Simultaneous dual protein labeling using a triorthogonal reagent

Article abstract of DOI:10.1021/ja403813b

Construction of heterofunctional proteins is a rapidly emerging area of biotherapeutics. Combining a protein with other moieties, such as a targeting element, a toxic protein or small molecule, and a fluorophore or polyethylene glycol (PEG) group, can imp

Full text of DOI:10.1021/ja403813b

Article
pubs.acs.org/JACS
Simultaneous Dual Protein Labeling Using a Triorthogonal Reagent
Mohammad Rashidian,†,‡ Sidath C. Kumarapperuma,†,§ Kari Gabrielse,§ Adrian Fegan,§
Carston R. Wagner,*,§ and Mark D. Distefano*,‡
§
‡Department of Chemistry, and Department of Medicinal Chemistry, University of Minnesota, Minneapolis, Minnesota 55455,
United States
S
* Supporting Information
ABSTRACT: Construction of heterofunctional proteins is a rapidly
emerging area of biotherapeutics. Combining a protein with other
moieties, such as a targeting element, a toxic protein or small molecule,
and a fluorophore or polyethylene glycol (PEG) group, can improve the
specificity, functionality, potency, and pharmacokinetic profile of a protein.
Protein farnesyl transferase (PFTase) is able to site-specifically and
quantitatively prenylate proteins containing a C-terminal CaaX-box amino
acid sequence with various modified isoprenoids. Here, we describe the
design, synthesis, and application of a triorthogonal reagent, 1, that can be
used to site-specifically incorporate an alkyne and aldehyde group simultaneously into a protein. To illustrate the capabilities of
this approach, a protein was enzymatically modified with compound 1 followed by oxime ligation and click reaction to
simultaneously incorporate an azido-tetramethylrhodamine (TAMRA) fluorophore and an aminooxy-PEG moiety. This was
performed with both a model protein [green fluorescent protein (GFP)] as well as a therapeutically useful protein [ciliary
neurotrophic factor (CNTF)]. Next, a protein was enzymatically modified with compound 1 followed by coupling to an azido-
bis-methotrexate dimerizer and aminooxy-TAMRA. Incubation of that construct with a dihydrofolate reductase (DHFR)−
DHFR−anti-CD3 fusion protein resulted in the self-assembly of nanoring structures that were endocytosed into T-leukemia cells
and visualized therein. These results highlight how complex multifunctional protein assemblies can be prepared using this facile
triorthogonal approach.
labeling of a Rab GTPase for fluorescence resonance energy
transfer (FRET) applications by applying chemoselective native
chemical ligation and oxime ligation simultaneously.15 A C-
terminal oxime was generated via expression of a C-terminal
thioester, while an N-terminal cysteine (for subsequent
ligation) was revealed by Tobacco etch virus (TEV)-catalyzed
proteolysis. In other work, Schultz and co-workers developed a
method for site-specific dual labeling of proteins for FRET
analysis based on the use of selective cysteine alkylation
combined with non-natural amino acid incorporation of a
ketone moiety.16 Park and co-workers successfully incorporated
two unnatural amino acids bearing ketone and alkyne groups
into a protein for analysis of protein dynamics using a related
nonsense suppression approach.17 Very recently, Chen and co-
workers designed and synthesized bifunctional sialic acid
analogues containing azide and alkyne moieties for incorpo-
ration of two distinct chemical reporters into cellular sialylated
glycans for FRET imaging.18 While useful, that method is
limited to sialylated cell surface glycans and requires metabolic
activation of the bifunctional sialic acid analogue to the
corresponding cytosine monophosphate (CMP) sugar prior to
incorporation.
INTRODUCTION
■
Over the course of the past decade, bioorthogonal chemical
methods have been developed for the site-specific chemical
modification of proteins and used to alter their properties and
function.1−7 For example, fluorophores can be site-specifically
attached to proteins as a biophysical or cellular localization tool,
while protein−polymer conjugation is a well-established
method for modulating the in vivo behavior of proteins.8−11
In addition, a number of groups have reported bioorthogonal
approaches for the construction of bifunctional protein
assemblies. Schultz and co-workers coupled two antibody Fab
fragments via an alkyne−azide cycloaddition click reaction
using non-natural mutagenesis techniques.12 Bertozzi and co-
workers used an enzymatic formyl generating strategy13 to
generate an aldehyde that was then converted to a cyclooctyne-
or azide-functionalized protein via oxime formation followed by
reaction with other azide-modified peptides or proteins. Ploegh
and co-workers used a variation of sortagging to create N−N
and C−C protein conjugates by preparing pairs of azide- and
alkyne-containing proteins that were then linked via click
reactions.14 In the above examples, proteins equipped with a
single bioorthogonal group were modified with a second small
molecule, polymer, or protein bearing a complementary
functional group.
Previously, our group and others have exploited the high
specificity of protein farnesyl transferase (PFTase) to site-
Recently, progress toward the introduction of multiple
functional groups into proteins has also been made. Wu and
co-workers developed a strategy for site-specific two-color
Received: April 16, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
specifically modify proteins.19−22 PFTase catalyzes the transfer
of an isoprenoid group from farnesyl diphosphate (FPP; Figure
1A) to a cysteineyl sulfur atom present in a tetrapeptide
Schrodinger). The PFTase crystal structure was prepared using the
default settings in the protein preparation wizard as part of the
Maestro 9.0 package. A receptor grid large enough to encompass the
entire binding site for compound 1 was generated from the prepared
PFTase enzyme. A standard precision docking parameter was set, and
100 ligand poses per docking were run. The 20 top poses with the
lowest docking score were then further subjected to postdocking
minimization. The conformations with overall lowest energy were
chosen for display for each probe. The Glide (version 5.5) and
Maestro (version 9.0) software used in docking studies was licensed
from Schrodinger, LLC (New York).
Enzymatic Studies of FPP Analogue 1 Using a Continuous
Fluorescence Assay. Enzymatic reaction mixtures contained tris-
(hydroxymethyl)aminomethane (Tris)·HCl (50 mM, pH 7.5), MgCl2
(10 mM), ZnCl2 (10 μM), dithiothreitol (DTT) (5.0 mM), 2.4 μM N-
dansyl-GCVIA (2), 0.040% (w/v) n-dodecyl-β-D-maltoside, PFTase
(80 nM), and varying concentrations of compound 1 (0−50 μM), in a
final volume of 250 μL. The reaction mixtures were equilibrated at 30
°C for 1 min, initiated by the addition of PFTase, and monitored for
an increase in fluorescence (λex, 340 nm; λem, 505 nm) for
approximately 20 min. The initial rates of formation of products
were obtained as slopes in IU/min using least-squares analysis.
Corrections were applied to all of the rate calculations based on the
difference between the fluorescence intensity of the prenylated
product and the starting peptide. Assuming 100% conversion, the
difference corresponds only to the fluorescence of the total amount of
the product. The slope was then divided by the fluorescence difference
followed by multiplying by the total concentration of peptide (2.4
μM), which then gives the rate of formation of product in μM/s. It
should be noted that the KM values reported here are actually apparent
KM values, because the measurements were performed in only a single
peptide concentration. The data were fit to a Michaelis−Menten
model (V = (([E0]kcat[S])/(KM + [S])) using a nonlinear regression
program, to determine kcat and KM (see Figure S4 of the Supporting
Information).
Enzymatic Incorporation of Compound 1 into Green
Fluorescent Protein (GFP)−CVIA (3).27 Enzymatic reaction
mixtures (10 mL) contained Tris·HCl (50 mM, pH 7.5), MgCl2 (10
mM), KCl (30 mM), ZnCl2 (10 μM), DTT (5.0 mM), compound 7
(2.4 μM), compound 1 (30−50 μM), and PFTase (80−200 nM).
After incubation at 30 °C for 4 h, the reaction mixture was
concentrated using an Amicon Centriprep centrifugation device (10
000 MW cutoff). Next, excess of compound 1 was removed through a
NAP-5 (Amersham) column using Tris·HCl (50 mM, pH 7.5) as the
eluant. The subsequent protein concentration was calculated by
ultraviolet (UV) absorbance at 488 nm (ε = 55 000 M−1 cm−1).
Initial Evaluation of Bifunctionalized GFP (4) with Amino-
oxy-dansyl (S10) and Azido-tetramethylrhodamine (TAMRA)
(S11). Coupling Reaction between Bifunctionalized GFP (4) with
Aminooxy-dansyl (S10). Aminooxy-dansyl (S10) [3.2 μL of 10 mM
solution in dimethyl sulfoxide (DMSO)] was added to 42 μL of
compound 4 (stock solution of 60 μM in 0.1 M phosphate buffer at
pH 7.0). Phosphate buffer (1 M, 2.5 μL, pH 7.0) was added, and the
reaction was initiated by adding 40 mM m-phenylendiamine as a
catalyst and allowed to proceed for 2 h at room temperature. The
mixture was then purified by a NAP-5 column to remove excess dye.
Liquid chromatography−mass spectrometry (LC−MS) analysis of the
sample showed only oxime-ligated protein, and no free aldehyde was
detected, suggesting that the reaction had proceeded to completion. A
gradient of 0−100% solvent A (H2O, 0.1% HCO2H) to B (CH3CN,
0.1% HCO2H) in 25 min was used for the LC−MS analysis.
Figure 1. (A) Structures of FPP and compound 1. (B) Schematic
representation of farnesylation reaction with a protein containing a
CaaX box at its C terminus.
sequence (denoted as a CaaX box) positioned at the C
terminus of a protein (Figure 1B). Importantly, CaaX-box
sequences, such as CVIA, can be appended to the C termini of
many proteins, rendering them efficient substrates for PFTase.
Because PFTase can tolerate many simple modifications to the
isoprenoid substrate,23−26 it can be used to introduce a variety
of functional groups into proteins; PFTase and some
bioorthoganol substrates are already commercially available.
Previously, we have showed that aldehyde-containing FPP
analogues and alkyne-containing FPP analogues can be
successfully incorporated into proteins using this strat-
egy.19,23,27 Consequently, we envisioned that enzymatic
incorporation of a substrate analogue containing both alkyne
and aldehyde functionality could be used to generate proteins
with two distinct orthogonal functional groups for subsequent
elaboration. This approach would enable site-specific and
simultaneous protein modification with two orthogonal groups,
which can be used to improve the specificity, functionality,
potency, and pharmacokinetic profile of the protein. In contrast
to the method reported by Chen and co-workers noted above,
our approach requires no metabolic incorporation and can be
applied to essentially any protein that can be successfully
expressed with a C-terminal CaaX-box fusion.18 To implement
this strategy, a triorthogonal molecule (1) was designed that
contained an allylic diphosphate, rendering it a substrate for
PFTase, as well as an aldehyde and alkyne. The compound was
synthesized in nine steps and shown to be a substrate. Next, to
showcase the utility of this molecule for multiple modification
reactions on a single polypeptide, it was used to selectively
functionalize a protein with both fluorophore and polyethylene
glycol (PEG) groups. Those two groups were chosen as
examples because fluorophores can act as biophysical or cellular
localization tools, while PEG moieties are useful for modulating
the pharmacokinetics of protein-based drugs. To further
highlight the capabilities of this strategy, a larger multifunc-
tional nanoscale construct was prepared. Compound 1 was
used to prepare a protein labeled with a fluorophore and a
chemical dimerizer, which was then used to self-assemble a
nanoring structure that was targeted to and readily internalized
by T-leukemia cells. This illustrates the power of this
triorthogonal strategy for self-organization of protein molecules
into higher order structures, which is an integral part of living
cells.
Coupling Reaction between Bifunctionalized GFP (4) with Azido-
TAMRA (S11). Azide-TAMRA (S11) (7 μL of 2.2 mM solution in
DMSO) was added to 100 μL of compound 4 [stock solution of 40
μM in phosphate buffer (PB)]. CuSO4 (1 mM), tris(2-carboxyethyl)-
phosphine (TCEP) (1 mM), and tris(benzyltriazolylmethyl)amine
(TBTA) (100 μM) were added, and the reaction was allowed to
proceed for 3 h. LC−MS analysis of the sample showed the presence
of both the clicked protein 4a and the free alkyne protein 4. The ratio
of free alkyne protein 4 to its respective clicked product 4a was ∼2,
EXPERIMENTAL SECTION
Docking. To model compound 1 in the active site of PFTase (PDB
file 1JCR), docking was performed using Glide (version 5.5,
■
B
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Docked structure of triorthogonal reagent 1 bound to PFTase. (Left) Docked structure of compound 1 superimposed on the structure of
FPP bound to PFTase (obtained from X-ray crystallographic data; PDB file 1JCR). FPP is shown in blue. Compound 1 is shown in green (C), red
(O), and orange (P). The protein surface is shown in gray. (Right) Docked structure of compound 1 bound to PFTase in a space-filling CPK
representation (same color scheme).
indicating that only ∼35% completion of the click reaction had been
achieved within this time and range of reactant concentrations. The
reaction was repeated using 50% more azido-TAMRA (S11) (225
μM) and was allowed to proceed for 16 h. LC−MS analysis of the
reaction mixture showed >95% conversion of compound 4 to its
respective clicked product 4a. A gradient of 0−100% solvent A (H2O,
0.1% HCO2H) to B (CH3CN, 0.1% HCO2H) in 25 min was used for
the LC−MS analysis.
