Synthesis of fluorescent microgonotropens (FMGTs) and their interactions with dsDNA

Article abstract of DOI:10.1016/S0968-0896(00)00116-4

A new class of microgonotropen compounds (FMGTs), which fluoresce upon binding to dsDNA, is introduced. The FMGTs consist of a minor groove binding moiety based upon Hoescht 33258 covalently attached to a polyamine chain capable of interacting with the phosphodiester backbone of dsDNA. The interactions of FMGTs with dsDNA were investigated by fluorescence and UV spectroscopy. Several different dsDNA oligomers were studied to determine the effect of binding site sequence on stoichiometry and binding affinity. The FMGTs were found to bind a dsDNA oligomer that contained the sequence 5'-AATTT-3' with FMGT:dsDNA stoichiometries equal to 2:1 or 3:1. Hoechst 33258 bound the same dsDNA oligomer with a 1:1 stoichiometry. The second and third order equilibrium constants for complexation were determined to be Log(K₁K₂)=17.9M^-2 and Log(K₁K₂K₃)=26.1M^-3, respectively, for two of strongest binding FMGTs. From thermal melting experiments ΔT(m) for Hoechst 33258 was determined to be 10°C while the ΔT(m) values for FMGTs ranged from 20-26°C indicating the greater stability of the latter. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00116-4

Bioorganic & Medicinal Chemistry 8 (2000) 1871±1880
Synthesis of Fluorescent Microgonotropens (FMGTs) and Their
Interactions with dsDNA
Alexander L. Satz and Thomas C. Bruice*
Department of Chemistry, University of California at Santa Barbara, Santa Barbara, CA 93106, USA
Received 21 January 2000; accepted 29 March 2000
AbstractÐA new class of microgonotropen compounds (FMGTs), which ¯uoresce upon binding to dsDNA, is introduced. The
FMGTs consist of a minor groove binding moiety based upon Hoescht 33258 covalently attached to a polyamine chain capable of
interacting with the phosphodiester backbone of dsDNA. The interactions of FMGTs with dsDNA were investigated by ¯uores-
cence and UV spectroscopy. Several di€erent dsDNA oligomers were studied to determine the e€ect of binding site sequence on
stoichiometry and binding anity. The FMGTs were found to bind a dsDNA oligomer that contained the sequence 50-AATTT-30
with FMGT:dsDNA stoichiometries equal to 2:1 or 3:1. Hoechst 33258 bound the same dsDNA oligomer with a 1:1 stoichiometry.
2
The second and third order equilibrium constants for complexation were determined to be Log(K1K2)=17.9 M and Log
(K1K2K3)=26.1 M 3, respectively, for two of strongest binding FMGTs. From thermal melting experiments ÁTm for Hoechst
33258 was determined to be 10 ꢀC while the ÁTm values for FMGTs ranged from 20±26 ꢀC indicating the greater stability of the
latter. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
MGTs were shown to be extraordinarily e€ective inhibi-
tors of the association of the transcription factor E2 factor
1 (E2F1), to its DNA promoter element.15 The most
active MGT, MGT-6a, was three orders of magnitude
more e€ective than distamycin.
Reagents capable of sequence selective recognition of
DNA, particularly small organic compounds capable of
binding to the minor groove of B-DNA, have drawn tre-
mendous attention due to their known biological activity
and diverse medicinal uses.1 12 Recent in vitro studies
indicate that these agents may in¯uence the regulation of
gene expression since they have been shown to inhibit
the binding of regulatory proteins to their DNA con-
sensus binding sites.13 21 These same studies indicate
that the inhibitory activity of often studied minor groove
binders as distamycin, netropsin, Hoechst 33258, etc.,
are rather limited to proteins which reside within the
minor groove. Detrimental to the potential ability of the
above mentioned minor groove binders to speci®cally
inhibit protein binding is their lack of interaction with
DNA outside the minor groove coupled with minimal
distortion of the DNA helix upon complexation.22 24
We now report studies of a minor groove binding moiety
based upon the widely used DNA ¯uorophore Hoechst
33258, known to bind A/T rich sequences of dsDNA.25 32
Hoechst 33258's increase in ¯uorescence emission
intensity upon binding dsDNA allows for expeditious
determination of ligand:dsDNA binding stoichiometry
and equilibrium constants for complexation.26,32,33
Results and Discussion
Synthesis
FMGTs were sythesized with a meta linker arm to direct
the polyamine moiety out of the minor groove (Scheme
1).8,9,34 The choice of 4-hydroxybutyrate as the linker
arm for the FMGTs was based upon its relative ¯ex-
ibility, length, and synthetic accessibility. Termination of
the linker arm with a carboxylic acid allows for coupling
to a polyamine chain using solution or solid-phase
chemistry. Polyamine chains were chosen based upon
synthetic accessibility and the results previously shown
by the tripyrrole peptide MGTs.35
We have investigated a novel class of minor groove
binding ligands, microgonotropens (MGTs), which
consist of an A/T-selective DNA minor groove binding
tripyrrole peptide and polyamine chains attached to the
central pyrrole that interact electrostatically with the
phosphodiester backbone of dsDNA (Fig. 1).13,16,25,26
*Corresponding author. Tel.: +1-805-961-2044; fax: +1-805-893-
4120; e-mail: tcbruice@bioorganic.ucsb.edu
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00116-4

1872
A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
Figure 1. The structures of FMGT-1, -2, -3, and -5, Ht-Ester, tripyrrole peptide MGT-7, and -6a, Hoechst 33258, and distamycin.
Mitsunobu36 type reaction of 3-hydroxybenzaldehyde (1)
with 4-hydroxybutyrate37 gave 2. Condensation of the
appropriate ortho diamine with 2 gave 3.38 Deprotection of
the carboxylic acid derivative by base hydrolysis yielded 4.
PyBOP139 was used for all peptide coupling reactions.
Synthesis of FMGTs-3 and -5 were accomplished using
standard solid-phase FMOC techniques. PyBOP1 cou-
pling reactions of 4 to the terminal amine of the poly-
arginine chain (3 or 5 peptides long) where achieved in
DMF using a 2-fold excess of 4.
most stable dsDNA complex with a ÁTm of 26ꢀC,
roughly 2.5 times larger than that for Hoechst 33258.
Experiments using the control sequence 20 indicate sub-
stantial duplex stabilization by the FMGTs although
ÁTm values for FMGT:21 complexes are still signi®cantly
greater. The ÁTm values for the FMGT:20 complexes
ranged from 11 to 19 ꢀC.
Measurement of ¯uorescence intensity upon titration of
an appropriately large concentration of dsDNA with
Hoechst 33258 provides an essentially linear relationship
between emission intensity and Hoechst 33258 con-
centration up to the point of binding site saturation. The
molar ratio (stoichiometry) of ligand to dsDNA is
determined by measuring the point of intersection of the
two slopes of the titration (Fig. 3). The stoichiometries
for Hoechst 33258:21 and FMGT:21 complexes were
determined to be 1:1 for Hoechst 33258, 2:1 for FMGT-
1, and -3, and 3:1 for FMGT-2, and -5. Attempts to
determine the stoichiometries of the Hoechst 33258 and
FMGT-1:20 complexes were made due to the 13 ꢀC
ÁTm value of the latter, however the ligand:20 com-
plexes were found to not be ¯uorescent. Thermal melt-
ing experiments and ¯uorescence spectroscopy provide
little insight into the nature of the FMGT:20 complexes.
Some possibilities include DNA intercalation and/or
minor groove binding with altered sequence selectivity.
Both of these alternative types of DNA binding have
Thermal melting experiments and ¯uorescence
spectroscopy
Table 1 lists the thermal melting temperatures of dsDNAs
20 and 21 when complexed with FMGTs (see Table 1 for
dsDNA sequences and Fig. 2 for thermal melting curves).
The dsDNA sequence 21 is a shortened variation of the
27 base pair derivative of the hamster dihydrofolate
reductase promoter which was the dsDNA sequence
used to determine the E2F1 inhibitory activity of tri-
pyrrole peptide MGTs.15 The dsDNA 20 is a derivative
of 21 that lacks an A/T rich sequence and was used as
the control in this study.
Thermal melting experiments gave a ÁTm for Hoechst
33258:21 of 10 ꢀC while ÁTm values for the FMGT:21
complexes ranged from 20±26 ꢀC. FMGT-2 formed the

A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
1873
Scheme 1. (a) 4-Hydroxybutyrate, (Ph)3P, DEAD, DCM; (b) 2-(3,4-Diaminophenyl)-6-(4-methyl-1-piperazinyl)benzimidazole, nitrobenzene,
120 ꢀC; (c) K2CO3, DMF, H2O, 90 ꢀC; (d) PyBOP1, tris(2-aminoethyl)amine, DMF; (e) PyBOP1, spermine, DMF; (f) PyBOP1, 4, DMF; (g) 95%
TFA, 2.5% triisopropylsilane, 2.5% water. Pbf=2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl.
Table 1. Thermal melting temperatures (ꢀC) for dsDNA 20 and 21a,b
20
21
No ligand
Hoechst 33258
FMGT-1
FMGT-2
FMGT-3
FMGT-5
57
60
70
76
68
68
55
65
77
81
76
77
aMelting points were determined by ®rst derivative analysis.
bThe dsDNA oligomers d(GCGACTGCAATTTCGACGTCC) and
d(CGACGACTGCAACGACGTCC) and are referred to as 21 and
20, respectively.
Figure 2. Thermal melting curves for dsDNA:ligand complexes.
Absorbance measurements were taken every 1ꢀC (for clarity only one
in ®ve data points are represented by a symbol). 0.80 mM dsDNA 21,
2.8 mM ligand when present.
been shown to occur with either Hoechst 33258 or related
analogues.40 43
Table 2 provides the relative molar ¯uorescence emis-
sion intensities of Hoechst 33258 and the FMGTs as
measured at the wavelength of maximum signal intensity,
and that complexation within the minor groove of
dsDNA limits water accessibility and a corresponding
increase in quantum yield results. As mentioned by a
reviewer, it is tempting to argue that the quantum yield
of ¯uorescence correlates with tighter or `deeper' bind-
ing (this idea has been previously raised by Robles and
McLaughlin, ref 44) and that the lower quantum yield
for the FMGTs results from weaker or `less deep' bind-
ing within the minor groove. Of course, this conclusion
can only be valid if Hoechst 33258 and the FMGTs
have equal quantum yields underneath the same condi-
tions. By taking the ¯uorescence spectra of Hoechst
33258 and FMGT-1 in methanol the two molecules
were shown to not be spectroscopically identical. The
solvent methanol was chosen because unlike in water
lmax, when excited at 354 nm (¯uorescence intensities for
the unbound ¯uorophores do not increase linearly with
increasing concentrations, especially at higher, mM, con-
centrations). The lmax was found to be approximately
480 nm for all unbound ligands and 470 nm for all
ligands when bound to dsDNA except Hoechst 33258
whose lmax blue shifts roughly 30 to 450 nm when bound.
