli (46,025 Da). The results indicated that
the secreted enzyme was modified only by cleavage of the
signal peptide. The sequence of COD enzyme was also used
to search against NCBI Conserved Domain Database (CDD
v2.08, 2004) and two types of conserved domains were found,
an N-terminal polysaccharide deacetylase domain and a C-
terminal type 3 chitin binding domain (ChtBD3). By searching
the NCBI database, the protein shows a high degree of similar-
ity (full length) to other Vibrio proteins, such as V. alginolyti-
cus H-8 (79.9% identity/85.9% similarity), V. vulnificus YJ016
VV2902 (85% identity/91% similarity), V. vulnificus CMCP6
VV11481 (85% identity/91% similarity), V. parahaemolyticus
RIMD2210633 VP2638 (80% identity/88% similarity), Pho-
tobacterium profundum SS9 PBPRA0494 (65% identity/79%
similarity). Considering this degree of conservation, it was sur-
prising to find that COD showed only low similarity to the Nod
B gene of Rhizobium SP NGR234 (17% identity/24.3% simi-
larity), which encodes a deacetylase for chitin oligosaccharides
that hydrolyzes the acetamido group at the nonreducing end of
the chain, rather than the penultimate GlcNAc residue as does
Characterization of enzymatic products
When individual N-acetylchitooligosaccharides (dp 2–6) were
treated with the deacetylase, a single product was isolated from
each (Figure 3B).
The reactions catalyzed by COD are summarized in the
scheme shown in Figure 3A, that is, the enzyme cleaves one
acetyl group from each of the oligosaccharides. The data for
this conclusion were obtained by mass spectrometry (ESI-MS),
which revealed that the mass of each product was 42 daltons
less than the corresponding starting material (Figure 3B, panels
To determine which N-acetyl group was removed, i.e., where
the single free amino group was located in each product, the
mono-deacetylated oligosaccharides were treated with nitrous
of the O-glycosidic bond via deamination and rearrangement,
which releases the glycon portion with a 2,5-anhydro-D-
mannose unit at the reducing end, together with the correspond-
ing aglycon fragment (Sashiwa et al. 1993; Tommeraas et al.
2001). Mass spectroscopic analysis of the reaction mixtures
revealed that HNO2 treatment of the mono-deacetylated
chitooligosaccharides released the disaccharide derivative
4-O-(N-acetyl-glucosaminyl)-2,5-anhydro-D-mannose (GM) as
the sole anhydro-mannose containing unit from each oligosac-
charide, together with the corresponding aglycon fragment
(Figure 3B and C and Table II). For example, treatment of
the mono-deacetylated disaccharide with HNO2 gave GM
as the sole product (Figure 3B, panel F); treatment of the
trisaccharide derivative with HNO2gave the GM disaccharide
and GlcNAc (G) (Figure 3B, panel G) (GlcNAc not shown);
and treatment of the tetra-, penta-, and hexasaccharides with
HNO2 resulted in the formation of GM, together with the
N-acetylchitooligosaccharides (GlcNAc)2, (GlcNAc)3, and
(GlcNAc)4respectively (Figure 3B, panels H to J). The results,
summarized in Table II, suggest that, regardless of the length of
the N-acetylchitooligosaccharides, enzymatic de-N-acetylation
by COD occurs specifically at the second GlcNAc unit from
the nonreducing end of the chain.
Kinetic properties of COD
Some kinetic properties of COD are illustrated in Figure 4.
The optimum pH for the enzyme was between pH 7 and 7.5,
depending on the buffer (Figure 4A). It exhibits more than 90%
environment. The optimum temperature in terms of activity was
45◦C, and it was 75% as active at 37◦C (Figure 4B). COD is
thermostable up to 50◦C while it loses almost all activity when
that COD was maximally active at low ionic strength, and the
activity gradually decreased as the ionic strength was increased;
it was 57% as active as the control at 0.5 M NaCl or 0.5 M KCl.
There was little effect on activity in the presence of 1 mM DTT
and 5 mM EDTA. On the other hand, the activity was sensitive
to metal ions, e.g., inhibition by 2 mM Ag+, Hg2+, Al3+, Co2+,
Cu2+, and Ni2+, but this inhibitory effect may result from the
6 His-tag which caused visible precipitation of COD at higher
concentrations of the protein.
COD exhibited no detectable activity with the monosaccha-
ride, GlcNAc, under the assay conditions described above, and
a barely detectable activity when GlcNAc was incubated for
prolonged periods with large amounts of the enzyme. A slight
but it was too low to obtain reliable kinetic constants. The COD
activity with the chitin oligosaccharides, (GlcNAc)n, n = 2–6,
is shown in Table III and Figure 4C –G. The isotopic assay was
used for these experiments, and CurveExpert 1.3 was used to
creases with increasing chain length. These values are reflected
catalytic efficiency of the enzyme for these substrates, kcat/KM.
The disaccharide, (GlcNAc)2, is the most active substrate, and
disaccharide is threefold more active than the trisaccharide, and
44-fold more active than the hexasaccharide.
We have reported that the chitin degradation system in V.
cholerae is very complex, involving many genes and enzymes
(reviewed in (Keyhani and Roseman 1999; Li and Roseman
2004)). The known pathway includes the following steps: (i)
Binding of the cells to chitin by what we have designated (Yu
et al. 1987, 1991; Li and Roseman 2004) a nutrient sensor.
(ii) Secretion of chitinases. (iii) Partial hydrolysis of the chitin
to chitooligosaccharides (see reviews (Keyhani and Roseman
1999)). (iv) Diffusion of the oligosaccharides through the outer
X Li et al.
Fig. 3. Evidence for reactions catalyzed by COD— (A). Schematic for general reaction catalyzed by COD. One N-acetamido group is hydrolyzed in each of the
Chitin oligosaccharide deacetylase (COD)
Fig. 3. (B). A–E, the ESI-MS profiles of the products obtained from the enzymatic deacetylation of the di-, tri-, tetra-, penta-, and hexasaccharides respectively. In
these spectra, the major peaks, M, signify the masses of each of the enzymatic products. To determine the location of the free amino groups in the products of the
COD reactions, each oligosaccharide was treated with nitrous acid as described in the text. This reaction results in conversion of each glucosamine residue with the
free amino group to 2,5-anhydro-D-mannose with simultaneous chain cleavage. Panels F–J show that each oligosaccharide yielded the same product, designated
GM. This product exhibits the mass of GlcNAc-2,5-anhydro-mannose. Some oligosaccharides gave an additional fragment in the nitrous acid reaction as follows:
Panel F, disaccharide gave only GM; Panel G, trisaccharide yielded GM + GlcNAc (not shown); Panel H, tetrasaccharide yielded GM + G2; Panel I,
pentasaccharide gave GM + G3; Panel J, hexasaccharide produced GM + G4. G2, G3, and G4 exhibit the masses corresponding to (GlcNAc)2, (GlcNAc)3, and
(GlcNAc)4respectively. (C). Schematic view of structures of each COD product and how these compounds reacted with nitrous acid. The abbreviation GlcN
denotes glucosamine (or GlcNH2).
specific porin (Keyhani, Li, Roseman 2000). (v) Hydrolysis of
the oligosaccharides to GlcNAc and (GlcNAc)2by a combi-
nation of two specific N-acetyl-glucosaminidases located in the
NAc linkage at the nonreducing termini in the tri- and higher
oligosaccharides, but is virtually inactive with (GlcNAc)2at the
pH of sea water or the growth medium (Keyhani and Roseman
Table II. Analysis of enzyme products
Substrate Enzyme productTheoretical mass (Da) Found ESI-MS (Da) m/z (M + H)+
GlcNAc-GlcNH2: ESI-MS: Calculated for C14H26N2O10, M (exact mass) = 382.16; found, 383.22 (M + H)+.
GlcNAc-GlcNH2-(GlcNAc): ESI-MS: Calculated for C22H39N3O15, M (exact mass) = 585.24; found, 586.32 (M + H)+.
GlcNAc-GlcNH2-(GlcNAc)2: ESI-MS: Calculated for C30H52N4O20, M (exact mass) = 788.32; found (m/z), 789.52 (M + H)+.
GlcNAc-GlcNH2-(GlcNAc)3: ESI-MS: Calculated for C38H65N5O25, M (exact mass) = 991.40; found (m/z), 992.69 (M + H)+.
GlcNAc-GlcNH2-(GlcNAc)4: ESI-MS: Calculated for C46H78N6O30, M (exact mass) = 1194.48; found (m/z), 1195.88 (M +H)+.
X Li et al.
Table III. Kinetic constants of CODa
aThe data in Figure 4 were used to determine the best fit kinetic constants by
the software program CurveExpert 1.3.
bThere was no detectable activity with GlcNAc under these assay conditions.
The activity with [3H]-acetyl labeled chitin was too low to accurately
determine Kmand Vmaxvalues.
1996b). (vi) (GlcNAc)2is a particularly important product be-
cause it is the key signal to a two component signaling system
that regulates expression of a host of genes (Li and Roseman
2004; Meibom et al. 2004). (vii) The mono- and disaccharides
are then translocated into the cytoplasm by specific transporters
catabolized (Roseman 1957; Comb and Roseman 1958; Bassler
et al. 1991; Yu et al. 1993; Park et al. 2000, 2002) via several
enzymes and converted to fructose-6-P, NH4+, and acetate.
Fig. 4. Some kinetic properties of COD—Unless otherwise indicated, all
assays were performed at 37◦C, and where product formation was proportional
to quantity of COD and time of incubation. (A) Effects of pH on COD activity.
Assayed by the colorimetric method (see text) with the following buffers: (O),
sodium citrate; (†), PIPES; ( ), potassium phosphate; (?), McIlwaine broad
range; (?) imidazole HCl; (♦), Hepes; (♦), sodium borate; (?),
glycine–NaOH. (B) Effect of temperature during assay on COD activity
(colorimetric assay). (C)–(G) Effect of substrate concentration on activity of
COD. Substrates were assayed by the isotopic method at the indicated
concentrations, and where less than 5% of the substrate was utilized.