Simultaneous Coupling Reaction between Bifunctionalized GFP
(4) with Azido-TAMRA S11 and Aminooxy-dansyl S10. Azido-
TAMRA (S11) (7 μL of 2.2 mM solution in DMSO) and aminooxy-
dansyl (S10) (7 μL of 10 mM solution in DMSO) were added to 100
μL of compound 4 (stock solution of 40 μM in PB). CuSO4 (1 mM),
TCEP (1 mM), TBTA (100 μM), and m-phenylenediamine (mPDA)
(40 mM) were added, and the reaction was allowed to proceed for 15
h. LC−MS analysis of the sample showed >90% conversion of the
click reaction and >99% oxime ligation on the protein. A gradient of
0−100% solvent A (H2O, 0.1% HCO2H) to B (CH3CN, 0.1%
HCO2H) in 25 min was used for the LC−MS analysis.
Modification of Bifunctionalized GFP (4) with Azido-TAMRA
(S11) and Aminooxy-PEG (S12). Azido-TAMRA (S11) (7 μL of 2.2
mM solution in DMSO) and aminooxy-PEG (S12) (3 kDa, from
Quanta BioDesign, Ltd.) (2 μL of 30 mM solution in DMSO) were
added to 100 μL of compound 4 (stock solution of 40 μM in PB).
CuSO4 (1 mM), TCEP (1 mM), TBTA (100 μM), and mPDA (40
mM) were added, and the reaction was allowed to proceed for 15 h.
Sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−
PAGE) analysis of the reaction revealed highly selective and almost
complete (>95%) conversions for both oxime and click reactions.
Prenylation of CNTF−CVIA (5) with Compound 1. Enzymatic
reaction mixtures (20 mL) contained Tris·HCl (50 mM, pH 7.5),
MgCl2 (10 mM), KCl (30 mM), ZnCl2 (10 μM), DTT (5.0 mM),
CNTF−CVIA (2.4 μM), compound 1 (20 μM), and PFTase (200
nM). After incubation at room temperature for 90 min, the reaction
mixture was analyzed by LC−MS to ensure complete prenylation
using the gradient conditions described above.
Simultaneous Coupling Reactions between Bifunctionalized
GFP (4) with Azido-bis-methotrexate (bis-MTX) (S14) and
Aminooxy-TAMRA (S15) To Yield TAMRA−GFP−Bis-MTX (6).
Aminooxy-TAMRA (S15) (7 μL of 2.2 mM solution in DMSO) and
azide-bis-MTX (S14) (7 μL of 10 mM solution in DMSO) were
added to 100 μL of compound 4 (stock solution of 40 μM in PB).
CuSO4 (1 mM), TCEP (1 mM), TBTA (100 μM), and mPDA (40
mM) were added to the mixture, and the reaction was allowed to
proceed for 15 h. LC−MS analysis of the sample using the gradient
conditions described above showed >90% conversion of the click
reaction and >99% oxime ligation on the bifunctionalized protein 4.
The mixture was then purified using a NAP-5 column to remove the
excess reagents.
Chemically Induced Self-Assembly of Dihydrofolate Reduc-
tase (DHFR) Proteins by TAMRA−GFP−Bis-MTX (6). To study
the self-assembly of dimeric DHFR (DD) proteins by TAMRA−
GFP−bis-MTX (6), an equimolar mixture of relevant dimeric DHFR
protein and compound 6 was mixed and incubated at room
temperature for 2 h. Next, the samples were analyzed by size-exclusion
chromatography (SEC) for the formation of higher order species using
our established high-performance size-exclusion chromatography (HP-
SEC) method.28 Briefly, 100 μL of the incubated sample (10 μM) was
injected onto a Superdex G75 size-exclusion column (10 × 300 mm)
and eluted with P500 buffer [0.5 M NaCl, 50 mM KH2PO4, and 1 mM
ethylenediaminetetraacetic acid (EDTA) at pH 7] at 0.5 mL/min. The
elution profile was monitored at 280 nm.
Confocal Microscopy. HPB-MLT cells (0.5 × 106) were treated
with compound 6 (1 μM) for control experiments and with self-
assembled nanostructures at either 4 or 37 °C for 1 h in RPMI media.
Cells were then pelleted by centrifugation (400g for 5 min). After
being washed twice with phosphate-buffered saline (PBS), cells were
incubated on Poly-Prep slides coated with poly-L-lysine (Sigma) at 4
or 37 °C for 30 min. Cells were then fixed with 4% paraformaldehyde
solution for 10 min and washed 3 times with PBS. Finally, cells were
treated with ProLong Gold Antifade reagent with 4′,6-diamidino-2-
phenylindole (DAPI) (Invitrogen), and a coverslip was applied. After
overnight incubation, they were imaged by fluorescence confocal
microscopy using an Olympus FluoView 1000 BX2 upright confocal
microscope.
Flow Cytometry. HPB-MLT cells (1 × 106) were treated with
self-assembled nanostructures (1, 0.5, and 0.1 μM) at 4 °C for 1 h in
PBS buffer [containing 0.05% bovine serum albumin (BSA) and 0.1%
sodium azide]. Cells were pelleted (400g for 10 min), washed twice,
and resuspended in the supplemented PBS, and their fluorescence was
analyzed with a FACS Calibur flow cytometer (BD Biosciences). For
the positive control experiment, HPB-MLT cells (1 × 106) were
incubated with 40 nM fluorescein isothiocyanate (FITC)-labeled
UCHT-1 (anti-CD3 monoclonal antibody). After 2 h of incubation,
cells were washed, resuspended, and analyzed by flow cytometry. A
negative control experiment was performed with self-assembled
nanostructures treated with CD3 negative Daudi B lymphoma cells
(1 × 106 cells).
