Bacterial surface engineering utilizing glucosamine phosphate derivatives as cell wall precursor surrogates

Article abstract of DOI:10.1002/chem.200801734

A study was conducted to examine glucosamine phosphate derivatives as cell wall precursor surrogates for bacterial surface display. N-acetylglu-cosamine-1- phosphate (GlcNAc-1-phosphate) derivate was used in the study as a bacterial cell wall precursor. It was observed during the study that the acetylation of the hydroxyl groups with high hydrophobicity can provide easy access to the cytoplasm of the bacteria. A flow cytometry was used to examine the incorporation of the derivatives into the bacterial cell wall. The study also used acetate buffer and biotin-PEO-hydrazide for washing the lactic acid bacteria. The incubation of lactic acid bacteria was carried out under the anaerobic conditions. This study can be used for development of oral vaccines.

Full text of DOI:10.1002/chem.200801734

DOI: 10.1002/chem.200801734
Bacterial Surface Engineering Utilizing Glucosamine Phosphate Derivatives
as Cell Wall Precursor Surrogates
Reiko Sadamoto,*[a, b] Takeshi Matsubayashi,[c] Masataka Shimizu,[c] Taichi Ueda,[c]
Shuhei Koshida,[b] Toshiaki Koda,[d] and Shin-Ichiro Nishimura[c, e]
Surface display on bacteria has attracted much attention
due to the potential applications in peptide library screen-
ing, and in the syntheses of various bioadsorbants, biosen-
sors, and oral vaccines.[1] Therefore, the cell-surface engi-
neering of bacteria has been an important theme in the field
of industry and in particular biotechnology. Several ap-
proaches to cell surface engineering have been explored
thus far. Genetic techniques have been used to express
target proteins as conjugates with their native surface pro-
teins.[2] However, the approaches are limited to the display
only of proteins and peptides.
the covalent attachment of various synthetic epitopes. Ber-
tozzi et al. originally proved that unnatural carbohydrate
substrates could be incorporated into an oligosaccharide to
introduce novel chemical reactivity to mammalian cell surfa-
ces.[3–6] In the case of bacteria, Tirrell and co-workers suc-
cessfully incorporated an azide group, using an azido amino
acid as a methionine surrogate, into an outer membrane
protein of Escherichia coli in which additional methionine
residues were engineered by site-directed mutagenesis.[5,7,8]
We previously developed an approach to display various
kinds of molecules on the bacterial cell surface using chemi-
cally synthesized cell-wall (peptidoglycan) precursors.[9,10]
Since the composition of peptidoglycan is strikingly similar
between bacterial strains,[11] our approach is applicable to a
variety of bacteria. In this method, chemically synthesized
cell-wall precursors, that is, UDP-MurNAc pentapeptide de-
rivatives, are incorporated into the cell wall through a bio-
synthetic pathway; this methods does not depend on genetic
modifications, thereby presenting a range of potential target
moieties, such as oligosaccharides,[9] beyond that of conven-
tional protein display. However, as the synthesis of the
UDP-MurNAc pentapeptide derivative is difficult and time-
consuming, the large-scale application of this system is prob-
lematic.
To overcome these problems, we focused on N-acetylglu-
cosamine-1-phosphate (GlcNAc-1-phosphate), which is an
early, simple intermediate in the peptidoglycan biosynthetic
route (Figure 1), and herein present our results. In bacteria,
GlcNAc-1-phosphate can be an intermediate of both the
MurNAc-pentapeptide and GlcNAc components of peptido-
glycans (see Figure 1). In bacterial cell-wall biosynthesis,
GlcNAc-1-phosphate is formed from glucosamine-1-phos-
phate,[12] whereas in mammalian cells, GlcNAc is directly
converted to the phosphorylated form via salvage path-
ways.[13] Therefore, we chose GlcNAc-1-phosphate, and not
GlcNAc, as the platform precursor for the introduction of
non-native reactive groups into the bacterial cell wall, and
designed compound 1, in which a ketone group was intro-
duced into the N-acetyl group, as a bacterial cell wall pre-
In contrast to the genetic approaches, a chemical ap-
proach based on the metabolic pathway has been utilized to
display a variety of molecules. Metabolic labeling of cell-sur-
face biomolecules with a non-native chemical tag allows for
[a] Dr. R. Sadamoto
Present address:
Ochadai Academic Production, The Glycoscience Institute
Ochanomizu University, 2-1-1, Otsuka, Bunkyo-ku
Tokyo 112-8610 (Japan)
Fax : (+81)3-5978-2581
E-mail: sadamoto.reiko@ocha.ac.jp
[b] Dr. R. Sadamoto, Dr. S. Koshida
Shionogi Laboratory of Biomolecular Chemistry
Graduate School of Advanced Life Science, Hokkaido University
Kita-21, Nishi-11, Sapporo 001-0021 (Japan)
[c] T. Matsubayashi, M. Shimizu, T. Ueda, Prof. S.-I. Nishimura
Laboratory for Bio-Macromolecular Chemistry
Graduate School of Advanced Life Science, Hokkaido University
Kita-21, Nishi-11, Sapporo 001-0021 (Japan)
[d] Prof. T. Koda
Laboratory of Embryonic and Genetic Engineering
Graduate School of Advanced Life Science, Hokkaido University
Kita-21, Nishi-11, Sapporo 001-0021 (Japan)
[e] Prof. S.-I. Nishimura
Drug-Seeds Discovery Research Laboratory
National Institute of Advanced Industrial Science and Technology
(AIST)
Sapporo 062-8517 (Japan)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801734.
10192
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 10192 – 10195

COMMUNICATION
cursor surrogate. We synthesized derivatives of GlcNAc-1-
phosphate 1–5, in which 1–4 each bear a ketone group. The
incorporation of the derivatives into the bacterial cell wall
Figure 1. Schematic presentation of the biosynthesis of the bacterial cell
wall.
was estimated by flow cytometry after labeling of the dis-
played ketone group with biotin-PEO-hydrazide 6 and, sub-
sequently, with streptavidin–Alexa Fluor 488 conjugate. The
selective reaction of ketones with the hydrazide group of 6
showed the effective incorporation of 2 into lactic acid
streptavidin–Alexa Fluor 488 conjugate was added and the
bacteria were washed with PBS buffer prior to analysis
using a flow cytometer.
ACHTUNGTRENNUNG
creased hydrophobicity, providing easier access to the cyto-
plasm of the bacteria. We expected that the protection of
the phosphate group by the cyclosaligenyl group would fur-
ther increase the hydrophobicity of 3. As the cyclosaligenyl
group is known to be removable in basic conditions above
pH 7,[14] the phosphate group was expected to be deprotect-
ed under the basic environment in bacteria. To confirm the
effect of phosphorylation in the bacterial cell wall pathway,
2-ketosugar 4, which does not possess a phosphate group,
was also tested. Peracetylated GlcNAc-phosphate 5 was
used as a negative control. The synthesis and characteriza-
tion of these compounds are described in the Supporting
Information.
First, we used the same strain (Lactobacillus plantarum
JCM1149) as was used in our previous report on the incor-
poration of UDP-MurNAc-pentapeptide derivatives.[9] Bac-
teria were incubated in Lactobacilli deMan Rogosa Sharpe
(MRS) broth at 378C for three hours, then bacteria were
collected by centrifuge (3000 rpm, 3 min.), resuspended in a
MRS Lactobacilli broth containing each test compound (1–
5, final concentration: 0, 5, 25, or 50 mm), and incubated for
a further 15 h. The lactic acid bacteria were washed with
The fluorescence histograms obtained from the flow cyto-
metric analysis were shifted as the concentration of com-
pound 2 in MRS media increased (Figure 2a). Figure 2b
shows the plot of the flow cytometry. FSC-A and SSC-A did
not change significantly, indicating that the shape of the bac-
teria was not changed by the addition of the precursor. In
order to estimate more accurately the changes in fluores-
cence of the labeling, fluorescence intensities were normal-
ized by dividing the intensity by size as shown in Figure 2c.