Fig. 5. Relationship between COD and Nod B—As explained in the text (see
Discussion), Nod factors are signaling molecules essential for the development
of the symbiotic relationships between nitrogen fixing bacteria, such as
Rhizobia, and their hosts, such as soybean plants. The Nod factors determine
the specificity of this interaction, and Nod B is a key enzyme in this process.
The figure shows that COD and Nod B catalyze equivalent reactions, removing
one acetyl group from the chitooligosaccharides. The enzymes differ in that
Nod B most generally acts on the tetra- and pentasaccharides, and sometimes
on the trisaccharide.
An alternate pathway for chitin catabolism has recently been
described that depends on the action of two enzymes molec-
ularly cloned from the hyperthermophilic archaeon Thermo-
coccus kodakaraensis KOD1. (i) A deacetylase that hydrolyzes
chitin oligosaccharide N-acetyl groups located at the nonre-
ducing termini. These products are identical (except for chain
length) to those of Nod B (see Figure 5) (ii) An exo β-
(GlcNAc)noligosaccharides products of the deacetylase. The
authors propose that the sequential action of the two enzymes
gives the final products, GlcNH2and GlcNAc (Tanaka et al.
Here, we describe a COD that specifically removes the N-
acetyl group from the penultimate GlcNAc residue of the chi-
spite this similarity (79.9% identity), the two enzymes exhibit
significantly different substrate specificities. The V. alginolyti-
cus enzyme is specific for N,N?-diacetylchitobiose giving the
disaccharide (GlcNAc-GlcNH2), while the V. cholerae enzyme
hydrolyzes a broader spectrum of chitooligosaccharides, yield-
ing the products (GlcNAc-GlcNH2-(GlcNAc)n = 0−4), although
(GlcNAc)2is kinetically the most active substrate.
The chitin catabolic enzymes are highly conserved in Vibrio
species. Thus, it is likely that the following genes encode COD
enzymes: VV2902 from V. vulnificus CMCP6, VV11481 from
V. vulnificus YJ016, VP2638 from V. parahaemolyticus RIMD
2210633, and PBPRA0494 from P. profundum SS9.
Asshown inFigure1,thecod gene isadjacent toaPTS trans-
porter operon. This operon is similar to the E. coli (GlcNAc)2
PTS transporter system, but transcription of cod and the PTS
transporter genes are in opposite directions. In Figure 1, the
annotations are those currently in the V. cholerae gene bank,
that is, the older annotations for the disaccharide transporter
proteins, CelA, CelB, CelC, and CelF. This nomenclature was
used when it was believed to be a cryptic, cellobiose specific
transporter that is expressed when the cells mutated. In fact,
it has been clearly shown that these genes are neither cryptic,
nor normally function in the catabolism of cellobiose. They are
involved in the catabolism of (GlcNAc)2(Keyhani and Rose-
man 1997; Keyhani, Rodgers, Demeler, Hansen, and Roseman
Chitin oligosaccharide deacetylase (COD)
2000; Keyhani, Boudker, Roseman 2000; Keyhani, Wang, Lee,
Roseman 2000; Keyhani, Bacia, Roseman 2000). The corrected
annotations and abbreviations for the genes are shown on the
bottom two lines of Figure 1. The microbial genome sequence
database (www.tigr.org) shows that the same gene cluster and
Vibrios and Photobacteria.
From microarray data, we know that VC1280 to VC1286
can be strongly induced by chitobiose, (GlcNH2)2, and by crab
shells, but not by (GlcNAc)2(Meibom et al. 2004). Considering
that COD was coinduced with the PTS transporter system by
chitobiose, it seems reasonable to suggest that the products of
COD, the oligosaccharides GlcNAc-GlcNH2-(GlcNAc)n = 0−4,
are the true substrates for this PTS transporter system, although
this hypothesis remains to be tested.
For the COD products to reach the cytoplasm intact requires
that they are not hydrolyzed during the steps outlined above,
but most especially by a very active outer membrane β-N-
acetylhexosaminidase (Jannatipour et al. 1987; Soto-Gil and
Zyskind 1989), and furthermore, they must be recognized by
the membrane transporter encoded by the genes in Figure 1.
These possibilities can be tested experimentally with mutants
constructed for these purposes.
Our final speculation may be the most interesting and poten-
tially important. The COD products closely resemble essential
intermediates in the biosynthesis of Nod factors (Figure 5).
These factors play a critical role in the development of symbi-
otic relationships between nitrogen fixing bacteria and plants.
To show their importance in plant/microbe relationships, it may
to 150,000 according to Google). The factors are derivatives of
chitin tri-, tetra-, and pentasaccharides. These oligosaccharides
are partially hydrolyzed by Nod B as shown in Figure 5. The
free amino groups are first coupled to fatty acids to provide
membrane anchors, and the precursors are then derivatized (fu-
cose, mannose, arabinose, acetate, sulfate, etc.) to form the Nod
factors that confer specificity on the symbionts.
Thus, there is a great deal of similarity between Nod B and
COD. Both act on chitin oligosaccharides, and both remove one
acetyl group. They differ primarily in the location of the acetyl
Nod B and COD (17% identity, 24% similarity). Nevertheless,
the structural similarities between the products generated by the
two enzymes, and the great importance of the Nod factors as
signaling molecules for such a complex process makes it rea-
by COD may be precursors of important signaling events yet to
Materials and methods
The following chemicals, reagents, and materials were
purchased from the indicated sources. Chitin, and N-
Chitin oligosaccharides (GlcNAc)n(n = 2–6) and the corre-
sponding chitosan oligomers (GlcNH2)n(n = 2–6) were from
Seikagaku America, Inc. (Rockville, MD) Reagents for bac-
terial growth media were from BD Biosciences (Palo Alto,
CA) and J.T. Baker (Phillipsburg, NJ). Reagents for molecu-
lar biology were obtained from New England Biolabs (Ipswich,
MA), Promega (Madison, WI), Stratagene (La Jolla, CA), and
Invitrogen (Carlsbad, CA). [3H]-N-Acetyl labeled chitooligo-
(Roseman and Ludowieg 1954; Roseman and Daffner 1956;
anhydride (TRK2-25mCi, 150–370 GBq/mmol) obtained from
Amersham Pharmacia Biotech (Piscataway, NJ). Electrophore-
sis gels were from Cambrex BioScience Rockland, Inc (Rock-
land, ME). Other buffers and reagents were of the highest purity
The N-terminal sequence of the purified enzyme was deter-
mined at the Biosynthesis and Sequencing Facility (Department
of Biological Chemistry, Johns Hopkins School of Medicine)
using an Applied Biosystems (Foster, CA) 475A protein se-
Two methods were used to determine protein concentrations,
as a standard, and the total nitrogen concentration by a reported
method (Jaenicke 1974) for the enzyme purified to apparent
All strains of V. cholerae were derived from V. cholerae
EI Tor N16961, the organism used for sequencing the V.
cholerae genome (Heidelberg et al. 2000). A mutant des-
ignated VCXB21 was used to determine the subcellular lo-
cation of the enzyme, and was constructed from the parent
strain as follows: (i) The lacZ gene (VC2338) was replaced
with a kanamycin cartridge from plasmid pNK2859 (Kleckner
et al. 1991). In this procedure, the primers used were: GalR-F,
5?-GGATCCTGGAACTGCTCATCAA CA-3?and GalR-R, 5?-
CCCGGGAGATCTTAAGGCTCTCTTT-3?; DR-F, 5?-AGATC
TCCCGGGT CGATATTGACCCAA-3?, and DR-R, 5?-AGA-
By bridge PCR, the deletion construct will delete a fragment
from 160 bp before and 16 bp after lacZ gene. The PCR prod-
uct was subcloned in the pGEM-T vector. A BamHI fragment
of kanamycin resistance gene from pNK2859 (originally from
Tn903) was inserted at BglII site in the middle of the construct.
Then the deletion construct was transferred into SmaI site of the
out. (ii) The two methods for selecting knock out mutants and
allelic exchange by the pMAKSACA plasmid are temperature
sensitive replication of the plasmid, and expression of the sacB
gene product from the suicide vector (Favre and Viret 2000)
which makes the cells sensitive to sucrose. Since V. cholerae
ferments sucrose, this would interfere with the sensitivity of the
selection process. Therefore, the SucIIBC (ScrA, VCA0653)
gene was partly deleted, so that the mutant could not ferment
The forward primer, 5?-GATAGGATCCGACTTGGAGTA
sucIIBC gene. The PCR fragment was subcloned in pGEM-T
vector, the NruI fragment removed and the resulting construct
religated. This treatment resulted in the deletion of 664 bp in
the middle of the sucIIBC gene (from 308 bp to 971 bp). The
deletion construct was then transferred to the BamHI site of
X Li et al.
conditional suicide vector- pMAKSACB. The resulting mutant
was selected by its inability to ferment sucrose.
V. cholerae strains were grown in either Luria Broth (LB) for
maintenance or in 50% artificial sea water (50%ASW-HEPES
buffered, pH 7.5) supplemented with 0.005% K2HPO4, 0.1%
NH4Cl, 0.5% D,L-lactate as carbon source, and the indicated
designated plasmid constructs were grown overnight with vig-
Two techniques were employed to measure enzyme activ-
ity. The first was a colorimetric method using 3-methyl-2-
benzothiazolinone hydrazone hydrochloride reagent (MBTH;
Sigma Chemical Co.) for quantization of amino groups in
the partially deacetylated oligosaccharides. The second method
consisted of using [3H]-CH3CO-labeled (GlcNAc)n as sub-
strates, and determining the quantity of [3H]- CH3COOH re-
leased by the enzyme.
product was treated with nitrous acid, which specifically deam-
inates the hexosamine residue, but not the N-acetyl blocked
hexosamine residues, yielding the corresponding anhydro sugar
(discussed below). The latter is very sensitive to treatment with
various sugar chromogenic reagents, such as anthrone sulfu-
ric acid (Horowitz et al. 1957), or the MBTH reagent (Tsuji
et al. 1969a, 1969b). In the latter assay, the colorimetric stan-
coefficients at 650 nm for each of the enzymatic products, i.e.,
gave the relative, not absolute concentration of each product. It
was a useful method for screening for enzymatic activity, and
for determining some of the kinetic values, such as the optimum
For routine assays, each incubation mixture consisted of the
following: a 50 µL reaction volume containing 2 mM N,N?-
diacetyl-chitobiose, 20 mM HEPES buffer, pH 7.0, and 200 ng
COD enzyme. After incubating at 37◦C for 15 min, reactions
were terminated by boiling for 4 min and the products assayed
by the MBTH method exactly as described (Tsuji et al. 1969a).
Product formation was proportional to the quantity of COD
in the incubation mixture, and to the time of incubation in all
Method 2: [3H]-Labeled Oligosaccharides Reactions were per-
formed in 200 µL volumes using 20 mM HEPES buffer, pH
7.0, 200 ng COD, 0.1 mg/mL bovine serum albumin (BSA)
and [3H]-(GlcNAc)n = 2−6(1–1000 µM) at 37◦C for 15 min.
Reactions were stopped by heating for 4 min at 100◦C. The
incubation mixtures (200 µL) were applied to 0.5 mL Dowex
AG1×8 resin columns (fluoride form, 50–100 mesh, Bio-Rad
(Hercules, CA)). Each column was washed with 5 mL water
to remove unreacted substrate, and the oligosaccharide prod-
uct; 0.5 mL aliquots were mixed with 3 mL Ultima-Gold XR
(Packard Instrument Co., Meriden, CT) solution for counting.
The desired product, labeled acetic acid, was eluted from the
ion-exchange resin columns with 3 mL of 0.7 M KCl. Aliquots
of the eluates (2 mL) were mixed with 3 mL Hionic Fluor so-
lution (Packard Instrument Co.). All samples were counted in
a Packard Liquid Scintillation Spectrometer. The quantity of
labeled substances in the water and KCl eluates was calculated
dpm/pmol. It should be noted that counting both the water and
KCl eluates provided a good check of the quantization, since
the sum should be equal to the total quantity of substrate in the
This method was used for determining the kinetic constants
given in Table III.
Construction of cod overexpression vector
The gene cod (VC1280) was molecularly cloned from V.
cholerae EI Tor N16961 genomic DNA by PCR with primers
(VC1280F- 5?-AACCATGGACAGTACCCCTAA GGGCA-3?;
3?). The PCR product was subcloned into the pGEM-T vec-
tor (Promega) and designated as pGEM-VC1280. The 1.3 kb
corresponding site of pET21d(+) (Novagen) giving the overex-
pression construct pET21d:VC1280, which now had a His tag
at the C-terminus of the cod gene.
Overexpression and purification of COD
E. coli strain BL21 (DE3) pLysS harboring pET21d:VC1280
was grown in LB medium overnight with 75 µg/mL ampicillin.
Fresh medium was inoculated with cells from the overnight
culture at a 1:50 dilution and the 2000 mL culture grown to
centration and the incubation continued at 30oC for 4–12 h. The
culture was first centrifuged at 3000 × g and the supernatant
was filtered through a 0.22 µm membrane (Nalgene) to re-
move any remaining cells. Macromolecules in the filtrate were
concentrated by ultrafiltration using a Centricon Plus-80 (Mil-
lipore) membrane with a 10,000 Da cut off. After dialyzing
against Buffer A (20 mM potassium phosphate buffer, pH 8.0,
0.2 M NaCl, 0.02% NaN3), the concentrated enzyme solution
(170 mL) was applied to a 20 mL Ni2+charged affinity column
(Sigma). The column was washed with five volumes of Buffer
A followed by three volumes of Buffer A containing 20 mM
imidazole. The enzyme was then eluted with an imidazole gra-
dient (20 mM–120 mM) in 400 mL Buffer A. Active fractions
with different degrees of purity (38 mL, 92 mL, and 113 mL)
phosphate buffer, pH 8.0). COD was further purified either by
chromatography on a DEAE-Sepharose 4B column (50 mL),
or by repeating the Ni2+charged affinity column (20 mL). For
the DEAE column, the enzyme was eluted with 0–0.3 M NaCl
gradients in buffer B. Activity appeared at about 80 mM NaCl.
The purified enzyme fractions were collected, concentrated,
dialyzed, and stored as 50 µL aliquots containing 2.53 mg/mL
enzyme solution in buffer B at −80◦C. The enzyme was stable
for months under these conditions.
Antibody preparation and immunophoresis assay
COD preparations that appeared to be homogeneous on SDS-
PAGE were used to generate rabbit antibodies by Covance Re-
search Products Inc (Denver, PA). A highly sensitive, and pre-
Chitin oligosaccharide deacetylase (COD)
to quantify COD by comparing rocket heights and areas with
varying amounts of standard purified enzyme.
Kinetics properties of COD
Unless otherwise specified, the standard assay conditions de-
scribed above were used with COD preparations that appeared
homogenous by SDS-PAGE. The colorimetric and/or the iso-
topic methods were used for the following determinations. As-
says were conducted under conditions where product formation
The following parameters were studied:
Effects of pH and Buffer Type The following buffers were tested
at 20 mM concentration: McIlvaine’s sodium phosphate– citric
acid broad-range buffer from pH 2.6–7.6, sodium citrate buffer
3.0–6.0, potassium phosphate buffer 6.0–8.0, imidazole–HCl
buffer 6.2–7.8, PIPES buffer 6.1–7.5, HEPES buffer 7.0–8.0,
glycine–NaOH buffer 8.6–10.6, borate buffer 8.1–10.5.
Effects of Ionic Strength, Temperature, and Metal Ions The ef-
fect of ionic strength on purified COD enzyme activity was
determined with 0 to 1 M NaCl or KCl in the reaction mix-
tures. The optimum temperature for the assay was determined
by incubating replicate reaction mixtures over the range 4◦C to
80◦C for 15 min. Thermal stability of COD was measured by
incubating the enzyme in the assay buffer without substrate at
the indicated temperatures for 30 min, stored on ice for 5 min,
warmed to 37◦C, and residual enzyme activity determined by
adding 2 mM N,N?-diacetylchitobiose to the reaction mixture.
The effects of metal ions, generally as the chloride salts,were
at 5 mM concentrations, DTT at 1 mM, and KAc from 0 to 0.2
M were tested as possible inhibitors.
Preparation of enzymatic reaction products
Chitin oligosaccharides (GlcNAc)n(n = 1–6), 5 mg each, were
incubated at 37◦C for 4 h with 62 µg pure COD enzyme in 20
mM pyridine–acetic acid buffer, pH 7.0, total volumes of 6 mL.
Chromatography of the reaction mixtures by TLC showed a vir-
tual complete conversion of substrates to products. (The devel-
oping solvent was 1-butanol:methanol:NH4OH:H2O (5:4:2:1)
v/v, which separates the fully acetylated and mono-deacetylated
chitooligosaccharides). The reaction mixtures were passed
through a Centricon filtration device (Millipore) with a 10,000
Da cut off to remove the enzyme. The filtrates were lyophilized
and the resulting sugars were used for mass spectra analyses
and chemical modification assays.
Subcellular localization of COD
Subcellular fractions of V. cholerae were obtained as described
(Miyazato et al.2003). Briefly,V. cholerae VCXB21 was grown
in 50 mL minimal medium (50% artificial sea water-ASW, 50
mM HEPES pH 7.5, 0.1% NH4Cl, 0.005% K2HPO4, 0.5% D,L-
lactate, and 50 µg/mL kanamycin) with or without 1 mM chi-
tobiose, (GlcNH2)2, as inducer at 37◦C. Mid-exponential phase
cells monitored by their absorbance at 540 nm were used for
fractionation. The cells were harvested at 3000 × g for 30 min,
and the supernatants collected (designated Extracellular Frac-
All remaining steps were conducted between 0–4◦C unless
otherwise indicated. The harvested cells were suspended in 8
mL 1 M sucrose in Buffer C (30 mM Tris-HCl, pH 8.0), to
which 80 µL 0.5 M EDTA and 80 µL of a 20 mg/mL lysozyme
solution were added. The suspensions were incubated on ice
for 40 min, after which MgCl2was added to a final concen-
tration of 75 mM. After centrifugation at 15,000 × g for 30
min, the resulting supernatants were collected and designated
the Periplasmic Fraction. The cell pellets were resuspended in
8 mL of 1 M sucrose in buffer C containing 75 mM MgCl2. The
suspensions were sonicated four times (1 min each) and then
subjected to two rounds of freezing and thawing. The unbroken
the supernatants centrifuged again at 113,000 × g for 1 h. The
high speed supernatants were termed the Cytoplasmic Fraction.
The pellets were suspended in 1 mL 10 mM PBS buffer (137
mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4, 1.4 mM KH2PO4)
to which 100 µL of a 10% Sarkosyl solution was added. The
suspensions were passed through a 22 gauge hypodermic sy-
ringe needle 5–10 times and were centrifuged for 15 min in a
desktop centrifuge at maximum speed (12000 × g). The Sarko-
syl soluble fractions were termed the Inner Membrane (IM)
Fraction. The pellets were resuspended in 500 µL 0.5% SDS in
PBS and designated the Outer Membrane Fraction (OM). All
fractions were concentrated to 0.5 mL by ultrafiltration, using
a 10,000 Da cut off membrane, and the COD content in each
was measured by rocket immunophoresis, with isolated COD
Protein samples were heated at 80◦C for 5 min in Laemmli
buffer (65 mM Tris/HCl, pH 6.8, 0.3% SDS, 10% glycerol, 5%
by electrophoresis in a 4–20% gradient gel. Gels were stained
with Coomassie Blue G-250.
The mass measurements of pure recombinant COD en-
zyme from E. coli was performed using a Voyager DE-
STR MALDI-TOF in the Mass Spectrometry/Proteomics Fa-
cility at Johns Hopkins University, School of Medicine
Center for Research Resources shared instrumentation grant 1S
10-RR 14702, the Johns Hopkins Fund for Medical Discovery
and the Johns Hopkins Institute for Cell Engineering.
Electron spray ionization mass spectrometry (ESI-MS) analysis
ple mass spectrometer at the University of Maryland Medical
Nitrous acid treatment of mono-deacetylated
The mono-deacetylated chitooligosaccharides (3 mg) were dis-
solved in 1 M acetic acid (300 µL) containing 1 mg sodium
nitrite. Aqueous HCl (0.1 M, 30 µL) was added to the solution,
maintained at 4◦C for 10 h, and passed through a Dowex 50W-
X8 column (H+form). The column was washed with an equal
volume (300 µL) of distilled water, and the eluants combined
and subjected to ESI-MS analysis.
X Li et al.
National Institutes of Health (NIH) (GM51215).
able a grant from New England Biolabs and Ms. Haijing Song
for technical support.
Conflict of interest statement
COD, Chitin oligosaccharide deacetylase; DEAE, Diethy-
2,5-anhydro-D-mannose; ESI-MS, Electron spray ioniza-
tion mass spectrometry; GlcNAc, N-acetyl-D-glucosamine;
IPTG, isopropyl β-D-thiogalactopyranoside; LB, Luria Broth;
MBTH, 3-methyl-2-benzothiazolinone hydrazone hydrochlo-
ride reagent; PTS, Phosphoenolpyruvate:glycose phospho-
transferase system; SDS-PAGE, sodium dodecyl sulfate-
polyacrylamide gel electrophoresis.
Bassler BL, Yu C, Lee YC, Roseman S. 1991. Chitin utilization by marine
bacteria: Degradation and catabolism of chitin oligosaccharides by Vibrio
furnissii. J Biol Chem. 266:24276–24286.
Bouma CL, Roseman S. 1996. Sugar transport by the marine chitinolytic bac-
terium Vibrio furnissii: Molecular cloning and analysis of the glucose and
N-acetylgucosamine permeases. J Biol Chem. 271:33457–33467.
Chakraborty S, Nair GB, Shinoda S. 1997. Pathogenic vibrios in the natural
aquatic environment. Rev Environ Health. 12:63–80.
Comb DG, Roseman S. 1958. Glucosamine metabolism. IV. Glucosamine-6-
phosphate deaminase. J Biol Chem. 232:807–827.
Favre D, Viret J-F. 2000. Gene replacement in gram-negative bacteria: the
pMAKSAC vectors. BioTechniques. 28:199–204.
Heidelberg JF, Eisen JA, Nelson WC, Clayton RA, Gwinn ML, Dodson RJ,
Haft DH, Hickey EK, Peterson JD, Umayam L, et al. 2000. DNA sequence
of both chromosomes of the cholera pathogen Vibrio cholerae. Nature.
oligosaccharides. I. Separation. J Am Chem Soc. 79:5046–5049.
Huq A, Small EB, West PA, Huq MI, Rahman R, Colwell RR. 1983. Ecological
relationships between Vibrio choleare and planktonic crustacean copepods.
Appl Environ Microbiol. 45:275–283.
Jaenicke L. 1974. A rapid micromethod for the determination of nitrogen and
phosphate in biological material. Anal Biochem. 61:623–627.
Jannatipour M, Soto-Gil RW, Childers LC, Zyskind JW. 1987. Translocation of
coli. J Bacteriol. 169:3785–3791.
diacetylchitobiose in Escherichia coli: Characterization of phospho-IIBChb
and of a potential transition state analogue in the phosphotransfer reactions
between the proteins IIAChband IIBChb. J Biol Chem. 275:33102–33109.
Keyhani NO, Boudker O, Roseman S. 2000. Isolation and characterization
of IIAChb, a soluble protein of the enzyme II complex required for the
transport/phosphorylation of N,N?-diacetylchitobiose in Escherichia coli. J
Biol Chem. 275:33091–33101.
Vibrio furnissii: Identification and molecular cloning of a chitoporin. J Biol
Keyhani NO, Rodgers ME, Demeler B, Hansen JC, Roseman S. 2000. Ana-
lytical sedimentation of the IIAChband IIBChbproteins of the Escherichia
coli N,N?-diacetylchitobiose phosphotransferase system: Demonstration of
Keyhani NO, Roseman S. 1996a. The chitin catabolic cascade in the ma-
rine bacterium Vibrio furnissii: Molecular cloning, isolation and char-
acterization of a periplasmic chitodextrinase. J Biol Chem. 271:33414–
Keyhani NO, Roseman S. 1996b. The chitin catabolic cascade in the marine
bacterium Vibrio furnissii: Molecular cloning, isolation and characteriza-
tion of a periplasmic b-N-acetylglucosaminidase. J Biol Chem. 271:33425–
Keyhani NO, Roseman S. 1997. Wild-type Escherichia coli grows on the chitin
disaccharide, N,N?-diacetylchitobiose, by expressing the cel operon. Proc
Natl Acad Sci USA. 94:14367–14371.
Keyhani NO, Roseman S. 1999. Physiological aspects of chitin catabolism in
marine bacteria. Biochim Biophys Acta. 1473:108–122.
Keyhani NO, Wang L-X, Lee YC, Roseman S. 1996. The chitin catabolic
cascade in the marine bacterium Vibrio furnissii: Characterization of
an N,N?-diacetyl-chitobiose transport system. J Biol Chem. 271:33409–
Keyhani NO, Wang L-X, Lee YC, Roseman S. 2000. The chitin disaccharide,
N,N?-diacetylchitobiose, is catabolized by Escherichia coli, and is trans-
ported/phosphorylated by the phosphoenolpyruvate:Glycose phosphotrans-
ferase system. J Biol Chem. 275:33084–33090.
Kleckner N, Bender J, Gottesman S. 1991. Uses of transposons with emphasis
on Tn10. Methods Enzymol. 204:139–180.
oligosaccharides and a novel two component chitin catabolic sensor/kinase.
Proc Natl Acad Sci USA. 101:627–631.
Meibom KL, Li X, Nielsen AT, Wu C-Y, Roseman S, Schoolnik GK. 2004.
The Vibrio cholerae chitin utilization program. Proc Natl Acad Sci USA.
Miyazato T, Toma C, Nakasone N, Yamamoto K, Iwanaga M. 2003. Molec-
ular analysis of VcfQ protein involved in Vibrio cholerae type IV pilus
biogenesis. J Med Microbiol. 52:283–288.
Nalin DR. 1976. Cholera, copepods, and chitinase. Lancet. 2:958.
Nalin DR, Daya V, Reid A, Levine MM, Cisneros L. 1979. Adsorp-
tion and growth of Vibrio cholerae on chitin. Infect Immun. 25:768–
Ohishi K, Murase K, Ohta T, Etoh H. 2000. Cloning and sequencing of the
deacetylase gene from Vibrio alginolyticus H-8. J Biosci Bioeng. 90:561–
Ohishi K, Yamagishi M, Ohta T, Motosugi M, Izumida H, Sano H, Adachi
K, Miwa T. 1997. Purification and properties of two deacetylases pro-
duced by Vibrio alginolyticus H-8. Biosci Biotech Biochem. 61:1113–
Park JK, Keyhani NO, Roseman S. 2000. Chitin catabolism in the marine bac-
terium Vibrio furnissii: Identification, molecular cloning, and characteriza-
tion of a N,N?-diacetylchitobiose phosphorylase. J Biol Chem. 275:33077–
Park JK, Wang L-X, Roseman S. 2002. Isolation of a glucosamine-specific
kinase, a unique enzyme of Vibrio cholerae. J Biol Chem. 277:15573–
Poulicek M, Jeuniaux C. 1982. Biomass and biodegradation of mollusk shell
chitin in some marine sediments. In: Hirano S, Tokura S, editors. Japanese
Society of Chitin and Chitosan, Tottori, Japan p. 192–196.
Roseman S. 1957. Glucosamine metabolism I. N-acetylglucosamine deacety-
lase. J Biol Chem. 226:115–124.
Roseman S, Daffner I. 1956. Colorimetric method for the determination of
glucosamine and galactosamine. Anal Chem. 28:1743–1746.
Roseman S, Ludowieg J. 1954. N-Acetylation of the hexosamines. J Am Chem
Sashiwa H, Saimoto H, Shigemasa Y, Tokura S. 1993. N-Acetyl group distri-
bution in partially deacetylated chitins prepared under homogeneous condi-
tions. Carbohydr Res. 242:167–172.
Soto-Gil RW, Zyskind JW. 1989. N,N?-diacetylchitobiase of Vibrio harveyi.
Primary structure, processing, and evolutionary relationships. J Biol Chem.
Tanaka T, Fukui T, Atomi H, Imanaka T. 2003. Characterization of an exo-
beta-D-glucosaminidase involved in a novel chitinolytic pathway from the
hyperthermophilic archaeon Thermococcus kodakaraensis KOD1. J Bacte-
Chitin oligosaccharide deacetylase (COD)
Tanaka T, Fukui T, Fujiwara S, Atomi H, Imanaka T. 2004. Concerted action of
diacetylchitobiose deacetylase and exo-beta-D-glucosaminidase in a novel
chitinolytic pathway in the hyperthermophilic archaeon Thermococcus ko-
dakaraensis KOD1. J Biol Chem. 279:30021–30027.
Tommeraas K, Varum KM, Christensen BE,Smidsrod O. 2001.Preparation and
characterisation of oligosaccharides produced by nitrous acid depolymeri-
sation of chitosans. Carbohydr Res. 333:137–144.
Tsuji A, Kinoshita T, Hoshino M. 1969a. Analytical chemical studies on amino
hydrazone hydrochloride. Chem Pharm Bull (Tokyo). 17:1505–1510.
Tsuji A, Kinoshita T, Hoshino M. 1969b. Microdetermination of hexosamines.
Chem Pharm Bull (Tokyo). 17:217–218.
Yu C, Bassler BL, Roseman S. 1993. Chemotaxis of the marine bacterium
Vibrio furnissii to sugar substrates of the phosphoenolpyruvate: glycose
phosphotransferase system. J Biol Chem. 268:9405–9409.
Yu C, Lee AM, Bassler BL, Roseman S. 1991. Chitin utilization by marine
bacteria: A physiological function for bacterial adhesion to immobilized
carbohydrates. J Biol Chem. 266:24260–24267.
Yu C, Lee AM, Roseman S. 1987. The sugar-specific adhesion/deadhesion
apparatus of the marine bacterium Vibrio furnissii is a sensorium that con-
tinuously monitors nutrient levels in the environment. Biochem Biophys
Res Comm. 149:86–92.
Zobell CE, Rittenberg SC. 1937. The occurrence and characteristics of chitin-
oclastic bacteria in the sea. J Bacteriol. 35:275–287.