RESULTS AND DISCUSSION
■
Design, Synthesis, and Enzymatic Reaction with a
Triorthogonal Protein Labeling Reagent. To obtain site-
selective protein labeling, triorthogonal compound 1 was
C
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 3. (A) Schematic representation of prenylation of GFP−CVIA (3) with FPP analogue 1 to yield bifunctionalized GFP (4). The deconvoluted
mass spectra are shown adjacent to the compound structures, showing successful prenylation of GFP 3. (B) Schematic representation of the click
reaction between bifunctionalized GFP (4) with azido-TAMRA (S11) to yield labeled GFP 4a. The deconvoluted mass spectrum of the clicked
product is shown adjacent to the product, showing a successful click reaction between compounds S11 and 4. (C) Schematic representation of the
oxime ligation reaction between bifunctionalized GFP (4) and aminooxy-dansyl (S10) to yield oxime GFP 4b. The deconvoluted mass spectrum is
shown adjacent to the oxime product, indicating a successful oxime ligation reaction. (D) Schematic representation of simultaneous click and oxime
reactions between bifunctionalized GFP (4) with aminooxy-dansyl (S10) and azido-TAMRA (S11) to yield GFP 4c. The deconvoluted mass
spectrum of the product is shown, showing successful simultaneous click and oxime reactions between compounds S10, S11, and 4.
designed. The molecule contains a geranyl substructure
including an allylic diphosphate to render it a substrate for
PFTase along with aldehyde and alkyne functional groups to
allow for two distinct bioorthogonal reactions. Docking of
compound 1 into the active site of PFTase indicates that
compound 1 can adopt a conformation similar to that of the
physiological substrate, FPP (left panel of Figure 2), and that
sufficient room exists to accommodate the 1,3,5-trisubstituted
aryl ring (right panel of Figure 2). Inspection of the docked
structure also suggests that ortho substitution to the phenoxide
oxygen could lead to unfavorable steric interactions with the
protein. Hence, for wild-type PFTase, the 1,3,5-substitution
pattern appears to be the optimal design.
Compound 1 was then synthesized from commercially
available geraniol and 3,5-dihydroxybenzaldehyde in nine steps
(see Figure S3 of the Supporting Information). In brief,
tetrahydropyranyl acetal (THP)-protected geraniol was initially
oxidized at C-8 to a terminal alcohol, followed by bromination
of the hydroxyl group using CBr4 and PPh3. The 3,5-alkyne−
aldehyde−functionalized phenol was prepared from 3,5-
dihydroxybenzaldehyde via a Sonogashira Pd−Cu-catalyzed
cross-coupling reaction29 in three steps. The modified phenol
was alkylated with the aforementioned bromide using K2CO3 as
the base. The THP group was removed, and the alcohol was
converted to the corresponding diphosphate via a direct
phosphorylation strategy employing (HNEt3)2HPO4 and
CCl3CN as the activating reagent. Subsequent purification by
reversed-phase high-pressure liquid chromatography (RP-
HPLC) produced the desired bifunctional aldehyde−alkyne
1
analogue 1, whose structure was confirmed by H nuclear
magnetic resonance (NMR), 31P NMR, and high-resolution
electrospray ionization mass spectrometry (HR-ESI−MS).
Initially, prenylation reactions containing a model peptide, N-
dansyl-GCVIA (2), analogue 1, and PFTase were performed,
and the reactions were monitored via a continuous
fluorescence-based enzyme assay, as previously reported,30
which demonstrated that compound 1 is an alternative
substrate for the enzyme. Next, a kinetic analysis of that
reaction was performed using varying concentrations of
compound 1 and N-dansyl-GCVIA (2) in the presence of
PFTase, showing that the enzymatic process obeyed saturation
kinetics. Steady-state kinetic parameters for the prenylation
reaction using the bifunctional aldehyde−alkyne analogue are
provided in Figure S4 of the Supporting Information. This
analysis revealed that the catalytic efficiency for alternative
substrate 1 is reduced relative to that of FPP, with a kcat/KM
D
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
value of 0.02 (relative to FPP). We found that a decrease in kcat
constituted the major reason for the diminished catalytic
activity with the analogue, with the kcat value for compound 1
(kcat = 0.0123 s−1) observed to be 42-fold lower than the kcat for
FPP (kcat = 0.52 s−1), while no substantial difference was
observed between the KM values for the analogue (KM = 2.52
μM) and FPP (KM= 1.71 μM). It should be noted that, while
the docking results shown above (left panel of Figure 2)
suggest that compound 1 binds in a conformation similar to
that of FPP, small differences are apparent in the vicinity of C-1
of the isoprenoid and the diphosphate, which may account for
the reduced reactivity. Nevertheless, despite this lower activity
with compound 1, using elevated levels of PFTase (200 nM), it
is relatively easy to drive these enzymatic reactions to
completion in 2 h.
Application of Triorthogonal Protein Labeling to an
Individual Protein. With the ability of analogue 1 to be
incorporated by PFTase established, we next evaluated the
utility of the analogue for selective protein modification.
Accordingly, compound 1 was incubated with GFP−CVIA (3)
in the presence of PFTase (4 h at room temperature). That
choice of reaction time was based on our earlier work with the
peptide substrate N-dansyl-GCVIA (2), where a 2 h reaction
resulted in complete conversion. Concentration by ultra-
centrifugation followed by SEC to remove excess compound
1 yielded pure bifunctionalized GFP (4) (Figure 3A).
Completion of the reaction was confirmed by LC−ESI−MS,
with the major peak corresponding to prenylated GFP (4) and
a minor peak corresponding to a negligible amount of
unreacted GFP−CVIA (3) (Figure 3A).
In previous studies, we had shown that aldehyde−GFP- and
alkyne−GFP-modified proteins could be derivatized to produce
oxime-linked or clicked products, respectively.23,27 In this study,
we explored simultaneous oxime and click reactions on a single
prenylated protein. First, to demonstrate the orthogonality of
the two reactions, a separate oxime ligation or click reaction
was carried out with the bifunctionalized protein. Thus, oxime
ligation on compound 4 was performed with an aminooxy-
dansyl reagent (S10) using mPDA as the catalyst.31 LC−MS
analysis of the crude ligation reaction mixture confirmed
complete conversion to the oxime-linked product 4b in less
than 1 h (Figure 3C). The click reaction was tested on
compound 4 with azido-TAMRA (S11) using TCEP, TBTA,
and copper as the catalyst to yield GFP−TAMRA 4a (Figure
3B). LC−MS analysis revealed that >80% conversion was
obtained within 12 h of reaction, using equimolar concen-
trations of prenylated protein and the coupling partner (azido-
TAMRA, S11) as used in the above oxime ligation reaction
(Figure 3B). Typically, the oxime ligation reaction proved to be
modestly more efficient than the click reaction. When the
oxime and click reactions were carried out simultaneously on
the modified protein to yield 4c using equimolar concentrations
of reagents (vide supra), quantitative conversion for the oxime
reaction and ∼90% conversion for the click reaction after 12 h
were observed (Figure 3D).
