Phosphopeptide-dependent labeling of 14 - 3 - 3 ζ proteins by fusicoccin-based fluorescent probes

Article abstract of DOI:10.1002/anie.201106995

Fluorescent combination: Cell-penetrating probes derived from the diterpene fusicoccin can form ternary complexes with 14-3-3 proteins and phosphopeptide ligands, whereupon the probes site-specifically attach a fluorescent tag onto the surface of the 14-3-3 proteins. Copyright

Full text of DOI:10.1002/anie.201106995

Angewandte
Chemie
DOI: 10.1002/anie.201106995
Protein–Protein Interactions
Phosphopeptide-Dependent Labeling of 14–3–3z Proteins by
Fusicoccin-Based Fluorescent Probes**
Michiko Takahashi, Akie Kawamura, Nobuo Kato, Tsuyoshi Nishi, Itaru Hamachi, and
Junko Ohkanda*
14–3–3 proteins play a critical role in serine/threonine kinase
dependent signaling pathways through protein–protein inter-
phosphorylated ligands would also enhance our understand-
ing of the role of 14–3–3 proteins in PPI networks.
actions (PPIs) with multiple phosphorylated ligands.[1]
A
The diterpene fusicoccin A[4] (FC; Figure 1A) is a phyto-
toxin that is produced by Phomopsis amygdali. FC activates
plant plasma membrane H+-ATPase through the formation of
a ternary complex with plant 14–3–3.[5] Recent collaborative
studies led by Kato[6] to semisynthetically modify FC resulted
in the identification of several potent FC analogues that are
active against human cancer cells, however, the biology
underlying this activity is not fully understood. Further
development of FC-based antitumor agents for medicinal
purposes will thus necessitate identification of their target(s)
and elucidation of their mechanism of action.
ligand-dependent 14–3–3 detection technique would facilitate
elucidation of 14–3–3-related intracellular signaling networks.
Herein, we describe phosphopeptide-dependent fluorescent
labeling of 14–3–3z by using cell-penetrating probes derived
from the diterpene natural product fusicoccin.
The 14–3–3 proteins are a family of dimeric conserved
regulatory proteins that are expressed in all eukaryotic cells.
Each of the highly helical monomers possesses a shallow
groove (approximately 25 ꢀ in length) that recognizes con-
sensus phosphopeptide motifs containing either phosphoser-
ine (pS) or phosphothreonine (pT) residues.[1b,c] Hundreds of
intracellular ligand proteins that possess 14–3–3 consensus
motifs have been identified, including proteins involved in
signaling and cell-cycle control, such as the Raf family, p53,
and Cdc25 phosphatases.[1b] A number of studies have
revealed that the disruption of 14–3–3 PPIs results in
suppression of tumor growth,[2] thus suggesting that 14–3–3
proteins could potentially serve as new therapeutic targets.
However, many details regarding intracellular regulatory
processes that involve 14–3–3 proteins remain unknown.
Low-molecular-weight agents that detect 14–3–3 proteins are
desirable for elucidating intracellular 14–3–3 functions, how-
ever, such chemical probes are not currently available. Recent
studies have successfully identified small molecule stabili-
zers[3a] and inhibitors[3b–c] of 14–3–3 PPIs, thus providing useful
reagents for cell-based analysis. We anticipated that develop-
ment of a cell-penetrating chemical probe that is capable of
detecting 14–3–3 proteins in response to their binding to
Wittinghofer and co-workers[7] reported the crystal struc-
ture of plant 14–3–3 bound to FC and a consensus phos-
phopeptide (QSYpTV) derived from the C terminus of
H+-ATPase. Their study revealed that FC binds to a hydro-
[*] M. Takahashi, A. Kawamura, Prof. N. Kato, Prof. T. Nishi,
Prof. J. Ohkanda
The Institute of Scientific and Industrial Research (ISIR)
Osaka University
8-1 Mihogaoka, Ibaraki, Osaka 567-0047 (Japan)
E-mail: johkanda@sanken.osaka-u.ac.jp
Prof. I. Hamachi
Department of Synthetic Chemistry and Biological Chemistry
Graduate School of Engineering, Kyoto University (Japan)
[**] This work was supported by the Japan Society for the Promotion of
Science (KAKENHI no22655055, J.O.) and the Eisai Research
Foundation (J.O.), and was partially supported by the Program for
the Promotion of Fundamental Studies in Health Sciences of the
National Institute of Biomedical Innovation of Japan (N.K.). We
thank Prof. S. Uchiyama and Dr. M. Noda (Osaka University) for
their assistance on MALDI-TOF mass measurements.
Figure 1. A) Chemical structure of fusicoccin A and crystal structure of
the ternary complex of plant 14–3–3, FC, and the phosphopeptide
QSYpTV (Ref. [7]). B) Schematic representation of the phosphopeptide-
dependent 14–3–3 protein-labeling strategy. Kd values are cited from
Ref. [7]. Q=glutamine, S=serine, V=valine, Y=tyrosine.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201106995.
Angew. Chem. Int. Ed. 2012, 51, 509 –512
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
509

.
Angewandte
Communications
phobic cavity adjacent to the phosphopeptide binding pocket
and forms the ternary complex (Figure 1A, right). Impor-
tantly, formation of the ternary complex significantly
increases the affinity of both FC and QSYpTV to 14–3–3 by
nearly two orders of magnitude. The key driving force behind
the formation of the complex is the notable van der Waals
interaction between the hydrophobic FC backbone and the
isopropyl side chain of the V residue at position i + 1, which is
adjacent to the pT residue. These results suggest that FC
stabilizes the PPI between 14–3–3 and any partner protein
that bears a consensus motif with a V residue at position i + 1,
therefore, appropriately functionalized FC derivatives should
be capable of detecting such interactions. As the first step
toward the development of such a functionalized FC deriv-
ative, we rationally designed and evaluated an FC-based
chemical probe that, after formation of a ternary complex
with a 14–3–3 protein and a phosphopeptide, covalently
attaches a fluorescent tag to the 14–3–3 surface. We then used
the probe to detect endogenous 14–3–3 in cancer cells.
We hypothesized that a chemically modified FC bearing a
fluorescent tag that is attached by a reactive spacer, and a
phosphopeptide that contains a V residue at position i + 1
would bind to 14–3–3 and form the ternary complex (Fig-
ure 1B). The formation of the complex would in turn trigger
the spacer to react with nucleophilic residues on the protein
surface to covalently attach the tag. Labeling through
formation of a ternary complex with the consensus peptide
should be kinetically more favorable than labeling with the
probe alone, because of the stabilizing effect imparted by van
der Waals interactions. Accordingly, selective labeling in the
presence of the phosphopeptide ligand would be achieved.
Among the seven human isoforms of 14–3–3, the z iso-
form was of great interest to us because it has been implicated
in lung[8a] and breast[8b] tumorigenesis. Histidine 164 (His164)
of 14–3–3z is located in a flexible loop region, thus suggesting
that it is vulnerable to labeling. We incorporated the recently
developed bioorthogonal phenylsulfonylmethylene moiety
(tosylate)[9] as the reactive spacer in order to make our
probe react with His164. Thus, we synthesized FC-based
fluorescent probes 1a and 1b by covalently linking the
transformed glycosyl moiety of the FC analogue ISIR-042[6] to
the reactive spacer, in which the terminus was attached to a
dansyl or a BODIPY group, respectively (Scheme 1; see the
Supporting Information for details). Based on a computer-
generated docking model, a spacer with an appropriate length
was chosen to bring the sulfonylmethylene moiety near
His164 of 14–3–3z (Figure 2A).
Scheme 1. Chemical structures of fluorescence-labeling reagents 1a
and 1b attached to the fusicoccin derivative used in this study. 2 is a
control compound.
Figure 2. A) Superimposed model of 1a (CPK), 14–3–3z protein (PDB
1YWT; blue ribbon, His164 highlighted in yellow CPK), and phospho-
peptide MARSHpSVPA (orange CPK; Y in the original structure was
modified to V). B) 14–3–3 proteins (40 mm) were incubated with
compound (80 mm) in the presence or absence of PMA2 (80 mm) at
358C for 18 h. The SDS-PAGE gel image was visualized by fluorescence
(Fl) and staining with Coomassie brilliant blue (CBB).
phosphopeptide until equimolarity with 14–3–3z was reached
(Figure S2), thus supporting a 1:1:1 model for the formation
of the ternary complex. In the case of compound 1a, the
formation of a 1:1 adduct between the fluorescent tag and 14–
3–3z was confirmed by using MALDI-TOF mass spectrom-
etry (Figure S3).
The reaction with 14–3–3s, which possesses an Asn
instead of a His at position 164, resulted in a negligibly
stained band (Figure 2B, lane 6), thus suggesting that labeling
predominantly involves His164 of 14–3–3z. This site specific-
ity was further confirmed by using site-directed mutagenesis
of 14–3–3z. Substitution of Ala for His at position 164
decreased the degree of labeling (Figure S4). Furthermore, a
control experiment involving compound 2 demonstrated that
removal of the FC anchor completely precludes labeling
(Figure 2B, lanes 7 and 8), thus further confirming that the
FC anchor is essential for binding to 14–3–3.
Following the synthesis of the probes, labeling was
evaluated by incubating compounds 1a,b and
2 with
recombinant 14–3–3z in the presence or absence of the
PMA2 phosphopeptide[5b] ETIQQSYpTV, followed by SDS-
PAGE (see Figure S1 in the Supporting Information for 1a,
and Figure 2B for 1b and 2). Clearly, 1a and 1b labeled 14–3–
3z in the presence of the peptide (Figure 2B, lane 3for 1b),
whereas no apparent fluorescent band was detected when the
peptide was absent during labeling (Figure 2B, lane 2 for 1b),
thus demonstrating that the formation of the ternary complex
is the determining factor for the labeling reaction. The yield of
the labeled protein increased with the concentration of the
We next evaluated whether compound 1b recognizes the
shape of the phosphopeptide ligand. We predicted that the
residue at position i + 1 of the phosphopeptide should have a
major influence on the formation of the ternary complex
510
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 509 –512

Angewandte
Chemie
because this residue is responsible for the critical van der
Waals interactions. To test this hypothesis, a series of
consensus RSHpSXP hexapeptides,[10] each containing one
of the 20 amino acids at position X (X = i + 1) were evaluated
for labeling (Figure 3). Use of peptides containing Val, Leu,
or Ile at position i + 1 resulted in high labeling efficiency,
Figure 4. Labeling of the 14–3–3 protein was performed using E. coli
cell lysate, 1b (20 mm), and PMA2 (20 mm) in HEPES (4-(2-hydroxye-
thyl)piperazine-1-ethanesulfonic acid) buffer (10 mm, pH 7.3) at 358C
for 18 h. a) Total amount of added protein given in mg. b) 14–3–3z
(10 mm) in which the His tag had been cleaved. mw: molecular-weight
marker, lane 0: His-14–3–3z. mw and lanes 0–6: CBB-stained gel
image; lanes 7–12: fluorescent gel image.
compound 1b (Figure 4, lanes 1 and 7), an apparent fluores-
cent band corresponding to His-14–3–3z was observed when
PMA2 peptide was added exogenously (Figure 4, lanes 8–9).
In order to confirm that the His6 tag in His-14–3–3z does not
interfere with the labeling site, we tested lysate mixed with
14–3–3z in which the His tag had been cleaved. Both 14–3–3z
and His-14–3–3z were almost equally labeled (Figure 4,
lanes 6 and 12), thus confirming that selectivity based on
His164 is not compromised by the His tag.
Figure 3. Various phosphopeptides (RSHpSXP, X=one of the 20
amino acids) were evaluated for use in the labeling reaction of 14–3–
3z with 1b under the same conditions as described in Figure 2. PMA2
was used as a positive-control standard. Top: fluorescent gel image
and CBB-stained gel image. Bottom: Relative intensity of the fluores-
cent bands normalized against PMA2. Standard deviations are given
for nꢀ2. H=histidine, P=proline, R=arginine.
Finally, we sought to determine whether our FC-based
probes could detect endogenous human 14–3–3 protein. We
first evaluated the ability of compound 1b to penetrate cells
by using adhesive lung adenocarcinoma epithelial A549 cells,
which are known to overexpress 14–3–3z.[8a] Compound 1b
was distributed throughout the cytosol, thus indicating that it
readily enters cells (Figure 5A). For the in-cell detection
experiment, we next chose a floating cell line, human
leukemic monocyte lymphoma U937 cells, as it provides a
sufficient amount of lysate for analysis. Cells were incubated
with 1b for 2 days in the presence of the protein kinase A
activators 3-isobutyl-1-methylxanthine (IBMX) and forsko-
lin.[11] The cell lysates were analyzed by fluorescence gel
imaging and Western blotting with human 14–3–3z antibody
(Figure 5B). No apparent fluorescent bands were detected in
lysates of cells treated with 1b in the absence of forskolin
(Figure 5B, lane 7). In contrast, in cells incubated under
hyperphosphorylation conditions, detection of endogenous
14–3–3 was dependent on the concentration of 1b (i.e.,
detection at 8 and 16 mm, no detection at 4 mm; Figure 5B,
lanes 8–10). Obviously, labeling of 14–3–3 with compound 1b
in U937 cells requires a high concentration of phosphorylated
proteins, which most likely include the partner protein(s) that
contribute to the formation of the ternary complex and
consequently trigger intracellular 14–3–3 labeling. To the best
of our knowledge, this is the first direct evidence indicating
that 14–3–3 proteins are the primary target of FCs in human
cells.
presumably because the hydrophobic and branched side
chains conform well to the FC backbone. The same trend was
observed with the peptide containing Thr at position i + 1,
furthermore, it was approximately eight times more efficient
at labeling than the peptide containing Ser, thus demonstrat-
ing that the methyl group of Thr contributes to the hydro-
phobic interaction with the FC moiety and compensates for
unfavorable hydration caused by the hydroxy group, which is
an effect that may be significant in the case of Ser. Surpris-
ingly, labeling also proceeded efficiently with peptides con-
taining Cys and Met at position i + 1, thus suggesting that the
sulfur atom does not interfere with the interaction. In
contrast, peptides containing small or bulky side chains,
such as Ala, Phe, Tyr, or Trp, were inefficient ligands,
presumably because fewer van der Waals contacts and steric
repulsion prevent binding of the FC anchor (Figure S5).
Interestingly, while peptides containing the basic residues Lys
and Arg were also effective ligands, peptides containing the
acidic residues Glu and Asp were not effective ligands. These
results suggest that electrostatic interaction with the acidic
14–3–3 protein might increase the affinity of the peptide
ligand for the protein groove.
We then examined the labeling selectivity of 14–3–3z by
using a cell lysate of E. coli, which expresses 14–3–3z with an
N-terminal His6 tag (His-14–3–3z; Figure 4). Although no
reaction was detected when the lysate was treated with only
In conclusion, we developed natural-product-based fluo-
rescent probes that are capable of labeling 14–3–3z in a site-
specific, 14–3–3-selective, and, most importantly, highly
ligand-dependent manner. The FC anchor precisely recog-
Angew. Chem. Int. Ed. 2012, 51, 509 –512
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
511

.
Angewandte
Communications
[1] a) M. B. Yaffe, K. Rittinger, S. Voilnia, P. R. Caron, A. Aitken,
H. Leffers, S. J. Gamblin, S. J. Smerdon, L. C. Cantley, Cell 1997,
91, 961 – 971; b) M. B. Yaffe, FEBS Lett. 2002, 513, 53 – 57; c) C.
Johnson, S. Crowther, M. J. Stafford, D. G. Campbell, R. Toth, C.
Mackintosh, Biochem. J. 2010, 427, 69 – 78.
[2] a) H. Hermeking, Nat. Rev. Cancer 2003, 3, 931 – 943; b) E.
Wilker, M. B. Yaffe, J. Mol. Cell. Cardiol. 2004, 37, 633 – 642, and
references therein.
[3] a) R. Rose, S. Erdmann, S. Bovens, A. Wolf, M. Rose, S. Hennig,
H. Waldmann, C. Ottmann, Angew. Chem. 2010, 122, 4223 –
4226; Angew. Chem. Int. Ed. 2010, 49, 4129 – 4132; b) H. Wu, J.
Ge, S. Q. Yao, Angew. Chem. 2010, 122, 6678 – 6682; Angew.
Chem. Int. Ed. 2010, 49, 6528 – 6532; c) V. Corradi, M. Mancini,
F. Manetti, S. Petta, M. A. Santucci, M. Botta, Bioorg. Med.
Chem. Lett. 2010, 20, 6133 – 6137.
Figure 5. A) A549 cells were incubated with 1b (50 mm) at 378C for
3 days. The sample was analyzed under a confocal microscope.
B) U937 cells were treated with IBMX and then incubated with
forskolin and 1b (see the Supporting Information for details). The cell
lysates were subjected to SDS-PAGE, and the gel was visualized by
Western blotting with human 14–3–3z antibody (left) and by fluores-
cence (right).
[4] A. Ballio, E. B. Chain, P. De Leo, B. F. Erlanger, M. Mauri, A.
Tonolo, Nature 1964, 203, 297.
[5] a) A. T. Fuglsang, S. Visconti, K. Drumm, T. Jahn, A. Stensballe,
B. Mattei, O. N. Jensen, P. Aducci, M. G. Palmgren, J. Biol.
Chem. 1999, 274, 36774 – 36780; b) F. Svennelid, A. Olsson, M.
Piotrowski, M. Rosenquist, C. Ottman, C. Larsson, C. Oecking,
M. Sommarin, Plant Cell 1999, 11, 2379 – 2391.
[6] a) N. Kato, T. Inoue, Y. Maruyama, H. Nitta, Y. Honma, PCT/
JP2010051537, 2010; b) K. Kawakami, M. Hattori, T. Inoue, Y.
Maruyama, J. Ohkanda, N. Kato, M. Tongu, T. Yamada, M.
Akimoto, K. Takenaga, T. Sassa, J. Suzumiya, Y. Honma,
unpublished results. See the Supporting Information for details.
[7] M. Wꢁrtele, C. J. Ottmann, A. Wittinghofer, C. Oecking, EMBO
J. 2003, 22, 987 – 994.
[8] a) T. Fan, R. Li, N. W. Todd, Q. Qiu, H-B. Fang, H. Wang, J. Shen,
R. Y. Zhao, N. P. Caraway, R. L. Katz, S. A. Stass, F. Jiang,
Cancer Res. 2007, 67, 7901 – 7906; b) C. L. Neal, J. Yao, W. Yang,
X. Zhou, N. T. Nguyen, J. Lu, C. G. Danes, H. Guo, K-H. Lan, W.
Hittelman, M-C. Hung, D. Yu, Cancer Res. 2009, 69, 3425 – 3432.
[9] a) S. Tsukiji, M. Miyagawa, Y. Takaoka, T. Tamura, I. Hamachi,
Nat. Chem. Biol. 2009, 5, 341 – 343; b) S. Tsukiji, H-x. Wang, M.
Miyagawam, T. Tamura, Y. Takaoka, I. Hamachi, J. Am. Chem.
Soc. 2009, 131, 9046 – 9054; c) H. Abe, E. Kool, Proc. Natl. Acad.
Sci. USA 2006, 103, 263 – 268.
nizes the structural difference of the residue at position i + 1
in the phosphopeptide, thus enabling selective 14–3–3 label-
ing which depends on the shape of the ligand. This recog-
nition-based selectivity of FCs will be an advantage further
development of probes to detect 14–3–3 PPIs in a partner-
protein-dependent manner. The FC derivative ISIR-042,
which was used as the anchor here, has been developed for
use as an antitumor agent.[6] The probes may also provide a
robust tool for elucidating the mechanisms of FC antitumor
activity, including identification of intracellular proteins
involved in the formation of the ternary complex. Work
toward these ends is currently in progress in our laboratory.
Received: October 4, 2011
Published online: November 21, 2011
[10] E. W. Wilker, R. A. Grant, S. C. Artim, M. B. Yaffe, J. Biol.
Chem. 2005, 280, 18891 – 18898.
[11] A. Chahdi, A. Sorokin, Mol. Cell. Biol. 2008, 28, 1679 – 1687.
Keywords: fluorescent probes · natural products ·
phosphopeptides · protein labeling ·
protein–protein interactions
.
512
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 509 –512