ARTICLE

pubs.acs.org/JACS

“Quenchbodies”: Quench-Based Antibody Probes That Show

Antigen-Dependent Fluorescence
Ryoji Abe,†,§ Hiroyuki Ohashi,‡ Issei Iijima,† Masaki Ihara,^,|| Hiroaki Takagi,§ Takahiro Hohsaka,† and
Hiroshi Ueda*,‡,^
†
School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan
§
ProteinExpress Company Ltd., 1-8-15 Inohana, Chuo-ku, Chiba 260-0856, Japan
‡
Department of Chemistry and Biotechnology, School of Engineering, The University of Tokyo, 7-3-Hongo, Bunkyo-ku,
Tokyo 113-8656, Japan
^
Department of Bioengineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan

bS Supporting Information
ABSTRACT: Here, we describe a novel reagentless fluorescent biosen-
sor strategy based on the antigen-dependent removal of a quenching
effect on a fluorophore attached to antibody domains. Using a cell-free
translation-mediated position-specific protein labeling system, we found
that an antibody single chain variable region (scFv) that had been
fluorolabeled at the N-terminal region showed a significant antigen-
dependent fluorescence enhancement. Investigation of the enhance-
ment mechanism by mutagenesis of the carboxytetramethylrhodamine
(TAMRA)-labeled anti-osteocalcin scFv showed that antigen-dependency was dependent on semiconserved tryptophan residues
near the VH/VL interface. This suggested that the binding of the antigen led to the interruption of a quenching effect caused by the
proximity of tryptophan residues to the linker-tagged fluorophore. Using TAMRA-scFv, many targets including peptides, proteins,
and haptens including morphine-related drugs could be quantified. Similar or higher sensitivities to those observed in competitive
ELISA were obtained, even in human plasma. Because of its versatility, this “quenchbody” is expected to have a range of applications,
from in vitro diagnostics, to imaging of various targets in situ.

’ INTRODUCTION As a possible application for PET-based biosensors, immuno-

The development of innovative fluorescence-based techni- logical detection of target molecules (immunoassay) is consid-
ques for probing molecular recognition is of major interest in ered as a very attractive area of research. It is an indispensable
biophysical chemistry. The naturally occurring amino acid tryp- technique for quantifying various molecules, and is utilized in a
tophan (Trp) is of particular interest in fluorescence-based work range of fields, from basic biological research to clinical diag-
on peptides and proteins. For example, Trp can serve as an nostics. Fluorescence-based assays that do not involve separation
efficient electron donor in photoinduced electron transfer (PET) steps are considered particularly useful, due to advantages such
reactions with certain dye molecules, a property conferred by the as short assay time and ease of handling. The reagentless fluo-
Trp indole side chain which is the most readily oxidized func- rescent biosensor approach, which is based on the fluorolabeled
tional group among all naturally occurring amino acids.1 antibody fragment whose fluorescence intensity is altered by
In further promising applications, PET-based biosensors have binding to its target (antigen), is probably the simplest example
been developed that use conformationally induced alterations in of such attempts.4,5 For example, Bedouelle et al.6 and Winter
PET efficiency upon binding for the specific detection of DNA or et al.7 used site-specifically labeled antibodies that demonstrated
RNA sequences or antibodies at the single-molecule level.2,3 In increased fluorescence upon binding to their target protein, using
contrast to F€orster resonance energy transfer (FRET)-based sys- the environment-sensitive dye 7-nitrobenz-2-oxa-1,3-diazole (NBD).
tems, in which long-range dipole
dipole interactions are probed, However, to construct such protein biosensors, the fluorolabel-
the above sensors require contact formation between the fluoro- ing position on the protein must be optimized, most typically in
phore and the guanosine or tryptophan residue. Depending on the or near complementarity determining regions (CDR) of the
reduction potential of the fluorophore used, efficient fluorescence antibody, and this often requires considerable time and labor.
quenching via PET can then occur. With careful design of con- Here, we show a novel strategy of PET-based biosensor for vari-
formationally flexible molecules and the use of appropriate fluoro- ous antigens, using a position-specific protein labeling methodology
phores, efficient single-molecule sensitive PET sensors can be
produced. Therefore, PET-based molecules offer an elegant alter- Received: July 2, 2011
native to conventional biosensors based on FRET processes. Published: October 06, 2011

r 2011 American Chemical Society 17386 dx.doi.org/10.1021/ja205925j | J. Am. Chem. Soc. 2011, 133, 17386–17394
Journal of the American Chemical Society ARTICLE

variable region genes (VH and VL) of the anti-BGP antibody

obtained from murine hybridoma KTM-219 cells were linked
via a gene encoding a (Gly4Ser)3 linker (Figure 1A). The in-
corporation of p-(TAMRA-aminocaproyl)-aminophenylalanine
(TAMRA-C6-AF) into the N-terminus of the scFv was carried
out using the amber suppression method with an N-terminal
ProX tag sequence (NH2-MSKQIEVNXSNE-COOH). This tag
is an optimized 12 amino acid sequence containing an amber
codon (X) at the ninth position to enhance the incorporation
of non-natural amino acids as well as protein expression.10
In addition, a His tag was fused at the C-terminus of scFv to
allow ready detection and purification. The amber codon was
decoded by a highly efficient amber suppressor tRNA derived
from Mycoplasma capricolum Trp1 tRNA,11 which had been amino-
acylated with TAMRA-C6-AF. The TAMRA-C6-AF-tRNA was
added to an Escherichia coli cell-free translation system, to-
gether with the scFv genes fused with a ProX coding sequence.
The translation products were resolved by SDS
PAGE and
detected by fluorescence imaging of the gel. A clear fluorescent
band showing TAMRA fluorescence was observed at the ex-
pected molecular weight (32 kDa) (Figure 1B). This result
indicated that scFv containing TAMRA at the N-terminal region
had been successfully expressed. The scFv was also analyzed by
Western blotting using an anti-His tag antibody. The full-length
Figure 1. Synthesis of scFv containing TAMRA at the N-terminal
protein was clearly observed in the presence of TAMRA-C6-AF-
region. (A) Illustration of the incorporation of TAMRA-C6-AF into tRNA, while a negligible amount of the protein was observed in
scFv in response to a UAG codon in a cell-free translation system. (B) the absence of the suppressor tRNA (Figure 1C). These results
Fluorescence image of SDS
PAGE for the expression of anti-BGP scFv indicated that the TAMRA-C6-AF-tRNA decoded the amber
containing TAMRA-C6-AF. Fluorescence of TAMRA shown in red was codon specifically, and that anti-BGP scFv had been site-speci-
detected with excitation at 532 nm and emission at 580 nm. Fluorescent fically and near-quantitatively labeled with TAMRA.
marker shown in green was detected with excitation at 488 nm and Antigen-Dependent Fluorescence of TAMRA-Labeled
emission at 520 nm. (C) Western blot analysis using an anti-His-tag Anti-BGP scFv. The TAMRA-labeled anti-BGP scFv was pur-
antibody. ified using nickel-affinity chromatography and its fluorescence
spectrum was measured in the absence and presence of a cognate
based on fluorolabeled aminoacyl tRNA and a cell-free transla- antigen BGP-C7 peptide (NH2-RRFYGPV-COOH). To our
tion system.8 During the attempts to make FRET-based biosen- surprise, the fluorescence spectrum showed a remarkable dose-
sor by this methodology, we fortuitously discovered an antigen- dependent increase (5.6-fold) in intensity upon addition of BGP-
dependent fluorescence enhancement of a single chain variable C7 (Figure 2A). A titration curve of the fluorescence intensity at
region fragment (scFv) of the antibody fluorolabeled at its 580 nm suggested that fluorescence increases with BGP-C7
N-terminal region. After investigation of the mechanism, it was peptide concentration (Figure 2B). Also, when the same mea-
found that this enhancement was observed due to the quenching surement was performed with full-length BGP, similar dose-
of used dye by the semiconserved Trp residues in scFv, probably dependent fluorescence increase (max 5.2-fold) was observed.
by PET mechanism. According to the proposed mechanism, the Using curve fitting, the ED50 values for the peptide and the pro-
phenomenon was considered to have generality to other anti- tein were estimated as 2.5  10
8 and 1.1  10
7 M, respec-
bodies, and it was experimentally confirmed to be indeed the case. tively, which compare well with the ED50 value obtained by
competitive ELISA (8.8  10
8 M for BGP-C7).9 The result
’ RESULTS indicates that the BGP peptide-binding activity of scFv had not
been affected by the incorporation of TAMRA-C6-AF at the
Position-Specific Incorporation of TAMRA-Labeled Amino N-terminal region. To test if the observed fluorescence change
Acid into Anti-BGP scFv. As the model antibody to apply could be used for imaging purposes, the fluorescence intensity of
position-specific fluorolabeling technology, we chose anti-human TAMRA-scFv was visualized in a 384-well microplate, in the pre-
osteocalcin (bone gla protein, BGP) scFv as the recognition unit sence and absence of the BGP peptide. The fluorogram obtained
for the bone-related disease marker, and as the model dye we confirmed that the TAMRA-labeled scFv could potentially be
took 5-carboxytetramethylrhodamine (TAMRA). This combi- used for quantitative imaging of the antigen in situ (Figure 2C).
nation was chosen because the variable region, Fv, of this anti- To further evaluate the binding specificity of TAMRA-scFv,
body has an interesting property that it is markedly stabilized by the same assay was performed with the somewhat longer pep-
bound BGP or its C-terminal epitope peptides.9 Also, we have tides BGP-C10 (NH2-EAYRRFYGPV-COOH, MW = 1257)
found in other experiments (Abe et al., in preparation) that and BGP-C10dV (NH2-EAYRRFYGP-COOH, MW = 1158).
position-specifically TAMRA-labeled anti-BGP Fv at the N- The former includes the C-terminal valine that is an essential
terminal region of heavy chain variable region VH demonstrates epitope of this antibody, while the latter lacks this residue.9
a modest but significant (1.7-fold) antigen-dependent fluo- Titration curves indicated that while BGP-C10 showed a similar
rescence intensity change. To prepare this construct, the two 5.8-fold increase to that observed for BGP-C7, no apparent
17387 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394
Journal of the American Chemical Society ARTICLE

Figure 2. Antigen-dependent fluorescence enhancement of TAMRA-labeled anti-BGP scFv. (A) Fluorescence spectra of TAMRA-scFv with excitation
at 550 nm in the absence and presence of BGP-C7 peptide. (B) Titration curve of the fluorescence intensity at 580 nm. The inten-
sities are relative values with respect to that in the absence of BGP-C7 peptide. (C) Fluorescence imaging of TAMRA-scFv on a microplate in the
presence of BGP-C7 peptide with excitation at 532 nm and emission at 580 nm. (D) Titration curves of TAMRA-scFv for BGP-C10 and BGP-C10dV
peptides.