Figure 4. (A) Schematic representation of oxime ligation and click
reaction of the bifunctionalized prenylated protein 4 with aminooxy-
PEG (3 kDa) (S12 and azido-TAMRA (S11) followed by analysis of
the reactions via SDS−PAGE. (B) SDS−PAGE analysis of the
aforementioned click and oxime reactions. Densitometry analysis on
the SDS−PAGE of the aforementioned reactions revealed highly
selective and almost complete (>95%) conversions for both oxime and
click reactions on the prenylated protein. The lower bands were
visualized by staining with Coomassie Blue, while the upper bands
were detected via in-gel fluorescence imaging of TAMRA. Lane 1,
reaction mixture containing protein 4, PEG−ONH2, and TAMRA−
N3; lane 2, reaction mixture containing protein 4 and TAMRA−N3;
lane 3, reaction mixture containing protein 4 and PEG−ONH2; and
lane 4, solution containing only protein 4. Densitometry analyses were
performed using ImageJ, version 1.46.
LC−MS analysis on the reaction mixture revealed complete
conversion of bifunctionalized protein 4 to GFP−TAMRA
(4a). Subsequently, the product was incubated with aminooxy-
PEG (3 kDa) (S12, from Quanta BioDesign) at pH 7 in the
presence of mPDA (40 mM) as a catalyst for 2 h. SDS−PAGE
analysis of the aforementioned reactions revealed highly
selective and almost complete (>95%) conversions for both
oxime and click reactions with the prenylated protein (Figure
4). The lower bands (Figure 4B) were visualized by staining
with Coomassie Blue, while the upper bands were detected via
in-gel fluorescence imaging of TAMRA. That data showed that
the reaction mixture including protein 4, aminooxy-PEG, and
azido-TAMRA contained a species with lower mobility in the
SDS−PAGE gel because of the higher mass obtained upon
addition of the PEG group; a dark band was also observed via
in-gel fluorescence imaging of TAMRA (lane 1 in Figure 4B)
that comigrated with the lower mobility species noted above.
When no PEG reagent was present in the reaction mixture, the
lower mobility band was absent (lane 3 in Figure 4B), and
when azido-TAMRA was not present in the reaction, only the
higher mobility species that migrated near the starting protein
was fluorescently labeled. Taken together, these results
establish that two different functionalities can be installed
simultaneously into a protein via two distinct bioorthogonal
reactions using this triorthogonal labeling strategy.
After optimizing the conditions for the oxime and click
reactions on the bifunctional prenylated protein 4, we next
evaluated the preparation of a PEG−GFP−TAMRA (4d)
(Figure 4A). We chose TAMRA and PEG as examples because
fluorophore can act as a biophysical or cellular localization tool,
while PEG can improve protein pharmacokinetics. Protein 4
was incubated first with azido-TAMRA (S11) under the
optimized click reaction conditions described above for 12 h.
Application of Triorthogonal Protein Labeling to a
Protein of Biomedical Importance. Having established the
utility of this method for C-terminal site-specific modification
with a model protein, GFP, we decided to illustrate its use by
modifying a biomedically important protein. CNTF is a
E
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
member of a class of proteins that promote neuron survival
during inflammatory events and can stimulate neurite out-
growth.32 Recently, CNTF has been studied extensively as a
possible therapeutic agent for slowing retinal degeneration.33,34
Thus, we elected to investigate the preparation of a
bifunctionalized CNTF using the new strategy reported here.
To accomplish this, purified CNTF−CVIA (5),31 a form of
CNTF engineered to contain a C-terminal CaaX box, was
prenylated with analogue 1 followed by labeling using azido-
TAMRA (S11) and PEGylation with aminooxy-PEG (S13).
LC−MS analysis (Figure 5) confirmed successful prenylation
and subsequent modification of CNTF by both azido-TAMRA
(S11) and aminooxy-PEG (S13).
antibodies (scFvs) and peptides, thus allowing them to be
targeted to cell surface receptors. In particular, we have
previously demonstrated that anti-CD3 scFv−DHFR2 fusion
proteins containing a 13 amino acid linker or a 1 amino acid
linker will self-assemble in the presence of a bis-MTX into bi-
and octavalent anti-CD3 CSANs, respectively. Further, we have
demonstrated that the anti-CD3 CSANs bind specifically to
CD3+ T-leukemia cells with dissociation constants comparable
to the parental monoclonal antibody (UCHT-1 mAb, Kd = 1.8
nM; bivalent anti-CD3 CSANs, Kd = 3.5 nM; and octavalent
anti-CD3, Kd = 0.93 nM). In addition, the anti-CD3 CSANs
undergo rapid clathrin-dependent endocytosis by T-leukemia
cells, which enables drug or oligonucleotide delivery.28,47−49 To
demonstrate the ability of the CSANs to carry out protein
delivery, we decided to explore the use of the dual-labeling
strategy reported here to link the bis-MTX element to a
protein. Accordingly, we designed and prepared anti-CD3-
CSANs capable of delivering a protein cargo into T-leukemia
cells. Thus, bis-MTX−GFP−TAMRA (6) was prepared. A
previously reported amino-bis-MTX trilinker48 (S16) was
coupled to 5-azidopentanoic acid using standard peptide
chemistry to obtain azido-bis-MTX (S14). Next, the bifunc-
tional prenylated protein 4 described above was incubated with
aminooxy-TAMRA (S15) and azido-bis-MTX (S14) for 15 h
using the optimized reaction conditions (vide supra).
Subsequent LC−MS analysis of the reaction mixture revealed
the successful simultaneous conjugation of the protein with
both aminooxy-TAMRA and azido-bis-MTX (Figure 6).
Figure 5. (A) Prenylation of CNTF−CVIA (5) with FPP analogue 1
followed by simultaneous fluorophore labeling and PEGylation of the
prenylated protein 5a via click and oxime ligations with azido-TAMRA
(S11) and aminooxy-PEG (S13) to yield compound 5b. (B−D)
Deconvoluted mass spectra are shown, showing successful prenylation
(B, pure CNTF; C, prenylated CNTF) and simultaneous click and
oxime reactions (D).
Application of the Triorthogonal-Protein-Labeling
Reagent in Macromolecular Protein Self-Assembly
(Figure 7). Self-organization of protein molecules into higher
order structures is an integral part of living cells. Unlike DNA-
and RNA-based nanostructures,35−38 the rules governing
protein self-assembly are only beginning to be defined. Much
of the previous effort to construct biomimetic protein
assemblies has been based on direct protein−protein
interactions through the incorporation of complementary
amino acid residue surface interactions.39 As an alternative,
chemical biological approaches for macromolecular protein
assembly based on metal coordination,40 polymer conjuga-
tion,41 enzyme−inhibitor interactions,42−44 and chemically
induced dimerization45 have begun to be explored.
In previous work, we reported on the design of chemically
self-assembled nanoring structures (CSANs) constructed from
dimeric forms of DHFR and bis-MTX.46 Upon mixing, the two
components spontaneously assemble into a mixture of compact
rings and polymeric structures that can be readily separated by
SEC. The ring structures have been studied by dynamic light
scattering, transmission electron microscopy (TEM), and
atomic force microscopy (AFM), and it has been shown that
the number of DHFR and MTX dimers in the ring can be
controlled by the length of the linker region that joins the two
DHFR monomers; because of the high affinity of MTX for
DHFR and, thus, high effective molarity, the cyclic forms resist
linearization in the absence of an exogenous inhibitor of
DHFR.46,47 The CSANs are highly stable exhibiting Tm values
between 62 and 65 °C (unpublished observation). The utility
of the nanorings has been expanded to incorporate single-chain
Figure 6. (A) Schematic representation of simultaneous oxime ligation
and click reactions of the bifunctionalized prenylated protein 4 with
azido-bis-MTX (S14) and aminooxy-TAMRA (S15). (B and C) ESI−
MS analysis of compounds 4 and 6 with the deconvoluted mass
spectra shown in the insets, respectively.
Overall, the conjugation reactions appear to be highly efficient,
because little free starting material was observed by LC−MS;
the high efficiency seen in this case is similar to results noted
above for other GFP and CNTF conjugates. With the bis-
MTX- and TAMRA-functionalized GFP protein 6 in hand, we
then evaluated the chemically induced self-assembly of this
protein with dimeric DHFR (DD) proteins by HP-SEC. After
simple mixing of the two entities and incubation at room
temperature for 2 h, the samples were analyzed by HP-SEC.
The elution profiles of the HP-SEC chromatograms indicated
the predominant formation of the internal monomeric species
with 13-DHFR−DHFR (13DD) and a much higher order
species with 1-DHFR−DHFR (1DD; see Figures S10 and S11
F
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
collected and concentrated by ultrafiltration, and their final
concentration was determined on the basis of the collected
peak area. This material was used for cellular internalization
studies (Figure 9). Similar SEC experiments and cell internal-
ization experiments were performed with assemblies prepared
from 1DDCD3 as well, which gave similar results to those
obtained with 13DDCD3 (see Figure S12 of the Supporting
Information for SEC data and Figures S13 and S14 of the
Supporting Information for cell experiments).
To study the functionality of the TAMRA−GFP−13DDCD3
and TAMRA−GFP−1DDCD3 CSANs (7 and 8, respectively),
we evaluated their ability to specifically deliver the protein
nanostructure to CD3+ T-leukemia cells. The cells were
incubated with compound 7 for 1 h (at 4 and 37 °C) and
evaluated by confocal laser scanning microscopy for cellular
internalization. Both the GFP and TAMRA chromophores
were excited individually to observe the GFP and TAMRA
emission. Additionally, FRET was observed between the two
fluorophores when GFP was excited and TAMRA emission was
monitored. First, control experiments in which CD3+ T-
leukemia cells were treated only with the GFP−TAMRA
conjugate 6 showed minimal fluorescence (Figure 9G). In
addition, no significant binding to CD3+ T-leukemia cells by
CSANs prepared with compound 6 was observed by FACS (see
Figure S15 of the Supporting Information) Confocal
microscopy imaging confirmed the delivery of the nanostruc-
tures (7 and 8) into the cells via an energy-dependent
endocytic process (Figure 9 and Figures S12−S14 of the
Supporting Information). At 4 °C, TAMRA fluorescence was
observed on the cell membranes (Figure 9A), and at 37 °C,
distinct punctate structures (Figure 9B) were observed within
the cells, indicating the energy-dependent uptake of both GFP
and TAMRA. Overlay of the GFP/TAMRA channels indicated
that the two fluorophores were present in the same location
(Figure 9H), while the positive FRET signal was consistent
with the GFP−TAMRA conjugate 6 having remained intact in
the cells during the time period of the experiment (24 h)
(panels I and J of Figure 9). The observed temperature-
dependent uptake of nanostructures at 37 °C and the punctate
structures are consistent with our previous observations with
anti-CD3 scFv CSANs.28
Figure 7. Application of the dual protein-labeling strategy in
macromolecular protein self-assembly. A high-avidity “effector−
antibody−fluorophore” conjugate for therapeutic cargo delivery and
tracking was designed and synthesized.
of the Supporting Information). Note that, in this nomencla-
ture system, 13DD is a single polypeptide chain incorporating
two DHFR domains separated by a 13 residue spacer, whereas
1DD is a similar dimer separated by a 1 residue spacer.28
Having confirmed the self-assembly with non-targeting DHFR
proteins, we then moved to construct the TAMRA−GFP−
DD−anti-CD3 protein nanostructures by mixing 13DD-anti-
CD3 (13DDCD3) and 1DD-anti-CD3 (1DDCD3) with
compound 6. These latter DHFR-based constructs are similar
to 13DD and 1DD, with the exception that they each have a
single-chain antibody (anti-CD3) fused to the C terminus of
the DHFR dimer; the presence of the anti-CD3 moiety allows
these constructs to be targeted to cells displaying the CD3
antigen. The assemblies containing the targeting proteins
generated a similar pattern in the SEC profile compared to the
constructs prepared from DHFR dimers lacking the anti-CD3
domain (Figure 8).
Consistent with what was observed with the corresponding
non-targeting DHFR proteins, TAMRA−GFP−DDCD3
CSANs prepared with 13DDCD3 (7) exhibited predominantly
internal monomer, dimers, and small amounts of higher order
species. The species eluting between 20 and 22.5 min (7) were
As was noted in earlier studies with anti-CD3 CSANs, flow
cytometric analysis (see Figures S15 and S16 of the Supporting
Information) revealed that the nanostructure binds specifically
to CD3+ HPB-MLT cells but not CD3− Daudi cells (see Figure
S19 of the Supporting Information). When compared, little
difference by flow cytometry was observed between the binding
by CSANs prepared from 13DDCD3 (7) and 1DDCD3 (8) to
CD3+ HPB-MLT cells (Figure 10 and Figures S17 and S18 of
the Supporting Information).
CONCLUSION
■
In summary, we have designed a triorthogonal reagent to site-
specifically modify a protein with two orthogonal groups that
can be subsequently functionalized in a single one-pot
procedure. Our approach relies on the selective tagging of
proteins containing an appended farnesyl transferase substrate
sequence. The incorporation of the bifunctional ethynyl-
hydroxybenzaldehyde into the farnesyl group facilitates the
facile labeling of proteins with two different moieties, thus
expanding the potential capabilities of the tagged protein. The
simultaneous tagging of the cytokine, CNTF, with PEG and a
fluorophore, suggests that the alteration and monitoring of the
Figure 8. Self-assembly of antibody nanorings 7 observed by SEC
(13DDCD3 with compound 6). Blue curve, bis-MTX−GFP−TAMRA
(6, peak A); green curve, monomeric 13DDCD3 (peak B); black
curve, induced oligomerization of 13DDCD3 with compound 6
indicating the major products as the internal monomer (peak C) and
dimer (peak D). The shaded area indicates the material collected and
used for subsequent biological studies. HP-SEC was performed using a
Superdex G75 column with a flow rate of 0.5 mL/min and monitoring
at 280 nm.
G
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 9. continued
leukemia cells at 37 °C: (C and E) compound 7 incubated with CD3+
T-leukemia cells for 1 h (C, DAPI and GFP channels; E, TAMRA
channel), (D and F) compound 7 incubated with CD3+ T-leukemia
cells for 24 h (D, DAPI and GFP channels; F, TAMRA channel), (G)
HPB-MLT cells treated with bis-MTX−GFP−TAMRA (6) for 1 h
(control; blue = DAPI nuclear stain), (H) overlay of the GFP and
TAMRA channels indicating the intact assembled protein (co-localized
yellow punctate), and (I and J) FRET between GFP and TAMRA
observed by exciting the cells with 488 nm laser and a BA560/620 nm
emission filter.
Figure 10. Determination of the dose-dependent CD3 receptor
binding of assembled protein nanostructures and comparison of the
TAMRA−GFP−13DDCD3 (7) and TAMRA−GFP−1DDCD3 (8)
bindings using mean fluorescence values obtained from the TAMRA
channel. HPB-MLT cells were treated with different concentrations of
self-assembled nanostructures, 7 or 8, at 37 °C for 1 h before flow
cytometry analysis. For the FACS measurements, 10 000 cells were
counted for each sample and the average mean fluorescence intensity
from three replicates is plotted here along with the standard error
(indicated by the error bars).
biodistribution properties and pharmacokinetic behavior of a
potential therapeutic protein can be easily coupled. In addition,
we have demonstrated that this approach can be used to direct
the assembly of protein complexes while monitoring their
interactions with cells. Future studies will examine the ability of
our trifunctional bioconjugation approach to enhance the
plasma half-life of CaaX-box PEGylated proteins by simulta-
neously monitoring their tissue distribution with an appended
fluorophore or radiolabel. It should also be possible to combine
the method reported here with other strategies, noted
above,15−18 for bioorthogonal protein labeling to further tailor
the properties of proteins or create more complex constructs
thereof.
ASSOCIATED CONTENT
■
S
* Supporting Information
Figure 9. (A and B) Internalization studies of TAMRA−GFP−
1DDCD3 (8) with HPB-MLT T-leukemia cells at either (left) 4 °C or
(right) 37 °C. Cells were treated for 1 h. Images indicate red punctate
structures (left) on the cell membrane at 4 °C or (right) within the cell
at 37 °C (TAMRA channel). Magnification = 60×. (C−J) Internal-
ization studies of TAMRA−GFP−13DDCD3 (7) with HPB-MLT T-
Detailed experimental procedures for synthesis of compound 1,
enzymatic studies of FPP analogue 1 as a substrate for PFTase,
synthesis of azido-bis-MTX (S14), and confocal microscopy
and flow cytometry images. This material is available free of
charge via the Internet at http://pubs.acs.org.
H
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(26) Subramanian, T.; Liu, S.; Troutman, J. M.; Andres, D. A.;
Spielmann, H. P. ChemBioChem 2008, 9, 2872−2882.
(27) Duckworth, B. P.; Xu, J.; Taton, T. A.; Guo, A.; Distefano, M. D.
Bioconjugate Chem. 2006, 17, 967−974.
(28) Li, Q.; So, C. R.; Fegan, A.; Cody, V.; Sarikaya, M.; Vallera, D.
A.; Wagner, C. R. J. Am. Chem. Soc. 2010, 48, 17247−17257.
(29) Sonogashira, K. J. Organomet. Chem. 2002, 653, 46−49.
(30) Pompliano, D. L.; Gomez, R. P.; Anthony, N. J. J. Am. Chem.
Soc. 1992, 114, 7945−7946.
(31) Rashidian, M.; Mahmoodi, M. M.; Shah, R.; Dozier, J. K.;
Wagner, C. R.; Distefano, M. D. Bioconjugate Chem. 2013, 24, 333−
342.
AUTHOR INFORMATION
■
Corresponding Authors
wagne003@umn.edu
diste001@umn.edu
Author Contributions
†Mohammad Rashidian and Sidath C. Kumarapperuma
contributed equally to this work.
Notes
The authors declare no competing financial interest.
(32) Ip, N. Y.; Yancopoulos, G. D. Annu. Rev. Neurosci. 1996, 19,
491−515.
ACKNOWLEDGMENTS
■
Financial support for these studies through National Institutes
of Health (NIH) Grants GM058842 (to Mark D. Distefano),
GM084152 (to Mark D. Distefano), and CA120116 (to
Carston R. Wagner) and the University of Minnesota is
gratefully acknowledged.
(33) Rhee, K. D.; Yang, X.-J. Adv. Exp. Med. Biol. 2010, 664, 647−
654.
(34) Wen, R.; Tao, W.; Li, Y.; Sieving, P. A. Prog. Retinal Eye Res.
2012, 31, 136−151.
(35) Rothemund, P. W. K. Nature 2006, 440, 297−302.
(36) Pistol, C.; Dwyer, C. Nanotechnology 2007, 18, 125305.
(37) Guo, P. Nat. Nanotechnol. 2010, 5, 833−842.
(38) Ke, Y.; Lindsay, S.; Chang, Y.; Liu, Y.; Yan, H. Science 2008, 319,
180−183.
REFERENCES
■
(1) Bertozzi, C. R. Acc. Chem. Res. 2011, 44, 651−653.
(2) Prescher, J. A.; Bertozzi, C. R. Nat. Chem. Biol. 2005, 1, 13−21.
(3) Mamidyala, S. K.; Finn, M. G. Chem. Soc. Rev. 2010, 39, 1252.
(4) Christman, K. L.; Broyer, R. M.; Tolstyka, Z. P.; Maynard, H. D.
J. Mater. Chem. 2007, 17, 2021.
(39) Matsuura, K.; Watanabe, K.; Matsuzaki, T.; Sakurai, K.;
Kimizuka, N. Angew. Chem., Int. Ed. 2010, 49, 9662−9665.
(40) Salgado, E. N.; Radford, R. J.; Tezcan, F. A. Acc. Chem. Res.
2010, 43, 661−672.
(5) Prost, L. R.; Grim, J. C.; Tonelli, M.; Kiessling, L. L. ACS Chem.
(41) Hu, M.; Qian, L.; Brinas, R. P.; Lymar, E. S.; Hainfeld, J. F.
̃
Biol. 2012, 7, 1603−1608.
(6) Wang, T.; Kartika, R.; Spiegel, D. A. J. Am. Chem. Soc. 2012, 134,
8958−8967.
(7) Park, S.; Yousaf, M. N. Langmuir 2008, 24, 6201−6207.
(8) Keppler, A.; Pick, H.; Arrivoli, C.; Vogel, H.; Johnsson, K. Proc.
Natl. Acad. Sci. U. S. A. 2004, 101, 9955−9959.
(9) Yin, J.; Liu, F.; Li, X.; Walsh, C. T. J. Am. Chem. Soc. 2004, 126,
7754−7755.
(10) Chen, I.; Ting, A. Y. Curr. Opin. Biotechnol. 2005, 16, 35−40.
(11) Discher, D. E.; Eisenberg, A. Science 2002, 297, 967−973.
(12) Kim, C. H.; Axup, J. Y.; Dubrovska, A.; Kazane, S. A.; Hutchins,
B. A.; Wold, E. D.; Smider, V. V.; Schultz, P. G. J. Am. Chem. Soc.
2012, 9918−9921.
(13) Hudak, J. E.; Barfield, R. M.; de Hart, G. W.; Grob, P.; Nogales,
E.; Bertozzi, C. R.; Rabuka, D. Angew. Chem., Int. Ed. 2012, 51, 4161−
4165.
Angew. Chem., Int. Ed. Engl. 2007, 46, 5111−5114.
(42) Chen, G.; Jiang, M. Chem. Soc. Rev. 2011, 40, 2254.
(43) Uhlenheuer, D. A.; Petkau, K.; Brunsveld, L. Chem. Soc. Rev.
2010, 39, 2817.
(44) Lai, J. R.; Fischbach, M. A.; Liu, D. R.; Walsh, C. T. J. Am. Chem.
Soc. 2006, 128, 11002−11003.
(45) Fegan, A.; White, B.; Carlson, J. C. T.; Wagner, C. R. Chem. Rev.
2010, 110, 3315−3336.
(46) Carlson, J. C. T.; Jena, S. S.; Flenniken, M.; Chou, T.; Siegel, R.
A.; Wagner, C. R. J. Am. Chem. Soc. 2006, 128, 7630−7638.
(47) Li, Q.; Hapka, D.; Chen, H.; Vallera, D. A.; Wagner, C. R.
Angew. Chem., Int. Ed. 2008, 47, 10179−10182.
(48) Fegan, A.; Kumarapperuma, S. C.; Wagner, C. R. Mol.
Pharmaceutics 2012, 9, 3218−3227.
(49) Gangar, A.; Fegan, A.; Kumarapperuma, S. C.; Wagner, C. R. J.
Am. Chem. Soc. 2012, 134, 2895−2897.
(14) Witte, M. D.; Cragnolini, J. J.; Dougan, S. K.; Yoder, N. C.;
Popp, M. W.; Ploegh, H. L. Proc. Natl. Acad. Sci. U. S. A. 2012, 109,
11993−11998.
(15) Yi, L.; Sun, H.; Itzen, A.; Triola, G.; Waldmann, H.; Goody, R.
S.; Wu, Y.-W. Angew. Chem., Int. Ed. Engl. 2011, 50, 8287−8290.
(16) Brustad, E. M.; Lemke, E. A.; Schultz, P. G.; Deniz, A. A. J. Am.
Chem. Soc. 2008, 130, 17664−17665.
(17) Kim, J.; Seo, M.-H.; Lee, S.; Cho, K.; Yang, A.; Woo, K.; Kim,
H.-S.; Park, H.-S. Anal. Chem. 2013, 85, 1468−1474.
(18) Feng, L.; Hong, S.; Rong, J.; You, Q.; Dai, P.; Huang, R.; Tan,
Y.; Hong, W.; Xie, C.; Zhao, J.; Chen, X. J. Am. Chem. Soc. 2013, 135,
9244−9247.
(19) Rashidian, M.; Dozier, J. K.; Lenevich, S.; Distefano, M. D.
Chem. Commun. 2010, 46, 8998.
(20) Weinrich, D.; Lin, P.-C.; Jonkheijm, P.; Nguyen, U. T. T.;
Schroder, H.; Niemeyer, C. M.; Alexandrov, K.; Goody, R.;
̈
Waldmann, H. Angew. Chem., Int. Ed. 2010, 49, 1252−1257.
(21) Gauchet, C.; Labadie, G. R.; Poulter, C. D. J. Am. Chem. Soc.
2006, 128, 9274−9275.
(22) Kim, M.; Kleckley, T. S.; Wiemer, A. J.; Holstein, S. A.; Hohl, R.
J.; Wiemer, D. F. J. Org. Chem. 2004, 69, 8186−8193.
(23) Rashidian, M.; Song, J. M.; Pricer, R. E.; Distefano, M. D. J. Am.
Chem. Soc. 2012, 134, 8455−8467.
(24) Placzek, A. T.; Gibbs, R. A. Org. Lett. 2011, 3576−3579.
(25) Labadie, G. R.; Viswanathan, R.; Poulter, C. D. J. Org. Chem.
2007, 72, 9291−9297.
I
dx.doi.org/10.1021/ja403813b | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX