TP-2Rho Is a Sensitive Solvatochromic Red-Shifted Probe for Monitoring the Interactions between CDK4 and Cyclin D

Article abstract of DOI:10.1002/cbic.201500641

Understanding the intricate steps of protein kinase regulation requires characterization of protein-protein interactions between the catalytic subunit, its regulatory partners and the substrate. Fluorescent probes are useful tools with which to study such

Full text of DOI:10.1002/cbic.201500641

DOI: 10.1002/cbic.201500641
Full Papers
TP-2Rho Is a Sensitive Solvatochromic Red-Shifted Probe
for Monitoring the Interactions between CDK4 and
Cyclin D
Morgan Pellerano+,[a] Delphine Naud-Martin+,[b, c] Marion Peyressatre,[a] Camille PrØvel,[a]
Marie-Paule Teulade-Fichou,[b, c] May Morris,*[a] and Florence Mahuteau-Betzer*[b, c]
Understanding the intricate steps of protein kinase regulation
requires characterization of protein–protein interactions be-
tween the catalytic subunit, its regulatory partners and the
substrate. Fluorescent probes are useful tools with which to
study such interactions and to gain insight into their affinities
and specificities. Solvatochromic probes, which display
changes in their fluorescence emission in response to changes
in the polarity of the medium, are particularly attractive. Here
we describe conjugation of a switchable fluorescent dye, TP-
2Rho, to peptide and protein derivatives of cyclin-dependent
kinase 4 (CDK4) and its application to characterization of the
interactions between the catalytic subunit of this kinase, its
regulatory partner cyclin D1 and a peptide substrate. We dem-
onstrate the sensitivity of TP-2Rho in relation to of those other
dyes used for monitoring peptide–protein and protein–protein
interactions. Moreover, we show that TP-Rho-labelled peptides
can be introduced into living cells to probe endogenous
CDK4/cyclin D.
Introduction
Monitoring of protein–protein and peptide–peptide interac-
tions is one of the many approaches that contributes to char-
acterization of the functions and regulation of protein partners
in biological signalling pathways. Acquisition of insight into
the mechanisms of binding and determination of dissociation
constants between protein subunits, enzymes and their sub-
strates, agonists or antagonists are essential to complement
structural studies. Likewise, development of new inhibitors for
therapeutic means requires determination of their binding
sites and affinities for their targets, as well as the characteriza-
tion of their specificity and selectivity profiles through comple-
mentary biochemical, biophysical and biological approaches.
Fluorescent probes are very useful tools with which to study
protein and peptide interactions. A wide variety of fluorescent
probes with distinctive biophysical properties that might make
them more suitable for in vitro or in vivo applications are avail-
able.[1] The emission properties of these solvatochromic probes
are modified upon modification of the polarity of the medium.
Among them, light-up probes have attracted considerable at-
tention. These dyes are not fluorescent in their free forms but
become fluorescent when bound to their targets. Their on/off
behaviour provides better contrast and therefore better sensi-
tivity to environmental changes.[2] For biological applications,
an ideal light-up probe requires additional photophysical and
physicochemical properties such as water solubility, red emis-
sion and brightness, and there is still an important need for
bright and red-emitting dyes.[3]
Recently we have developed a new family of vinyltriphenyla-
mine dyes and shown that the switchable dyes of this class are
highly sensitive fluorescent probes for probing DNA by cellular
imaging.[4] By changing their acceptor moieties from cationic
to anionic variants, with the aim of labelling proteins rather
than DNA, we synthesized the TP-Rho class of dyes.[5] These
retain the lighting-up properties of the cationic dyes, display-
ing very efficient fluorescence enhancement (F/F0 ꢀ80 for TP-
2Rho) when complexed with human serum albumin (used as
a model protein).[5] Moreover, the two-photon absorption abili-
ties of TP-Rho dyes, as well as their emission properties (labs
ꢀ600 nm), are an asset for live-cell imaging. Indeed, infrared
excitation causes less damage to the sample than UV/Vis and
avoids cellular endogenous fluorescence.[6] These water-soluble
and cell-permeant p-conjugated molecules display optical
properties that make them attractive candidates for protein-
and peptide-based applications.
[a] M. Pellerano,+ M. Peyressatre, Dr. C. PrØvel, Dr. M. Morris
Institut des BiomolØcules Max Mousseron–IBMM-CNRS-UMR 5247
FacultØ de Pharmacie, UniversitØ Montpellier 1
15, avenue Charles Flahault, 34093 Montpellier (France)
E-mail: may.morris@univ-montp1.fr
[b] D. Naud-Martin,+ Dr. M.-P. Teulade-Fichou, Dr. F. Mahuteau-Betzer
Institut Curie, PSL Research University
CNRS, INSERM, UMR9187-U1196
91405 Orsay (France)
E-mail: florence.mahuteau@curie.fr
[c] D. Naud-Martin,+ Dr. M.-P. Teulade-Fichou, Dr. F. Mahuteau-Betzer
UniversitØ Paris Sud
UniversitØ Paris-Saclay, CNRS, UMR9187-U1196
91405 Orsay (France)
[+] These authors contributed equally to this work.
Supporting information for this article can be found under http://
dx.doi.org/10.1002/cbic.201500641.
Cyclin-dependent kinases (CDKs) form a family of hetero-
dimeric kinases that are central to regulation of cell cycle pro-
gression, but also play major roles in a wide variety of biologi-
This article is part of a joint Special Issue between ChemBioChem and
ChemMedChem on Protein–Protein Interactions
ChemBioChem 2016, 17, 737 – 744
737
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
cal functions, including transcriptional regulation, neurological
functions and metabolism.[7] Moreover, these kinases are tar-
gets of interest for cancer therapeutics because their hyperacti-
vation, contributing to sustain uncontrolled cell proliferation,
has been documented in a wide variety of cancers.[8] Sequen-
tial activation of these kinases throughout the cell cycle is sub-
ject to several levels of control, which have been investigated
through complementary biological, biochemical, biophysical
and structural studies.[9] The assembly between CDKs and cy-
clins, initially investigated between CDK2 and cyclin A, was
shown to involve contacts between both the N-terminal and
the C-terminal lobes of CDK2 and cyclin A. This process was
described as two-step mechanism that first involves a high-af-
finity interaction between the C helix of the CDK and the a5
helix of the cyclin, thereby inducing a rotation of the N-termi-
nal lobe of the CDK and consequent alignment of the ATP
binding site with the catalytic cleft, followed by conformational
reorganization of the C lobe and in particular a positional shift
of the CDK activation loop, thereby favouring substrate bind-
ing.[9d,h,10] Although this interaction has been extensively stud-
ied in the case of CDK2 and cyclin A, there are very few studies
relating to the biochemistry of interactions between CDK4 and
cyclin D.
Figure 1. CDK4/cyclin D1 and CDK2/cyclin A. A) Stick representation of the
“open conformation” of the crystal structure of CDK4/CycD1 (PDB 2W9F),
the major interface between CDK4 (green) and cyclin D1 (cyan), consisting
of the C helix (PISTVRE) and the a5 helix of cyclin D1, highlighted in yellow
and dark blue, respectively. B) Stick representation of the “closed conforma-
tion” of the crystal structure of CDK2/cyclin A3 (PDB 1FIN), with the interface
between CDK2 (green) and cyclin A3 (cyan) mediated by contacts between
the C helix (PSTAIRE)/a5 helix of cyclin A3, highlighted in yellow and dark
blue, respectively, as well as by additional contacts in the C lobe of CDK2.
with cyclin D1, and to the Rb peptide, and their comparative
titration against cyclin D1 protein, a peptide derived from the
main interface of cyclin D1 with CDK4, and CDK4 monomer or
CDK4/cyclin D heterodimer, respectively. We show that this dye
is a much more sensitive probe for monitoring these peptide–
protein and protein–protein interactions than other standard
dyes. Finally we show that a TP-2Rho-conjugated peptide de-
rived from CDK4 can be efficiently internalized into living cul-
tured cells and provides a means of localizing endogenous
CDK4/cyclin D.
This heterodimeric kinase is a key regulator of the G1 phase,
integrating mitogenic signals through the Ras signalling path-
way to induce exit from quiescence and progression through
G1 by phosphorylation of several transcriptional regulators, in
particular the retinoblastoma protein family members (p107,
p130, pRb), together with CDK6.[11] This phosphorylation fur-
ther induces the de-repression of E2F transcription factors, al-
lowing the transcription of genes required for G1/S transition. Results and Discussion
Hyperactivation of CDK4/cyclin D1 associated with genetic
Synthesis of the TP-2Rho-maleimide reagent
mutation, amplification or overexpression of either subunit has
been widely documented in a great variety of human cancers,
including breast cancer, neuroblastoma, mantle cell lymphoma
and multiple myelomas, melanoma and lung cancers, contribu-
ting to enhancing and sustaining cancer cell proliferation.[8b]
This kinase has thus been identified as a relevant target for an-
ticancer therapeutics.[12]
We chose to label peptides and proteins of interest on unique
cysteine residues with a maleimide-conjugated derivative of
TP-2Rho. We designed our conjugated compound with a C4
spacer. The TP-2Rho maleimide reagent (TP-2Rho-Mal) was pre-
pared in six steps from (4-methoxyphenyl)diphenylamine
(Scheme 1). Vilsmeier–Haack formylation followed by the de-
protection of the phenolic function led to compound 3 in very
good yields. The nucleophilic substitution of 1,4-dibromobu-
tane by phenol 3 afforded the dialdehyde 4. A furan-protected
maleimide (to prevent maleimide polymerization) group was
introduced through a nucleophilic substitution on bromide 4.
Compound 5 was deprotected (retro-Diels–Alder) without fur-
ther characterization, and dialdehyde 6 was obtained in a very
good yield. The last step was a Knoevenagel reaction on both
aldehyde groups to provide TP-2Rho-Mal.
The crystal structures of CDK4/cyclin D1 and CDK4/cyclin D3
suggest that these complexes adopt an “open conformation”,
with the interaction between CDK4 and its regulatory cyclin
subunit appearing to be essentially mediated by the C helix
(PISTVRE) of CDK4 and the a5 helix of the D-cyclin, in contrast
to the “closed conformation” of CDK2/cyclin A (Figure 1).[13]
However, a recent study addressing the conformational dy-
namics of CDK4/cyclin D inferred that this complex could
adopt a conformation resembling that of CDK2/cyclin A.[14]
In this study we have chosen to investigate the utility of the
two-branch anionic dye, TP-2Rho, to characterize peptide–pro-
tein and protein–protein interactions between the catalytic
cyclin-dependent kinase CDK4 and its regulatory partner cy-
clin D1, and to compare the affinities of monomeric CDK4 and
of the CDK4/cyclin D1 complex towards a peptide derived
from retinoblastoma (Rb) substrate. Specifically, we report on
the conjugation of TP-2Rho to the catalytic subunit of the
CDK4, to a peptide derived from the main interface of CDK4
Photophysical properties of TP-2Rho-Mal
The photophysical properties of this compound were studied
in PBS buffer and in glycerol (Table 1) and compared with
those of TP-2Rho. As expected of the TP-Rho series, the com-
pound was poorly fluorescent in aqueous medium (Figure 2).
In glycerol, however, both TP-2Rho-Mal and TP-2Rho are fluo-
rescent, displaying yellow emission with good quantum yields
ChemBioChem 2016, 17, 737 – 744
www.chembiochem.org
738
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Scheme 1. Synthesis of TP-2Rho-Mal. a) POCl3, DMF (78%); b) BBr3, CH2Cl2 (92%); c) Cs2CO3, DMF (41%); d) K2CO3, DMF, 408C, 15 h (63%); e) toluene, reflux,
12 h (95%); f) NH4OAc, H3COOH, 1158C, 16 h (35%).
through restriction of molecular motions. The same fluores-
Table 1. Photophysical properties of TP-2Rho-Mal and TP-2Rho.
cence restoration can be expected for our probes (in monitor-
ing of protein–protein interactions), because immobilization in
glycerol can mimic the confinement of the probe between two
peptide or protein partners. Both dyes display very large
Stokes shifts (from 100 to 150 nm) in buffer and in glycerol.
Compounds
e [molÀ1 LcmÀ1
PBS glycerol PBS
]
labs/lem
glycerol PBS
FF
glycerol
TP-2Rho-Mal 42000 52000
TP-2Rho 53500 76300
505/650 513/618 <0.005 0.04
505/633 500/615 <0.005 0.09
Peptide–peptide interactions
In order to determine whether the TP-2Rho probe could be
used to characterize the interaction between the C helix (PIST-
VRE) of CDK4 and the a5 helix of cyclin D1, we labelled a pep-
tide derived from the PISTVRE helix (PISTVRE peptide) with this
probe and titrated it against increasing concentrations of a
peptide derived from the a-helix of cyclin D1 (a5 D1 peptide).
For comparison we also labelled the PISTVRE peptide with
other traditional probes, including FITC, BODIPY, coumarin,
rhodamine, Cy3 and Cy5. As shown in Figure 3, TP-2Rho exhib-
ited a tenfold increase in fluorescence upon titration with the
a5 D1 peptide, whereas the other probes exhibited much
lower fluorescence enhancement. Moreover, curve fitting and
calculation of the dissociation constants revealed essentially
similar values when titration was performed with TP-2Rho-,
FITC- or BODIPY-labelled PISTVRE peptide of 11–12 mm, and a
somewhat lower affinity for the Cy5-labelled peptide, with a
dissociation constant value of (23Æ13) mm (Table 2). Moreover,
upon interaction with a5 D1 peptide, we observed a shift of
the PISTVRE-labelled TP-2Rho fluorescence maximum emission
from 616 to 608 nm (Figure 3C).
Figure 2. Absorption (c) and emission (····) spectra of 2 mm TP-2Rho-Mal
in PBS buffer in red (labs/emmax: 505/650 nm; note that 30 mm was used to
determine emission spectrum in PBS) and in glycerol in blue (labs/emmax: 513/
618 nm).
from 0.04 to 0.09. This behaviour means that our probe is
responsive to the constraints of environment, glycerol being
a viscous solvent commonly used to restore fluorescence
ChemBioChem 2016, 17, 737 – 744
www.chembiochem.org
739
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
between these peptides and the lower (by twofold) dissocia-
tion constant determined for the Cy5-PISTVRE/D1 peptide sug-
gest that TP-2Rho, FITC and BODIPY might contribute to pro-
moting the interaction between two peptides that exhibit
a natural tendency to interact when surrounded by their com-
plete protein environments but are less prone to interact in an
unstructured form in solution.
Peptide–protein interactions
We next asked whether TP-2Rho could similarly serve to report
on the interaction between the PISTVRE helix-derived peptide
of CDK4 and cyclin D1. As shown in Figure 4A, titration of the
TP-2Rho-labelled-PISTVRE peptide with glutathione S-transfer-
ase (GST)-cyclin D1 yielded a 19-fold increase in fluorescence
emission intensity of this probe, with a Kd value of (351Æ
94) nm. In comparison, the TAMRA-labelled PISTVRE only yield-
ed a three- to fourfold increase in fluorescence enhancement
at saturation, with a Kd value of (152Æ51) nm. As in the case
of the peptide–peptide experiments, these results reveal that
TP-2Rho is a highly sensitive probe for monitoring interactions
between peptides and proteins. Moreover, titration experi-
ments show that the PISTVRE helix of CDK4 interacts with 34-
fold tighter affinity with cyclin D1 than with the a5 helix alone.
This finding is in agreement with structural data that show
how the PISTVRE helix interacts with an overall broader surface
of cyclin D1 rather than just with the residues of the a5 helix
(i.e., that the residues of the PISTVRE helix “sit within a pocket”
of cyclin D1; Figure 4B, C). The difference in dissociation con-
stant values determined for this interaction with TP-2Rho- or
Figure 3. Titration of the fluorescently labelled PISTVRE peptide against
a5 D1 peptide. A) Titration of PISTVRE labelled variously with TP-2Rho
(200 nm, lex =510 nm, lem =614 nm), FITC (200 nm, lex =483 nm, lem
530 nm), BODIPY (200 nm, lex =477 nm, lem =525 nm), coumarin (200 nm,
lex =398 nm, lem =484 nm), rhodamine (200 nm, lex =550 nm, lem
605 nm), Cy3 (200 nm, lex =530 nm, lem =580 nm) or Cy5 (200 nm, lex
=
=
=
610 nm, lem =675 nm) against a5 D1 peptide. B) Histogram representation
of maximum fluorescence enhancement upon titration against 30 mm a5 D1
peptide. C) Normalized fluorescence spectrum of TP-2Rho-labelled PISTVRE
peptide before (blue) and after (red) incubation with aD1 peptide.
Table 2. Dissociation constants calculated for interaction between PIST-
VRE and a5 D1 peptides.
Labelled peptide
Kd [mm]
Labelled peptide
Kd [mm]
PISTVRE-TP-2Rho
PISTVRE-FITC
PISTVRE-BODIPY
PISTVRE-coumarin
12Æ2
12Æ5
11 Æ4
–
PISTVRE-rhodamine
PISTVRE-Cy3
PISTVRE-Cy5
–
–
23Æ13
Taken together, these data indicate that TP-2Rho is a highly
sensitive probe for monitoring the interaction between two
peptides, in comparison with other frequently used dyes such
as FITC, BODIPY, coumarin, rhodamine, Cy3 and Cy5. Moreover
TP-2Rho does not affect the interaction between two peptides
derived from the natural occurring primary protein–protein in-
terface of CDK4/cyclin D any more than FITC or BODIPY. This
having been said, the micromolar affinities of the interaction
Figure 4. A) Titration of PISTVRE peptide labelled either with TP-2Rho
(200 nm, lex =510 nm, lem =614 nm) or with TAMRA (200 nm, lex =535 nm,
lem =585 nm) against GST-cyclin D1. B) Stick/surface representation of the
PISTVRE peptide (yellow)/a5 D1 peptide (blue) interaction. C) Stick/surface
representation of the PISTVRE peptide (yellow)/cyclin D1 protein (cyan) inter-
action.
ChemBioChem 2016, 17, 737 – 744
www.chembiochem.org
740
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
TAMRA-labelled peptide is indicative of different contributions
of these dyes to stabilizing the interaction. Nevertheless, these
values are within the same sub-micromolar range, and not
considerably different from those determined in the peptide/
peptide titration experiments described above.
To investigate the potential of TP-2Rho to monitor peptide–
protein interactions further, we titrated a TP-2Rho-labelled
peptide derived from the substrate of CDK4, retinoblastoma
protein, with GST-CDK4 and with the GST-CDK4/cyclin D com-
plex.
Figure 5 reveals a significantly greater affinity of the Rb sub-
strate for the complex, with a dissociation constant of (50Æ
9) nm, than for monomeric GST-CDK4, with a calculated disso-
ciation constant of (1470Æ28) nm. Moreover, binding of CDK4/
cyclin D1 to the Rb-labelled peptide induced a much greater
Figure 6. Titration of GST-CDK4 mutant (C78 C135) labelled either with TP-
2Rho (200 nm, lex =510 nm, lem =614 nm) or with Cy3 (200 nm, lex
530 nm, lem =580 nm) against GST-cyclin D1.
=
probe than Cy3, because maximum fluorescence emission at
saturation reached 13-fold with TP-2Rho, in comparison with
twofold with Cy3. Titration of TP-2Rho-labelled GST-CDK4
mutant with GST-cyclin D1 yielded a Kd value of (123Æ47) nm.
A somewhat greater Kd value of (445Æ88) nm was calculated
for the interaction between the Cy3-labelled GST-CDK4 mutant
and GST-cyclin D1. This finding suggests either that TP-2Rho
might be contributing to the interaction, or that Cy3 might
somehow be preventing efficient interactions. Nonetheless,
these values are within the same sub-micromolar range indica-
tive of an interaction between CDK4 and cyclin D1 that is 50–
100 times tighter than that between the individual peptides
derived from the PISTVRE helix of CDK4 and the a5 helix of cy-
clin D1.
Figure 5. Titration of TP-2Rho-labelled (200 nm, lex =510 nm, lem =614 nm)
Rb substrate peptide against GST-CDK4 and against GST-CDK4/cyclin D1.
TP-2Rho-labelled PISTVRE peptide colocalizes with intra-
cellular target proteins
fluorescence enhancement than monomeric CDK4: 15-fold
versus sixfold, respectively, at 400 nm. These findings are in
line with previous studies that reported that CDK2/cyclin A has
a higher affinity than monomeric CDK2 for its histone sub-
strate, because the cyclin promotes a significant conformation-
al change of CDK2, thereby favouring substrate binding in the
catalytic cleft. However, it should be noted that the structure
of monomeric CDK4 has not been solved, and that therefore
there is no means of assessing whether cyclin D1 binding indu-
ces conformational reorganization of CDK4. Our results are
therefore the first to indicate that this conformational change
is likely to occur, because the Rb substrate binds CDK4/cy-
clin D1 with 30-fold greater affinity and a threefold greater
fluorescence enhancement than monomeric CDK4.
Finally, we speculated as to whether TP-2Rho might, when
conjugated to the PISTVRE peptide, serve as a probe to coloc-
alize endogenous cyclin D1 in HeLa cells. As expected, the
PISTVRE-TP-2Rho-labelled peptide did not enter cells. We
therefore used a Pep-PISTVRE peptide, which bears an N-termi-
nal fusion of the Pep1 cell-penetrating peptide[15] to the se-
quence of PISTVRE, and found that it penetrated readily both
into the cytoplasm and into the nucleus when overlaid onto
HeLa cells (Figure 7A). To address the issue of whether this
peptide indeed colocalized with cyclin D1 and CDK4, we per-
formed indirect immunofluorescence on HeLa cells overlaid
with the TP-2Rho-labelled Pep-PISTVRE peptide (Figure 7B).
These experiments revealed a significant overlay of the TP-
2Rho signal and of the signal revealed with Alexa647-labelled
secondary antibodies against the rabbit polyclonal anti-CDK4
or anti-cyclin D1 antibodies.
Protein–protein interactions
We finally addressed whether TP-2Rho was applicable for mon-
itoring of protein–protein interactions. To this end we labelled
a mutant of GST-CDK4 that contains only two cysteine residues
(C78 and C135, above and below the C a-helix) instead of four.
This form of GST-CDK4 was labelled either with TP-2Rho or
with Cy3 and titrated against cyclin D1 (Figure 6). Once again,
these experiments revealed that TP-2Rho was a more sensitive
Conclusions
Although small-molecule fluorescent probes offer the promise
of monitoring peptide–peptide and protein–protein interac-
tions,[16] they are not always suited to respond in a robust,
ChemBioChem 2016, 17, 737 – 744
www.chembiochem.org
741
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 7. TP-2Rho-labelled Pep-PISTVRE peptide penetrates into HeLa cells and colocalizes with intracellular CDK4/cyclin D1 in HeLa cells. A) TP-2Rho Pep-
PISTVRE peptide (2 mm) was found, on imaging with Cy3 filters, to have readily penetrated HeLa cells (centre), whereas the TP-2Rho-labelled PISTVRE peptide
(left) and TP-2Rho (right) had not. Nuclear staining with Hoechst 33342 is shown in the upper panels, whereas the lower panels represent fluorescence emis-
sion of TP-2Rho with use of the Cy3 filter of the microscope. B) Anti-CDK4 and anti-cyclin D1 immunofluorescence performed on HeLa cells overlaid with
2 mm TP-2Rho Pep-PISTVRE peptide. Upper panels show the colocalization of the TP-2Rho-labelled Pep-PISTVRE peptide with CDK4. Lower panels show its co-
localization with cyclin D1. Anti-CDK4 and anti-cyclin D1 immunostainings were revealed with Alexa647 secondary antibodies.
sensitive fashion. Moreover, traditional fluorescent probes are
based on modular core dyes that in some cases respond to en-
vironmental changes, yet do not exhibit significant on/off re-
sponses.[17] More recent developments in the chemistry of
small-molecule fluorescent probes have yielded compounds
that are essentially dark prior to conjugation or binding to
protein/peptide scaffolds.[18] Such dyes make highly sensitive
probes for studies directed towards dissecting interactions be-
tween peptides and proteins and requiring selective tools to
quantify, to compare and to highlight differences in affinity
and specificity.
cyclin D1 yielded an 18-fold increase in fluorescence emission
intensity of this probe, with a Kd value of (351Æ94) nm, indi-
cating that the surface provided by cyclin D1 protein offers
a broader surface of interaction with the PISTVRE peptide than
the a5 D1 peptide. A third titration experiment performed
with a GST-CDK4 protein bearing two TP-2Rho probes labelled
above and below the PISTVRE helix and GST-cyclin D1 yielded
a Kd value of (123Æ47) nm with a 13-fold increase in fluores-
cence. This again reveals a tighter interaction between CDK4
and cyclin D1.
Furthermore, because there is no structural information
available for monomeric CDK4, it is difficult to identify any con-
formational changes that might arise upon its interaction with
cyclin D1. Nevertheless, these changes can be inferred from
the major differences in fluorescence enhancement of the TP-
2Rho probe conjugated to a substrate peptide, associated with
a net increase in its affinity for CDK4/cyclin D1 relative to mon-
omeric CDK4, emphasizing the utility of the highly sensitive
TP-2Rho probe in dissection of the intricate steps of protein
kinase regulation through assembly of a catalytic and regulato-
ry subunit.
Here we have described the synthesis and conjugation of
a switchable fluorescent vinyltriphenylamine dye, TP-2Rho, in
order to characterize the mechanism of interaction between
the catalytic kinase subunit CDK4 and its regulatory subunit cy-
clin D1, using peptide and protein derivatives of this heterodi-
meric kinase as well as a peptide substrate derived from Rb.
Throughout this study we have clearly demonstrated that TP-
2Rho is a superior dye, in terms of sensitivity, to other dyes tra-
ditionally used for monitoring peptide–protein and protein–
protein interactions. Moreover, our results provide new infor-
mation, complementary to structural studies on this hetero-
dimeric complex, relating to the interactions that mediate as-
sembly of CDK4/cyclin D1 and its ability to bind its substrate.
Titration of the TP-2Rho-labelled-PISTVRE peptide with the
complementary peptide interface of cyclin D1 yielded a dissoci-
ation constant of 12 mm and 15-fold increase in fluorescence.
In comparison, titration of this same peptide with GST-
Finally, we have successfully employed TP-2Rho-labelled
peptides to probe endogenous CDK4/cyclin D1 in living cells,
indicating the potential of TP-2Rho for cellular imaging appli-
cations.
ChemBioChem 2016, 17, 737 – 744
www.chembiochem.org
742
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
(d, J=9 Hz, 2H), 4.02 (t, J=6 Hz, 2H), 3.51 (t, J=6.5 Hz, 2H), 2.16–
1.92 ppm (m, 4H); 13C NMR (75 MHz, CDCl3): d=190.62, 157.58,
152.21, 138.23, 131.41, 131.06, 129.03, 122.18, 116.09, 67.30, 33.47,
29.55, 27.99 ppm; MS (ES+): m/z: 453.0.
Experimental Section
General: All chemicals were purchased from Sigma–Aldrich or
Acros Organics and were used as received. All other chemicals
were used as received from the appropriate suppliers. Preparative
flash chromatography was carried out with Merck silica gel (Si 60,
35–70 mm). NMR spectra were recorded with a Bruker Advance 300
spectrometer; 1H NMR spectra were recorded at 300 MHz and
13C NMR spectra at 75 MHz. Chemical shifts are reported in ppm
downfield from TMS (d=0.00) for 1H NMR and from the central
CDCl3 resonance (d=77.16) or [D6]DMSO resonance (d=39.52) for
13C NMR. HPLC was carried out with a Waters Alliance instrument
equipped with a photodiode array detector and a Xterra MS
column with a linear gradient of acetonitrile versus water contain-
ing 0.1% FA ranging from 10 to 55% over 5 min and then to
100% over 6 min at a flow rate of 1 mLminÀ1. Mass spectrometry
was performed at the Institut Curie for Electrospray Ionization (ESI).
HRMS was performed at the Institut de Chimie des Substances Na-
turelles (ICSN) in Gif-sur-Yvette.
4’,4’’-Bisformyl-4-(4-maleimidobutoxy)triphenylamine
(6):
3a,4,7,7a-Tetrahydro-2H-4,7-epoxyisoindole-1,3-dione ([6253–28–7],
487 mg, 2.95 mmol, 2 equiv) and potassium carbonate (407 mg,
2.95 mmol, 2 equiv) were added under argon to a solution of dial-
dehyde 4 (667 mg, 1.47 mmol, 1 equiv) in anhydrous DMF (40 mL).
The resulting mixture was stirred for one hour at room tempera-
ture and overnight at 408C. Solvents were removed under reduced
pressure. Water (20 mL) was added to the residue, and the ob-
tained aqueous phase was extracted twice with CH2Cl2. The organ-
ic phase was washed twice with water, dried over MgSO4 and con-
centrated. After inspection by LC-MS, the protected maleimide 5
was dissolved in toluene (50 mL) and stirred at 1208C overnight.
After concentration, purification by flash chromatography using
CombiFlash (CH2Cl2/EtOH 100 to 98:2) afforded the title compound
(354 mg, 0.75 mmol, 51%) as a yellow oil. 1H NMR (300 MHz,
CDCl3): d=9.87 (s, 2H), 7.75 (d, J=8.5 Hz, 4H), 7.15 (d, J=8.5 Hz,
4H), 7.09 (d, J=9 Hz, 2H), 6.91 (d, J=9 Hz, 2H), 6.71 (s, 2H), 3.99
(m, 2H), 3.61 (m, 2H), 1.80 ppm (m, 4H); 13C NMR (75 MHz, CDCl3):
d=190.66, 170.98, 157.61, 152.22, 138.15, 134.26, 131.42, 131.02,
129.02, 122.17, 116.12, 67.53, 37.62, 26.60, 25.47 ppm; MS (ES+): m/
z: 469.2; HRMS (ESI+): m/z calcd for C28H24N2O5: 469.1763; found
469.1768
4,4’-Bisformyl-4’’-methoxytriphenylamine (2): POCl3 (10 mL,
16.7 g, 109 mmol, 15 equiv) was added dropwise, at 08C under Ar,
over 15 min to DMF (9 mL, 7.9 g, 109 mmol, 15 equiv). The solution
was stirred at 08C for 1 h. 4-Methoxytriphenylamine (2 g,
7.26 mmol) was added, and the solution was stirred at 958C for 4 h
30 min. The dark mixture was poured into an ice bath and stirred
for 30 min. The solution was basified with NaOH (1m). The aque-
ous phase was extracted with CH2Cl2, and the organic phase was
washed twice with water, dried over MgSO4 and concentrated
under reduced pressure to give a black oil. Purification by flash
chromatography using CombiFlash (AcOEt/cyclohexane 10:80 to
70:30) afforded the title compound (1.72 g, 5.19 mmol, 71%). CAS
TP-2Rho-maleimide: Dialdehyde 6 (180 mg, 0.38 mmol), rhoda-
mine-3-acetic acid (162 mg, 0.84 mmol, 2.2 equiv) and ammonium
acetate (119 mg, 1.53 mmol, 4 equiv) were dissolved in acetic acid
(50 mL). The solution was stirred at 110 8C overnight. The reaction
mixture was allowed to cool to room temperature and concentrat-
ed under vacuum until approximatively 10 mL of acetic acid re-
mained. Ether was then added to the solution, and a dark red pre-
cipitate formed. The solid was filtered, washed with ether and
dried to afford the expected product (142 mg, 0.17 mmol, 46%) as
1
number [149676-16-4]; H NMR (300 MHz, CDCl3): d=9.87 (s, 2H),
7.75 (d, J=9 Hz, 4H), 7.16 (d, J=9 Hz, 4H), 7.11 (d, J=9.0 Hz, 2H),
6.94 (d, J=9.0 Hz, 2H), 3.84 ppm (s, 3H).
4,4’-Bisformyl-4’’-hydroxytriphenylamine (3): BBr3 (2.7 mL, 7.21 g,
28.79 mmol, 3 equiv) was added dropwise at 08C to a solution of 2
(3.18 g, 9.6 mmol, 1 equiv) in CH2Cl2 (20 mL). The reaction mixture
was then allowed to warm slowly to RT and stirred for 3 h. The so-
lution turned from yellow to dark. The mixture was then slowly
poured into an ice bath. The aqueous phase was extracted twice
with CH2Cl2. The combined organic phases were washed twice
with water, dried over MgSO4 and concentrated to give a green
foamy solid. Purification by flash chromatography using Combi-
Flash (AcOEt/cyclohexane 10:90!60:40) afforded the title com-
pound (2.8 g, 8.82 mmol, 92%) as a yellow solid. CAS number
[336619-78-4]; 1H NMR (300 MHz, CDCl3): d=9.88 (s, 2H), 7.76 (d,
J=8.5 Hz, 4H), 7.17 (d, J=8.5 Hz, 4H), 7.07 (d, J=8.5 Hz, 2H),
6.89 ppm (d, J=8.5 Hz, 2H).
1
a dark red powder. M.p. 244–2468C; H NMR (300 MHz, DMSO): d=
7.75 (s, 2H), 7.60 (d, J=9 Hz, 4H), 7.14 (m, 6H), 7.01 (m, 4H), 4.54
(s, 4H), 4.00 (t, J=5 Hz, 2H), 3.48 (t, J=6 Hz, 2H), 1.72–1.64 ppm
(m, 4H); 13C NMR (75 MHz, DMSO): d=192.82, 171.12, 167.16,
166.62, 157.04, 148.51, 137.33, 134.45, 132.79, 132.56, 128.93,
126.75, 122.20, 119.22, 116.08, 67.13, 46.27, 36.82, 25.98,
24.74 ppm; HRMS (ESI+): m/z calcd for C38H30N4O9S4: 815.0974;
found 815.1003.
Peptide synthesis and labelling: Peptides used in this study were
purchased from GL Biochem (Shanghai, P.R. China), labelled on
their single cysteine residues with maleimide dye derivatives, in
the presence of a five- to tenfold excess of dye overnight, and
then purified from free dye on NAP-5 columns (GE Healthcare).
4’,4’’-Bisformyl-4-(4-bromobutoxy)triphenylamine (4): A solution
of caesium carbonate (4.1 g, 12.6 mmol, 5 equiv) and 3 (0.8 g,
2.52 mmol, 1 equiv) in anhydrous DMF (30 mL) was stirred at room
temperature for 15 min. 1,4-Dibromobutane (4.5 mL, 37.5 mmol,
15 equiv) was then added, and the resulting mixture was stirred at
room temperature for 2 h. The reaction mixture was then concen-
trated under reduced pressure. Water (20 mL) was added to the
residue, and the obtained aqueous phase was extracted twice with
CH2Cl2. The organic phase was washed twice with water, dried
over MgSO4 and concentrated to give a yellow oil. Purification by
flash chromatography using CombiFlash (AcOEt/cyclohexane 10:90
to 60:40) afforded the title compound (475 mg, 1.05 mmol, 42%)
PISTVRE peptide: GGGCPISTVREVALLRRL
a5 D1 peptide: SIRPEELLQMELLLVNKLKWNLC
Rb peptide: YKFCSSPLRIPG
Pep-PISTVRE peptide: KETWWETWWTEKK GGGCPISTVREVALLRRL
Protein expression and purification of CDKs: Recombinant CDKs
and cyclins were expressed in Escherichia coli by IPTG induction
and purified by chromatography as described previously.[9h] GST-
CDKs or cyclins were first purified on GSTTrap columns, followed
by gel filtration chromatography on HiLoad 16/600 Superdex 75
preparatory grade.
1
as a yellow oil. H NMR (300 MHz, CDCl3): d=9.88 (s, 2H), 7.76 (d,
Design and engineering of CDK4 mutant: The cDNA sequence of
J=8.5 Hz, 4H), 7.16 (d, J=8.5 Hz, 4H), 7.11 (d, J=9 Hz, 2H), 6.92
mouse CDK4 was cloned into the pGex6P1 vector (GE Healthcare)
ChemBioChem 2016, 17, 737 – 744
www.chembiochem.org
743
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[1] X. Li, X. Gao, W. Shi, H. Ma, Chem. Rev. 2014, 114, 590–659.
[2] a) M. E. Vµzquez, J. B. Blanco, B. Imperiali, J. Am. Chem. Soc. 2005, 127,
1300–1306; b) G. Loving, B. Imperiali, J. Am. Chem. Soc. 2008, 130,
13630–13638.
[3] a) I. A. Karpenko, M. Collot, L. Richert, C. Valencia, P. Villa, Y. Mely, M.
Hibert, D. Bonnet, A. S. Klymchenko, J. Am. Chem. Soc. 2015, 137, 405–
412; b) I. A. Karpenko, R. Kreder, C. Valencia, P. Villa, C. Mendre, B. Mouil-
lac, Y. Mely, M. Hibert, D. Bonnet, A. S. Klymchenko, ChemBioChem
2014, 15, 359–363.
[4] a) C. Allain, F. Schmidt, R. Lartia, G. Bordeau, C. Fiorini-Debuisschert, F.
Charra, P. Tauc, M. P. Teulade-Fichou, ChemBioChem 2007, 8, 424–433;
b) B. Dumat, G. Bordeau, E. Faurel-Paul, F. Mahuteau-Betzer, N. Saettel,
G. Metge, C. Fiorini-Debuisschert, F. Charra, M. P. Teulade-Fichou, J. Am.
Chem. Soc. 2013, 135, 12697–12706.
[5] B. Dumat, G. Bordeau, A. I. Aranda, F. Mahuteau-Betzer, Y. El Harfouch,
G. Metge, F. Charra, C. Fiorini-Debuisschert, M. P. Teulade-Fichou, Org.
Biomol. Chem. 2012, 10, 6054–6061.
5’-BamHI 3’-EcoRI. Cysteine residues 202 and 215 were mutagen-
ized to serine, according to the QuickStart protocol (Stratagene),
and DNA extracted from mutant E. coli colonies was sequenced to
confirm mutations and to verify that no undesired mutations had
occurred.
Fluorescence titration experiments: Fluorescence titration assays
were performed in 96-well plates in potassium phosphate [200 mL,
KH2PO4/K2PO4 (50 mm), pH 7.4] and NaCl (150 mm) solution with
use of a Clariostar spectrofluorimeter (BMG). The fluorescently la-
belled peptides or proteins (200 nm) were excited at 510 nm, and
emission signals were acquired at 614 nm unless stated otherwise.
Data analysis was performed with the aid of the GraFit Software
(Erithacus, Ltd., Horley, UK). Experiments were performed in tripli-
cate, and error bars indicate standard deviations from averages.
Cell culture, internalization and microscopy: Cell culture media,
serum and antibiotics were purchased from Invitrogen. HeLa cells
were cultured in Dulbecco’s modified Eagle’s medium (DMEM)+
Glutamax supplemented with foetal calf serum (FCS, 10%), penicil-
[6] a) W. R. Zipfel, R. M. Williams, W. W. Webb, Nat. Biotechnol. 2003, 21,
1369–1377; b) B. G. Wang, K. Konig, K. J. Halbhuber, J. Microsc. 2010,
238, 1–20.
[7] a) M. Malumbres, M. Barbacid, Trends Biochem. Sci. 2005, 30, 630–641;
b) S. Lim, P. Kaldis, Development 2013, 140, 3079–3093.
[8] a) U. Asghar, A. K. Witkiewicz, N. C. Turner, E. S. Knudsen, Nat. Rev. Drug
Discovery 2015, 14, 130–146; b) M. Peyressatre, C. PrØvel, M. Pellerano,
M. Morris, Cancers 2015, 7, 179.
[9] a) D. O. Morgan, Annu. Rev. Cell Dev. Biol. 1997, 13, 261–291; b) A. J.
Obaya, J. M. Sedivy, Cell. Mol. Life Sci. 2002, 59, 126–142; c) A. Satyanar-
ayana, P. Kaldis, Oncogene 2009, 28, 2925–2939; d) P. D. Jeffrey, A. A.
Russo, K. Polyak, E. Gibbs, J. Hurwitz, J. Massague, N. P. Pavletich, Nature
1995, 376, 313–320; e) A. A. Russo, P. D. Jeffrey, N. P. Pavletich, Nat.
Struct. Mol. Biol. 1996, 3, 696–700; f) N. R. Brown, M. E. M. Noble, A. M.
Lawrie, M. C. Morris, P. Tunnah, G. Divita, L. N. Johnson, J. A. Endicott, J.
Biol. Chem. 1999, 274, 8746–8756; g) F. Heitz, M. C. Morris, D. Fesquet,
J. C. Cavadore, M. Doree, G. Divita, Biochemistry 1997, 36, 4995–5003;
h) M. C. Morris, C. Gondeau, J. A. Tainer, G. B. Divita, J. Biol. Chem. 2002,
277, 23847–23853.
lin G (sodium salt, 100 unitsmLÀ1) and streptomycin (100 mgmLÀ1
at 378C under an atmosphere containing CO2 (5%).
)
TP-2Rho-labelled PISTVRE peptide or Pep-PISTVRE peptide (2 mm)
was overlaid onto HeLa cells grown to 50–60% confluency for 1 h
at 378C and then for another hour in the presence of DMEM com-
plemented with serum (10%).
Cells were then fixed with formalin in phosphate-buffered saline
(PBS) for 10 min and washed in PBS, and the nuclei were stained
with Hoechst 33342. Indirect immunofluorescence was performed
with rabbit polyclonal anti-CDK4 (C-22 from Santa Cruz) at 1:250
dilution or anti-cyclin D1 antibody (C-20 from Santa Cruz) at 1:250
dilution in blocking buffer [BSA (4%), goat serum (4%), Triton
(0.1%)] overnight, washing, subsequent incubation with Alexa647-
conjugated secondary antibodies at 1:500 dilution for 1 h, washing
again in PBS, staining of nuclei with Hoechst 33342, washing in
water and mounting on glass slides in Mowiol. Fluorescent cells
were observed with a Zeiss microscope equipped with a CoolSnap
camera and images were acquired with the aid of MetaMorph soft-
ware. Excitation band/dichroic/emission band filters for imaging
fluorescent signals were as follows: Hoechst 33342: 340–380/400/
450–490 nm. Alexa647-conjugated antibodies were imaged by
using the Cy5 filters: 590–650/660/662–737 nm. TP-2Rho fluores-
cence was imaged by using the Cy3 filters: 525–565/565/572–
647 nm.
[10] H. L. De Bondt, J. Rosenblatt, J. Jancarik, H. D. Jones, D. O. Morgant, S.-
H. Kim, Nature 1993, 363, 595–602.
[11] a) A. Ho, S. F. Dowdy, Curr. Opin. Genet. Dev. 2002, 12, 47–52; b) S. J.
Baker, E. P. Reddy, Genes Cancer 2012, 3, 658–669.
[12] a) M. Malumbres, M. Barbacid, Nat. Rev. Cancer 2001, 1, 222–231; b) Y.
Dobashi, A. Goto, M. Fukayama, A. Abe, A. Ooi, Int. J. Cancer 2004, 110,
532–541; c) O. Gautschi, D. Ratschiller, M. Gugger, D. C. Betticher, J.
Heighway, Lung Cancer 2007, 55, 1–14; d) M. Puyol, A. Martin, P. Dubus,
F. Mulero, P. Pizcueta, G. Khan, C. Guerra, D. Santamaria, M. Barbacid,
Cancer Cell 2010, 18, 63–73; e) E. A. Musgrove, C. E. Caldon, J. Barra-
clough, A. Stone, R. L. Sutherland, Nat. Rev. Cancer 2011, 11, 558–572;
f) K. E. Sheppard, G. A. McArthur, Clin. Cancer Res. 2013, 19, 5320–5328.
[13] a) P. J. Day, A. Cleasby, I. J. Tickle, M. O’Reilly, J. E. Coyle, F. P. Holding,
R. L. McMenamin, J. Yon, R. Chopra, C. Lengauer, H. Jhoti, Proc. Natl.
Acad. Sci. USA 2009, 106, 4166–4170; b) T. Takaki, A. Echalier, N. R.
Brown, T. Hunt, J. A. Endicott, M. E. M. Noble, Proc. Natl. Acad. Sci. USA
2009, 106, 4171–4176.
Acknowledgements
[14] N. Floquet, M. G. S. Costa, P. R. Batista, P. Renault, P. M. Bisch, F. Raussin,
J. Martinez, M. C. Morris, D. Perahia, Biophys. J. 2015, 109, 1179–1189.
[15] M. C. Morris, J. Depollier, J. Mery, F. Heitz, G. Divita, Nat. Biotechnol.
2001, 19, 1173–1176.
[16] A. Hawe, M. Sutter, W. Jiskoot, Pharm. Res. 2008, 25, 1487–1499.
[17] L. D. Lavis, R. T. Raines, ACS Chem. Biol. 2008, 3, 142–155.
[18] a) M. P. Bruchez, Curr. Opin. Chem. Biol. 2015, 27, 18–23; b) L. M. Wy-
socki, L. D. Lavis, Curr. Opin. Chem. Biol. 2011, 15, 752–759; c) L. D.
Lavis, R. T. Raines, ACS Chem. Biol. 2014, 9, 855–866.
This work was supported by the CNRS (Centre National de la Re-
cherche Scientifique) and a grant from the Institut de Recherche
contre Cancer (INCA) to M.C.M. This work has benefited from the
facilities and expertise of the Small Molecule Mass Spectrometry
platform of IMAGIF (Centre de Recherche de Gif: www.ima-
gif.cnrs.fr), and of the Montpellier RIO imaging facility
(www.mri.cnrs.fr) at the Centre de Recherches en Biochimie Mac-
romolØculaire, Montpellier.
Keywords: biosensors · CDK4/cyclin D kinase · peptide–
peptide/protein interactions · protein–protein interactions ·
solvatochromic fluorescent probes
Manuscript received: November 26, 2015
Accepted article published: March 4, 2016
Final article published: March 17, 2016
ChemBioChem 2016, 17, 737 – 744
www.chembiochem.org
744
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim