Incorporation of fluorescent-labeled non-α-amino carboxylic acids into the N-terminus of proteins in response to amber initiation codon

Article abstract of DOI:10.1246/bcsj.20090320

Incorporation of non-natural amino acid derivatives containing fluorescent groups into proteins is a useful method for protein analyses. Here, we investigated the incorporation of fluorescent-labeled non-α-amino carboxylic acids into the N-terminus of proteins in response to the UAG initiation codon. A series of TAMRA-labeled amino carboxylic acids were synthesized and attached to an amber suppressor initiator tRNA derived from Escherichia coli initiator tRNA. Fluorescent-labeled amino carboxylic acids were successfully incorporated into the N-terminus of streptavidin by adding TAMRA-acylated initiator tRNAs to an E. coli cell-free translation system, although the incorporation efficiency differed depending on the amino carboxylic acid chain length. Based on this observation, 3-aminopropionic acid derivatives labeled with BODIPY, rhodamine, and cyanine fluorophores were designed and synthesized. The fluorescent-labeled 3-aminopropionic acid derivatives developed using BODIPY and rhodamine dyes could be incorporated with good efficiency. On the other hand, 6-aminohexanoic acid was effectively incorporated when labeled with cyanine dyes. The present study demonstrates that translation initiation can accept a wide variety of non-natural substrates and provides a useful method for N-terminal-specific labeling of proteins with various fluorophores.

Full text of DOI:10.1246/bcsj.20090320

546
Bull. Chem. Soc. Jpn. Vol. 83, No. 5, 546–553 (2010)
© 2010 The Chemical Society of Japan
Incorporation of Fluorescent-Labeled Non-¡-Amino Carboxylic Acids
into the N-Terminus of Proteins in Response to Amber Initiation Codon
Masanori Miura, Norihito Muranaka, Ryoji Abe, and Takahiro Hohsaka*
School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi 923-1292
Received December 4, 2009; E-mail: hohsaka@jaist.ac.jp
Incorporation of non-natural amino acid derivatives containing fluorescent groups into proteins is a useful method
for protein analyses. Here, we investigated the incorporation of fluorescent-labeled non-¡-amino carboxylic acids into
the N-terminus of proteins in response to the UAG initiation codon. A series of TAMRA-labeled amino carboxylic acids
were synthesized and attached to an amber suppressor initiator tRNA derived from Escherichia coli initiator tRNA.
Fluorescent-labeled amino carboxylic acids were successfully incorporated into the N-terminus of streptavidin by adding
TAMRA-acylated initiator tRNAs to an E. coli cell-free translation system, although the incorporation efficiency differed
depending on the amino carboxylic acid chain length. Based on this observation, 3-aminopropionic acid derivatives
labeled with BODIPY, rhodamine, and cyanine fluorophores were designed and synthesized. The fluorescent-labeled
3-aminopropionic acid derivatives developed using BODIPY and rhodamine dyes could be incorporated with good
efficiency. On the other hand, 6-aminohexanoic acid was effectively incorporated when labeled with cyanine dyes. The
present study demonstrates that translation initiation can accept a wide variety of non-natural substrates and provides a
useful method for N-terminal-specific labeling of proteins with various fluorophores.
Fluorescence labeling is a very useful method for analyzing
incorporation in response to the AUG initiation codon competes
with the incorporation of methionine or formylated methionine
by endogenous initiator tRNA, which decreases the labeling
yield. In contrast, using an amber codon UAG, originally a stop
codon, as an initiation codon has helped to achieve quantita-
tive labeling without competition with endogenous initiator
tRNA.18-20 Alternatively, a reconstituted cell-free translation
system without addition of methionine has been used for N-
terminal-specific incorporation of non-natural amino acids and
peptide derivatives.21 The initiation codon-mediated method
can achieve rapid and quantitative N-terminal-specific labeling,
independent of N-terminal amino acid sequence. These proper-
ties are advantageous compared to other N-terminal-specific
labeling methods such as those using protein splicing,22,23
leucyl/phenylalanyl-tRNA-protein transferase,24 and sortase.25
In addition to ¡-amino acid derivatives, we reported the
incorporation of BODIPYFL-labeled 3-aminopropionic and
6-aminohexanoic acids without ¡-amino groups26 and p-
aminophenylalanine derivatives double-labeled with both
BODIPYFL and biotin27 into the N-terminus of proteins in
response to the UAG initiation codon. These results suggest
that various non-natural molecules are accepted as substrates
during translation initiation. We found that p-substituted
phenylalanine derivatives are efficient substrates for peptide
elongation.5 Based on this substrate specificity, p-aminophen-
ylalanine derivatives labeled with fluorophores,6,7 biotin,28 and
poly(ethylene glycol) chains29 at the p-amino groups have been
designed and successfully incorporated into proteins. As
with translation elongation, it will be worthwhile to identify
substrate specificities during translation initiation in order to
design non-natural molecules that can be efficiently incorpo-
rated into the N-terminus of proteins.
protein structure and function. Chemical modification using
fluorescent molecules has generally been used for protein
fluorescence labeling. However, most proteins have multiple
reactive residues such as lysine and cysteine on their surfaces,
and therefore site-directed fluorescence labeling is not easily
achieved. Chemical modification of engineered cysteines
facilitates site-directed labeling, although quantitative labeling
remains difficult. Fusion gene expression with green fluores-
cent protein (GFP) derivatives is an alternative method for site-
directed fluorescence labeling, but the large molecular size of
these derivatives may influence the structure and function of
fused proteins.
On the other hand, site-directed and quantitative incorpo-
ration of fluorescent groups into proteins using non-natural
amino acid mutagenesis1-4 has been developed. Various
fluorescent non-natural amino acids have been incorporated
in response to an amber codon and four-base codons.5-12 This
method enables the introduction of fluorescent groups at
desired positions. However, incorporation into inappropriate
positions may significantly damage proteins. N- or C-Terminal-
specific fluorescence labeling is desirable as a general method
available for labeling a wide variety of proteins without
causing significant damage.
Ribosome-mediated N-terminal-specific labeling has been
developed using initiator tRNAs. Initiator Met-tRNAs whose
¡-amino groups are acylated with non-natural groups can
introduce non-natural methionine derivatives into the N-
terminus of proteins using cell-free translation systems.13 Using
this method, methionine derivatives that have various fluoro-
phores14-16 and biotin17 labeled at their ¡-amino groups
have been successfully incorporated. However, the N-terminal
Published on the web April 29, 2010; doi:10.1246/bcsj.20090320

M. Miura et al.
Bull. Chem. Soc. Jpn. Vol. 83, No. 5 (2010)
547
O
O
NH
O
Cell-free
translation
AUC
UAG
Figure 1. Schematic illustration of the incorporation of fluorescent-labeled non-¡-amino carboxylic acids into the N-terminus of
proteins in response to a UAG initiator codon.
In this study, we investigated the incorporation of various
fluorescent-labeled non-¡-amino carboxylic acids into the
N-terminus of proteins in response to the UAG initiation
codon in an E. coli cell-free translation system (Figure 1).
Tetramethylrhodamine (TAMRA)-labeled amino carboxylic
acids containing various alkyl chains were synthesized and
their incorporations into the N-terminus of proteins were
evaluated. In addition, the incorporation of amino carboxylic
acid derivatives labeled with various fluorophores was shown
to be useful for the N-terminal-specific fluorescence labeling of
proteins.
Boc-amino carboxylic acid cyanomethyl ester. Acylation of 5¤-
O-phosphoryl-2¤-deoxycytidylyl(3¤-5¤)adenosine (pdCpA) was
carried out by adding Boc-amino carboxylic acid cyanomethyl
ester (8.8 ¯mol) to 44 mM DMF solution of pdCpA tetrabutyl-
ammonium salt (15 ¯L, 0.66 ¯mol). The resulting solution was
incubated at room temperature for 1 h. Diethyl ether (600 ¯L)
was added to the solution, and the precipitate was collected by
centrifugation. The resulting precipitate was washed twice with
diethyl ether and dried under vacuum. The pellet was dissolved
in trifluoroacetic acid (TFA; 200 ¯L) and placed on ice for
15 min to remove the Boc-group. After evaporating TFA by
vacuum centrifugation, the pellet was washed twice with
diethyl ether (600 ¯L) and dried under vacuum.
Experimental
Materials. Succinimide esters of TAMRA, AlexaFluor488,
BODIPYFL, BODIPY558, 5-carboxyrhodamine110 (CR110),
and 5-carboxyfluorescein (FAM) were purchased from
Molecular Probes (Eugene, OR, USA). Cy3 and Cy5 mono-
reactive dyes were from GE Healthcare (Piscataway, NJ, USA).
T4 RNA ligase was from Takara BIO (Osaka, Japan). E. coli
S30 extract for a linear template, alkaline phosphatase-labeled
anti-mouse IgG, and MagneHis Ni-Particles were from
Promega (Madison, WI, USA). RTS E. coli linear tem-
plate generation set was from Roche Diagnostics (Basel,
Switzerland). Anti-His tag antibody was from Novagen (La
Jolla, CA, USA). Lysyl endopeptidase was from Wako
Chemicals (Osaka, Japan). XTerra C18 and XBridge C18
columns were from Waters (Milford, MA, USA). Poros R2/10
column was from Applied Biosystems (Foster, CA, USA).
ZipTip C18 was from Millipore (Milford, MA, USA).
Synthesis of TAMRA-Labeled Aminocarboxyl-Dinucleo-
tides. Aqueous 0.05 M NaHCO3 (60 ¯L) was added to a
mixture of 2.2 mM dimethylsulfoxide (DMSO) solution of
aminocarboxyl-pdCpA (30 ¯L, 66 nmol), 50 mM 5-TAMRA
succinimide ester (3 ¯L, 150 nmol), and DMSO (27 ¯L). The
mixture was incubated on ice for 3 h and neutralized by adding
1 M acetic acid (4.5 ¯L). TAMRA-labeled aminocarboxyl-
pdCpA was purified using reverse-phase HPLC (XBridge C18,
2.5 ¯m, 4.6 © 20 mm, 1.5 mL min¹1, 0-100% methanol in
0.38% formic acid, over 10 min). The product was identified
by electrospray ionization time-of-flight mass spectrometry
(ESI-TOF MS) (Mariner, Applied Biosystems). TAMRA-AC2-
¹
pdCpA, calculated for (M ¹ H) 1104.265, found 1104.322;
¹
TAMRA-AC3-pdCpA, calculated for (M ¹ H) 1118.281,
found 1118.236; TAMRA-AC4-pdCpA, calculated for (M ¹
¹
H) 1132.297, found 1132.209; TAMRA-AC5-pdCpA, calcu-
¹
Synthesis of Aminocarboxyl-Dinucleotides.
Dioxane
lated for (M ¹ H) 1146.312, found 1146.326; TAMRA-AC6-
¹
solution of (Boc)2O (Boc = t-butyloxycarbonyl) (1.5 equiv,
1.68 mmol) (200 ¯L) was gradually added to a mixture of
amino carboxylic acid (1.12 mmol) and NaHCO3 (2 equiv,
2.24 mmol) in water/dioxane (1:1 v/v, 4 mL) on ice. The
mixture was stirred for 18 h on ice, acidified to pH 2 with
aqueous 5% KHSO4, and extracted with ethyl acetate. The
extract was washed three times with aqueous 5% KHSO4, once
with saturated aqueous NaCl, and dried over sodium sulfate.
The solvent was evaporated to give Boc-amino carboxylic
acid. To a mixture of the resulting Boc-amino carboxylic acid
and triethylamine (500 ¯L) in acetonitrile (2 mL), chloroaceto-
nitrile (200 ¯L) was gradually added on ice. The mixture was
stirred for 18 h on ice, acidified to pH 2 with aqueous 5%
KHSO4, and extracted with ethyl acetate. The extract was
washed three times with aqueous 5% KHSO4, three times with
aqueous 4% NaHCO3, once with saturated aqueous NaCl, and
dried over sodium sulfate. The solvent was evaporated to give
pdCpA, calculated for (M ¹ H) 1160.328, found 1160.322;
¹
TAMRA-AC7-pdCpA, calculated for (M ¹ H) 1174.344,
found 1174.290; TAMRA-AC12-pdCpA, calculated for (M ¹
¹
H) 1244.422, found 1244.437; TAMRA-X-AC12-pdCpA;
¹
calculated for (M ¹ H) 1357.506, found 1357.505; TAMRA-
¹
X-Met-pdCpA, calculated for (M ¹ H) 1291.369, found
¹
1291.321; TAMRA-X-AF-pdCpA, calculated for (M ¹ H)
1322.407, found 1322.418.
Aminopropionyl-pdCpA
derivatives
labeled
with
AlexaFluor488, BODIPYFL, BODIPY558, CR110, FAM,
Cy3, and Cy5, and aminohexanoyl-pdCpA derivatives
labeled with Cy3 and Cy5 were synthesized in a similar
manner using the corresponding succinimide esters. ESI-TOF
¹
MS; AlexaFluor488-AC3-pdCpA, calculated for (M ¹ H)
1222.133, found 1222.129; BODIPYFL-AC3-pdCpA, calcu-
¹
lated for (M ¹ H) 980.2476, found 980.2482; BODIPY558-
¹
AC3-pdCpA, calculated for (M ¹ H) 1118.281, found

548
Bull. Chem. Soc. Jpn. Vol. 83, No. 5 (2010)
Protein Labeling with Amino Carboxylic Acids
¹
1118.293; FAM-AC3-pdCpA, calculated for (M ¹ H)
Mass Analysis.
The incorporation of TAMRA-AC3,
TAMRA-AC4, and CR110-AC3 was confirmed by matrix-
assisted laser desorption ionization time-of-flight mass spec-
trometry (MALDI-TOF MS). The cell-free translation reaction
mixture (50 ¯L) was loaded onto the MagneHis Ni-particles
(20 ¯L). The beads were washed five times with 200 ¯L of
wash buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl), and the
His-tagged protein was eluted with 25 ¯L of elution buffer
(10 mM Tris-HCl, pH 7.5, 500 mM imidazole). The eluent was
mixed with 0.01 units of lysyl endopeptidase and incubated at
30 °C for 6 h. The resulting peptide fragments were desalted
and concentrated using ZipTip C18 and eluted with a matrix
solution containing saturated ¡-cyano-4-hydroxycinnamic acid
in 1:1 mixture of acetonitrile and 0.1% TFA. Mass measure-
ment was performed by MALDI-TOF MS (Voyager DE-Pro,
Applied Biosystems) in the positive mode using angiotensin II
as an external calibrant.
1064.187, found 1064.202; Cy3-AC3-pdCpA, calculated for
(M ¹ H) 1318.335, found 1318.305; Cy5-AC3-pdCpA, cal-
culated for (M ¹ H) 1344.351, found 1344.387; Cy3-AC6-
pdCpA, calculated for (M ¹ H) 1361.311, found 1361.320;
¹
¹
¹
¹
Cy5-AC6-pdCpA, calculated for (M ¹ H) 1387.349, found
1387.284.
Preparation of Aminocarboxyl-tRNAs.
An E. coli
initiator tRNA containing a CUA anticodon and lacking the 3¤-
terminal dinucleotide was synthesized as described previous-
ly.26 Ligation of the truncated initiator tRNA and fluorophore-
labeled aminocarboxyl-pdCpAs was carried out in a reaction
mixture (10 ¯L) containing 5.5 mM HEPES-Na (pH 7.5),
1 mM ATP, 15 mM MgCl2, 3.3 mM dithiothreitol (DTT),
¹1
2 ¯g mL
bovine serum albumin (BSA), 0.25 mol tRNA,
2.2 nmol of fluorescent-labeled aminocarboxyl-pdCpA in
DMSO (1 ¯L), and T4 RNA ligase (30 units). The reaction
mixture was incubated at 4 °C for 12 h, except for TAMRA-X-
AF which was incubated for 2 h in order to prevent hydrolysis
of the ester bond between the amino acid and tRNA. After
incubation, potassium acetate (pH 4.5) was added to a final
concentration of 0.3 M. The tRNA was isolated by ethanol
precipitation and dissolved in pre-chilled 1 mM potassium
acetate (pH 4.5). The acylation yield was determined using
reverse-phase HPLC (Poros R2/10, 4.6 © 100 mm), flow rate =
Results and Discussion
Synthesis of TAMRA-Labeled Aminocarboxyl-tRNAs.
To investigate the effects of linker structures on the incorpo-
ration of fluorescent-labeled amino carboxylic acids, a variety
of TAMRA-labeled amino carboxylic acids were designed as
shown in Figure 2. TAMRA was chosen as a model fluoro-
phore because of good characteristic properties, such as high
fluorescence quantum yield, long emission wavelength, and
high photostability. TAMRA-ACn, where n is 2-8 and 12,
contained alkyl linkers of n carbon atoms. TAMRA-X-AC12
contained an aminohexyl linker in addition to an AC12 linker.
TAMRA-EG8 containing an ethylene oxide linker of 8 atoms
was also used to examine the influence of linker flexibility. In
addition, methionine (TAMRA-Met and TAMRA-X-Met) and
p-aminophenylalanine derivatives (TAMRA-X-AF) labeled at
the ¡- and p-amino groups, respectively, were also examined.
TAMRA-labeled amino carboxylic acids were attached to a
dinucleotide (pdCpA) to prepare acyl-tRNAs by a chemical
aminoacylation method.30,31 For this purpose, aminocarboxyl-
pdCpA derivatives were first synthesized via cyanomethyl
esters after which they were coupled with TAMRA succinimide
ester. This synthetic route was advantageous for increasing the
reaction yield using expensive fluorophore succinimide esters
and synthesizing aminocarboxyl-pdCpAs labeled with various
fluorophores. All products were purified by HPLC, identified
by ESI-TOF MS, and ligated with a synthetic E. coli initiator
tRNA containing the CUA anticodon.26 HPLC analysis of the
ligated tRNAs using a Poros R2/10 column16 showed that
the ligation reaction efficiencies, independent of the amino
carboxylic acid types used, were about 80%.
¹1
1.0 mL min with a linear gradient of 0-40% acetonitrile in
0.1 M triethylammonium acetate (pH 7.0) over 20 min.
Cell-Free Translation.
Streptavidin mRNA containing
the UAG initiation codon and His-tag at N- and C-termini,
respectively, was prepared as described previously.26 The
mRNA and each fluorescent-labeled aminocarboxyl-tRNA
were added to an E. coli cell-free translation system. Trans-
lation was performed in a reaction mixture (10 ¯L) containing
55 mM HEPES-KOH (pH 7.5), 210 mM potassium glutamate,
6.9 mM ammonium acetate, 12 mM magnesium acetate,
1.7 mM DTT, 1.2 mM ATP, 0.28 mM GTP, 26 mM phosphor-
(enol)pyruvate, 1 mM spermidine, 1.9% poly(ethylene glycol),
¹1
35 ¯g mL folinic acid, 0.1 mM of 20 standard amino acids,
8 ¯g mRNA, 80 pmol fluorescent-labeled aminocarboxyl-
tRNA, and an E. coli S30 extract (2 ¯L). The reaction mixture
was incubated at 37 °C for 60 min. The reaction mixture
(10 ¯L) was then mixed with 10 ¯g of RNase A (10 ¯L) and
incubated at 37 °C for 15 min to digest the remaining acyl-
tRNA. The resulting solution (2 ¯L) was mixed with 2©
sample buffer for sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE) (10 ¯L) and water (8 ¯L), and
the sample (5 ¯L) was subjected to 20% SDS-PAGE after
boiling. The fluorescent bands on the SDS-PAGE gel were
visualized and quantified using
a
fluorescence scanner
Incorporation of TAMRA-Labeled Amino Carboxylic
Acids into the N-Terminus of Streptavidin. The incorpo-
ration of TAMRA-labeled amino carboxylic acids into proteins
was examined using a streptavidin gene containing the UAG
initiation codon and His-tag at N- and C-termini, respectively.
When the UAG initiation codon is decoded by the initiator
tRNA acylated with TAMRA-labeled amino carboxylic acids,
TAMRA-containing streptavidin will be obtained. On the other
hand, when TAMRA-labeled amino carboxylic acids are not
incorporated into the protein, no proteins will be produced
because the translation initiation does not occur. Therefore, the
(FMBIO-III, Hitachi Software Engineering, Tokyo, Japan).
The same gel was analyzed by Western blot using an anti-His
tag antibody and alkaline phosphatase-labeled anti-mouse IgG.
The band intensities were quantified using Scion Image
program (Scion Corporation, Frederick, Maryland, USA), from
which the yields of the full-length streptavidin were determined
using a calibration curve prepared by measuring the band
intensities of serial dilutions (100, 50, 25, 12.5, 6.25, and
3.13%) of the full-length streptavidin expressed from a wild-
type mRNA.

M. Miura et al.
Bull. Chem. Soc. Jpn. Vol. 83, No. 5 (2010)
549
O
O
H
N
H
TAMRA−AC2
TAMRA−AC3
TAMRA−AC4
TAMRA−AC5
TAMRA−AC6
TAMRA−AC7
TAMRA−AC8
TAMRA−AC12
TAMRA−X−AC12
TAMRA−EG8
TAMRA−Met
N
O
OH
O
R
R
R
R
O
OH
O
S
H
N
OH
N
H
OH
O
H
N
O
S
H
R
R
OH
O
R
N
TAMRA−X−Met
N
OH
H
O
N
H
OH
O
TAMRA−X−AF
O
H
N
O
OH
H
N
R
R
OH
O
NH2
R
N
H
N
H
OH
O
R =
O
H
N
COOH
R
R
OH
O
H
N
(CH3)2N
O
N(CH3)2+
OH
O
H
R
N
N
OH
H
O
Figure 2. Structures of TAMRA-labeled non-¡-amino carboxylic acids and ¡-amino acids.
incorporation of TAMRA-labeled amino carboxylic acids can
be evaluated by judging the expression level of the TAMRA-
containing streptavidin.
the increase in the alkyl chain length. These results supported
that the TAMRA-labeled amino carboxylic acids were evi-
dently incorporated into the N-terminus of the protein. The
mass spectra also showed weak peaks with mass values lower
by 14 and 44 than the TAMRA-containing peptides, which
could be identified as TAMRA-containing peptides lacking a
methyl group and a carboxyl group of TAMRA, respectively,
because TAMRA succinimide ester also gave weak peaks
with mass values lower by 14 and 44 possibly due to
photodegradation of TAMRA during the laser-ionization in
the MALDI-TOF MS measurement.
Fluorescence band intensities were quantified to compare the
incorporation efficiencies of TAMRA-labeled amino carboxylic
acids (Figure 3b), assuming that the fluorescence intensity was
not affected by the type of amino carboxylic acids. The
incorporation efficiencies differed depending on the structure
and chain length of amino carboxylic acids. From TAMRA-
AC3 to AC7, the incorporation efficiency decreased with
increasing chain length. AC12 and X-AC12 derivatives also
showed decreasing efficiencies with increasing chain length.
However, TAMRA-AC8 showed an inordinately high effi-
ciency. EG8 derivative containing the same 8-atom chain as
AC8 was incorporated with an efficiency comparable to AC8,
suggesting that linker flexibility was not a significant factor.
TAMRA-Met and X-Met derivatives showed higher incorpo-
ration efficiencies than most of the amino carboxylic acid
derivatives.
The streptavidin mRNA was expressed in an E. coli cell-free
translation system in the presence of TAMRA-labeled amino-
carboxyl-tRNAs. Cell-free translation products were separated
by SDS-PAGE and visualized by fluorescence imaging, and
the same gel was analyzed by Western blot using anti-His
tag antibody. The fluorescence image showed that fluorescent
bands around 17 kDa, which corresponded to monomeric
streptavidin, were observed for all TAMRA-labeled amino
carboxylic acids (Figure 3a), though TAMRA-AC2 gave a very
faint band. The Western blot also showed that full-length
translation products were obtained in the presence of TAMRA-
labeled aminocarboxyl-tRNAs. On the other hand, no bands
were observed in the absence of the amber suppressor initiator
tRNA, indicating that the UAG initiation codon was not
decoded by any tRNA other than the added initiator tRNAs.
These results indicated that the streptavidins obtained in
the presence of TAMRA-labeled aminocarboxyl-tRNAs were
quantitatively labeled with TAMRA at the N-terminus.
The translation products obtained in the presence of
TAMRA-AC3-tRNA and TAMRA-AC4-tRNA were analyzed
by MALDI-TOF MS after purification on Ni-NTA beads and
digestion with lysyl endopeptidase. As expected, a mass peak
corresponding to TAMRA-AC3-containing N-terminal pep-
tides (TAMRA-AC3-Asp-Pro-Ser-Lys; calculated for MH+
929.4, found 929.5) was observed (Figure 4a). For TAMRA-
AC4, a mass value higher by 14 was observed (calculated for
MH+ 943.4, found 943.5; Figure 4b), which corresponded to
The relationship between the structure of amino carboxylic
acids and incorporation efficiency suggests the presence of a
binding space for TAMRA near the tRNA, for derivatives with

550
Bull. Chem. Soc. Jpn. Vol. 83, No. 5 (2010)
Protein Labeling with Amino Carboxylic Acids
TAMRA−AC3−Asp−Pro−Ser−Lys
a)
a)
(kDa)
39.0 -
29.0 -
20.1 -
14.2 -
6.5 -
*
**
47.5 -
32.5 -
25.0 -
800
800
800
880
960
1040
1040
1040
1120
1120
1120
1200
1200
1200
m z-1
16.5 -
6.5 -
b)
TAMRA−AC4−Asp−Pro−Ser−Lys
b)
3
2
1
0
*
**
880
960
m z-1
Figure 3. Incorporation of TAMRA-labeled amino car-
boxylic acids into the N-terminus of streptavidin in
response to the UAG initiation codon. (a) Fluorescence
image of 20% SDS-PAGE gel with excitation at 532 nm
and emission at 580 nm, and Western blot analysis of the
same gel using anti-His tag antibody. (b) Fluorescence
band intensities for TAMRA-labeled streptavidin observed
in the SDS-PAGE gel fluorescence image. Results are
mean « SDs of three assays.
c)
CR110−AC3−Asp−Pro−Ser−Lys
*
less than 8-carbon chains. In TAMRA-AC8, EG8, and much
longer derivatives, TAMRA may be located at another binding
space far from the tRNA or may be surface-exposed. The
incorporation of TAMRA-X-AC12 with a 19-atom chain
revealed that very large non-natural molecules are potentially
acceptable substrates during ribosomal initiation.
The highly efficient incorporation of TAMRA-Met and
X-Met may have been due to the binding of the methionine
side chain to the binding site originally meant for formyl-
methionine. Nonetheless, TAMRA-AC3 and AC8 showed
nearly comparable efficiencies for the methionine derivatives
with similar length linkers, suggesting that fluorescent-labeled
amino carboxylic acids can be efficiently incorporated without
contributions from the binding of the methionine side chain and
¡-amino or amide groups.
880
960
m z-1
Figure 4. MALDI-TOF MS measurements of the lysyl
endopeptidase-digested N-terminal peptide fragments
(Asp-Pro-Ser-Lys) labeled with TAMRA-AC3 (calculated
for MH+ 929.4, observed 929.6), TAMRA-AC4 (calcu-
lated 943.4 for MH+, found 943.5), and CR110-AC3
(calculated 873.3 for MH+, found 873.5). Peaks indicated
by asterisks (Ã and ÃÃ) could be identified as labeled
peptide fragments lacking a carboxyl group and a methyl
group, respectively, which were generated possibly due to
photodegradation of the fluorophores during the laser-
ionization.
The inefficient incorporation of TAMRA-AC2 may be due to
the short alkyl groups without side-chains being inappropriate
for TAMRA to be accepted by the translation machinery. In
addition, the ¡-amide bond may have destabilized the amino-
acyl linkage between TAMRA-AC2 and tRNA as suggested
in previous study,26 and enhanced the hydrolysis of this
linkage. To evaluate the stability of the ester bond between the
TAMRA-labeled amino carboxylic acids and the initiator
tRNA, the hydrolysis of TAMRA-AC2- and TAMRA-AC3-
tRNAs in the same pH and temperature conditions (pH 7.5 at
37 °C) as for the cell-free translation was monitored using
reverse-phase HPLC (Figure 5). The hydrolysis yields of
TAMRA-AC2-tRNA were 10% at 1 h and 95% at 24 h, while
TAMRA-AC3-tRNA showed negligible hydrolysis at 1 h and
36% hydrolysis at 24 h. These results showed that the ¡-amide
bond evidently destabilized the ester bond. However, the result
that 90% of TAMRA-AC2-tRNA remained after 1 h was not
enough to explain the inefficient incorporation of TAMRA-
AC2. Therefore, the destabilized ester bond may only partially
contribute to the inefficient incorporation of TAMRA-AC2.

M. Miura et al.
Bull. Chem. Soc. Jpn. Vol. 83, No. 5 (2010)
551
The concentration dependence of incorporation efficiencies
of TAMRA-labeled aminocarboxyl-tRNAs was examined.
Nearly the same amounts of TAMRA-labeled streptavidin
were obtained for all derivatives in the presence of 40,
80, and 160 pmol of TAMRA-labeled aminocarboxyl-tRNAs
in 10 ¯L of cell-free translation, indicating that tRNA con-
centration (80 pmol per 10 ¯L of cell-free translation) was
sufficient.
100
80
60
40
Incorporation of 3-Aminopropionic Acids with Various
20
Fluorophores.
We reported that BODIPYFL-labeled
3-aminopropionic acid is efficiently incorporated into the
N-terminus of proteins. The present study also showed that
3-aminopropionic acid is an effective framework for incorpo-
rating TAMRA. These results suggested that the aminopro-
pionic acid framework would be effective for introducing
various fluorophores. To demonstrate this, 3-aminopropionic
acid derivatives labeled with various fluorophores, such as
AlexaFluor488, 5-carboxyrhodamine110 (CR110), 5-carboxy-
fluorescein (FAM), BODIPY dyes, and Cy dyes (Figure 6a),
were synthesized and their incorporations into the N-terminus
of streptavidin were examined.
0
0
6
12
Time / h
18
24
Figure 5. Time course of the hydrolysis of TAMRA-AC2-
tRNA (open circle) and TAMRA-AC3-tRNA (filled circle)
in 10 mM Tris-HCl (pH 7.5) at 37 °C. The remaining
TAMRA-tRNAs were monitored by reverse-phase HPLC.
R
R
a)
O
O
O
R
O
R
S
F
F
F
F
B
B
N
N
N
N
N+
N+
N
R
N
-O3S
-O3S
BODIPY558
O
BODIPYFL
SO3-
SO3-
Cy3
Cy5
O
O
R
R
O
COOH
R =
COOH
N
H
OH
COOH
H2N
O
NH2+
H2N
O
NH2+
HO
O
O
SO3-
SO3-
FAM
CR110
AlexaFluor488
b)
c)
25
20
15
10
5
Ex / Em
(nm)
488 / 520
532 / 580
0
635 / 670
anti-His tag
Figure 6. Incorporation of various fluorescent-labeled 3-aminopropionic acids into the N-terminus of streptavidin in response to the
UAG initiation codon. (a) Structures of fluorescent-labeled 3-aminopropionic acids. (b) Fluorescence image of 20% SDS-PAGE gel
with excitation and emission at 488 and 520 nm (green), 532 and 580 nm (red), and 635 and 670 nm (blue), and Western blot of the
same SDS-PAGE gel using anti-His tag antibody. (c) Incorporation efficiencies of the fluorescent-labeled 3-aminopropionic acids
into the N-terminus of streptavidin determined by comparing band intensities on Western blot with those of serial dilutions of wild-
type streptavidin. Results are mean « SDs of three assays.

552
Bull. Chem. Soc. Jpn. Vol. 83, No. 5 (2010)
Protein Labeling with Amino Carboxylic Acids
Fluorescence images of SDS-PAGE showed that fluores-
cent-labeled 3-aminopropionic acids, other than Cy3 deriva-
tive, were successfully incorporated into the N-terminus of
streptavidin (Figure 6b). It should be noted that smear bands
were observed for BODIPYFL and BODIPY558 at the
lower molecular-weight range on the SDS-PAGE, which were
derived from low-molecular-weight BODIPY-labeled amino-
propionic acids due to the tendency of the hydrophobic
BODIPY dyes to adsorb to the polyacrylamide gel. For
AlexaFluor488 and FAM, weak bands were observed at the
lower molecular-weight range, which may be generated by any
irregular termination in the peptide elongation process. A
similar truncated polypeptide had been observed on Western
blotting using an anti-T7 tag antibody for wild-type streptavi-
din containing T7 tag at the N-terminus.5 The truncated by-
products generated in the peptide elongation process would not
affect the incorporation of the aminocarboxylic acids in the
initiation process and, in addition, could be removed by
purification using Ni chelate affinity chromatography. How-
ever, the generation of the truncated by-products slightly
decreased the yields of the full-length proteins. From the
fluorescence band intensities, the relative yields of the
truncated polypeptides were estimated to be 5 and 8%
compared to the full-length proteins for AlexaFluor488 and
FAM, respectively. Optimization of the nucleotide sequence
may be effective to suppress the irregular termination of the
peptide elongation.
MALDI-TOF MS analysis of the lysyl endopeptidase-
digested translation product for CR110-AC3 showed that the
expected mass peak was evidently observed (calculated for
MH+ 873.3, found 873.5; Figure 4c) as in the case of TAMRA.
The mass spectrum also showed a weak peak with mass value
lower by 44, which could be identified as a CR110-containing
peptide lacking a carboxyl group of CR110 due to photo-
degradation during the laser-ionization. Unlike the case of
TAMRA, CR110, which contained no methyl group in the
fluorophore, gave no peak with mass value lower by 14,
supporting the idea that the mass value lower by 14 observed
for TAMRA is due to the lack of the methyl group.
TAMRA Cy3
Cy5
Ex / Em
(nm)
532 / 580
635 / 670
anti-His tag
Figure 7. Incorporation of TAMRA-, Cy3-, and Cy5-
labeled 3-aminopropionic acid (AC3) and 6-aminohexanoic
acid (AC6) derivatives into the N-terminus of streptavidin
in response the UAG initiation codon. Fluorescence image
of 20% SDS-PAGE gel with excitation and emission at
532 and 580 nm (red) and 635 and 670 nm (blue), and
Western blot of the same SDS-PAGE gel using anti-His tag
antibody.
carboxyl chain length. To test this possibility, Cy3- and
Cy5-labeled 6-aminohexanoic acid derivatives (Cy3- and
Cy5-AC6) were synthesized and their incorporations were
investigated.
A fluorescence image of SDS-PAGE and Western blotting
showed that the incorporations of Cy3 and Cy5 were enhanced
by attaching them to the aminohexanoic acid (Figure 7), and
the incorporation efficiencies estimated by Western blot were
about 3%. The aminohexyl chain may be effective in delivering
Cy dyes to the above-mentioned binding space which was
far from the acyl-tRNAs or to surface-exposed sites. This
finding will be valuable for designing amino carboxylic acids
containing large fluorophores.
Incorporation into Other Proteins. The incorporation of
TAMRA-AC3 was examined for proteins other than streptavi-
din. As shown in Figure 8, TAMRA-AC3 was successfully
incorporated into maltose-binding protein (MBP) and inter-
leukin 2 (IL2) in response to the UAG initiator codon,
suggesting that fluorescent-labeled non-¡-amino acids can be
incorporated into the N-terminus of proteins regardless of N-
terminal amino acid sequence. A weak band observed at the
lower molecular-weight range may be generated by any
irregular termination in the peptide elongation process as in
the case of the incorporation of AlexaFluor488 and FAM into
streptavidin.
The relative yields of the proteins were quantified by
comparing the band intensities on Western blot with those
of serial dilutions of wild-type (Figure 6c). Fluorescein and
rhodamine derivatives showed about 10-20% yields, and
TAMRA derivative showed the highest yield (22%). Cy5
derivative showed very low yield, whereas a distinct fluores-
cent band was observed, possibly due to its bright fluorescence.
These results indicate that various fluorescent groups are
accepted as substrates during translation initiation, and the
incorporation efficiency depended on the structure of the
fluorescent groups. In contrast, Cy3 and Cy5 derivatives were
incorporated with very low efficiencies, suggesting that the
large molecular size of Cy dyes may cause steric hindrance
with the ribosomal initiation complex.
Conclusion
The present study demonstrates that fluorescent-labeled non-
¡-amino carboxylic acids are successfully incorporated into the
N-terminus of proteins in response to the UAG initiation
codon. This indicates that translation initiation can accept
a wide variety of non-natural substrates. The results that
incorporation efficiency depends on the linker length of amino
carboxylic acids and structure of fluorophores suggest that
Influence of Chain Length on Incorporation of Cy
Dyes.
In the case of TAMRA, an 8-aminooctanoic acid
derivative (TAMRA-AC8) was incorporated more efficiently
than amino carboxylic acids with shorter alkyl chains. This
raised the possibility that Cy dyes, which originally have a
hexyl linker, may be incorporated by increasing the amino-

M. Miura et al.
Bull. Chem. Soc. Jpn. Vol. 83, No. 5 (2010)
553
Reference
C. J. Noren, S. J. Anthony-Cahill, M. C. Griffith, P. G.
Schultz, Science 1989, 244, 182.
J. D. Bain, E. S. Diala, C. G. Glabe, T. A. Dix, A. R.
Chemberlin, J. Am. Chem. Soc. 1989, 111, 8013.
T. Hohsaka, Y. Ashizuka, H. Murakami, M. Sisido, J. Am.
Chem. Soc. 1996, 118, 9778.
1
2
3
4
5
T. Hohsaka, Bull. Chem. Soc. Jpn. 2004, 77, 1041.
T. Hohsaka, D. Kajihara, Y. Ashizuka, H. Murakami, M.
Sisido, J. Am. Chem. Soc. 1999, 121, 34.
H. Murakami, T. Hohsaka, Y. Ashizuka, K. Hashimoto, M.
Sisido, Biomacromolecules 2000, 1, 118.
T. Hohsaka, N. Muranaka, C. Komiyama, K. Matsui, S.
6
7
Takaura, R. Abe, H. Murakami, M. Sisido, FEBS Lett. 2004, 560,
173.
8
D. Kajihara, R. Abe, I. Iijima, C. Komiyama, M. Sisido, T.
Hohsaka, Nat. Methods 2006, 3, 923.
V. W. Cornish, D. R. Benson, C. A. Altenbach, K. Hideg,
9
W. L. Hubbell, P. G. Schultz, Proc. Natl. Acad. Sci. U.S.A. 1994,
91, 2910.
10 G. Turcatti, K. Nemeth, M. D. Edgerton, U. Meseth, F.
Talabot, M. Peitsch, J. Knowles, H. Vogel, A. Chollet, J. Biol.
Chem. 1996, 271, 19991.
11 B. E. Cohen, T. B. McAnaney, E. S. Park, Y. N. Jan, S. G.
Boxer, L. Y. Jan, Science 2002, 296, 1700.
12 R. D. Anderson, III, J. Zhou, S. M. Hecht, J. Am. Chem.
Soc. 2002, 124, 9674.
13 W. Kudlicki, O. W. Odom, G. Kramer, B. Hardesty, J. Mol.
Biol. 1994, 244, 319.
14 V. Ramachandiran, C. Willms, G. Kramer, B. Hardesty,
J. Biol. Chem. 2000, 275, 1781.
15 B. McIntosh, V. Ramachandiran, G. Kramer, B. Hardesty,
Biochimie 2000, 82, 167.
16 S. Gite, S. Mamaev, J. Olejnik, K. Rothschild, Anal.
Biochem. 2000, 279, 218.
17 M. Taki, S. Y. Sawata, K. Taira, J. Biosci. Bioeng. 2001, 92,
149.
18 U. Varshney, U. L. RajBhandary, Proc. Natl. Acad. Sci.
U.S.A. 1990, 87, 1586.
19 S. Mamaev, J. Olejnik, E. K. Olejnik, K. J. Rothschild,
Anal. Biochem. 2004, 326, 25.
20 J. Olejnik, S. Gite, S. Mamaev, K. J. Rothschild, Methods
2005, 36, 252.
21 Y. Goto, H. Suga, J. Am. Chem. Soc. 2009, 131, 5040.
22 B. Schuler, L. K. Pannell, Bioconjugate Chem. 2002, 13,
1039.
23 V. Muralidharan, T. W. Muir, Nat. Methods 2006, 3, 429.
24 M. Taki, H. Kuroiwa, M. Sisido, ChemBioChem 2008, 9, 719.
25 T. Yamamoto, T. Nagamune, Chem. Commun. 2009, 1022.
26 N. Muranaka, M. Miura, H. Taira, T. Hohsaka,
ChemBioChem 2007, 8, 1650.
27 T. Egusa, K. Shiraga, T. Chiba, Y. Tokuda, I. Iijima, T.
Hohsaka, Bull. Chem. Soc. Jpn. 2010, 83, 176.
28 T. Watanabe, N. Muranaka, I. Iijima, T. Hohsaka, Biochem.
Biophys. Res. Commun. 2007, 361, 794.
Figure 8. Fluorescence image of SDS-PAGE gel for the
incorporation of TAMRA-AC3 into the N-terminus of
streptavidin, maltose-binding protein (MBP), and inter-
leukin 2 (IL2) with excitation at 532 nm and emission at
605 nm.
various non-natural molecules can be incorporated into the
N-terminus using appropriately designed structure of non-¡-
amino carboxylic acids. Relatively long chains would be
required for non-natural molecules that originally have long
alkyl chains, as in the case of Cy dyes.
This study also demonstrates that the initiation codon-
mediated method allows rapid, quantitative, and neighboring
sequence-independent labeling. These advantages are not
achieved by other chemical or enzymatic methods including
those using protein splicing, leucyl/phenylalanyl-tRNA-
protein transferase, and sortase, which require particular amino
acid residues at the N-terminus and are unable to achieve fully
quantitative labeling. The present method will be useful for
double-fluorescent-labeling and FRET analysis of proteins by
combining four-base codon-mediated incorporation of fluores-
cent-labeled non-natural ¡-amino acids which allows quanti-
tative labeling at specific internal positions.8
On the other hand, low-productivity of the cell-free trans-
lation system is a disadvantage of the present method that must
be overcome for future applications. Nonetheless, the use of
bright fluorophores can partially compensate for the low-yield
of the labeled proteins. In addition, the result that non-¡-
aminoacyl-tRNAs are more stable against hydrolysis than ¡-
aminoacyl-tRNAs will be advantageous for increasing the
yield of fluorescent-labeled proteins in an E. coli dialysis cell-
free translation system or wheat-germ system, which require
long incubation periods.
29 N. Shozen, I. Iijima, T. Hohsaka, Bioorg. Med. Chem. Lett.
2009, 19, 4909.
30 S. M. Hecht, B. L. Alford, Y. Kuroda, S. Kitano, J. Biol.
Chem. 1978, 253, 4517.
31 S. A. Robertson, J. A. Ellman, P. G. Schultz, J. Am. Chem.
Soc. 1991, 113, 2722.
We thank Dr. Kenzo Fujimoto for MALDI-TOF MS
measurements. This work was supported by Grants-in-Aid
for Scientific Research (B) (No. 19370063) and Scientific
Research on Innovative Areas (No. 20107005) from the
Ministry of Education, Culture, Sports, Science and Technol-
ogy, Japan.