increase in fluorescence was observed in the presence of BGP-

C10dV (Figure 2D). This clearly suggests that the binding
specificity as well as the affinity of scFv for BGP is preserved,
even in the presence of the N-terminal TAMRA dye.
Quenching of TAMRA Fluorescence by Trp Residues. Since
the BGP peptides tested did not contain moieties capable of en-
hancing the TAMRA fluorescence, and no other dyes that might
cause FRET between them were present either in the antigen or
the antibody, we reasoned that the observed fluorescence en-
hancement mechanism involved electron transfer, rather than
FRET. It has been reported previously that a number of organic
fluorophores are quenched by Trp residues, either in solution or
in a polypeptide. This quenching is thought to occur via a PET
mechanism.12
14 We therefore investigated the possibility that
the observed fluorescence intensity change was attributable to
intramolecular quenching by Trp residues in scFv. The VH frag-
ment of this antibody has four Trp residues, namely, Trp33H,
Trp36H, Trp47H, and Trp103H (using the Kabat numbering Figure 3. (A) The model of TAMRA-scFv. Structure for scFv was built
scheme15), while the VL fragment has a single residue, Trp35L. using WAM antibody modeling server (http://antibody.bath.ac.uk)
Almost all of these Trp residues, with the exception of Trp33H, with CONGEN side chain building and Accessibility profile screen
are located in the framework region, and are generally conserved methods. Trp residues, Trp33H, Trp36H, Trp47H, Trp103H, and Trp35L
in antibodies derived from various origins. A three-dimensional are colored green. (B) Titration curves of the Trp-to-Phe mutants.
model (Figure 3A) suggested that Trp47H and Trp103H con- Fluorescence intensities are shown as relative values with respect to
tribute to the hydrophobic interaction with VL, and that Trp33H those in the absence of the antigen. (C) Normalized titration curves.
contributes to the interaction with the BGP peptide. Trp36H and
Trp35L are located close to the interacting surfaces of VH and VL, of fluorescence were diminished, depending on the mutation,
respectively. although different behaviors were observed for different
To evaluate the potential role of the Trp residues on antigen- mutants (Figure 3B,C). Compared with the basal intensity
dependent fluorescence quenching, five Trp-to-Phe point mu- without antigen, the fluorescence enhancements observed on
tants containing TAMRA-C6-AF were prepared, and the antigen- addition of antigen were 3.7-, 4.7-, 2.5-, 3.5-, and 3.1-fold for
dependent fluorescence intensities were measured. All mutants the W33HF, W36HF, W47HF, W103HF, and W35LF mutants,
showed significantly attenuated enhancements in fluorescence respectively. These values are all considerably less than the value
intensity upon addition of the BGP-C7 peptide, when compared to observed for the wild-type scFv (5.6-fold). It is worth noting that
the wild-type scFv. Both the magnitude and the antigen-dependency a mutant of the Trp residue most distant to the TAMRA (W35LF)
17388 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394
Journal of the American Chemical Society ARTICLE

Table 1. Effect of Antigen and Denaturant (7 M GdnHCl, 100 mM DTT) on TAMRA Fluorescence

λmax normalized intensity at λmax

PBST + BGP-C7 PBST + denaturant (A) + BGP-C7 (B) + denaturant recovery ratio (%)a

5-TAMRA 573 580 1.0 1.1 -

w.t. 580 585 5.0 5.5 100
W33HF 580 585 3.8 4.2 100
W36HF 580 585 4.0 4.6 96
W47HF 580 585 2.2 2.6 93
W103HF 580 585 3.0 3.3 100
W35LF 580 585 2.4 3.0 88
a
Recovery ratio = [A/(B/1.1)]  100.

Table 2. Fluorescence Lifetimes τi, and Corresponding residues participate in quenching, the observed quenching was
Amplitudes, ai, of TAMRA Labeled Anti-BGP scFv in the supposed to be dynamic rather than due to static interaction. In
Absence and Presence of 1 μM BGP-C7 and Denaturant 7 M the case of dynamic quenching, faster fluorescence decay will be
GdnHCl and 100 mM Dithiothreitol observed in the presence of quenchers.13 When fluorescence
lifetime measurement was performed for TAMRA-scFv in the
τ1(ns)/a1 τ2(ns)/a2
presence or absence of BGP-C7 peptide or denaturant, a sig-
no BGP-C7 1.20/0.48 3.92/0.52 nificant difference in the amplitude of shorter lifetime species
+ BGP-C7 1.46/0.21 3.75/0.79 (1.2
1.46 ns) was observed (Table 2, Figure S1). In the pre-
Denaturant 3.14/1.00 sence of antigen or denaturant, longer lifetime species (3
4 ns,
similar to reported values for free Trp16,17) dominated. However,
showed markedly reduced response, while fully maintaining its in the absence of these agents, the amplitude of shorter lifetime
antigen dependency (i.e., binding affinity). increased to almost half, clearly suggesting the occurrence of
To further confirm the antigen-dependent removal of quench- dynamic quenching probably due to a PET mechanism.
ing mechanism proposed to explain the observed fluorescence Fluorescence Correlation Spectroscopy Analysis. Depend-
enhancement, the fluorescence of TAMRA-scFv proteins under a ing on its concentration, scFv molecules can form dimers or
denaturing condition was investigated. As summarized in Table 1, other higher order species.18 Since TAMRA can form quenched
the fluorescence of free TAMRA and the proteins was measured dimer,19 there is a small possibility that the addition of antigen
in a standard buffer (PBS containing 0.05% Tween 20, PBST) induced dissociation of a TAMRA-scFv oligomer to monomers,
or that containing 7 M guanidine hydrochloride (GdnHCl) and thus, resulted in increased species that emit brighter fluores-
100 mM dithiothreitol (DTT). While free TAMRA showed cence. To rule out this possibility, fluorescence correlation spec-
a modest (1.1-fold) fluorescence increase in the denaturant, the troscopy (FCS) was utilized to measure the diffusion time of
wild-type TAMRA-scFv showed a 5.5-fold increment, with a TAMRA-scFv, as a measure of its average molecular size in solu-
slight red shift in emission wavelength. The result is in accor- tion. As shown in Figure S2, compared with TAMRA-scFv alone,
dance with our hypothesis that the quenching occurs in the the addition of BGP-C7 or intact BGP resulted in increased
folded protein, and this was removed by the denaturation, al- diffusion time, which means slower diffusion probably due to in-
lowing the fluorescence to be revealed. On the other hand, the creased molecular weight and possibly due to larger molecular
addition of BPG-C7 at saturating concentration to the wild-type size of scFv
antigen complex with exposed TAMRA dye. From
TAMRA-scFv led to a 5.0-fold increment in fluorescence, sup- this result and low scFv concentration used throughout this
porting the “antigen-dependent release from the quenched state” study, it is highly unlikely that the observed fluorescence increase
hypothesis. The mutant TAMRA-scFv proteins showed dimin- was due to antigen-dependent increase of fluorescent monomers.
ished fluorescence increments upon denaturation, and also in the Effect of Dye Mobility. The observed fluorescence quench-
presence of saturated antigen. The observed increments were ing may be resulted from the long, flexible linkage of the VH
similar to each other (shown as recovery ratio in Table 1), indi- N-terminus and the fluorophore. To confirm the effect of linker
cating the active role of each Trp residue in the observed quench- on quenching, flexible peptide linkers of (Gly3Ser)n (n = 1
3)
ing. These results suggest that in the absence of antigen, the were inserted between ProX tag and anti-BGP scFv to increase
TAMRA fluorophore is in close proximity to these Trp residues. the mobility of the fluorophore (Figure 4A). Fluorescence spec-
This includes the Trp35L residue near the VH/VL interface, tral measurements showed that the antigen-dependent fluores-
which may be able to interact with distant TAMRA due to the cence increase was also observed for all the linker lengths while
ProX tag sequence and the flexible aminohexyl linker between the response was slightly decreased for longer linkers (Figure 4B).
the scFv and the fluorophore. We postulate that the addition of To evaluate the effect of the linker length further, an anti-
antigen promotes the closure of the VH/VL interface, preventing bisphenol A scFv20 was used as another recognition unit. The
the interaction of TAMRA with the Trp residues, and therefore anti-bisphenol A scFv has six Trp residues including four con-
removing the quenching. served ones, which were expected to contribute to the fluores-
Fluorescence Lifetime Measurement. To verify the occur- cence quenching as in the case of anti-BGP scFv (Table 3). The
rence of quenching and its release, fluorescence lifetime mea- anti-bisphenol A scFv genes were constructed and expressed
surement of TAMRA-scFv was performed. Since multiple Trp as TAMRA-labeled proteins in a similar manner. Fluorescence
17389 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394
Journal of the American Chemical Society ARTICLE