The normalized fluorescence intensity when incubated with
compounds 1–5 is summarized in Figure 2c. When adding
precursor 2 or 3, the fluorescence intensity was increased as
a function of concentration. These increases in fluorescence
suggested that the acetylated precursor 2 was incorporated
into the bacterial cell wall even at low concentrations,
whereas the more hydrophobic precursor 3, in which the
phosphate group was protected, was only incorporated at
higher concentrations (Figure 2c). The level of incorporation
of 3 was about half that of precursor 2, which possessed an
unprotected phosphate group. The lower level of incorpora-
tion of 3 is possibly due to the deprotection process not pro-
ceeding in the cytoplasm as expected. No increase in fluo-
rescence was observed for compound 1, probably due to its
highly hydrophilic nature. The negative control 5, which had
no ketone group, showed no increase in fluorescence, indi-
AHCTUNGTRENNUNG
acetate buffer (pH 4), and biotin-PEO-hydrazide (6), which
binds specifically to the ketone group, was added. After
washing with phosphate-buffered saline (PBS) (pH 7.2),
Chem. Eur. J. 2008, 14, 10192 – 10195
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10193

R. Sadamoto et al.
cating that the increase in fluorescence in 2 and 3 was not
due to non-specific binding of the precursors to the labeling
molecules. In the case of 4, which has a ketone group but no
phosphate group at C-1 position, the fluorescence intensity
did not increase. In bacterial cell-wall biosynthesis, the N-
acetyl group of GlcNAc is once removed and transformed
into glucosamine-1-phosphate. Thus, the ketone group of
phosphate-free precursor 4 might be removed before incor-
poration into the cell wall. Notably, the fact that 2, which
has a phosphate group, was incorporated and 4, which does
not have a phosphate group, was not incorporated, indicates
that incorporation proceeded via cell-wall biosynthesis and
not due to non-specific binding.
moieties.[15] However, our method for the surface display of
the reacting ketone group using the cell-wall precursor 2 tar-
gets the carbohydrate polymer chain, which is a common
structure among bacteria; therefore, it is applicable to a
wide variety of bacteria.
To quantify the absolute number of chemically accessible
ketone groups on the bacterial surface, the following experi-
Figure 3. Normalized fluorescence intensity of lactic acid bacteria. The x
axis is concentration of 2 in the growth culture and the y axis is normal-
&
ized fluorescence intensity (fluorescence intensity/FSC-A). : Lactobacil-
&
&
lus plantarum JCM1149; : Weissella confusa JCM1093; : Lactobacillus
plantarum JCM11125.
ments were performed. Lactic acid bacteria (Lactobacillus
plantarum JCM1149) were incubated in the presence of 2
and labeled with the streptavidin–Alexa Fluor 488 conju-
gate, as described above, for preparation of the sample for
flow cytometry analysis. The bacteria were washed and then
resuspended in PBS buffer and, using an aliquot of the sus-
pension, the number of bacteria in the suspension was deter-
mined by colony counting. The remaining bacteria were
lyzed by addition of a lyzing buffer containing 1.3 mm Tris,
and 0.28 mm EDTA (pH 8.0), and were analyzed by fluores-
cence spectrophotometer. The results were then compared
to the data obtained using the buffer containing free fluores-
cence dye. By dividing the fluorescence intensity by the
number of bacteria, the number of accessible ketone groups
was calculated to be more than 105 per bacterium. We as-
sumed that the bacteria have a columnar shape (length 5 mm
and diameter 1 mm, surface area: 6.6 mm2) that is covered
with a peptidoglycan monolayer since the avidin can only
access the ketone groups on the outer layer of the bacterial
surface. Assuming the molecular area of the repeating unit
of peptidoglycan (GlcNAc-MurNAc pentapeptide) is 1 nm2,
there are about 6.6ꢁ106 repeating units on the bacterium
surface. Based on this simplified model, the ratio of ketone-
attached repeating units in the peptidoglycan is roughly
1.5%.
Figure 2. Flow cytometric analysis of lactic acid bacteria. Lactic acid bac-
teria (Lactobacillus plantarum JCM1149) were incubated with GlcNAc-
1-phosphate derivative 2 (brown: 0 mm; blue: 5 mm; green: 25 mm; red:
50 mm) and labeled with biotin-PEO-hydrazide and streptavidin–Alexa
Fluor 488 conjugate. a) Fluorescence histograms: fluorescence intensity
(Alexa Fluor 488 A) versus the number of bacteria. b) Size (FSC-A)
versus complexity (SSC-A). c) Concentration of GlcNAc-1-phosphate de-
rivatives versus normalized fluorescence intensity (fluorescence intensity/
FSC-A).
Next, we used other strains of bacteria (Lactobacillus
plantarum JCM11125 and Weissella confusa JCM1093) to
check the versatility of incorporation of 2 into the cell wall
using the same procedure as that described above. The re-
sults are shown in Figure 3 together with the results for Lac-
tobacillus plantarum JCM1149. Increases in fluorescence in-
tensity due to the binding of dye-conjugated avidin were ob-
served for both strains. These strains have different types of
bacterial cell wall in that while the main carbohydrate struc-
ture is the same, there are small variations in the peptide
As the natural cell wall precursor, GlcNAc-1-phosphate,
is originally made from glucose,[11] the incorporation of the
artificial cell-wall precursor 2 is expected to be accelerated
when the availability of natural glucose is limited To reduce
the glucose concentration in the growth media, phosphate-
buffer media (PBM; see the Experimental Section)[16] con-
taining 2 and 0 to 50 mm glucose was used instead of the
MRS broth. The PBM media contained only artificial ingre-
10194
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 10192 – 10195

Bacterial Surface Engineering
COMMUNICATION
Incorporation experiments: The precursors synthesized were sterilized by
filtration before being added to the growth media for bacterial incuba-
tion. Lactic acid bacteria were incubated for 8 h under anaerobic condi-
tions and the cell culture was diluted with MRS broth (BD) to give OD
values (at 600 nm) of 0.05. The diluted culture (270 mL) was then incubat-
ed for 3 h. After the addition of 30 mL of MRS broth containing 2 (0, 1.5,
7.5, or 15 mm), the bacteria were incubated for a further 15 h. An aliquot
of the culture (30 mL; suspended before collection) was washed with ace-
tate buffer (pH 4, 3ꢁ100 mL), and then 100 mm biotin–PEO-hydrazide
(15 mL), which binds specifically to the ketone group, was added before
further incubation for 2 h at 408C. The cells were collected by centrifuga-
tion (3000 g, 3 min), washed PBS buffer (pH 7.2, 3ꢁ100 mL), and then
dients, apart from peptone. The bacteria were able to grow
even in the glucose-free PBM medium for a short period,
such as 12 h, although the growth speed was dramatically re-
duced to only 5% of that in MRS broth. On the basis of
flow cytometric analyses (Figure 4), fluorescence due to the
incorporated ketone group at 0 mm glucose was found to be
three-fold larger than that obtained at other glucose concen-
trations. Incorporation of 2 in MRS broth was similar to
that in PBM media containing 10, 25, or 50 mm glucose. We
consider that artificial precursor 2 is incorporated into the
cell wall much less effectively compared with native glucose;
therefore, the complete removal of glucose was necessary to
enhance the incorporation of precursor 2 into the bacterial
cell wall. These results also support the hypothesis that pre-
cursor 2 is metabolically incorporated through the cell wall
synthetic pathway.
streptavidin–Alexa Fluor 488 conjugate (Molecular Probes, 28 mgmLÀ1
,
100 mL) was added before incubation for 20 min at 48C. After washing
two times with PBS buffer (100 mL), PBS buffer (300 mL) was added and
the bacteria were analyzed using a flow cytometer (FACScant, BD). Ten
thousand signal counts were analyzed using FlowJo software (Tree star
Inc.). As a change in size was observed for bacteria incubated in media
supplemented with the test compounds, the fluorescence intensity was
normalized using average of FITC-A/average of FSC-A.
Acknowledgements
This work was supported by Grant-in-Aid for Scientific Research (KA-
KENHI) 18350081, the Naito Foundation, and Hayashi Memorial Foun-
dation for Female Natural Scientists. The analysis of flow cytometry was
carried out with instruments at the Open Facility, Hokkaido University
Sousei Hall. We thank Ms. Seiko Oka of the Center for Instrumental
Analysis, Hokkaido University, for providing the ESI-MS data.
Keywords: carbohydrates · glycoconjugates · peptidoglycans
Figure 4. Fluorescence intensity of lactic acid bacteria (JCM 1149) after
incubation in PBM broth containing 0, 10, 25, or 50 mm glucose, and la-
beled using the same procedure as described in Figure 2.
[1] J. F. M. L. Seegers, Trends Biotechnol. 2002, 20, 508–515.
[2] A. Sergeeva, M. G. Kolonin, J. J. Molldrem, R. Pasqualini, W. Arap,
Adv. Drug Delivery Rev. 2006, 58, 1622–1654.
[3] D. J. Vocadlo, H. C. Hang, E.-J. Kim, J. A. Hanover, C. R. Bertozzi,
Proc. Natl. Acad. Sci. USA 2003, 100, 9116–9121.
[4] L. K. Mahal, K. J. Yarema, C. R. Bertozzi, Science 1997, 276, 1125–
1128.
[5] A. J. Link, M. K. S. Vink, N. J. Agard, J. A. Prescher, C. R. Bertozzi,
D. A. Tirrell, Proc. Natl. Acad. Sci. USA 2006, 103, 10180–10185.
[6] H. C. Hang, C. R. Bertozzi, J. Am. Chem. Soc. 2001, 123, 1242–
1243.
[7] A. J. Link, M. K. S. Vink, D. A. Tirrell, J. Am. Chem. Soc. 2004, 126,
10598–10602.
[8] A. J. Link, D. A. Tirrell, J. Am. Chem. Soc. 2003, 125, 11164–11165.
[9] R. Sadamoto, K. Niikura, T. Ueda, K. Monde, N. Fukuhara, S.-I.
Nishimura, J. Am. Chem. Soc. 2004, 126, 3755–3761.
[10] R. Sadamoto, K. Niikura, P. S. Sears, H. Liu, C.-H. Wong, A. Suk-
somcheep, F. Tomita, K. Monde, S.-I. Nishimura, J. Am. Chem. Soc.
2002, 124, 9018–9019.
In conclusion, GlcNAc-1-phosphate derivative 2 can be
used as a practical precursor for bacterial surface display.
However, using GlcNAc derivative 4, which has no phos-
phate group, the ketone group cannot be displayed on the
bacterial surface. These results suggest that the incorpora-
tion of 2 occurred via cell-wall biosynthesis. By limiting the
glucose in the growth media, the degree of incorporation
was increased about three-fold. Further, this method was su-
perior to our prior system using UDP-MurNAc pentapep-
tide derivatives and is applicable to large-scale in vivo stud-
ies aimed at the development of oral vaccines.
Experimental Section
[11] G. Seltmann, O. Holst, The Bacterial Cell Wall, Springer, New York,
2002.
[12] J. van Heijenoort, Nat. Prod. Rep. 2001, 18, 503–519.
[13] A. Varki, FASEB J. 1991, 5, 226–235.
[14] U. Muus, C. Kranz, T. Marquardt, C. Meier, Eur. J. Org. Chem.
2004, 1228–1235.
[15] K. H. Schleifer, O. Kandler, Bacteriol. Rev. 1972, 36, 407–477.
[16] J. B. Velꢂzquez, C. Sieiro, J. Cansado, P. Calo, E. Longo, P. Blanco,
T. G. Villa, J. Ind. Microbiol. 1993, 12, 368–372.
General: NMR spectra were recorded on a Bruker AMX-500 spectrome-
ter. All NMR measurements were carried out at 278C in CDCl3, D2O, or
[D4]MeOH. Unless otherwise noted, reagents were obtained from com-
mercial suppliers, Wako Pure Chemical Industries Ltd. (Osaka, Japan),
Tokyo Kasei Kogyo Co. (Tokyo, Japan), and Aldrich Chemical Co. (Mil-
waukee, WI), and were used without further purification. Synthetic de-
tails are described in the Supporting Information.
Composition of PBM: Peptone (5.0 g), ammonium sulfate (2.0 g), bipo-
tassium phosphate (1.4 g), monopotassium phosphate (0.6 g), citric acid
(0.1 g), magnesium sulfate (0.12 g), sodium chloride (5.0 g), calcium chlo-
ride (0.7 g), manganese sulfate (0.07 g), d-glucose (0–50.0 mm).
Received: August 21, 2008
Published online: October 15, 2008
Chem. Eur. J. 2008, 14, 10192 – 10195
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10195