Synthesis of cell-penetrating S-galactosyl-oligoarginine peptides as inducers of recombinant protein expression under the control of lac operator/repressor systems

Article abstract of DOI:10.1002/cplu.201300130

The synthesis of glycodendrimers and glycopoly(oxazoline)s as inducers of recombinant protein expression has recently been reported; however, these compounds induced the expression of only small amounts of the green fluorescence protein (GFP), which was used as the model recombinant protein, because of their poor ability to penetrate the Escherichia coli cell membrane. Therefore, S-galactosyl-oligo(Arg) conjugates have now been synthesized to overcome this problem. Following in vivo expression of GFP induced by each of the S-galactosyl (Arg)_n constructs (n=5, 6, 8) at the T5 promoter in E. coli for 18 hours, we visually observed that the cultures fluoresced green light when excited with UV light. The fluorescent intensities for these cultures were greater than that found for a control culture, which indicates that the peptides had induced GFP expression. Quantitative fluorescent measurements also supported the observations that the peptides were better inducers of GFP expression than the galactosyl dendrimers and the poly(oxazoline)s and the natural inducer lactose. Because the level of GFP expression was directly related to the number of arginine moieties in each peptide, we propose that the number of arginine moieties is responsible for how well each peptide passes through the E. coli membrane, which affects the expression level. A similar tendency was observed when the T7 promoter was placed upstream from the gene for an artificial extracellular matrix protein and the S-Gal-oligo(Arg) peptides were used as inducers. To assess how the distance between two galactosyl moieties as well as how the multivalent effect (cluster effect) in an oligo(Arg) inducer affects the expression level of GFP, we synthesized a conjugate of Lys(Arg)₈ (Lys=lysine) and two S-galactosyls, which enhanced the expression of GFP in comparison with that obtained for S-Gal(Arg)₈. Copyright

Full text of DOI:10.1002/cplu.201300130

CHEMPLUSCHEM
FULL PAPERS
DOI: 10.1002/cplu.201300130
Synthesis of Cell-Penetrating S-Galactosyl–Oligoarginine
Peptides as Inducers of Recombinant Protein Expression
under the Control of lac Operator/Repressor Systems**
Yukina Mizuta, Akinori Takasu,* and Masahiro Higuchi[a]
The synthesis of glycodendrimers and glycopoly(oxazoline)s as
inducers of recombinant protein expression has recently been
reported; however, these compounds induced the expression
of only small amounts of the green fluorescence protein (GFP),
which was used as the model recombinant protein, because of
their poor ability to penetrate the Escherichia coli cell mem-
brane. Therefore, S-galactosyl–oligo(Arg) conjugates have now
been synthesized to overcome this problem. Following in vivo
expression of GFP induced by each of the S-galactosyl (Arg)n
constructs (n=5, 6, 8) at the T5 promoter in E. coli for
18 hours, we visually observed that the cultures fluoresced
green light when excited with UV light. The fluorescent intensi-
ties for these cultures were greater than that found for a con-
trol culture, which indicates that the peptides had induced
GFP expression. Quantitative fluorescent measurements also
supported the observations that the peptides were better in-
ducers of GFP expression than the galactosyl dendrimers and
the poly(oxazoline)s and the natural inducer lactose. Because
the level of GFP expression was directly related to the number
of arginine moieties in each peptide, we propose that the
number of arginine moieties is responsible for how well each
peptide passes through the E. coli membrane, which affects
the expression level. A similar tendency was observed when
the T7 promoter was placed upstream from the gene for an ar-
tificial extracellular matrix protein and the S-Gal–oligo(Arg)
peptides were used as inducers. To assess how the distance
between two galactosyl moieties as well as how the multiva-
lent effect (cluster effect) in an oligo(Arg) inducer affects the
expression level of GFP, we synthesized a conjugate of Lys-
(Arg)8 (Lys=lysine) and two S-galactosyls, which enhanced the
expression of GFP in comparison with that obtained for S-Gal-
(Arg)8.
Introduction
The RNA polymerase of bacteriophage T5/T7 has a stringent
specificity for its own promoters. In 1961, Jacob and Monod
suggested the use of the Escherichia coli lactose (lac) operon
as a model for gene regulation.[1] This model system is still
used to study how a structural gene set can be coordinately
transcribed or repressed, depending upon the metabolites
found in the intercellular environment. The lac repressor,
which is the protein product of lacI binds to the lac operator
in the absence of an inducer, for example, the naturally occur-
ring inducer lactose. When the repressor is bound to the oper-
ator, transcription of lacZ, lacY, and lacA, which encode b-galac-
tosidase, lac permease, and a transacetylase, respectively, does
not occur (Figure 1). The lac repressor promoter contains
360 amino acids and associates to form a homotetramer of
154520 Da.[2] Each monomer contains one saccharide binding
site. Notably, isopropyl b-d-thiogalactoside (IPTG)—which is
Figure 1. The T5 expression system under the control of a lac promoter/op-
erator system.
not a substrate for b-galactosidase, but a molecular mimic of
allolactose derived from lactose, which may be the “true” in-
ducer—acts as a gratuitous inducer and turns on transcription
of the lactose operon through its interaction with the lac re-
pressor.[3,4] IPTG permeates E. coli without the assistance of the
lac permease.[5]
[a] Y. Mizuta, Prof. A. Takasu, M. Higuchi
Department of Frontier Materials
Nagoya Institute of Technology
Gokiso-Cho, Showa-Ku
Nagoya 466-8555 (Japan)
Fax: (+81)52-735-5342
E-mail: takasu.akinori@nitech.ac.jp
Much attention is currently being devoted to glycobiology
and glycochemistry topics in biomedicine and biochem-
istry,[6–14] especially as they relate to interactions between pro-
[**] lac=lactose
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cplu.201300130.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 677 – 683 677

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Scheme 1. Scheme for S-Gal–oligo(Arg) peptide synthesis (DIPCI=diisopropylcarbodiimide, HOAt=1-hydroxybenzotriazole).
teins and carbohydrates. Carbohydrate–protein[6–12] and carbo-
hydrate–carbohydrate[13,14] interactions have been well charac-
terized, and, notably, such interactions are often strengthened
by the presence of multiple binding sites. Certain glycopoly-
mers in which multiple saccharide residues are incorporated
into their polymer backbones have enhanced binding affinities
toward their targeted proteins relative to their monomeric, sac-
charide-containing building block. This property has been as-
cribed to multivalent recognition, that is, the cluster effect.[6–12]
Notably, certain glycopolymers, in which the saccharide spac-
ing is random, strongly bind their target proteins.[6–12] Aoi
et al.[15,16] used a dendrimer skeleton and Matsuura et al.[17,18]
reported alternative strategies to prepare periodic glycosylated
oligonucleotides (20-mers) as a means of controlling the three-
dimensional arrangements of the pendant saccharides. These
carbohydrate-containing compounds bound strongly to certain
lectins. We have also described the binding affinities of a-heli-
cal peptides that contained a pendant saccharide linked to the
peptide backbone through an O-glycoside linkage[19] (a model
for mucine-type glycoproteins and glycopeptides) or through
an N-glycoside linkage.[20,21] Recent progress involving glyco-
conjugates for biomedical applications, for example, drug-de-
livery systems, antibacterial activity, inhibition of viral infection,
and the spread of malignant tumors and the human immuno-
deficiency virus has been remarkable with many excellent re-
sults reported.[6]
as we are aware. Recently, we synthesized S-galactosyl den-
drimers[22] and S-galactosyl poly(oxazoline)s,[23] which we ex-
pected to act as inducers of recombinant protein synthesis
through the cluster effect; however, their limited ability to per-
meate the E. coli membrane precludes their use as inducers.
For this study, we designed S-galactosyl–oligo(arginine) (S-Gal–
oligo(Arg)) conjugates (Scheme 1), including one containing
the “magic arginine number” eight (R8),[24–26] to use as new
types of lac operon inducers.
Results and Discussion
We synthesized the S-Gal–(Arg)n conjugates (n=5, 6, or 8)
using solid-phase methods (Scheme 1). However, compound 1,
which contains a protected galactoside moiety, was synthe-
sized from penta-O-acetyl-b-d-galactopyranoside and 3-mer-
captopropionic acid as reported previously.[22,23] The yield was
1
91%, and its structure was confirmed by H NMR spectroscopy.
After preparing the three protected arginine peptides through
solid-phase fluorenylmethyloxycarbonyl chloride (Fmoc)
chemistry, 1 was coupled to each of the resin-bound peptides
through their N termini. O-Acetyl deprotection of the galacto-
syl hydroxyl groups was accomplished using hydrazine mono-
hydrate, following the release of the glycopeptides from the
resin and the deprotection of the arginine side chains by tri-
fluoroacetic acid. The yields of the peptides were 60–69%.
1H NMR spectroscopy (200 MHz; Figure S1 in the Supporting
Information) and MALDI-TOF mass spectrometry (Figure S2)
showed the expected structure and that the configuration of
the anomeric carbon was b in all cases. The deacetylated prod-
ucts were completely water soluble and their conformations,
as indicated by their room-temperature CD spectra, which con-
As far as we know, there have been only a few studies con-
cerning the design of new glycoconjugates that can act as in-
ducers of recombinant protein expression under the control of
a lac promoter/operator/repressor system.[22,23] Furthermore,
no reports have focused on the cluster effect of saccharides
and allosteric effects of the repressor proteins until now, as far
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 677 – 683 678

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
tained distinct minimums at 197 nm and less pronounced min-
imums at 222 nm (Figure S3), were primarily aperiodic struc-
tures with possibly a small amount (ꢀ5%) of a helices (a heli-
ces increase the cell-membrane permeability of poly(Arg) and
oligo(Arg) constructs[26]).
also supported the theory that the cell-penetrating peptide is
a key moiety. However, since oligoarginine peptides are well
known to enhance gene transfection efficiency in various cell
lines,[24–26] we now had to eliminate the possibility that oligoar-
ginine peptides can enhance protein expression in the E. coli
cell. The model peptides containing no galactosyl moiety were
prepared (n=8) and used as the inducer for the expression of
GFP under the same conditions. After expression for 18 h, the
green fluorescence was not detected (F/OD600 =5) and we con-
cluded that there is the possible synergistic effect of the galac-
tosyl moiety and the oligoarginine peptides as we expected.
Next, we used a pET expression system that included the
gene for an artificial extracellular matrix (ECM) protein (aECM-
CS5-ELF-F) that contains a fibronectin cell-binding domain
(CS5) and an elastin sequence (ELF).[29,30] To assess the ability of
the S-Gal–oligo(Arg) peptides to act as inducers, aECM-CS5-
ELF-F[29,30] has been expressed in E. coli using a T7 ex-
pression system. We have also examined how aECM-
CS5-ELF-F mutants, containing phenylalanine (Phe)
analogues instead of Phe, affect the adhesive behav-
ior of human umbilical vein endothelial cells
(HUVEC).[30] The T7 promoter and its downstream
target gene reside in a pET plasmid. Residual T7 poly-
merase activity is inhibited by T7 lysozyme, which is
constitutively expressed at low levels, and for which
its gene is present in either the pLysS or pLysE plas-
mid (Lys=lysine; Figure S4).[31,32] The pET system can
produce large amounts of protein (up to 100 mgLÀ1
of culture medium). For this study, the S-Gal–oligo-
(Arg) peptides and IPTG was used as inducers of
aECM-CS5-ELF-F expression[30] that was controlled by
a T7 promoter upstream from a lac operator. Whole
cell lysates and purified aECM-CS5-ELF-F were sub-
jected to sodium dodecyl sulfate (SDS) polyacryl-
amide gel electrophoresis (PAGE). Substantial
amounts of a protein with a calculated molecular
weight of 42600 that corresponds to aECM-CS5-ELF-F
Assays that relied on green fluorescent protein (GFP) fluores-
cence in E. coli cells containing pQE9-GFP2/pREP4 were per-
formed to evaluate the abilities of the S-Gal–oligo(Arg) pep-
tides to act as inducers of GFP expression. For each assay,
when the optical density (OD600 nm) of a culture (5 mL) was 0.6,
one of the S-Gal–oligo(Arg) peptides or IPTG (each at 1 mm)
was added. After an 18 hours induction, the visible fluores-
cence of each culture was observed (irradiation at 365 nm,
Figure 2), quantified by fluorescence spectrometry (excitation
at 395 nm and emission at 509 nm), and normalized to the
OD600 nm of each culture (F/OD600 nm; Figure 2).[27,28]
Figure 2. Normalized fluorescent intensities (F/OD600 nm) of E. coli cultures after induction
with IPTG or an S-Gal–oligo(Arg) peptide.
were seen, which indicates that the lac operator was
activated by the S-Gal–oligo(Arg) peptides in the T7
expression system. The expressed protein concentra-
As expected, the normalized fluorescence of the culture was
relatively large (F/OD600 =64) when IPTG was the inducer.
When using the S-Gal–oligo(Arg) peptides as inducers, fluores-
cence at 509 nm was detected (F/OD600 between 16 and 24),
which is smaller than that found for IPTG, although larger than
that for lactose (F/OD600 =5). The fluorescent measurements
also showed that the S-Gal–oligo(Arg) peptides were better in-
ducers than were the galactosyl dendrimers (F/OD600 =6–10)[22]
and the poly(oxazoline)s (F/OD600 =6–11).[23] Because GFP ex-
pression increased as the number of arginine moieties in the
S-Gal–oligo(Arg) peptides increased, it seems that an increase
in the number of arginine moieties improved their passage
through the E. coli cell membrane. Induction by deacetylated
1, prepared by reaction of 1 and hydrazine monohydrate (20:1
molar ratio, hydrazine/galactosyl moiety) was also performed
under the same conditions. The F/OD600 of 14 was lower than
that induced by S-Gal–oligo(Arg)n (n=8; F/OD600 =24), which
tions were greater than that of the 10 mgLÀ1 culture and the
yields are summarized in Figure 3.
The expression levels of aECM-CS5-ELF-F were smaller (20–
39 mgLÀ1) when the S-Gal–oligo(Arg) peptides were used as
inducers, than when IPTG served as the inducer (82 mgLÀ1),
but greater than that found when lactose was used as the in-
ducer (10 mgLÀ1). Because aECM-CS5-ELF-F expression in-
creased as the number of arginine moieties in the S-Gal–oligo-
(Arg) peptides increased, an increase in the number of arginine
moieties apparently improved their passage through the E. coli
cell membrane as had been shown for the GFP/T5 expression
system.
1
The 600 MHz H NMR spectra of the proteins purified from
the T7 expression system using S-Gal(Arg)8 or IPTG as the in-
ducer are the same, which indicates that aECM-CS5-ELF-F[30,33]
had been expressed in both systems (Figure 4, lower panel).
Lower critical solution temperature (LCST) values were also
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 677 – 683 679

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
1 (Scheme 2). This inducer showed the expression of GFP (F/
OD600 =40; see also Figure 2, bottom), which is higher than
that obtained for S-Gal–oligo(Arg)8 (F/OD600 =24). The X-ray
structure of the lac repressor[3] shows that the distance be-
tween two sugar-binding sites is approximately 2.4 nm. The in-
tersugar distance (between the C4 carbon atoms in the two
galactosyl moieties) in a Corey–Pauling–Koltun (CPK) model of
the conjugate would be between 0.8 and 2.2 nm depending
on the conformation. Therefore, to better estimate the dis-
tance between the two galactosyl moieties, an energy minimi-
zation using a semiempirical molecular orbital method[39] was
performed for a derivative of Lys(Arg)3 that has 1 covalently at-
tached to the N terminal and the Ne of the lysine, Gal-S-
(CH2)2CONLys[NCO(CH2)2-S-Gal](Arg)3-COOH. After energy mini-
mization, the intersugar distance, as defined above, was
1.0 nm (Figure S5), which is shorter than the 2.4 nm separation
found for the lac repressor binding sites,[3] thereby indicating
that the enhanced expression level induced by the S-diGal-
(Arg)8 peptide can be ascribed, in part, to the cluster (multiva-
lent) effect. It seems that the abilities of the peptides synthe-
sized for this study to induce protein expression mainly
depend on their ability to penetrate E. coli.
Figure 3. Yield of aECM-CS5-ELF-F expressed in 1 L of an E. coli culture after
induction with IPTG or an S-Gal–oligo(Arg) peptide from a T7 expression
system under lac repressor control.
Conclusion
For the study reported herein, we prepared a set of S-Gal–
oligo(Arg) peptides and a di-S-Gal-Lys-(Arg)8 peptide using
solid-phase methods that represent a new type of inducer for
protein expression under the control of a lac operator/promot-
er/repressor system. Excellent galactosyl functionalities (cou-
pling yield >99%) were obtained. Their structures were veri-
fied by NMR spectroscopy and MALDI-TOF mass spectrometry.
The abilities of these compounds to act as inducers were veri-
fied by the fluorescence associated with GFP expression in
E. coli by using a T5 promoter. A similar tendency was ob-
served when the T7 promoter was placed upstream from the
gene for an artificial extracellular matrix protein. These funda-
mental results provide new possibilities for a recombinant pro-
tein expression system that is better than IPTG, although they
are not yet as effective as IPTG.
Figure 4. Above: LCST measurements (protein concentration, 1 mgmLÀ1 in
H2O). Below: 600 MHz H NMR spectra of proteins expressed after addition
1
of IPTG (black) or S-Gal(Arg)8 (red) (solvent, [D6]DMSO).
Experimental Section
measured using 1 mg of aqueous expressed protein per millili-
ter. The temperature was increased at a rate of 18CminÀ1
while monitoring the percent transmission at 350 nm. A prop-
erty that distinguishes elastin-like polypeptides from other
types of polypeptides is that they are miscible at all concentra-
tions in water below their LCST value.[29,30,33–38] For the protein
preparations both LCST values were 348C (Figure 4, upper
panel).
Materials
Pentaacetyl b-d-galactose was purchased from Sigma–Aldrich
Corp. 3-Mercaptopropionic acid was purchased from Wako Corp.
Fmoc/2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl (Pbf)-pro-
tected Arg-Clear acid resin and Fmoc-Arg(Pbf) were purchased
from the Peptide Institute, Inc. 1-Hydroxybenzotriazole (HOAt) and
N,N’-diisopropylcarbodiimide (DIPCI) were purchased from Wata-
nabe Chemical Industries. Piperidine, trifluoroacetic acid, hydrazine
monohydrate, boron trifluoride diethyl ether complex, THF, and all
other reagents were analytical grade and acquired from Peptide In-
stitute Inc., Nacalai Tesque, Wako, and Sigma-Aldrich; these com-
pounds were used without further purification.
Finally, to assess how the distance between and the pres-
ence of two galactosyl moieties in an oligo(Arg) inducer as
well as the multivalent effect (cluster effect) affect the expres-
sion levels of GFP, we synthesized a conjugate of Lys(Arg)8 and
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 677 – 683 680

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Scheme 2. Strategy for the synthesis of Gal-S-(CH2)2CONLys[NCO(CH2)2-S-Gal](Arg)8-COOH.
3.91 (t, J=6.1 Hz, 1H; H-5), 3.99–4.28 (m, 2H; H-6), 4.55 (d, J=
9.9 Hz, 1H; H-1), 5.05 (dd, J=3.3, 9.8 Hz, 1H; H-3), 5.23 (t, J=
9.9 Hz, 1H; H-2), 5.43 ppm (d, J=3.1 Hz, 1H; H-4).
Measurements
1H and 13C NMR spectra were recorded at 278C using a Bruker
DPX200 spectrometer (200 MHz for 1H and 50 MHz for 13C) or
a Bruker DPX-600 spectrometer (600 MHz for 1H). Chemical shifts
were referenced to the tetramethylsilane signal (d=0 ppm).
MALDI-TOF mass spectra were recorded using a JMS-S3000 mass
spectrometer (JEOL), with dithranol as the matrix agent. CF3COONa
was used to generate monosodium-cationized ions of the oligo-
(Arg) derivatives ([M+Na]+). CD spectra were measured using
0.02 mgmLÀ1 aqueous samples of the S-Gal–oligo(Arg) peptides,
a 1 mm quartz cell, and a Jasco J820 K spectropolarimeter. LCST
values were measured for 1 mgmLÀ1 aqueous samples of the re-
combinant, artificial ECM protein aECM-CS5-ELF-F that contains a fi-
bronectin cell-binding domain (CS5) and an elastin sequence
(ELF).[29,30] The temperature was increased at a rate of 18CminÀ1
while measuring the percent transmission of a sample at 350 nm
with a V-550 spectrometer.
Preparation of S-Gal–oligo(Arg) peptides
Pbf-protected (Arg)n peptides (n=5, 6, 8) were first prepared by
solid-phase synthesis using an EYELA Solid Organic Synthesizer
CCS-150M, Fmoc-(Pbf)Arg-resin (0.50 g, 0.0176 mol), and Fmoc
chemistry. Next, 1 (0.230 g, 0.528 mmol, 3 equiv) was coupled to
the N terminal of each peptide while it was still bound to the resin
by Fmoc chemistry. Removal of the acetylated S-Gal–oligo(Arg)
preparations from the resin and deprotection of the arginine side
chains were performed by adding ice-cold trifluoroacetic acid
(28.5 mL) and H2O (1.5 mL) to each ice-cold peptide solution and
then each mixture was stirred at room temperature for 1.5 h. Each
mixture was then filtered, and each filtrate was subjected to re-
duced pressure by rotary evaporation to remove any remaining tri-
fluoroacetic acid. Next, each filtrate was poured into cold diethyl
ether (200 mL) to precipitate the product, which was, in all cases,
a white powder. Each ether/product mixture was centrifuged to
isolate the product, which was then dried. To remove the acetyl
groups, the acetyl-protected S-Gal–oligo(Arg) peptides were indi-
vidually dissolved in THF containing hydrazine monohydrate (20:1
molar ratio, hydrazine/galactosyl moiety). After the solutions had
been left standing at 08C for 16 h, acetone was added into each
solution to quench the reaction. The solvent was completely re-
moved under vacuum, and the residue was dissolved in H2O. Each
preparation was dialyzed against water (2 L) in seamless cellulose
tubing (Spectra/Por Biotech Cellulose Ester, cut-off molecular
weight of 1000) for 3 days, with the dialysate replaced every 12 h.
For each preparation, a white powder was obtained upon lyophili-
zation. The yields were 60–69%.
Semiempirical molecular orbital calculation
A model of Gal-S-(CH2)3CONLys[NCO(CH2)3-S-Gal](Arg)3-COOH was
subjected to a semiempirical molecular orbital calculation by the
AM1 molecular orbital (MO) method[27] in MOPAC 2000 to investi-
gate the preferred distance between the two galactosyl moiet-
ies.[40] As a starting conformation, the backbone of the (Arg)3 chain
was modeled as a planar zigzag structure.
Preparation of 3-(2,3,4,6-tetra-O-acetyl-b-d-galactopyrano-
sylthio)propionic acid (1)
Compound 1 was prepared (91% yield) as described (Scheme 1).[22]
1H NMR (CDCl3): d=1.99, 2.06, 2.16, 2.17 (4s, 12H; COCH3), 2.77 (t,
J=6.7 Hz, 2H; SCH2CH2COO), 2.96 (t, J=6.0 Hz, 2H; SCH2CH2COO),
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 677 – 683 681

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
48C). Precipitates were removed by centrifugation (20100g,
60 min, 28C), and the clarified retentates were lyophilized. The
purity of the proteins and the uniformity of their molecular masses
were confirmed by SDS-PAGE and 1H NMR spectroscopy.
Expression of GFP in E. coli by a T5 expression system under
the control of a lac promoter/operator/repressor system
E. coli K10 cells were transformed with a pQE9-GFP2/pREP4 vector
containing the gene for GFP for which expression was under the
control of a lac promoter/operator/repressor system.[22,23] The cells
were cultured at 378C in M9 medium (5 mL) supplemented with
0.2% (w/v) glucose, 35 mgLÀ1 thiamine, 0.1 mm MgSO4, 0.1 mm
CaCl2, the 20 common amino acids (4 mgmLÀ1 each), 25 mgLÀ1 ka-
namycin, and 20 mgLÀ1 chloramphenicol to an OD600 nm of 0.6,
after which the culture was divided into aliquots for the expression
experiments. GFP expression was assessed at 378C by adding one
of the S-Gal–oligo(Arg) peptides, IPTG, or lactose into a culture
(each at final concentration of 1 mm) and incubating the cultures
for 4 h, after which the fluorescence of each was visualized by UV-
light irradiation (395 nm). A culture to which no peptide or inducer
had been added served as the negative control. To quantify ex-
pression levels, the fluorescence of each culture was measured at
509 nm (excitation at 395 nm) using a Hitachi F-2700 spectrometer.
The fluorescence (F) of each culture, expressed as relative light
units, was normalized to the number of cells (F/OD600 nm).[27,28] At
least three independent experiments were carried out to check the
reproducibility.
Synthesis of S-(Gal)2-Lys-(Arg)n peptide (n=8) containing
two galactoside residues
Solid-phase chemistry, as described above was used to prepare
Lys(Fmoc)2-(Arg)8(Pbf) starting from Fmoc-(Pbf)Arg-resin (0.50 g,
0.0176 mol). Next, after removal of the Fmoc groups, 1 (0.460 g,
1.056 mmol) was coupled to the N terminal of the peptide and the
lysine side-chain amino group while the peptide was still bound to
the resin using the procedure described above. The acetylated di-
S-Gal-peptide product was removed from the resin while on ice by
addition of an ice-cold mixture of trifluoroacetic acid (28.5 mL) and
H2O (1.5 mL), after which the mixture was stirred at room tempera-
ture for 1.5 h with simultaneous deprotection of the arginine side
chains. Next, the mixture was filtered, and the filtrate was subject-
ed to reduced pressure to remove any remaining trifluoroacetic
acid. Then the filtrate was poured into ice-cold diethyl ether
(200 mL) and the mixture was centrifuged to isolate the precipi-
tate, which was then dried to give the acetylated peptide as
a white powder. Next, to deprotect the galactosyl hydroxyl groups,
the product was dissolved in THF containing hydrazine monohy-
drate (20:1 molar ratio, hydrazine/galactosyl moiety) and was left
standing at 08C for 16 h, after which acetone was added to the so-
lution, and the mixture was stirred for 1 h to quench the reaction.
The solvent was evaporated under vacuum, and the residue was
dissolved in H2O. The solution was dialyzed as described above
using seamless cellulose tubing (Spectra/Por Biotech Cellulose
ester, cut-off molecular weight of 1000) for 3 days. A white powder
was obtained by lyophilization with a yield of 54%.
Expression of aECM-CS5-ELF-F in E. coli by a T7 expression
system
The pET28-CS5-ELF-PheRS* plasmid,[29,30] in which a linker sequence
encoding a T7 tag, a hexahistidine tag, and an enterokinase cleav-
age site was cloned into a pET28 plasmid between its Nco I and
Xho I sites, was transformed into the phenylalanine-auxotrophic
E. coli BL21(DE3) strain AF (HsdS gal (ncIts857 ind1 Sam7 nin5
lacUV5-T7 gene 1) pheA), which was constructed in the Tirrell labo-
ratory. This E. coli strain AF-IQ[pET28-CS5-ELF-PheRS*] served as the
expression system.[29,30] Initially, samples (300 mL) of M9 minimal
medium supplemented with 0.4% (w/v) glucose, 35 mgLÀ1 thia-
mine, 0.1 mm MgSO4, 0.1 mm CaCl2, the 20 common amino acids
(40 mgLÀ1 each), 25 mgmLÀ1 kanamycin, and 20 mgLÀ1 chloram-
phenicol were inoculated with 5 mL of the same medium contain-
ing the expression strain that had been cultured at 378C overnight.
Next, 1 L preparations of the same medium were each inoculated
in culture (300 mL). When the OD600 nm of a culture was between
0.7 and 1.0, a medium shift was performed. To do so, each culture
was centrifuged for 7 minutes at 20100g, 48C, and the supernatant
was removed. Cells were resuspended in the supplemented M9
medium described above that was deficient in Phe. Protein expres-
sion was then induced by addition of an S-Gal–oligo(Arg) peptide
or IPTG (final concentration, 1 mm). After 10 min, Phe was added at
concentrations between 50 and 250 mgLÀ1 as previously de-
scribed.[30] Cells were cultured for an additional 4 h, and protein ex-
pression was monitored by SDS PAGE using a normalized OD600 of
0.5 per sample. The aECM-CS5-ELF-F purification scheme, which
takes advantage of its inverse temperature transition, has been re-
ported.[33–38] Briefly, wet cell masses were each dispersed in TEN
buffer (10 mm tris(hydroxymethyl)aminomethane (Tris)-HCl/1 mm
ethylenediaminetetraacetate (EDTA), pH 8.0) at a concentration of
1 gmLÀ1, frozen, and thawed at 48C with 1.0 mgmLÀ1 deoxyribonu-
clease (DNase), 10 ngmLÀ1 ribonuclease (RNase), and 50 ngmLÀ1
phenylmethylsulfonyl fluoride (PMSF) added. aECM-CS5-ELF-F sam-
ples, which were found in the pellets of the whole-cell lysates after
centrifugation (20100g, 60 min, 378C), were resuspended in 4m
urea. The solutions were centrifuged (20100g, 60 min, 48C) to
remove nonprotein debris and then dialyzed against water (3 days,
Acknowledgements
We are grateful to Professor David A. Tirrell (California Institute
of Technology) and Dr. Inchan Kwon (Virginia University) for the
gift of the E. coli K10 GFP expression system, as well as helpful
suggestions and discussion. This study was funded by the Minis-
try of Education, Culture, Sports, Science and Technology of
Japan (Grant-in-Aid for Development Scientific Research, no.
24550132) and Japan Science and Technology (JST) “Research for
Promoting Technological Seeds”.
Keywords: bacteria
·
biosynthesis
·
fluorescence
·
glycopeptides · protein expression
[1] F. Jacob, J. Monod, J. Mol. Biol. 1961, 3, 318–356.
[2] W. Gilbert, B. Mꢁller-Hill, Proc. Natl. Acad. Sci. USA 1966, 56, 1891–1898.
[3] M. Lewis, G. Chang, N. C. Horton, M. A. Kercher, H. C. Pace, M. A. Schu-
macher, R. G. Brennan, P. Lu, Science 1996, 271, 1247–1254.
[4] K. Matthews, Science 1996, 271, 1245–1246.
[5] A. Khlebnikov, J. D. Keasling, Biotechnol. Prog. 2002, 18, 672–674.
[6] Y. C. Lee, R. T. Lee, Neoglycoconjugates: Preparation and Applications,
Academic Press Inc., San Diego, 1994.
[7] A. Kobayashi, T. Akaike, K. Kobayashi, H. Sumitomo, Makromol. Chem.
Rapid Commun. 1986, 7, 645–650.
[8] R. Roy, F. C. Troopper, J. Chem. Soc. Chem. Commun. 1988, 1058–1060.
[9] D. D. Manning, X. Hu, P. Beck, L. L. Kiessling, J. Am. Chem. Soc. 1997,
119, 3161–3162.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 677 – 683 682

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
[10] S.-I. Nishimura, T. Furuike, K. Matsuoka, K. Maruyama, K. Nagata, N.
Nishi, S. Tokura, Macromolecules 1994, 27, 4876–4880.
[11] G. Wulff, J. Schmid, T. Venhoff, Macromol. Chem. Phys. 1996, 197, 259–
274.
[12] K. Yamada, M. Minoda, T. Miyamoto, Macromolecules 1999, 32, 3553–
3558.
[13] N. Kojima, S. Hakomori, J. Biol. Chem. 1991, 266, 17552–17558.
[14] K. Matsuura, R. Oda, H. Kitakouji, M. Kiso, K. Kitajima, K. Kobayashi, Bio-
macromolecules 2004, 5, 937–941.
[26] Y. Takechi, H. Yoshii, M. Tanaka, T. Kawakami, S. Aimoto, H. Saito, Lang-
muir 2011, 27, 7099–7107.
[27] H. Lebeau, S. Lindow, J. Bacteriol. 2001, 183, 6752–6762.
[28] K.-S. Ko, J. Kruse, N. L. Pohl, Org. Lett. 2003, 5, 1781–1783.
[29] I. S. Carrico, S. A. Maskarinec, S. C. Heilshorn, M. L. Mock, J. C. Liu, P. J.
Nowatzki, C. Franck, G. Ravichandran, D. A. Tirrell, J. Am. Chem. Soc.
2007, 129, 4874–4875.
[30] A. Takasu, S. Kondo, A. Ito, Y. Furukawa, M. Higuchi, T. Kinoshita, I.
Kwon, Biomacromolecules 2011, 12, 3444–3452.
[15] K. Aoi, K. Ito, M. Okada, Macromolecules 1995, 28, 5391–5393.
[16] K. Aoi, K. Tsutsumiuchi, A. Yamamoto, M. Okada, Tetrahedron 1997, 53,
15415–15427.
[17] T. Hasegawa, S. Kondoh, K. Matsuura, K. Kobayashi, Macromolecules
1999, 32, 6595–6603.
[31] A. H. Rosenberg, B. N. Lade, D. Chui, S. Lin, J. J. Dunn, F. W. Studier, Gene
1987, 56, 125–135.
[32] F. W. Studier, A. H. Rosenberg, J. J. , Dunn, J. W. Dubendorff, Meth. Enzy-
mol. 1990, 185, 60–89.
[33] E. R. Welsh, D. A. Tirrell, Biomacromolecules 2000, 1, 23–30.
[34] A. Panitch, T. Yamaoka, M. J. Fournier, T. L. Mason, D. A. Tirrell, Macromo-
lecules 1999, 32, 1701–1703.
[18] K. Matsuura, M. Hibino, Y. Yamada, K. Kobayashi, J. Am. Chem. Soc.
2001, 123, 357–358.
[19] A. Takasu, T. Houjyou, Y. Inai, T. Hirabayashi, Biomacromolecules 2002, 3,
775–782.
[35] S. C. Heilshorn, K. A. DiZio, E. R. Welsh, D. A. Tirrell, Biomaterials 2003,
24, 4245–4252.
[20] S. Horikoshi, A. Takasu, T. Houjyou, Y. Inai, T. Hirabayashi, Polym. Prepr.
Am. Chem. Soc. Div. Polym. Chem. 2002, 43, 1110 – 1111.
[21] A. Takasu, S. Horikoshi, T. Hirabayashi, Biomacromolecules 2005, 6,
2334–2342.
[22] A. Takasu, T. Makino, T. Hirabayashi, J. Polym. Sci. Part A 2009, 47, 310–
314.
[36] P. J. Nowatzki, D. A. Tirrell, Biomaterials 2004, 25, 1261–1267.
[37] J. C. Liu, S. C. Heilshorn, D. A. Tirrell, Biomacromolecules 2004, 5, 497–
504.
[38] J. C. Liu, D. A. Tirrell, Biomacromolecules 2008, 9, 2984–2988.
[39] M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P. Stewart, J. Am. Chem.
Soc. 1985, 107, 3902–3909.
[23] A. Takasu, H. Kojima, J. Polym. Sci. Part A 2010, 48, 5953–5960.
[24] S. Futaki, T. Suzuki, W. Ohashi, T. Yagami, S. Tanaka, K. Ueda, Y. Sugiura,
J. Biol. Chem. 2001, 276, 5836–5840.
[40] MOPAC 2000, version 1.0, Fujitsu Ltd., Tokyo, Japan, 1999.
Received: April 12, 2013
[25] T. Suzuki, S. Futaki, M. Niwa, S. Tanaka, K. Ueda, Y. Sugiura, J. Biol. Chem.
2002, 277, 2437–2443.
Published online on May 23, 2013
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 677 – 683 683