Figure 4. (A) Schematic structure of scFv containing flexible peptide linkers. (B) Titration curves of TAMRA-labeled anti-BGP scFvs with and without
peptide linkers with excitation at 550 nm. Fluorescence intensities are relative values with respect to those in the absence of the antigen. (C) Titration
curves of TAMRA-labeled anti-bisphenol A scFvs with and without peptide linkers with excitation at 550 nm. Fluorescence intensities are relative values
with respect to those in the absence of the antigen.

Table 3. Locations of Trp Residues in the scFvs Useda

residue number

CDRH1 CDRH3 CDRL3

H33 H34 H35 H36 ... H47 ... H95 ... H103 ... L35 ... L47 ... L91 L92 ... L94

αBGP W I H W W S W W L T T V
αBisphenol A Y Y W W W V W W I S W I
αHEL Y W S W Y W W W L S N W
αBSA/HSA A M A W W S W W L A D S
αEstradiol T I H W W Y W W W W S Y
αMorphine W I E W W W W W L W Y N
a
Numbered according to Kabat.15

the anti-BGP scFv (Figure 4C). Since this weak response may be
due to suboptimal interaction between the fluorophore and Trp
residues, elongation of the peptide linker was attempted. The
insertion of (Gly3Ser)n linkers, in this case, markedly enhanced
antigen-dependent fluorescence increase, and the scFv contain-
ing the longest (Gly3Ser)5 linker showed 2.0-fold fluorescence.
From the fitting of titration curves, the EC50 was estimated as
2.0  10
8 M, which was approaching to the IC50 value deter-
mined by competitive ELISA for the original scFv (1.4  10
9 M).
Taken together, the results suggest that the insertion of pep-
tide linker between the fluorophore and scFv affects the fluores-
cence quenching and its antigen-dependency. Although the exact
reason of antibody-specific difference in response is unclear, this
may be resulted from the difference in the optimal orientation
of fluorophore to the differently positioned Trp residues in each
scFv.
Application of the Quenchbody Strategy to Other scFvs.
We propose to name this position-specifically fluorolabeled scFv
Figure 5. Titration curves of TAMRA-labeled scFvs Fluorescence
intensities are relative values with respect to those in the absence of as a “quenchbody” after the quench phenomenon associated with
the antigen. (A) HyHEL-10 for HEL, (B) 29IJ6 for BSA and HSA, (C) it. Since many Trp residues are highly conserved in antibodies of
ES1-11 for estradiol. (D) Microplate-based imaging assay for BGP-C7 various origins (see below), we expected that this approach might
peptide either in 50% human plasma in PBST or in PBST buffer be broadly applicable to other antibody/antigen pairs. To further
performed with TAMRA-labeled anti-BGP scFv. The TAMRA-scFv investigate this potential, we prepared quenchbodies for other
for A
C contain (Gly3Ser)2 linker, while that for D contains no linker. antigens. First, hen egg lysozyme, 17β-estradiol, and serum
albumins (SAs) were used as model antigens. Examination of
spectra showed an increase in fluorescence intensity upon the the primary sequence of these antibodies revealed that a number
addition of bisphenol A, although the observed response was of Trp residues were conserved among them (Trp36H, Trp47H,
much weaker (1.1-fold for the scFv without linker) than that of Trp103H, and Trp35L) (Table 3). While anti-hen egg lysozyme
17390 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394
Journal of the American Chemical Society ARTICLE

50% human plasma to that in PBST. The maximum fluorescence

was as high as 7-fold of that observed in the absence of anti-
gen. This result compares well with our previous open sandwich
enzyme-linked immunosorbent assay (OS-ELISA) for BGP,
which showed considerable signal reduction in serum-derived
samples without pretreatment to remove SAs.22
Application to Morphine Detection. Sensitive detection,
identification, and confirmation of opiates are considered highly
important in view of public health and crime prevention. How-
ever, most conventional analytical methods for opiates such as
morphine or heroin are laborious and time-consuming. As an-
other practical application of quenchbody technology, rapid
detection of opiates was attempted. On the basis of the published
nucleotide sequence of an scFv (VL-VH) recognizing morphine-
6-glucuronide (M6G),23 a quenchbody for M6G was prepared
and investigated for its fluorescence spectra upon addition of
morphine and related drugs (Figure 6B and Figure S3).
As shown in the dose
response curves in Figure 6A, the
constructed quenchbody reacted most strongly with codeine,
and also heroin and morphine in this order with the calculated
EC50 values of 1.9, 3.7, and 5.2  10
8 M, respectively. Although
the maximum fluorescence increase was about 1.5-fold over the
background, 1
100 ng/mL of these compounds were detectable
within 3 min after reaction start. On the other hand, naloxone
wherein the N-methyl group observed in opiates is substituted
Figure 6. (A) Titration curves of TAMRA-labeled anti-morphine scFv with N-propenyl, as well as cocaine, ketamine, and methamphet-
for morphine and related drugs. Fluorescence intensities are relative amine, showed negligible fluorescence increase, probably due to
values with respect to those in the absence of the antigen. (B) Molecular the absence of this N-methyl as the essential epitope.
structure of the drugs with its molecular weight. To compare the antigen dose-dependency of quenchbody
fluorescence with that of antigen binding, we performed com-
petitive ELISA using immobilized morphine-BSA and free mor-
HyHEL-10 lacks Trp47H, and anti-SA 29IJ6 contains only the phine, and the amount of bound quenchbody was evaluated
conserved Trp residues, the number of Trp residues present in using peroxidase-conjugated anti-His5 antibody. From the curve-
each scFv made us optimistic that antigen-dependent fluores- fitting, the IC50 for morphine was obtained as 26 ( 12 nM
cence would be observed. (Figure S4), which was a similar value to the EC50 of quenchbody
To allow the performances of these antibodies to be compared fluorescence. These data suggest that the anti-morphine quench-
under the same conditions, TAMRA-labeled quenchbodies con- body could detect as little as parts per billion (ppb) range of
taining the (Gly3Ser)2 linker between the ProX tag and the scFv opiates by just mixing the sample and measuring its fluores-
were prepared as described above. The fluorescence intensity for cence, which obviates the needs of long and dangerous reaction/
the purified proteins in the absence and presence of their respec- washing steps inevitable in ELISA.
tive antigen were evaluated by fluorescence spectral measure-
ments on titration with the antigens. As shown in Figure 5A
C,
all the quenchbodies showed antigen-dependent fluorescence ’ DISCUSSION
enhancements. However, the extent of enhancement was vari- Our working model of the antigen-dependent quenchbody
able with the different scFvs The anti-estradiol scFv showed the fluorescence observed with this system is as follows. In the ab-
strongest antigen-dependent increase in fluorescence (4.5-fold sence of antigen, TAMRA within the ProX tag sequence pen-
increase), while the anti-lysozyme and anti-SA scFvs showed etrates between the VH/VL interface, and interacts with Trp
weaker responses (1.3- and 1.5-fold, respectively). However, residues by hydrophobic and/or π
π stacking interactions
it is worth noting that the calculated ED50 for the lysozyme was (Figure 3A). This interaction leads to a quenching electron
7.5  10
9 M, which is comparable to the reported Kd values of transfer from the Trp residues to the dye due to transient contact
the Fv and scFv (5.0  10
9 and 1.2  10
8 M, respectively).21 between each Trp residue and the chromophore.13,14 Binding of
The ED50 measured for human SA (∼10
5 M) was sufficiently an antigen, such as the BGP peptide, to the scFv induces tighter
lower than its reference range of 3.4
5.4 g/dL (0.5
0.8 mM) in complexation of VH and VL, and thus releases the TAMRA dye
serum (MedlinePlus, http://www.nlm.nih.gov/medlineplus/ from interactions with the Trp residues near the VH/VL interface.
ency/article/003480.htm) to indicate its potential utility as a Among the five tryptophans found to be responsible for the
diagnostic reagent. quenching of the TAMRA-labeled anti-BGP quenchbody,
Measurement of BGP Peptide in Human Plasma. To fur- four (Trp36H, Trp47H, Trp103H, and Trp35L) are located
ther demonstrate the utility of quenchbodies in clinical diagnos- in the framework region, and are highly conserved among
tic applications, the performance of the quenchbody in a plasma immunoglobulins of various origins. According to a database of
sample was investigated using a fluorescence imager. As shown in >35 000 nonredundant immunoglobulin sequences (Abysis, http://
Figure 5D, the TAMRA-quenchbody for BGP showed an almost bioinf.org.uk/), the rate of conservation for the three H chain
indistinguishable dose
response for the BGP-C7 peptide in tryptophans is 99.0%, 93.3%, and 97.0%, respectively, and that for
17391 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394
Journal of the American Chemical Society ARTICLE

