Sequence-selective DNA recognition with peptide-bisbenzamidine conjugates

Article abstract of DOI:10.1002/chem.201300519

Transcription factors (TFs) are specialized proteins that play a key role in the regulation of genetic expression. Their mechanism of action involves the interaction with specific DNA sequences, which usually takes place through specialized domains of the protein. However, achieving an efficient binding usually requires the presence of the full protein. This is the case for bZIP and zinc finger TF families, which cannot interact with their target sites when the DNA binding fragments are presented as isolated monomers. Herein it is demonstrated that the DNA binding of these monomeric peptides can be restored when conjugated to aza-bisbenzamidines, which are readily accessible molecules that interact with A/T-rich sites by insertion into their minor groove. Importantly, the fluorogenic properties of the aza-benzamidine unit provide details of the DNA interaction that are eluded in electrophoresis mobility shift assays (EMSA). The hybrids based on the GCN4 bZIP protein preferentially bind to composite sequences containing tandem bisbenzamidine-GCN4 binding sites (TCAT×AAATT). Fluorescence reverse titrations show an interesting multiphasic profile consistent with the formation of competitive nonspecific complexes at low DNA/peptide ratios. On the other hand, the conjugate with the DNA binding domain of the zinc finger protein GAGA binds with high affinity (K_D≈12 nM) and specificity to a composite AATTT×GAGA sequence containing both the bisbenzamidine and the TF consensus binding sites. Bound by recognition: Conjugation of fragments of bZIP and zinc finger transcription factors to bisbenzamidines allows selective DNA recognition of composite DNA sequences containing the original target site of the peptide fragment in tandem with A/T-rich sequences targeted by the minor-groove binder unit (see figure). The fluorogenic aza-bisbenzamidine is exploited for the spectroscopic characterization of the DNA-recognition process. Copyright

Full text of DOI:10.1002/chem.201300519

FULL PAPER
DOI: 10.1002/chem.201300519
Sequence-Selective DNA Recognition with Peptide–Bisbenzamidine
Conjugates
Mateo I. Sꢀnchez, Olalla Vꢀzquez, M. Eugenio Vꢀzquez,* and Josꢁ L. MascareÇas*[a]
Abstract: Transcription factors (TFs)
are specialized proteins that play a key
role in the regulation of genetic expres-
sion. Their mechanism of action in-
volves the interaction with specific
DNA sequences, which usually takes
place through specialized domains of
the protein. However, achieving an ef-
ficient binding usually requires the
presence of the full protein. This is the
case for bZIP and zinc finger TF fami-
lies, which cannot interact with their
target sites when the DNA binding
fragments are presented as isolated
monomers. Herein it is demonstrated
that the DNA binding of these mono-
meric peptides can be restored when
conjugated to aza-bisbenzamidines,
which are readily accessible molecules
that interact with A/T-rich sites by in-
sertion into their minor groove. Impor-
tantly, the fluorogenic properties of the
aza-benzamidine unit provide details of
the DNA interaction that are eluded in
electrophoresis mobility shift assays
(EMSA). The hybrids based on the
GCN4 bZIP protein preferentially bind
to composite sequences containing
tandem bisbenzamidine–GCN4 binding
sites (TCAT·AAATT). Fluorescence
reverse titrations show an interesting
multiphasic profile consistent with the
formation of competitive nonspecific
complexes at low DNA/peptide ratios.
On the other hand, the conjugate with
the DNA binding domain of the zinc
finger protein GAGA binds with high
affinity (KD ꢀ12 nm) and specificity to
a composite AATTT·GAGA sequence
containing both the bisbenzamidine
and the TF consensus binding sites.
Keywords: DNA recognition · fluo-
rescence · molecular recognition ·
oligonucleotides · peptides
Introduction
small peptide motifs, high-affinity DNA binding requires the
full protein domain and, in many cases, the concerted action
of multiple binding components.[5] Therefore, the prepara-
tion of small synthetic mimics of DNA-binding proteins is
extremely challenging. Chemists have been able to engineer
relatively small peptides capable of interacting with good af-
finity to specific dsDNA sequences.[6] Our own approach for
the design of such minimized systems has relied on the con-
jugation of small peptide fragments derived from the DNA
binding domains of TFs to tripyrrolic distamycin derivatives
that bind to the minor groove of A/T-rich DNA sequences.[7]
Although the strategy is effective, the synthesis of the tripyr-
role moiety requires more than seven steps, and the han-
dling of relatively unstable aminopyrrole synthetic inter-
mediates. Therefore, the use of other minor groove binding
moieties that could be synthetically more accessible repre-
sents an important goal. In this context, we were attracted
by propamidine and its derivatives, well-known A/T-rich
DNA binders that display a much simpler structure than
that of distamycin-related polyamides.[8] Herein we describe
the synthesis of a hybrid between the basic region of GCN4
and propamidine, and demonstrate by electrophoresis mobi-
lity shift assays (EMSA) that this conjugate is capable of
binding to designed DNA sites with good selectivity. More
importantly, following our recent discovery of aza-bisbenz-
Gene expression is regulated by the action of specialized
proteins called transcription factors (TFs) that bind to spe-
cific DNA regulatory sequences,[1] and thereafter promote
the assembly of the multiprotein complex directly responsi-
ble for the initiation of transcription.[2] As a consequence of
their fundamental role in the regulation of gene expression,
it is not surprising that alterations in the activity of TFs are
at the origin of many diseases, including cancer.[3] Therefore,
the development of non-natural agents that can emulate or
interfere with the double stranded DNA (dsDNA) recogni-
tion of TFs remains a major goal in biological chemistry.
These agents might have a great impact on fundamental and
applied biological research, and even lead to the develop-
ment of gene-targeted therapies.[4] It is well-established that
although the DNA-binding of TFs is mediated by relatively
[a] M. I. Sꢀnchez, Dr. O. Vꢀzquez, Prof. M. E. Vꢀzquez,
Prof. J. L. MascareÇas
Departamento de Quꢁmica Orgꢀnica and
Centro Singular de Investigaciꢂn en Quꢁmica Biolꢂxica e
Materiais Moleculares (CIQUS)
Universidade de Santiago de Compostela
15782 Santiago de Compostela (Spain)
Fax: (+34)981-595-012
AHCTUNGTREGaNNNU midines as highly sensitive fluorogenic minor groove DNA
E-mail: eugenio.vazquez@usc.es
binding agents,[8a,9] we also describe the synthesis of conju-
gates between one of these minor-groove binders and frag-
ments of the GCN4 and GAGA proteins, two representative
joseluis.mascarenas@usc.es
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201300519.
Chem. Eur. J. 2013, 19, 9923 – 9929
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9923

members of the bZIP and zinc-finger families of transcrip-
tion factors. In these conjugates, the aza-bisbenzamidine de-
rivative not only promotes the interaction of the peptide
with the DNA, but also acts as a sensing device that allows
spectroscopic monitoring of the DNA binding process in
solution.
DNA binding studies of GCN4–pr: DNA binding experi-
ments were carried out by using standard nondenaturing
EMSA protocols in polyacrylamide gels[11] with double
stranded oligonucleotides containing a composite binding
site including the consensus recognition sequences of the
GCN4 basic region and the propamidine binder, in tandem
(TCAT·AAATTT). As shown in Figure 1 (lanes 1–4), incu-
Results and Discussion
Synthesis of the GCN4–propamidine hybrid (GCN4–pr):
The design of the GCN4–propamidine conjugate was based
on the GCN4 peptide fragment comprising residues Asp226
to Gln248, which has been identified as the smallest peptide
that retains specific DNA recognition properties as a
dimer,[10] and was used in our previous studies with the tri-
pyrrole conjugates.[7e] The propamidine derivative containing
an appropriate linker for conjugation (3) was readily synthe-
sized in three steps from commercially available p-fluoro-
benzonitrile (Scheme 1). The resulting propamidine-amino
Figure 1. EMSA GCN4–pr DNA binding assays in Tris-HCl buffer. Lanes
1–4: target AP1hs·AT dsDNA (50 nm); lanes 2–4: 100, 300, 500 nm
GCN4–pr; lanes 5–8: control AP1hs·GC dsDNA (50 nm); lanes 6–8: 100,
300, 500 nm GCN4–pr; lanes 9–12: control GC·AT DNA (50 nm); lanes
10–12: 100, 300, 500 nm GCN4–pr. dsDNA sequences (binding site in ital-
ics, only one strand is shown): AP1hs·AT: 5’-CGAACG TCAT AAATTT
CCTC-3’; AP1hs·GC: 5’-ACGAACG TCAT GGCC GCCTC-3’; GC·AT:
5’-ACGAAG GCGGC AAATT CCTC-3’. Products were resolved by
PAGE on a 10% nondenaturing polyacrylamide gel and 0.5ꢄTBE buffer
over 40 min at RT and analyzed by staining with SyBrGold (Molecular
Probes: 5 mL in 50 mL of 1ꢄTBE) for 10 min, followed by fluorescence
visualization.
bation of the target dsDNA oligonucleotide with the
GCN4–pr hybrid leads to the formation of a relatively clean
retention band, consistent with the formation of the desired
complex in which the peptide derivative binds in the major
groove of its target sequence (TCAT) and the propamidine
is inserted in the adjacent A/T-rich minor groove
(AAATTT).
Consistent with this interaction model, circular dichroism
experiments confirmed that the peptide chain folds into an
a-helix upon interaction with the DNA, as expected for spe-
cific recognition (see the Supporting Information). A con-
trol oligonucleotide lacking the A/T-rich site failed to pro-
vide DNA–peptide complexes (Figure 1, lanes 5–8). Incuba-
tion with oligonucleotides that did not exhibit a consensus
peptide binding site (Figure 1, lanes 9–12) resulted in the
formation of complexes of lower affinity that are more re-
tained in the gel. These complexes most probably arise from
nonspecific DNA interactions involving the highly charged
basic region, whereas the propamidine unit is inserted in the
minor groove.[7d,e] Taken together, these results suggest that
the propamidine is a viable alternative to the distamycin de-
rivatives previously used as minor-groove binders, although
the performance in terms of affinity and selectivity seems
poorer.
Scheme 1. Top: synthesis of the propamidine derivative 3 for conjugation
with peptide. Bottom: key steps in the synthesis of the conjugate between
3 and a minimal basic region of the GCN4 transcription factor (Asp226
to Gln252, see the Supporting Information; only the key Glu required
for the coupling is indicated). The allyl removal and the coupling steps
are done while the peptide is still attached to the solid support.
derivative was conjugated to the peptide fragment while it
was still attached to the resin, and was selectively deprotect-
ed at Glu245, which replaces a native Arg residue in that
position. After standard deprotection/cleavage and reverse-
phase HPLC purification, the expected conjugate GCN4–pr
was obtained in a good overall yield.
Design and synthesis of bisbenzamidine conjugates of
GCN4 and GAGA transcription factors: At this stage, we
discovered that aza-bisbenzamidines, which feature a nitro-
gen instead of an oxygen atom at the para-position of the
benzamidinium moiety, are weakly fluorescent in aqueous
solvents but exhibit a great increase in the quantum yield of
9924
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9923 – 9929

DNA Recognition with Peptide–Bisbenzamidine Conjugates
FULL PAPER
their fluorescence emission when bound to A/T-rich sites in
DNA.[9] We, therefore, envisioned the construction of a new
type of hybrid featuring these chromophores as minor
groove binding moieties, which might allow us to gather in-
formation on the DNA interaction of the hybrids in solution
using spectroscopic methods. In particular, we decided to
use as minor groove binding unit the phenol derivative 5,
which can be synthesized in a single step from commercial
products, and contains a hydroxyl handle to engineer the
connection to the peptide chains. With this compound at
hand, we did not only approach the construction of the
hybrid with the basic region of GCN4, but also a conjugate
with a truncated module of the zinc finger protein GAGA
(Ser28 to Phe58), a 30 residue peptide that by itself is inca-
pable of showing significant DNA affinity.[7a,12]
The design of the GCN4-bisbenzamidine hybrid (GCN4–
bb) relies on the same Arg245!Glu245 mutation of the
GCN4 DNA binding domain fragment previously used for
GCN4–pr (Figure 2). In this case, the conjugate was obtain-
Scheme 2. Top: synthesis of the bisbenzamidine aromatic derivative 6 by
reductive amination of 5-hydroxyisophthalaldehyde, introduction of the
aminoalkylic side chain in the phenolylc oxygen, and acidic deprotection.
Bottom: key steps in the synthesis of GAGA–bb. The black circle repre-
sents the solid resin used in the peptide synthesis.
Figure 2. Left: schematic representation of the simultaneous interaction
of the GCN4 basic region peptide (gray cylinder) and the bisbenzamidine
(inserted block), indicating the Glu245 residue selected for conjugation.
Right: representation of a simultaneous interaction of the DNA-binding
domain of the GAGA transcription factor fragment in the major groove
and the bisbenzamidine in the minor groove. In this case Lys44 is select-
ed as an appropriate conjugation position. The models used to design the
synthetic targets were based on the reported structures of the DNA bind-
ing complexes of the GCN4 and GAGA transcription factors[13] and that
of pentamidine,[14] as described in earlier reports of distamycin hybrids.[7]
ylic acid for attachment of the bisbenzamidine. Once the
carboxylic acid functionality was set in the peptide scaffold,
the bisbenzamidine 6 was coupled to the solid-phase linked
peptide with HATU as activating reagent. A standard cleav-
age and deprotection step involving treatment with a TFA
cocktail, followed by reverse-phase HPLC purification gave
the desired conjugate GAGA–bb in reasonable overall yield
of approximately 19%.
ing by coupling the aza-bisbenzamidine amine 6, instead of
the previously used propamidine-amine 3 (see the Support-
ing Information). Compound 6 was readily constructed by a
reductive amination reaction between 4 and 5-hydroxyiso-
DNA binding studies of GCN4–bb and GAGA–bb conju-
gates: Having synthesized the desired peptide conjugates,
we first assessed their DNA binding ability using standard
EMSA experiments under nondenaturing conditions. Thus,
incubation of the double stranded oligonucleotide
AP1hs·AT, containing the target composite binding site
(TCAT·AAATT) at room temperature with increasing
amounts of GCN4–bb led to the appearance of two new
bands in addition to that corresponding to the free DNA
(Figure 3, top, lanes 1–5). The faster, more intense band, is
consistent with the expected compact hybrid-DNA complex
(band a), whereas the slower migrating band could result
from a complex in which the bisbenzamidine unit is inserted
into the minor groove of the A/T-rich region, whereas the
extended peptide makes nonspecific contacts with the DNA
ACHTUNGTRENNUNGphthalate, followed by alkylation of the phenolic oxygen in
5 with a Boc-protected derivative of N-(5-iodopentyl)pro-
pane-1,3-diamine, and a final TFA deprotection. Therefore,
the synthesis of bisbenzamidine 6 is straightforward, and in-
volves only two separate steps from commercial and inex-
pensive starting materials.
The zinc finger conjugate was prepared from a peptide
containing the truncated zinc finger unit of the GAGA tran-
scription factor in which the Arg44 was replaced by Lys.
This Lys44 was introduced with its side chain protected with
an orthogonal alloc group that could be selectively removed
in the solid phase (Scheme 2). The deprotected Lys side
chain was then derivatized with glutaric anhydride to in-
crease the linker length and simultaneously install a carbox-
Chem. Eur. J. 2013, 19, 9923 – 9929
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9925

M. E. Vꢀzquez, J. L. MascareÇas et al.
and hence provides complementary information that could
not be derived from EMSA studies. Thus, incubation of a
solution of GCN4–bb (0.25 mm) with increasing concentra-
tions of a double stranded oligonucleotide (AP1hs·AT) con-
taining the composite consensus binding site (TCA-
T·AAATT) led to a notable and progressive increase in the
fluorescence emission (lexc =329 nm). Curiously, the result-
ing titration curve exhibited an atypical shape, consisting of
a spike in the emission intensity at low DNA/hybrid ratios,
followed by a gradual increase as in a conventional satura-
tion curve, at higher DNA concentrations (Figure 4, left).
Figure 3. Top: EMSA assays of DNA recognition by GCN4–bb. Lanes 1–
5: target AP1hs·AT dsDNA (50 nm); lanes 2–5: 100, 300, 500, 800 nm
GCN4–bb; lanes 6–10: control AP1m·AT dsDNA (50 nm); lanes 7–10:
100, 300, 500, 800 nm GCN4–bb. Bottom: EMSA assays of DNA recogni-
tion by GAGA–bb. Lanes 1–7: target dsDNA AT·GAG (50 nm); lanes 2–
7: 100, 200, 300, 400, 500, 600 nm GAGA–bb; lanes 8–10: control
ATm·GAG dsDNA (50 nm); lanes 9 and 10: 300, 500 nm GAGA–bb.
dsDNA sequences (binding site in italics): AP1hs·AT: 5’-CGAACG
TCAT AAATTT CCTC-3’; AP1m·AT: 5’-CGAACG TGCT AAATTT
CCTC-3’; AT·GAG: 5’-GACGG AATTT GAGAG CGTCG-3’;
ATm·GAG: 5’-GACC GGGCC GAGAG TACGT-3’. EMSA assays of
DNA recognition by GCN4–bb were done under the same conditions as
for GCN4–pr; in the case of GAGA–bb the buffer used was slightly
different (see the Supporting Information for details).
Figure 4. Left: fluorescence emission titration of a solution of GCN4–bb
(0.25 mm) with dsDNA containing the target sequence (AP1hs·AT) in Tris-
HCl buffer (20 mm, pH 7.5), NaCl (100 mm; 208C). Solid line: a nonlin-
ear fit to a mixed model considering a 1:1 complex and competitive spe-
cies resulting from nonspecific binding; dashed line: the best fit to a
simple 1:1 binding mode, discarding points 2 to 9 in the titration. Right:
titration of a 0.5 mm solution of GAGA–bb with DNA containing the
composite recognition site (AT·GAG). Solid line: the nonlinear fit to a
1:1 binding mode for AT·GAG and a mixed mode including competitive
nonspecific binding for the G/C-rich oligonucleotide.
(Figure 3 top, band b). This assignment is supported by the
observation of a similarly retarded band when GCN4–bb is
incubated with an oligonucleotide featuring a mutated pep-
tide binding site (AP1m·AT; Figure 3, top, lanes 6–10).[7e] As
expected, incubation of GCN4–bb with a G/C-rich or the
AP1hs·GC control oligonucleotide did not show any new
bands (see the Supporting Information).
In the case of the conjugate GAGA–bb, we observed a
single retarded band when treated with the AT·GAG oligo-
nucleotide containing the composite binding site. This is
consistent with the formation of the expected specific com-
plex between the hybrid and the DNA (Figure 3, bottom,
lanes 1–7). No retarded bands were observed when this
This type of multiphasic profile has been previously ob-
served in reverse titrations of cationic proteins with DNA,[15]
and could be explained in terms of the competitive forma-
tion of multiple nonspecific complexes with the DNA
oligomers at low DNA/ligand ratios. Upon increasing the
DNA concentration, the 1:1 binding mode becomes predom-
inant, and the curve follows a more familiar titration profile,
consistent with the results observed in the gel-shift experi-
ments (note that the EMSA experiments were performed as
forward titrations: constant DNA concentration and addi-
tion of the peptide).
In contrast with the relatively complex binding profile dis-
played by the GCN4–bb hybrid, fluorescence titration of a
solution of the zinc-finger conjugate GAGA–bb (0.50 mm)
with its target DNA (AT·GAG), containing the composite
binding sequence (AATT·GAG), could be fitted to a simple
1:1 binding mode with an apparent KD of approximately
12 nm, (Figure 2, right).[16] Titration of this same hybrid with
a G/C-rich double stranded DNA, resulted in a multiphasic
profile with a very small increase in the emission intensity
from the minor-groove binder, in line with the expected low
same hybrid was incubated with
a control dsDNA
(ATm·GAG), lacking the minor groove recognition sequence
(Figure 2, bottom, lanes 8–10); likewise, incubation with
DNA containing the consensus minor groove binding se-
quence but a mutated peptide binding site failed to show
any clear binding by EMSA (see the Supporting Informa-
tion). The different behavior of the two conjugates (GCN4–
bb and GAGA–bb) could be explained by the presence of a
much higher number of basic residues in the case of the
GCN4 hybrid, which promote electrostatic interactions with
the negatively charged dsDNA, and hence the production of
competitive less-specific DNA complexes.
Steady-state fluorescence studies: The presence of the aza-
bisbenzamidine in the new conjugates allows the DNA-bind-
ing process to be monitored by fluorescent spectroscopy,
9926
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9923 – 9929

DNA Recognition with Peptide–Bisbenzamidine Conjugates
FULL PAPER
EAARR SRAE245 KLQ; Ser28 to Phe58 of GAGA: SQSEQ PATCP
ICYAV IK44QSR NLRRH LELRHF.
affinity for this oligonucleotide sequence. Control experi-
ments with the bisbenzamidine 6, which also contains a
charged side chain with two amine groups, are consistent
with this analysis, demonstrating low affinity for G/C-rich
oligonucleotides and high affinity binding for A/T-rich sites
(see the Supporting Information).
N-(3-{[(tert-Butoxy)carbonyl](5-oxopentyl)amino}propyl)carbamate 4-[2-
({5-[(3-aminopropyl)amino]pentyl}amino)-3-(4-carbamimidoylphenoxy)-
propoxy] benzene carboximidamide (3): The amine 2[9b] (90 mg,
0.13 mmol), and tert-butyl (3-((tert-butoxycarbonyl)amino)propyl)(5-oxo-
pentyl)carbamate (39 mg, 0.10 mmol) were dissolved in MeOH (1.25 mL)
in a round-bottom flask. NaBH3CN (8 mg, 0.12 mmol) was added over
the solution, and the mixture was stirred for 5 h at RT. The solvent was
removed under reduced pressure, and the residue was purified by prepa-
rative reverse-phase chromatography. The fractions containing the de-
sired product were collected, concentrated and freeze-dried to obtain the
desired conjugate. The isolated Boc-protected compound was dissolved
in CH2Cl2 (1 mL) and cooled to 08C. TFA (1 mL) was added dropwise,
and the resulting solution was stirred at 08C for 1 h, and at RT for a fur-
ther 2 h. The solvent was removed under reduced pressure, and the resid-
ual TFA was removed by co-distillation with CH2Cl2. The freeze-dried
white powder was identified as the desired product (3; 65.0 mg,
0.08 mmol, 65% overall yield for the two-step process). 1H NMR
(400 MHz, [D6]DMSO): d=1.38 (m, 2H), 1.62 (m, 2H), 1.71 (m, 2H),
1.91 (m, 2H), 2.88 (m, 4H), 2.99 (m, 2H), 3.16 (m, 2H), 4.08 (m, 1H),
4.54 (d, J=4.5 Hz, 4H), 7.21 (d, J=8.9 Hz, 4H), 7.87 (d, J=8.8 Hz, 4H),
8.03 (s, 2H, NH), 8.89 (s, 1H, NH), 9.23 (s, 4H), 9.34 (s, 4H), 9.53 ppm
(s, 1H, NH); 13C NMR (400 MHz, [D6]DMSO): d=23.5 (CH2), 24.2
(CH2), 25.4 (CH2), 25.6 (CH2), 36.6 (CH2), 44.3 (CH2), 46.2 (CH2), 47.0
(CH2), 55.7 (CH), 65.3 (CH2), 115.4 (CH), 121.0 (C), 139.6 (CH), 159.2
(q, C TFA), 162.2 (C), 165.2 ppm (C); ESI+-MS: [M+H] calcd for
C25H40N7O2 470.3238; found 470.3236.
Taken together, these results demonstrate that bisbenz-
ACHTUNGTRENNUNGamidines are synthetically straightforward minor groove
binding handles for the construction of functional conjugates
with TF peptide fragments, which, by themselves, are not ca-
pable of binding to the DNA. In addition to the thermody-
namic stabilization of otherwise unstable complexes, these
anchors display marked fluorogenicity that allowed us to ob-
serve molecular associations that are not evident in regular
gel-shift studies.
Conclusion
In summary, we have demonstrated that conjugation of frag-
ments of transcription factors to bisbenzamidines allows the
selective recognition of relatively long DNA sequences, con-
taining composite sites of the original TF target sequence
and A/T-rich sites targeted by the minor-groove binder.
Moreover, the fluorogenic nature of the minor-groove
binder allows the monitoring of the DNA recognition pro-
Synthesis of 4-({[3-({5-[(3-aminopropyl)amino]pentyl}oxy)-5-{[(4-carb-
AHCTUNGTERGaNNUN mimidoyl phenyl)amino]methyl}phenyl]methyl}amino)benzene-1-carb-
ACHUTNGRENUoNG xACHUTTGNRENiNUGN midamide (6): Potassium tert-butoxide (36.4 mg, 0.324 mmol, 4 equiv)
ACHTUNGTRENNUNGcess by fluorescence spectroscopy. Therefore, the combina-
was added to a solution of the bis-amino benzamidine 5[8a] (50 mg,
tion of standard EMSA analysis and fluorescence titrations
provides a more exact account of the processes taking place
when these hybrids interact with the DNA. Given the ready
accessibility of this minor-groove binder, its optical sensing
properties and the nanomolar affinities and selectivity ex-
hibited by the conjugates, we expect a great future for this
DNA binding strategy.
0.081 mmol) in dry DMSO (1.62 mL). After 30 min, tert-butyl-[3-
[(tert-butoxycarbonyl)amino]propyl](5-iodopentyl)carbamate
(46.3 mg,
0.097 mol, 1.2 equiv) was added in portions. The reaction mixture was
stirred under Ar at RT for 2 h. The crude reaction was directly purified
by preparative reverse-phase chromatography (Bꢅchi Sepacore; gradient:
15% B, 5 min; 15!95% B, 30 min). The combined fractions were con-
centrated and freeze-dried.
The isolated Boc-protected amine was dissolved in CH2Cl2 (2 mL) and
cooled to 08C. TFA (2 mL) was added dropwise and the resulting solu-
tion was stirred at 08C for 1 h and at room temperature for other 2 h.
The solvent was removed under reduced pressure, and the residual TFA
was removed by co-distillation with CH2Cl2. The residue was purified by
preparative reverse-phase chromatography (Bꢅchi Sepacore; gradient:
0% B, 5 min; 0!50% B, 30 min). The freeze-dried solid was identified
as the desired product (6) as a white powder (56.8 mg, 0.072 mmol, 89%
overall yield for the two-step process). 1H NMR (400 MHz, [D6]MeOD):
d=1.32 (m, 2H), 1.55 (m, 2H), 1.77 (m, 2H), 2.07 (m, 2H), 3.05 (m,
2H), 3.11 (m, 2H), 3.94 (t, J=6.1 Hz, 2H), 4.36 (s, 4H), 6.68 (d, J=
8.9 Hz, 4H), 6.79 (s, 2H), 6.94 (s, 1H), 7.57 ppm (d, J=8.9 Hz, 4H);
13C NMR (400 MHz, [D4]DMSO): d=24.2 (CH2), 25.3 (CH2), 27.0 (CH2),
29.9 (CH2), 37.8 (CH2), 45.8 (CH2), 47.5 (CH2), 49.0 (CH2), 68.5 (CH2),
112.8 (CH), 113.1 (CH), 114.3 (C), 118.9 (CH), 130.6 (CH), 142.3 (C),
155.4 (C), 161.2 (C), 167.1 ppm (C); ESI+-MS: [M+H] calcd for
C30H43N8O 531.3554; found 531.3559.
Experimental Section
General synthetic procedures: All reagents were from commercial sour-
ces. DMF and TFA were purchased from Scharlau, CH2Cl2 from Panreac,
CH3CN from Merck. The rest of reagents were from Sigma–Aldrich.
When indicated, reactions were monitored by analytical RP-HPLC with
an Agilent 1100 series LC/MS with an Eclipse XDBC18 (4.6ꢄ150 mm,
5 mm) analytical column. Standard conditions for analytical RP-HPLC
consisted of an isocratic regime during the first 5 min, followed by a
linear gradient from 5 to 75% of solvent B for 30 min at a flow-rate of
1 mLminꢁ1 (A: water with 0.1% TFA, B: CH3CN with 0.1% TFA).
Compounds were detected by UV absorption. Amidine derivatives 2, 3
and 5 were purified on a Bꢅchi Sepacore preparative system consisting of
a pump manager C-615 with two pump modules C-605 for binary solvent
gradients, a fraction collector C-660, and UV photometer C-635. Purifica-
tion was carried out by using reverse phase linear gradients of MeOH/
H2O 0.1% TFA in 30 min with a flow-rate of 30 mLminꢁ1, by using a
prepacked preparative cartridge (150ꢄ40 mm) with reverse phase RP18
silica gel (Bꢅchi order #54863). The fractions containing the products
were freeze-dried, and their identity confirmed by ESI-MS and NMR
spectroscopy. Compounds were isolated as TFA salts. The peptide conju-
gates GCN4–pr, GCN4–bb and GAGA–bb were purified by analytical
RP-HPLC by using an Eclipse XDBC8 (4.6ꢄ150 mm, 5 mm) analytical
column, following standard HPLC purification conditions. The sequences
of the peptides are: Asp226 to Gln248 of GCN4: DPAAL KRANT
Peptide synthesis: All peptide synthesis reagents and amino acid deriva-
tives were from GL Biochem (Shanghai) and Novabiochem; amino acids
were used as protected Fmoc amino acids with the standard side chain
protecting scheme, except for the orthogonally protected Fmoc-Lys-
ACHUTNGREN(NUG alloc)-OH and Fmoc-GluACHTUTGNREN(NUGN alloc)-OH, which were purchased from
Bachem. C-terminal amide peptides were synthesized on a 0.1 mmol
scale by using an Fmoc-PAL-PEG-PS resin from Applied Biosystems. All
solvents were dry and of synthesis grade. Peptides were synthesized by
using an automatic PS3 peptide synthesizer from Protein Technologies.
Amino acids were coupled in fourfold excess by using HBTU/HOBt as
activating agent. Each amino acid was incubated with the HBTU/HOBt
Chem. Eur. J. 2013, 19, 9923 – 9929
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9927

M. E. Vꢀzquez, J. L. MascareÇas et al.
mixture for 30 s in DMF before being added onto the resin. Peptide
bond-forming couplings were conducted for 30 to 45 min.
buffer (20 mm pH 7.5), NaCl (100 mm), aliquots of approximately 200 mm
stock solutions of the respective dsDNAs were successively added, and
the fluorescence spectrum was recorded after each addition. In the case
of the titration of GAGA–bb before adding the first aliquot of dsDNA,
5 equiv of ZnCl2 was added to make the complex.
Deprotection of the temporal Fmoc protecting group: Deprotection of the
temporal Fmoc was performed by treating the resin with 20% piperidine
in DMF solution for 10 min. N-terminal acetylation: The N-acetylation
coupling was performed immediately after Fmoc deprotection by treating
the resin with Ac2O/(DIEA/DMF 0.195m) 2:1. Allyl and Alloc removal:
The selective deprotection of the glutamic acid residues was carried out
as following: approximately 3.5 mmol (25 mg of resin) of the peptide at-
tached to the solid support were treated with PPh3 (1.5 equiv), N-methyl
morpholine (10 equiv) and phenylsilane (10 equiv) in CH2Cl2 (400 mL) at
Electrophoretic mobility shift assays: EMSAs were performed with a
BIO-RAS Mini Protean gel system, powered by an electrophoresis
power supplies PowerPac basic model, maximum power 150 V, frequency
50.60 Hz at 140 V (constant V).
RT for 15 min. Subsequently PdACTHNUTRGNE(NUG OAc)2 (0.3 equiv) was added and the re-
sulting mixture was stirred, overnight. The selective deprotection of the
lysine was carried out by treating the peptide attached to the resin with
Acknowledgements
PdACHTUNGTRENNUNG(PPh3)4 (1 equiv), morpholine (190 equiv) in H2O/CH2Cl2 2%
We thank the financial support provided by the Spanish grants Consolid-
er Ingenio 2010 (SAF2010-20822-C02, CTQ2009-14431/BQU, CTQ2012-
31341, CSD2007-00006), the Xunta de Galicia (GRC2010/12, INCITE09
209 084PR, PGIDIT08CSA-047209PR). M.I.S. thanks the Ministerio de
Educaciꢂn, Cultura y Deporte for his PhD fellowship.
(400 mL). The resin was filtered and washed with DMF (2ꢄ1.5 mL,
2 min), diethyl dithiocarbamate (DEDTC, 25 mg in 5 mL of DMF, 2ꢄ
1.5 mL, 5 min), DMF (2ꢄ1.5 mL, 2 min) and CH2Cl2 (2ꢄ5 mL, 2 min).
Cleavage/deprotection step: The resin containing the desired peptide was
washed with CH2Cl2 (2ꢄ1.5 mL, 2 min) and filtered.
A mixture of
CH2Cl2 (50 mL), H2O (25 mL), triisopropylsilane (25 mL, TIS), and TFA
(900 mL) for every 40 mg of resin was added to the resin, and the mixture
was shaken for 2.5 h. The resin was filtered and the liquid was added
over 10 volumes of cold Et2O. After 5 min the suspension was centrifuged
and the resulting white residue was washed twice with Et2O and dried
under an Ar current.
[1] C. W. Garvie, C. Wolberger, Mol. Cell 2001, 8, 937–946.
[2] a) D. S. Latchman, Eukaryotic Transcription Factors, Elsevier,
London, 2004; b) N. M. Luscombe, S. E. Austin, H. M. Berman,
J. M. Thornton, Genome Biol. 2000, 1, 1–10.
[3] a) J. E. Darnell, Nat. Rev. Cancer 2002, 2, 740–749; b) A. T. Look,
Science 1997, 278, 1059–1064; c) P. Brennan, R. Donev, S. Hewama-
na, Mol. Biosyst. 2008, 4, 909–919; d) P. K. Vogt, Oncogene 2001,
20, 2365–2377.
[4] a) P. P. Pandolfi, Oncogene 2001, 20, 3116–3127; b) R. Pollock, M.
Giel, K. Linher, T. Clackson, Nat. Biotechnol. 2002, 20, 729–733;
c) C. Denison, T. Kodadek, Chem. Biol. 1998, 5, R129–45; d) R.
Yaghmai, G. Cutting, Mol. Ther. 2002, 5, 685–694; e) A. K. Mapp,
Org. Biomol. Chem. 2003, 1, 2217–2220.
[5] a) H. C. Nelson, Curr. Opin. Genet. Dev. 1995, 5, 180–189; b) R.
Moretti, A. Z. Ansari, Biochimie 2008, 90, 1015–1025; c) D. J. Segal,
C. F. Barbas, Curr. Opin. Chem. Biol. 2000, 4, 34–39.
Synthesis of GCN4–pr: The resin-bound selectively deprotected GCN4
peptide (25 mg, aprox. 3.5 mmol) was shaken for 1 h in DMF and then fil-
tered. A mixture of HATU (3 mg, 7.90ꢄ10ꢁ3 mmol) in DMF (170 mL)
and DIEA (28 mL, of a 0.5m solution in DMF) was added to the resin,
shaken for 5 min, and the resin was washed with DMF (2ꢄ1.5 mL,
2 min). Finally,
a
mixture of functionalized propamidine
3
(ꢀ5 mg,
ꢀ7 mmol) in DIEA (56 mL, of a 0.5m solution in DMF) was added to the
resin and shaken for 2 h. The resin was washed with DMF (2ꢄ1.5 mL,
2 min) and filtered. The peptide deprotection and resin cleavage step was
carried out following standard conditions, and the resulting crude mixture
was purified by RP-HPLC obtaining the desired hybrid (3.0 mg,
1.74 mmol, overall yield considering the peptide synthesis ꢀ25%). Ana-
lytical data of the purified products: MS: MALDI-TOF [M+H+]: calcd
for C134H229N49O36: 3100.7; found: 3101.2; tR =22.0 min. Peptide GCN4–
bb was synthesized following the same procedure but using the amine 6
for the coupling reaction (see the Supporting Information for experimen-
tal details).
[6] a) N. J. Zondlo, A. Schepartz, J. Am. Chem. Soc. 1999, 121, 6938–
6939; b) R. L. Stafford, H.-D. Arndt, M. L. Brezinski, A. Z. Ansari,
P. B. Dervan, J. Am. Chem. Soc. 2007, 129, 2660–2668; c) M. E.
Vꢀzquez, A. M. CaamaÇo, J. L. MascareÇas, Chem. Soc. Rev. 2003,
32, 338–349; d) E. Pazos, J. Mosquera, M. E. Vazquez, J. L. Mascar-
eÇas, ChemBioChem 2011, 12, 1958–1973.
[7] a) O. Vꢀzquez, M. E. Vꢀzquez, J. B. Blanco-Canosa, L. Castedo,
J. L. MascareÇas, Angew. Chem. 2007, 119, 7010–7014; Angew.
Chem. Int. Ed. 2007, 46, 6886–6890; b) J. B. Blanco, V. I. Dodero,
M. E. Vꢀzquez, M. Mosquera, L. Castedo, J. L. MascareÇas, Angew.
Chem. 2006, 118, 8390–8394; Angew. Chem. Int. Ed. 2006, 45, 8210–
8214; c) J. B. Blanco, M. E. Vꢀzquez, L. Castedo, J. L. MascareÇas,
ChemBioChem 2005, 6, 2173–2176; d) J. B. Blanco, O. Vazquez, J.
Martinez-Costas, L. Castedo, J. L. MascareÇas, Chem. Eur. J. 2005,
11, 4171–4178; e) J. B. Blanco, M. E. Vꢀzquez, J. Martinez-Costas,
L. Castedo, J. L. MascareÇas, Chem. Biol. 2003, 10, 713–722;
f) M. E. Vꢀzquez, A. M. CaamaÇo, J. Martꢁnez-Costas, L. Castedo,
J. L. MascareÇas, Angew. Chem. 2001, 113, 4859–4861; Angew.
Chem. Int. Ed. 2001, 40, 4723–4725; g) M. E. Vꢀzquez, A. M. Caa-
maÇo, L. Castedo, D. Gramberg, J. L. MascareÇas, Tetrahedron Lett.
1999, 40, 3625–3628; h) M. E. Vꢀzquez, A. M. CaamaÇo, L. Caste-
do, J. L. MascareÇas, Tetrahedron Lett. 1999, 40, 3621–3624.
[8] a) M. I. Sꢀnchez, O. Vazquez, J. Martinez-Costas, M. E. Vazquez,
J. L. MascareÇas, Chem. Sci. 2012, 3, 2383–2387; b) I. Haq, Arch. Bi-
ochem. Biophys. 2002, 403, 1–15; c) C. M. Nunn, S. Neidle, J. Med.
Chem. 1995, 38, 2317–2325.
Synthesis of GAGA–bb: The resin-bound GAGA peptide (28 mg, aprox.
2.3 mmol) was suspended in DMF (1 mL), shaken for 1 h and then fil-
tered. A mixture of succinic anhydride (127 mL, 0.1m in DMF, 18 mmol,
8 equiv), DIEA (36 mL, 0.5m in DMF, 18 mmol, 8 equiv) and DMAP
(28 mL, 0.2m in DMF, 5.6 mmol, 2 equiv) was added, and the resulting
resin suspension was shaken for 1 h at RT. The resin was then filtered
and washed with DMF (3ꢄ0.6 mL). A mixture of HATU (172 mL, 52 mm
in DMF, 9.1 mmol, 4 equiv) and DIEA (36 mL, 0.5m in DMF, 18 mmol,
8 equiv) was added. The resulting mixture was shaken for 5 min and fil-
tered. A solution of the bis-benzamidene 6 (9.1 mg in 200 mL of DMF,
4 equiv) and DIEA (36 mL, 0.5m in DMF, 18 mmol, 8 equiv) was added.
The mixture was shaken for 2.5 h after which the resin was filtered and
washed with DMF (3ꢄ0.6 mL, 5 min) and Et2O (2ꢄ0.5 mL, 5 min). The
cleavage step was performed following standard conditions, and the de-
sired peptide conjugate was purified by RP-HPLC. Analytical data of
the purified products: MS: MALDI-TOF [M+H+]: calcd for
C
195H304N60O49S2: 4337.0; found: 4337.2; tR =21.4 min.
Fluorescence spectroscopy: Measurements were made with a Jobin–Yvon
Fluoromax-3, (DataMax 2.20) coupled to a wavelength electronics LFI-
3751 temperature controller, by using the following settings: increment,
1.0 nm; integration time, 0.2 s; excitation slit width, 3.0 nm; emission slit
width, 6.0 nm; excitation wavelength, 329 nm. The emission spectra were
acquired from 345 to 500 nm at 208C. All titrations were made following
the same procedure: to 1 mL of either compound (50 nm in the case of 6,
250 nm in the case of GCN4–bb or 500 nm for GAGA–bb) in Tris-HCl
[9] a) O. Vꢀzquez, M. I. Sꢀnchez, J. L. MascareÇas, M. E. Vꢀzquez,
Chem. Commun. 2010, 46, 5518–5520; b) O. Vꢀzquez, M. I.
Sꢀnchez, J. Martꢁnez-Costas, M. E. Vꢀzquez, J. L. MascareÇas, Org.
Lett. 2010, 12, 216–219; c) M. I. Sꢀnchez, O. Vꢀzquez, M. E.
Vꢀzquez, J. L. MascareÇas, Chem. Commun. 2011, 47, 11107–11109;
d) J. Bordello, M. I. Sꢀnchez, M. E. Vꢀzquez, J. L. MascareÇas, W.
9928
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 9923 – 9929

DNA Recognition with Peptide–Bisbenzamidine Conjugates
FULL PAPER
Al-Soufi, M. Novo, Angew. Chem. 2012, 124, 7659–7662; Angew.
Chem. Int. Ed. 2012, 51, 7541–7544.
[14] K. J. Edwards, T. C. Jenkins, S. Neidle, Biochemistry 1992, 31, 7104–
7109; PDB ID: 1D64.
[10] a) R. V. Talanian, C. J. McKnight, R. Rutkowski, P. S. Kim, Bio-
chemistry 1992, 31, 6871–6875; b) R. V. Talanian, C. J. McKnight,
P. S. Kim, Science 1990, 249, 769–771.
[11] a) L. M. Hellman, M. G. Fried, Nature Protocols 2007, 2, 1849–1861;
b) D. Lane, P. Prentki, M. Chandler, Microbiol. Rev. 1992, 56, 509–
528.
[12] Residue numbering taken from the corresponding PDB structures.
[13] a) P. Kçnig, T. J. Richmond, J. Mol. Biol. 1993, 233, 139–154;
b) T. E. Ellenberger, C. J. Brandl, K. Struhl, S. C. Harrison, Cell
1992, 71, 1223–1237; c) J. G. Omichinski, P. V. Pedone, G. Felsen-
feld, A. M. Gronenborn, G. M. Clore, Nat. Struct. Mol. Biol. 1997, 4,
122–132; PDB IDs: 1YSA and 1YUI for the GCN4–DNA and
GAGA–DNA complexes, respectively.
[15] a) J. J. Hollenbeck, M. G. Oakley, Biochemistry 2000, 39, 6380–6389;
b) C. Portela, F. Albericio, R. Eritja, L. Castedo, J. L. MascareÇas,
ChemBioChem 2007, 8, 1110–1114; c) A. V. Fedorova, I.-S. Chan,
J. A. Shin, Biochim. Biophys. Acta 2006, 1764, 1252–1259.
[16] The relatively low emission intensity of the benzamidine fluoro-
phore precludes accurate titrations at ideal low nm concentration,
and therefore the reported Kd value should be taken as approxima-
tion: a) M. R. Eftink, Methods Enzymol. 1997, 278, 221–257; b) V. J.
LiCata, A. J. Wowor, Methods Cell Biol. 2008, 84, 243–262.
Received: February 8, 2013
Published online: June 18, 2013
Chem. Eur. J. 2013, 19, 9923 – 9929
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9929