Trp 35L is 98.0%, for 12 483 chains including kappa and lambda ’ CONCLUSION
chains. These values, as well as the number of Trp residues in- We reported a discovery of novel biosensing principle utilizing
volved, ensure that almost all Fv could potentially be successfully fluorescence quenching of a labeled dye by intrinsic Trp residues
incorporated into a quenchbody. It is interesting that while the in antibody Fv region. Compared with conventional fluoro-
side chain of Trp35L is located in the protein core, and therefore immunoassays, the quenchbody assay is simple, and requires no
is not exposed to solvent in any antibody structure, it plays a additional reagents to perform the detection. In addition, com-
considerable role in the quenching effect observed in the anti- pared with OS-ELISA using Fv molecules, scFv-based quench-
BGP quenchbody. Presumably, the dynamics of the scFv’s poly- bodies generally show higher sensitivity due to the linkage of the
peptide backbone might result in transient escape of this residue two variable region fragments. Another merit of this scFv-based
from the buried hydrophobic core, and/or penetration of hydro- system is their ready availability, as specific scFvs can be directly
phobic TAMRA dye into the vicinity of this residue. obtained from phage-display libraries of various origins. Because
In the anti-BGP quenchbody, additional quenching by Trp33H of the simplicity of the system, the application of quenchbodies is
was observed. This CDR2 residue is likely to be in the antigen- not likely to be limited to in vitro diagnostics, but may also be
binding site, and therefore, the presence of TAMRA in close applied to imaging in situ or in vivo. In future, the development of
proximity to this residue would be prohibited by antigen binding, more robust quenchbody production methods, such as in vivo
which therefore inhibits quenching and enhances fluorescence. It incorporation of non-natural amino acids,30 will further enlarge
can be speculated that the different fluorescence dynamic ranges the scope of their applications.
observed for the quenchbodies based on different antibodies is
due to differences in their Trp residue content, both in number
and in positions. As shown in Table 3, some scFvs contain Trp ’ EXPERIMENTAL METHODS
residues in addition to the conserved ones. For example, anti-
Materials. In-Fusion Advantage PCR Cloning Kit was from Takara
estradiol scFv, whose quenchbody showed a large fluorescence
Bio (Otsu, Japan). A pIVEX2.3d vector and RTS 100 E. coli Disulfide Kit
increase, has two additional Trp residues in VL. These may con- were from Roche Diagnostics (Basel, Switzerland) or 5-Prime GmbH
tribute to enhancing the quenching effect. On the other hand, the (Hamburg, Germany). C-terminal peptides for human osteocalcin (bone
anti-SA scFv obtained from a synthetic phage library, Tomlinson gla protein, BGP) (BGP-C7, NH2-RRFYGPV-COOH, MW = 894;
J,24 has no Trp other than the conserved ones, and its quench- BGP-C10, NH2-EAYRRFYGPV-COOH, MW = 1257; BGP-C10dV,
body demonstrated a lower response. Nonetheless, the obtained NH2-EAYRRFYGP-COOH, MW = 1158) were obtained from Gen-
results overall suggest that the conserved Trp residues are script (Piscaway, NJ). Bovine serum albumin (BSA) was from Bovogen
sufficient per se for an antigen-dependent fluorescence quench- (Essendon, Australia). Hen egg lysozyme (HEL) was from Wako
ing effect to be observed. Therefore, the present strategy can be (Tokyo, Japan). Human serum albumin (HSA) and 17β-estradiol were
expected to be effective for various scFvs, including those with from Sigma (Saint Louis, MO). Pooled normal human plasma was from
minimal conserved Trp residues. An improved response is likely Innovative Research (Northville, MI).
for those scFvs containing additional residues, depending on Construction of scFv Genes. An expression vector, pROX-BGP-
their number and positions. scFv, harboring a T7 promoter-controlled scFv gene of anti-BGP
The putative working mechanism also suggests that antibodies C-terminal fragment KTM-2199 fused with an N-terminal ProX tag
that show a higher antigen-dependent stabilization (larger signal containing an amber codon (ATG TCT AAA CAA ATC GAA GTA
change in OS-ELISA that detects the interaction between VH AAC TAG TCT AAT GAG) and a C-terminal His tag was constructed
and VL)25 are likely to show more fluorescence activation. This is by overlap PCR. A 15-amino acid linker with three Gly4-Ser repeats was
because these antibodies have a higher chance of having a solvent- inserted between the VH and VL. The VH chain was amplified using the
exposed VH/VL interface. However, while fluorescence activa- 50 -primer (CTTTAAGAAGGAGATATACCATGTCTAAACAAATC-
tion was observed for the anti-SA scFv, the same Fv showed a GAAGTAAACTAGTCTAATGAGACCCAAGTAAAGCTGCAGCA-
minimal response to BSA in a phage-based OS-ELISA.26 It is GTC) and the 30 -primer (CCAGAGCCACCTCCGCCTGAACCGC-
likely that a small antigen-dependent change in VH/VL interac- CTCCACCGCTCGAGACGGTGAC), and the VL chain was amplified
tion strength is sufficient to obtain an observable change in using the 50 -primer (CAGGCGGAGGTGGCTCTGGCGGTGGCG-
fluorescence intensity. This implies that quenchbodies can be GATCTGACATTGAGCTCACCC) and the 30 -primer (TGATGA-
applied not only to small molecule detection, but also to the TGAGAACCCCCCCCCCGTTTTATTTCCAG). The amplified VH
detection of various protein antigens, even if they contain several and VL genes were linked by overlap PCR using the VH 30 -primer and the
VL 50 -primer, and the construct was then cloned into NcoI- and SmaI-
Trp residues.
digested pIVEX2.3d using an In-Fusion PCR cloning kit.
PET is recognized as a useful mechanism to alter the fluores-
A wild-type (nonlabeled) scFv was constructed by replacing the TAG
cence of organic dyes and their derivatives using small molecules codon with a TTT codon in the ProX tag. For substitution of tryptophan
such as metal ions or reactive oxygen species.27,28 Only (Trp) residues with phenylalanine (Phe), TGG codons for Trp33,
recently, it has been found to be useful in the detection of Trp36, Trp47, and Trp103 in the VH gene, and Trp35 in the VL gene
natural biomolecules such as guanine29 and tryptophan.13 were each replaced by a TTT codon. For scFv genes encoding anti-BGP
Several organic fluorochromes have been shown quenched and anti-bisphenol A,20 the corresponding genes were cloned in place
by tryptophans, either by an intermolecular or intramolecular of anti-BGP scFv, with an additional sequence encoding N-terminal
PET mechanism. For example, ATTO 655-labeled streptavidin (G3S)0
5 linker. For scFv genes encoding anti-hen egg lysozyme,31
was efficiently quenched probably due to the spatial proximity bovine serum albumin (BSA),26 and estradiol (Fujioka et al., in pre-
of the fluorolabeled Lys residues to Trp residues in streptavi- paration), the corresponding genes were cloned in place of anti-
din. The observed quenching effect was markedly reduced in BGP scFv, with an additional sequence encoding N-terminal (G3S)2
the presence of streptavidin’s ligand biotin.13 When compared linker. The anti-morphine-3-glucuronide scFv gene was synthesized by
with this system, the quenchbody system has the benefit of a Mr. Gene GmbH (Regensburg, Germany) according to the published se-
large range of detectable targets. quence for clone E3,23 assuming that the undisclosed heavy chain FR4

17392 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394

Journal of the American Chemical Society ARTICLE

sequence was derived of J4 segment. The gene for VL-VH type scFv with Fluorescence lifetime measurements were performed on a TemPro
N-terminal (G3S)2 and internal (G4S)4 linkers was inserted to pROX- system (Horiba Jobin-Yvon, Japan) using the time-correlated single-
BGP-scFv in place of anti-BGP scFv as above. photon counting (TCSPC) technique. The excitation source was a
Preparation of Aminoacyl tRNA. TAMRA-C6-AF-pdCpA was 495 nm SpectraLED with a pulse length of ∼1 ns at a repetition rate of
synthesized as described previously.10 This was ligated to an amber 1 MHz. Emission at >600 nm was collected using a long-pass filter SCF-
suppressor tRNA, derived from M. capricolum Trp1 tRNA without the 50S-60R (SIGMA KOKI), and the decay parameters were determined
30 dinucleotide, by chemical ligation as described previously.11,32 The by least-squares deconvolution.
aminoacyl-tRNAs can be obtained as commercially available reagents Abbreviations. BGP, bone gla protein (osteocalcin); BPA, bis-
(CoverDirect tRNA reagents for site-directed protein labeling, Protein phenol A; CDR, complementarity determining region; dNTPs, deoxy-
Express, Chiba, Japan). ribonucleotide triphosphate; ELISA, enzyme-linked immunosorbent
Cell-free Transcription/Translation. The incorporation of assay; FCS, fluorescence correlation spectroscopy; Fv, antibody variable
TAMRA-C6-AF into the N-terminal region of scFv was performed region fragments; PBS, phosphate buffered saline; PBST, PBS contain-
using an RTS 100 E. coli Disulfide kit. The reaction mixture (50 μL) ing 0.05% Tween 20; PCR, polymerase chain reaction; scFv, single chain
comprised 7 μL of an amino acid mix, 1 μL of methionine, 7 μL of the antibody variable region fragment.
reaction mixture, 25 μL of activated E. coli lysate, 5 μL of plasmid DNA
(500 ng), and 5 μL of TAMRA-C6-AF-tRNA (0.8 nmol). All the
’ ASSOCIATED CONTENT
reagents used, with the exception of the plasmid and the tRNA, were
provided in the RTS 100 E. coli Disulfide kit. The reaction mixture was
incubated at 20 C with shaking on RTS ProteoMaster (Roche Diag-
bS Supporting Information. Decay curves obtained in the
fluorescence lifetime measurements, FCS data and methods,
nostics, Basel, Switzerland) at 600 rpm for 2 h, and subsequently at 4 C fluorescence spectra for TAMRA-scFv specific to M3G, and
without shaking for 16 h. An aliquot of the reaction mixture (0.5 μL) was competition ELISA result for morphine. This material is available
applied to 15% SDS
PAGE33 and the gel was visualized using a fluo-
free of charge via the Internet at http://pubs.acs.org.
rescence scanner FMBIO-III (Hitachi, Tokyo, Japan). The gel was also
analyzed by Western blot analysis using anti-His tag (Novagen, La Jolla, CA)
and alkaline phosphatase-labeled anti-mouse IgG (Promega, Madison, WI).
’ AUTHOR INFORMATION
To purify scFv, the reaction mixture (50 μL) was diluted in wash Corresponding Author
buffer (20 mM phosphate, 0.5 M NaCl, 60 mM imidazole, 0.1% [email protected]
polyoxyethylene(23)lauryl ether, pH 7.4) to a final volume of 400 μL,
and applied to a His Spin Trap Column (GE Healthcare, Piscataway, NJ). Present Addresses
)

After incubation at room temperature for 15 min, the column was washed Department of Bioscience and Biotechnology, Faculty of Agri-
three times with wash buffer. The labeled scFv proteins were eluted with culture, Shinshu University, Nagano, Japan.
two 200 μL volumes of wash buffer containing 0.5 M imidazole. The
eluate was passed through an UltraFree-0.5 centrifugal device (Millipore,
Billerica, MA) and equilibrated with phosphate buffered saline Tween-20 ’ ACKNOWLEDGMENT
(PBST, 10 mM phosphate, 137 mM NaCl, 2.7 mM KCl, 0.05% Tween-
We are indebted to T. Ueda, Y. Ohya, N. Tomioka, T.
20, pH 7.4) to allow buffer exchange and concentration of the protein.
The concentration of the labeled scFv protein was determined by
Suganuma, and S. Ebisu for valuable suggestions, T. Shinoda
comparing the fluorescence intensities of a known concentration of for anti-BGP gene, H. Ohkawa for anti-BPA gene, I. Tomlinson
TAMRA dye (Anaspec, Fremont, CA) and the sample under denaturing for anti-SA gene, and Y. Fujioka for anti-E2 gene. We also thank
conditions in 7 M guanidine hydrochloride, pH 7.4. Y. Mizuta, N. Teshigawara and M. Nakamura in Central Customs
Fluorescence Measurements of TAMRA-labeled scFvs. For Lab., Ministry of Finance, Japan, for their help in morphine
fluorescence spectral measurements of anti-BGP scFv, the purified scFv measurement. This project was partly supported by City-area
(2 μg/mL, 25 μL) was diluted in 200 μL of PBST containing 0.2% BSA, and program from MEXT, Japan, a Grant-in-Aid for Scientific Research
BGP peptide was added by titration in a 3  3 mm quartz cell (GL Sciences, (B20360368) from JSPS, Japan, a Grant-in-Aid for Scientific
Tokyo, Japan). After each addition of the antigen, the solution was incu- Research on Innovative Areas (20107005) from MEXT, Japan,
bated at 25 C, 70 min prior to the fluorescence spectral measurements. and by the Global COE Program for Chemistry Innovation,
Anti-lysozyme HyHEL-10, anti-estradiol, and anti-M3G scFvs (2 μg/ MEXT, Japan.
mL, 25 μL) were diluted in PBST containing 1% BSA. Anti-SA scFv
29IJ6 (2 μg/mL, 25 μL) was diluted in PBST containing 0.2% gelatin.
The antigens were added by titration at 5 min intervals, except for anti-
’ REFERENCES
M3G, to which antigens were added at 3 min intervals. (1) Jones, G.; Lu, L. N.; Vullev, V.; Gosztola, D. J.; Greenfield, S. R.;
Fluorescence spectra were measured from 565 to 700 nm, with Wasielewski, M. R. Bioorg. Med. Chem. Lett. 1995, 5, 2385–2390.
excitation at 550 nm at 25 C on a FluoroMax-4 (Horiba Jobin-Yvon, (2) Neuweiler, H.; Schulz, A.; Vaiana, A. C.; Smith, J. C.; Kaul, S.;
Kyoto, Japan). Excitation and emission slit widths were set to 5.0 nm. Wolfrum, J.; Sauer, M. Angew. Chem., Int. Ed. 2002, 41, 4769–4773.
(3) Knemeyer, J. P.; Marme, N.; Sauer, M. Anal. Chem. 2000, 72,
The ED50 values were calculated by curve fitting of the observed
3717–3724.
fluorescence intensities at the maximum emission wavelength, using a
(4) North, J. R. Trends Biotechnol. 1985, 3, 180–186.
sigmoidal dose
response model based on ImageJ software (http:// (5) Pollack, S. J.; Nakayama, G. R.; Schultz, P. G. Science 1988,
rsbweb.nih.gov/ij/). 242, 1038–1040.
For fluorescence imaging of anti-BGP scFv, the purified scFv (2 μg/ (6) Renard, M.; Belkad, L.; Hugo, N.; England, P.; Altschuh, D. l.;
mL, 6.25 μL) was mixed with various concentrations of the antigen in Bedouelle, H. J. Mol. Biol. 2002, 318, 429–442.
PBST (50 μL) containing BSA (0.2%), or in 50% human plasma from (7) Jespers, L.; P.Bonnert, T.; Winter, G. Protein Eng., Des. Sel. 2004,
which the endogenous antigen had been removed,22 in a 384-well micro- 17, 709–713.
plate (Olympus, Tokyo, Japan). The plate was visualized with excitation at (8) Hohsaka, T.; Kajihara, D.; Ashizuka, Y.; Murakami, H.; Sisido, M.
532 nm and emission at 580 nm using a fluorescence scanner. J. Am. Chem. Soc. 1999, 121, 34–40.

17393 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394

Journal of the American Chemical Society ARTICLE

(9) Lim, S.-L.; Ichinose, H.; Shinoda, T.; Ueda, H. Anal. Chem. 2007,
79, 6193–6200.
(10) Abe, R.; Shiraga, K.; Ebisu, S.; Takagi, H.; Hohsaka, T. J. Biosci.
Bioeng. 2010, 110, 32–38.
(11) Taira, H.; Matsushita, Y.; Kojima, K.; Shiraga, K.; Hohsaka, T.
Biochem. Biophys. Res. Commun. 2008, 374, 304–308.
(12) Buschmann, V.; Weston, K. D.; Sauer, M. Bioconjugate Chem.
2003, 14, 195–204.
(13) Marme, N.; Knemeyer, J.-P.; Sauer, M.; Wolfrum, J. Bioconju-
gate Chem. 2003, 14, 1133–1139.
(14) Vaiana, A. C.; Neuweiler, H.; Schulz, A.; Wolfrum, J.; Sauer, M.;
Smith, J. C. J. Am. Chem. Soc. 2003, 125, 14564–14572.
(15) Kabat, E. A.; Wu, T. T.; Perry, H. M.; Gottesman, K. S.; Foeller,
C. Sequences of Proteins of Immunological Interest, 5th ed.; U.S. Govern-
ment Printing Office: Bethesda, MD, 1991.
(16) Edman, L.; Mets, U.;€ Rigler, R. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 6710–6715.
(17) Wennmalm, S.; Edman, L.; Rigler, R. Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 10641–10646.
(18) Raag, R.; Whitlow, M. FASEB J. 1995, 9, 73–80.
(19) Ogawa, M.; Kosaka, N.; Choyke, P. L.; Kobayashi, H. ACS
Chem. Biol. 2009, 4, 535–546.
(20) Nishi, K.; Takai, M.; Morimune, K.; Ohkawa, H. Biosci.,
Biotechnol., Biochem. 2003, 67, 1358–1367.
(21) Tsumoto, K.; Nakaoki, Y.; Ueda, Y.; Ogasahara, K.; Yutani, K.;
Watanabe, K.; Kumagai, I. Biochem. Biophys. Res. Commun. 1994, 201,
546–551.
(22) Ihara, M.; Yoshikawa, A.; Wu, Y.; Takahashi, H.; Mawatari, K.;
Shimura, K.; Sato, K.; Kitamori, T.; Ueda, H. Lab Chip 2010, 10, 92–100.
(23) Dillon, P. P.; Manning, B. M.; Daly, S. J.; Killard, A. J.;
O’Kennedy, R. J. Immunol. Methods 2003, 276, 151–161.
(24) de Wildt, R. M.; Mundy, C. R.; Gorick, B. D.; Tomlinson, I. M.
Nat. Biotechnol. 2000, 18, 989–994.
(25) Ueda, H.; Tsumoto, K.; Kubota, K.; Suzuki, E.; Nagamune, T.;
Nishimura, H.; Schueler, P. A.; Winter, G.; Kumagai, I.; Mahoney, W. C.
Nat. Biotechnol. 1996, 14, 1714–1718.
(26) Aburatani, T.; Ueda, H.; Nagamune, T. J. Biochem. 2002, 132,
775–782.
(27) Silva, A. P. d.; Fox, D. B.; Moody, T. S.; M.Weir, S. Trends
Biotechnol. 2001, 19, 29–34.
(28) Koide, Y.; Urano, Y.; Kenmoku, S.; Kojima, H.; Nagano, T.
J. Am. Chem. Soc. 2007, 129, 10324–10325.
(29) Torimura, M.; Kurata, S.; Yamada, K.; Yokomaku, T.; Kamagata,
Y.; Kanagawa, T.; Kurane, R. Anal. Sci. 2001, 17, 155–160.
(30) Sakamoto, K.; Murayama, K.; Oki, K.; Iraha, F.; Kato-Murayama,
M.; Takahashi, M.; Ohtake, K.; Kobayashi, T.; Kuramitsu, S.; Shirouzu,
M.; Yokoyama, S. Structure 2009, 17, 335–344.
(31) Neri, D.; Momo, M.; Prospero, T.; Winter, G. J. Mol. Biol. 1995,
246, 367–373.
(32) Iijima, I.; Hohsaka, T. ChemBioChem 2009, 10, 999–1006.
(33) Laemmli, U. K. Nature 1970, 227, 680–685.

17394 dx.doi.org/10.1021/ja205925j |J. Am. Chem. Soc. 2011, 133, 17386–17394