Compared to Hoechst 33258 the FMGTs show a
weaker ¯uorescence signal and less of a blue shift in
lmax upon forming a ligand:21 complex. It is believed
that ¯uorophores such as Hoechst 33258 ¯uorescence
weakly in aqueous solution due to quenching by water

1874
A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
Table 2. Fluorescence emission of ligand and ligand: 21 complexes
F(intensity)
unbound
liganda
F(intensity)
ligand:21x
complexesb
Hoechst 33258
FMGT-1
FMGT-2
FMGT-3
FMGT-5
Ht-Ester
Hoechst 33258 in methanol
FMGT-1 in methanol
96
10
80
51
36
27
6200
2400
2300
4000
2600
3500
7200 (lmax=466 nM)
520 (lmax=480 nM)
1
aFluorescence (arbitrary units), adjusted to mM ligand, determined
at lmax as given in the text. All measurement made in bu€ered, pH 7.2,
aqueous solution except for the last two entries in which methanol was
used as the solvent.
1
bFluorescence (arbitrary units), adjusted to mM complex dsDNA:
ligandx, where x describes the stoichiometry of binding as given in the
text.
is calculated by eq (4). eq (4) is derived from eq (3). For
2:1 or 3:1 binding, the mathematical solution for ‰LŠ is
f
more complex than the quadratic solution given in eq
(2). Plots of ¯uorescence versus ‰LŠ are used to calculate
f
the equilibrium constants by curve ®tting with eqs (5) or
(6) for 2:1 or 3:1, respectively. The derivation and use of
eq (5) has been discussed by our laboratory.26
!
X
K ‰LŠ
f
F ˆ
Èf
ꢀ1†
1 ‡ K ‰LŠ
f
where ÆÈf is the total ¯uorescence intensity upon
saturation of dsDNA binding sites with ligand, K is the
equilibrium constant, and ‰LŠ is the concentration of
f
ligand free in solution.
!
r
2
4‰LŠ
T
‰LŠ ˆ 0:5 b ‡ b ‡
ꢀ2†
f
K
where b ˆ ‰DNAŠ
‰LŠ ‡ 1=K and ‰DNAŠ is the
Figure 3. Representative ¯uorescent titrations of a constant concentra-
tion of dsDNA, 50 mM, with ligand for determination of ligand:21
stoichiometries. Samples were excited at 354 nM and their ¯uorescence
emission scanned between 400 and 500 nM. Plots shown are of signal
intensity subtracted by background intensity. (a) Titration with Hoechst
33258. (b) Titration with FMGT-1. (c) Titration with FMGT-2.
T
total concentration of dsDNA in the sample.
T
T
X
‰LŠ
Bound
F ˆ
Èf
ꢀ3†
n‰DNAŠ
T
the ¯uorophores give a strong signal. The results in
Table 2 show that the structural di€erences between
Hoechst 33258 and FMGT-1 signi®cantly a€ect their
where ‰LŠ is the concentration of ligand bound to
Bound
dsDNA, and n is the stoichiometry of binding.
¯uorescence spectra in both quantum yield and lmax
.
n‰DNAŠTF
P
‰LŠ ˆ ‰LŠ
ꢀ4†
ꢀ5†
f
T
Èf
Determination of binding constants33
!
For cases in which binding stoichiometries were 1:1,
association constants were calculated by curve ®tting
analysis of the isothermal binding curves using either eqs
(1) and (2) or eqs (1) and (4). Equations (1) and (2) can be
used to ®t the plots of ¯uorescence versus total con-
centration of ligand (‰LŠT) added to a constant con-
centration of dsDNA. Alternatively, the expression of
eq. (1) can be employed to ®t plots of ¯uorescence ver-
2
f
X
X
0:5K1‰LŠ ‡K1K2‰LŠ
f
F ˆ
Èf
2
f
1 ‡ K1‰LŠ ‡K1K2‰LŠ
f
!
2
f
3
f
1=3K1‰LŠ ‡2=3K1K2‰LŠ ‡K1K2K3‰LŠ
f
F ˆ
Èf
2
f
3
f
1 ‡ K1‰LŠ ‡K1K2‰LŠ ‡K1K2K3‰lŠ
f
ꢀ6†
sus concentration of unbound ligand (‰LŠf). Where ‰LŠ
f

A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
1875
Table 3 lists the calculated equilibrium constants for
complexation for FMGTs with 21 as determined by
curve ®tting using the appropriate theoretical model.
The combined equilibrium constants LogꢁK1K2† for 2:1
binding, or LogꢁK1K2K3† for 3:1 binding, are reported
since separation of the individual equilibrium constants
was not possible. Figure 3 shows representative iso-
thermal binding curves and their ®ts. The combined
equilibrium constants for complexation of 21 by FMGT-
1 ꢁK1K2† is roughly an order of magnitude larger than
FMGT-3's and two orders of magnitude larger than
Hoechst Ester's. The binding curves for FMGT-3 and
Hoechst Ester show an `S' shape by visual inspection, an
indication of cooperative binding presumably as a
dimer. FMGT-1 also shows an `S' shaped binding curve
although less pronounced than for FMGT-3 and
Hoechst Ester. The cooperative binding of Hoechst Ester
shows that the polyamine chain is not the cause of the
higher order binding. The multiplied through equili-
brium constants for complexation of 21 by FMGT-2
ꢁK1K2K3† is two orders of magnitude larger than that of
FMGT-5's. As for the 2:1 binders, FMGT-2 and -5 both
have an `S' shaped binding curve, although it is more
noticeable for FMGT-2 (Fig. 4b). In the determination
of equilibrium constants of complexation it was assumed
that complexation of 21 outside of the A/T rich site had
no signi®cant e€ect on the observed ¯uorescent signal
since ligand:20 complexes do not ¯uoresce and their
complexes are less stable than those with 21.
2 whose complexes with 21 had stoichiometries of 1:1,
2:1, and 3:1, respectively.
The equilibrium constant for Hoechst 33258 changes
little between 21 and 20b. The K1K2 term for FMGT-1
is an order of magnitude less for 20b than for 21. The
K1K2K3 term for FMGT-2 decreases by two orders of
magnitude when complexed with 20b. The decreases in
the combined equilibrium constants for complexation of
the FMGTs is dicult to explain without assuming that
complexation occurs solely within the A/T rich sequence.
It is likely that higher order binding by FMGTs occur
solely within the A/T rich sequence of 21 in a coopera-
tive manner as a dimer or trimer. The e€ect of binding
site size on equilibrium constant for complexation was
investigated using dsDNA 20b d(GGACGTCGAATT-
GCAGTCGC) which contains a binding site of four
contiguous base pairs compared to ®ve for 21. It was
assumed that monomeric, dimeric, and trimeric binding
modes would be a€ected di€erently by a reduction in
size of the A/T rich sequence. The oligomer 20b was
titrated with either Hoechst 33258, FMGT-1, or FMGT-
Table 3. Combined equilibrium constants for complexation between
FMGTs and dsDNA oligomers 20b and 21a
Ligand
Equilibrium constants
for complexation with 20bb
Equilibrium constants
for complexation with 21c
1
1
2
3
2
3
Ht33258
FMGT-1
FMGT-2 Log(K1K2K3)=23.7 M
FMGT-3
FMGT-5
Ht-Ester
Log(K1)=9.2 M
Log(K1)=9.6Æ0.3 M
Log(K1K2)=17.9Æ0.1 M
Log(K1K2K3)=26.1Æ0.4 M
Log(K1K2)=17.0Æ0.1 M
Log(K1K2K3)=23.6Æ0.4 M
2
Log(K1K2)=16.6 M
3
2 b
Log(K1K2)=15.8 M
aDetermined by nonlinear least squares ®tting of an isothermal bind-
ing curve with the appropriate theoretical model. Equilibrium con-
stants are given as Log(K1), Log(K1K2), or Log(K1K2K3) for 1:1, 2:1,
or 3:1 binding stoichiometries respectively since separation of the
individual equilibrium constants was not possible. 21=d(GCGAC-
TGCAATTTCGACGTCC) and 20b=d(GGACGTCGAATTGCA-
GTCGC).
Figure 4. Isothermal binding curves generated by¯uorescent titrations of
a constant concentration of dsDNA with ligand. Samples, 1±5 nM 21,
were titrated with concentrated (mM) ligand solutions. Samples were
excited at 354 nm and their ¯uorescence emission monitored at the
appropriate wavelength. Fluorescence intensities have been subtracted by
the background value. The x-axis gives the concentration of unbound
bThe combined equilibrium constants were determined from a single
isothermal titration.
ligand ‰LŠ as calculated by eq (4) (see text). Data points are then ®t using
f
cCombined equilibrium constants were calculated as the average of at
least three trials. Error values were derived from the standard devia-
tions between trials.
the theoretical model described by eqs (1), (5), or (6) and a nonlinear least
squares curve ®tting routine. (a) 1 nM 21 titrated with Hoechst 33258. (b)
2 nM 21 titrated with FMGT-2. (c) 1 nM 21 titrated with FMGT-3.

1876
A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
Then it can be concluded that the monomer, dimer, and
trimer are successsively more sensitive to binding site
size. The 3:1 binding reported here, if it is occurring
within a single binding site, is certainly not usual. To this
time the only report of greater than a 2:1 stoichiometry
for any minor groove binder has been by Blasko and
Bruice45 who reported the binding of distamycin to
dsDNA in a 4:1 stoichiometry.
been found to bind dsDNA in greater than 1:1 stoichio-
metries and this knowledge has been used to generate
numerous molecules capable of altered sequence
selectivity.4,48 50 A major advance in attempts to obtain
active G/C recognition by minor groove binders was the
®nding that distamycin could form stacked dimers capable
of binding to the wider G/C containing minor grooves.22,51
The possibility that a slight change in Hoechst 33258
leads to higher order binding allows for the incorpora-
tion of bisbenzimidazole type moieties in analogous side-
by-side binding motifs. In addition to the bene®t of ¯uo-
rescence emission, bisbenzimidazole containing minor
groove binders may exhibit better or at least di€erent
sequence selectivity compared to their pyrrole peptide
counterparts.
The dsDNA 12 d(CCGGAATTCCGG) also contains an
A/T rich binding site of 4 base pairs. The equilibrium
constants for complexation of Hoechst 33258 with
dsDNA 12 has been previously published and is within an
order of magnitude of the value obtained in this labora-
tory of Log(K1)=9.1 M .
1 46 This value di€ers little from
the values obtained for the oligomers 20b or 21 as given
in Table 3. Surprisingly, the stoichiometry for the FMGT-
1
The combined equilibrium constant for complexation of
two FMGTs can be compared to the equilibrium con-
stant for complexation of one Hoechst 33258 by the
same dsDNA by use of eq (7) (derived from ligand:
dsDNA equilibrium equations).
1:12 complex was found to be 1:1 with Log(K1)=9.7 M
even though its binding site is identical to 20b. FMGT-1
was found to be the only FMGT which bound 12 with a
stoichiometry di€erent from 21. The equilibrium con-
stants for complexation between 12 and FMGT-2 and -
3
3 were determined to be Log(K1K2K3)=24.8 M and
Â
Ã
‰FMGT 1Š2K1K2
‰Ht33258ŠK
Log(K1K2)=16.2 M 2, respectively. Analogous to this
result, it has been reported26 that Hoechst 33258 binds the
hexadecamer 16, which contains the sequence 50-AAA
TTT-30 in a 2:1 stoichiometry (this is the only reported
instance of Hoechst 33258 binding dsDNA with a stoi-
chiometry greater than 1:1). However, Hoechst 33258
binds to the dodecamer 12b (CGCAAATTTGCG) with a
stoichiometry of 1:1, even though the binding site is
identical to that of 16.27,29,45,47 It is postulated that the
di€erences in base pairs outside the ligand binding site
a€ect the binding stoichiometry.45 This may also explain
the di€erences in binding stoichiometries for FMGT-1
with 20 and 12.
DNA : ꢀFMGT 1†
‰DNA : Ht33258Š
2
ˆ
ꢀ7†
At 1 mM unbound ligand concentrations the ratio of
[DNA:FMGT-1]/[DNA:Ht33258] would be ꢂ200:1. At
nM concentrations the ratio is ꢂ1:1. Replacing [Ht
33258] with [Protein] estimates the ability of FMGTs to
inhibit the binding of a protein with a known equili-
brium constant (the inhibition of major groove binding
proteins may be more complex than shown by eq (7)).
Equation (7) predicts that the FMGTs would be more
e€ective inhibitors of protein binding than Hoechst
33258 at concentrations greater than a nM. Studies
measuring the ability of FMGTs to inhibit the binding
of transcription factors are currently underway.
Conclusions
The thermal melting experiments done in this study
clearly demonstrate the increased stability of FMGT:
dsDNA complexes relative to Hoechst 33258. Less clear
is the nature of FMGT binding. It appears likely how-
ever that FMGTs bind the A/T rich sequence of 21 in a
cooperative manner leading to highly ¯uorescent com-
plexes. Complexation of 21 outside of the A/T rich
sequence likely occurs following saturation of the A/T
rich site and leads to non-¯uorescent complexes of an
unknown motif. Presumably, any higher order complex
formed within the A/T rich site occurs through a side-
by-side binding motif since there is no other way to ®t
two or three minor groove binders into a ®ve base pair
binding site. Why the parent compound Hoechst 33258
does not also form the higher order complexes is not
known although it would have to be assumed that the
phenolic hydroxyl plays an important role. De®nitive
proof of side-by-side binding would require structural
investigations which are outside the scope of this study.
Experimental
Materials
DNA binding studies. Puri®ed DNA oligomers were pur-
chased from the Biomolecular Resource Center, Uni-
versity of California at San Francisco. Hoechst 33258 and
0.05 wt.% 3-(trimethylsilyl)-propionic-2,2,3,3-d4 acid,
sodium salt in D2O were bought from Aldrich and used
without further puri®cation. Doubly distilled water was
used for all solutions.
Organic synthesis. Anhydrous solvents, other solvents,
and most reagents were bought from Aldrich Chemical
Company and used without further puri®cation. Deut-
erated solvents were bought from either Aldrich Che-
mical Company or Cambridge Isotope Laboratories.
Some reagents used for solid-phase chemistry including
Rink amide MBHA resin, FMOC-Arg(Pbf)-OH, and
PyBOP1 were bought from Novabiochem and used
without further puri®cation. 4-Hydroxybutyrate37 and 2-
The possibility of side-by-side binding is very interesting
because of the motifs importance in sequence selective
dsDNA recognition. Distamycin and its analogues have

A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
1877
(3-nitro-4-aminophenyl)-6-(4-methyl-1-piperazinyl)-benzi-
midazole8 were prepared according to the published
methods.
in 0.01 M potassium phosphate, 0.01 M NaCl, pH 6.9
or 7.2 bu€er unless explicitly stated otherwise. Solutions
were mixed by turning the cuvette upside down several
times and then allowing the solution to equilibrate for
one minute. The solutions were excited at 354 nm. Fluo-
rescence titrations were obtained by titrating a constant
concentration of dsDNA, between 1 and 50 nm, with a
relatively concentrated solution of ligand, between 0.2
and 40 mM. Emission spectra scanning the region from
400 to 500 nm were taken for each data point. The
background ¯uorescence intensity (bu€er and dsDNA)
was always subtracted from each data point.
Procedures (ligand:dsDNA experiments)
UV±vis spectroscopy. UV±vis spectra and thermal
denaturation experiments were obtained on a Cary 100
Bio UV±vis spectrophotometer equipped with a tem-
perature programmable cell block. All solutions were
®ltered with a 0.2 mm ®lter. Constant temperature UV±
vis spectra were acquired at 25 ꢀC using a 1 cm path
length quartz cuvette and were taken in 0.2% aqueous
glacial acetic acid. Molar extinction coecients were
calculated by a single addition of a known quantity of
compound (determined by NMR spectroscopy with an
internal standard). Due to the previously reported obser-
vation that Hoechst 33258 does not obey the Beer±Lam-
bert equation, the concentrations at which each molar
extinction coecient was determined is provided.30
Procedures (organic synthesis)
1
General. H and 13C NMR spectra were obtained on a
Varian Unity Inova 400 or 500 spectrometer at 400 MHz
and 100 or 150 MHz respectively. TLC was carried out
on silica gel (KIESELGER 60 F254) glass backed com-
mercial plates and visualized by UV light. Fast atom
bombardment mass spectra, HRMS and LRMS, were
obtained on a VG analytical, VG-70E Double Focusing
Mass Spectrometer, with an Ion Tech Xenon Gun FAB
source, and an OPUS/SIOS data interface and acquisi-
tion system. Electrospray ionization mass spectra (ESI)
were obtained on a VG Fisons, VG Platform 2, using an
ESI source, and a Mass Lynx data acquisition system.
Preparative HPLCs were acquired using a Hewlett Pack-
ard series 1050 HPLC equipped with a diode array detec-
tor. For preparative separations an Alltech Macrosphere
300A, C8, silica, 7 micron, 250Â10 mm reverse-phase
column was used at 3.0 mL/min solvent ¯ow. Signals
were monitored at 260 and 343 nm concurrently. For
analytical separations an Alltech Macrosphere 300 RP,
C18, 7 micron, 250Â4.6 mm column was used at 1.3 mL/
min solvent ¯ow in all cases except for FMGT-7 were
an Alltech Macrosphere C8, 300A, 7 micron, 250Â
4.6 mm column was used.
Thermal melting experiments were obtained by observing
the absorbance of samples at 260 nm while changing the
block temperature of the spectrophotometer from 15 to
95 ꢀC in increments of 0.5 ꢀC/min. Sample absorption at
260 nm was between 0.7 and 0.8 at 24 ꢀC in all cases.
The melting points of ligand:dsDNA complexes were
determined by addition of 3.5 equiv of ligand to a
dsDNA sample. All thermal melting experiments were
done using 0.01 M potassium phosphate, 0.01 M NaCl,
pH 6.9 or 7.2 bu€er. All aqueous solutions were mixed
with Chelex to remove any trace metal impurities and
then ®ltered. DNA was annealed with its com-
plementary sequence by heating in bu€er to 95±99 ꢀC
for about 10 min and slowly allowing the solution to
cool to room temperature over several hours. Molar
extinction coecients for dsDNA were approximated
1
1
using A260= 16800 M (G/C basepair) and A260
1
=
13600 M (A/T base pair) 1; 12 (A260=1.9Â105), 20
(A260=3.1Â105), 20b (A260=3.1Â105), and 21 (A260
=
3.2Â105). Melting points were determined by ®rst
3 - (4 - Hydroxybutyrate) - benzaldehyde (2). 3-Hydroxy-
benzaldehyde (1 g, 8.2 mmol), 4-hydroxybutyrate (1.6 g,
8.2 mmol), and triphenyl phosphine (3.4 g, 13.1 mmol)
were dissolved in 35 mL anhydrous dichloromethane
(DCM) and cooled in an ice bath under an argon atmo-
sphere with stirring. Diethylazodicarboxylate (DEAD)
(2.3 g, 13.1 mmol) was slowly added to the reaction ves-
sel by syringe. Reaction progress was monitored by
TLC (100% DCM, silica) and stirred for 18 h at room
temperature. Solvent was removed by evaporation
under reduced pressure and the crude product mixture
was puri®ed by ¯ash column chromatography (100%
DCM, silica). (942 mg, clear liquid, yield 39%). Rf 0.48
(silica, 100% DCM). 1H NMR (CDCl3) d 2.17 (m, 2H,-
CH2-), 2.602 (t, 2H, J=7.2 Hz,-CH2-C(ˆO)-), 4.07 (t,
2H, J=6.1 Hz, Ar-O-CH2-), 5.15 (s, 2H, Ar-CH2-), 7.12±
7.16 (ArH), 7.34±7.36 (ArH), 7.42±7.46 (ArH), 9.970 (s,
1H,-C(ˆO)-H) ppm; 13C NMR (CDCl3) d 19.71 (-C-C-C-
), 25.95 (-C-C(ˆO)-), 61.60 (Ar-O-C-), eleven aromatic
signals are detected, 108.02+117.077+118.72+123.49+
123.5+123.80+125.27+131.07+132.99+154.57+ 168.10,
187.34 (-C(ˆO)-H) ppm. IR data (n(cm 1)): nAromatic C-
H= 3089+3034, nAliphatic C-H=2928+2847, nAldehydic C-
H=2730, nC=0=1735+1698. LRMS (EI): 192, 177, 164,
derivative analysis.
Fluorescence spectroscopy. Quantitative determination
of the concentrations of the FMGT and Hoechst 33258
solutions were achieved using NMR spectroscopy (Var-
ian Unity Inova 400). An exact volume (usually 100 mL)
of 0.05 wt.% 3-(trimethylsilyl)-propionic-2,2,3,3-d4 acid,
sodium salt was added to puri®ed compound dissolved
in an exact amount of either D2O or D3COD or a mix-
ture of the two solvents. Integration of relative peak
intensities of the FMGT and the standard allowed for
quantitative determination of both [FMGT] and total
number of moles of compound in the NMR sample.
Fluorescence spectra were obtained on a Perkin±Elmer
LS50B ¯uorimeter equipped with a constant tempera-
ture water bath. All measurements were taken at 24 ꢀC
using matched quartz cuvettes. All bu€ers and solutions
used were ®ltered with a 0.2 mm ®lter. For all measure-
ments except for those involving 12, aqueous solutions
were mixed with Chelex to remove any trace metal
impurities and then ®ltered after readjusting the pH
with dilute aqueous HCl. All measurements were taken

1878
A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
105, 91m/z. HRMS (FAB): 299.12829 (calcd 299.128334)
(M+H)+1 m/z.
removed by evaporation under vacuum to give product
plus trace amounts of starting material. 10 mL of 5%
aqueous ammonium hydroxide was added to the pro-
duct mixture which resulted in a colloidal solution.
Washing the solution with ethyl acetate resulted in a
clear aqueous layer. Drop-wise addition of concentrated
HCl to the aqueous layer resulted in precipitation of
product upon neutralization of the solution. Solid pro-
duct was then collected and reconstituted in 0.01 M
triethylamine acetate, pH 7 with 30% acetonitrile. The
product precipitated out after 24 h at 4 ꢀC as the acetate
salt (420 mg, 86%, yellow solid) Rf 0.52 (100% metha-
meta-(4-Hydroxybutyrate)-Hoechst, acetate salt (Ht-Ester).
2-(3-Nitro-4-aminophenyl)-6-(4-methyl-1-piperazinyl)-
benzimidazole (154 mg, 0.48 mmol) was hydrogenated
according to literature procedure to form the reactive
diortho amine 2-(3,4-diaminophenyl)-6-(4-methyl-1-piper-
azinyl)benzimidazole. The diortho amine was dried
under vacuum for several hours and then without pur-
i®cation added to 10mL nitrobenzene. To this solution
was added 3-(4-hydroxybutyrate)-benzaldehyde (143mg,
0.48mmol). The solution was stirred at 140 ꢀC under
argon for 24 h.38 The reaction progress was monitored
by TLC (silica, with a gradual addition or triethyl amine
and methanol in DCM). When TLC showed the dis-
appearance of most of the 3-(4-hydroxybutyrate)-ben-
zaldehyde, the product was precipitated out with
hexanes giving a brownish solid. Successive washes with
hexanes removed most residual nitrobenzene. The crude
product was then puri®ed by ¯ash chromatrography
(silica, ethyl acetate:methanol:triethyamine, 16:8:1). The
solid product was collected and then redissolved in
methanol and precipitated out of solution with ether.
The product was redissolved in dilute aqueous acetic
acid and lyopholized. (117mg, 41%, yellow-brown solid).
Pure by analytical HPLC analysis. Rf 0.45 (ethylacetate:-
1
nol). H NMR (DMSO-d6+CD3OD) d 1.905 (s, ace-
tate), 2.01 (m, 2H,-CH2-), 2.25 (s, 3H, CH3-N), 2.42 (t,
J=7.3 Hz, 2H,-CH2-COR), 3.1±3.2 (broad triplet, 4H,
(-CH2)2NMe, piperazine), 3.8±4.2 (obscured by H-O-D
absorption,-(CH2)2N-Ar, piperazine), 4.11 (triplet atop
H-O-D absorption, J=6.4 Hz, Ar-O-CH2-), 6.94 (m,
1H, Ar H), 6.96±7.06 (bs, 1H, Ar H), 7.08 (m, 1H, Ar
H), 7.43-7.49 (m, 2H, Ar H), 7.70 (d, J=8.3 Hz, 1H, Ar
H), 8.10±8.12 (m, 2H, Ar H), 8.03 (m, 1H, Ar H), 8.63
(s, 1H, Ar H) ppm; 13C NMR (DMSO-d6+CD3OD) d
21.24 (Acetate), 24.52 (-C-C-C(ˆO)OH), 30.48 (-C-C-
C(ˆO) OH), 45.71+45.86 (Ar-NC-2, piperazine), 55.01
(-C2NMe, piperazine), 67.04 (Ar-O-C-), The following
ten signals account for all expected unsubstituted aro-
matic carbon absorptions, 112.30+116.66+119.10+
121.20+124.87+130.33+131.22+147.92+152.61+
159.12, 172.26 (acetate), 174.40 (-C(=O)OH) ppm.
Mass spectrum (FAB): 511 (M+H)+1. HRMS (FAB):
511.245173 (calculated value=511.245756). Fluores-
cence emission spectrum: broad emission peak centered
at 475 nm. UV spectrum: lmax=256 and 343 nm,
1
methanol:triethylamine, 16:8:1). H NMR (CD3OD) d
1.91 (s, acetate), 2.14 (m, 2H,-CH2-), 2.46 (s, 3H, CH3N-),
2.60 (t, J=7.2 Hz, 2H,-CH2-COR), 2.80 (broad triplet,
J=4.8 Hz, 4H, (-CH2)2NMe, piperazine), 3.25 (broad
triplet, J=4.5 Hz, 4H, (-(CH2)2N-Ar, piperazine), 4.11
(t, J=6.1 Hz, 2H, Ar-O-CH2-), 5.12 (s, 2H,-CH2-,
benzylic), 7.04 (m, 1H, Ar H), 7.13 (d, J=2.1 Hz, 1H,
Ar H), 7.23±7.33 (m, 3H, Ar H), 7.43 (t, J=8.3 Hz, 1H,
Ar H), 7.50 (d, J=8.8Hz, 1H, Ar H), 7.67±7.71 (m, 2H,
Ar H), 7.96 (m, 1H, Ar H), 8.27 (bs, 1H, Ar NH) ppm; 13C
NMR (CD3OD) d 22.89 (Acetate), 26.33 (-C-C-C(ˆO)
OH), 32.18 (-C-C-C(ˆO)OH), hidden by solvent peak
(Ar-NC-2, piperazine), 56.48 (-C2NMe, piperazine), 67.82
(-O-C-Ph, benzylic carbon), 68.66 (Ar-O-C-), The follow-
ing thirteen signals account for all expected unsubstituted
aromatic carbon absorptions, 114.13+117.06+118.84+
120.82+126.56+129.70+130.02+131.89+132.34+
149.92+154.23+155.61+161.46, three low intensity sig-
nals possibly due to substituted aromatic carbons were
detected at 103.02+123.35+138.15, 175.21 (acetate),
l
min=287 nm, E343=2.2Â104 at [7.97Â10 6].
FMGT-1. meta-(4-Hydroxybutyric acid)-Hoechst ace-
tate salt (4) (50 mg, 0.1 mmol) was dissolved in 1 mL
dilute HCl(aq) and lyopholized to attain the HCl salt.
The yellow powder was then added to tris(2-aminoethy-
l)amine (190 mg, 1.3 mmol) and PyBOP1 (0.15 mmol,
78 mg) in 0.5 mL DMF to give a slightly colloidal solu-
tion. The solution was stirred at room temperature for 40
h and the reaction progress monitored by TLC (silica gel
using a gradient of concentrated aqueous ammonium
hydroxide in methanol). The solution was evaporated
under vacuum to give a thick brown oil.
178.16 (-C(ˆO)OH) ppm. LRMS (FAB): 601 (M+H)+1
.
Puri®cation of a fraction of the product mixture was
achieved by preparative HPLC reverse-phase chroma-
tography (solvent A=0.1 M triethylamine acetate, pH
7, solvent B=100% methanol, solvent C=0.1% tri-
¯uoroacetic acid with 5% acetonitrile). The product was
found to elute very slowly with solvent A. Thus, a steep
gradient of methanol up to 100% was used in concert
with solvent A to remove most impurities. Then the
column was ¯ushed with solvent A, then ¯ushed with
solvent C, and an increasing gradient of methanol was
used in concert with solvent C until the product eluted.
Evaporation of solvent gave a thick oil which was dis-
solved in a methanol/ether mixture and the product was
precipitated out with drop wise addition of 12 N HCl
and collected by centrifugation to give a bright-yellow
solid. Product purity was shown by analytical HPLC.
HRMS (FAB): 601.293279 (calculated mass=601.292
714). Analytical HPLC (reverse-phase, C18) 0.1% tri-
¯uoroacetic acid (TFA) aqueous with gradient of 0±100%
Acetonitrile in 20 min, product eluted at 13.4 min.
Fluorescence emission spectrum: broad emission peak
centered at 470 nm. UV spectrum: lmax=255 and
343 nm, lmin=289 nm, E343=2.8Â104 at [8.72Â10 6].
meta-(4-Hydroxybutyric acid)-Hoechst acetate salt (4).
meta-(4-Hydroxybutyrate)-Hoechst (515mg, 0.86mmol)
was dissolved in 15 mL DMF/H2O (3/1) and excess
potassium carbonate was added. The solution was
heated with stirring at 110 ꢀC for ꢂ3 h until TLC
showed complete disappearance of the ester (silica, ethyl
acetate:methanol:triethylamine, 16:8:1). The solvent was

A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
1879
Quanti®cation of product quantity was achieved by
NMR spectroscopy using an internal standard which
showed puri®cation of 6.2Â106 mol of product as its
HCl salt. 1H NMR (D2O) d 2.02 (m, 2H,-CH2-), 2.45 (t,
J=7.34 Hz, 2H,-CH2-CONHR), 2.74 (bs, 2H,-C-CH2-
NR2), 2.90 (bs, 4H,-RN(CH2-C-N)2 ), 3.00 (s, CH3-N-),
3.0-3.1 ( m, (-CH2)2- NMe, piperazine), 3.13 (bs,-RN(C-
CH2-N)2 ), 3.36 (triplet, J=6.1 Hz, 2H, OˆC-N-CH2-),
3.6±3.8 (m, 4H,-(CH2)2N-Ar, piperazine), 3.82 (bs, 2H,
AR-O-CH2-), integration of all 6 peaks between 6.74 and
7.55 gives 11 protons, this accounts for 1,-(OˆC)NH-,
and 10, Ar H, protons, 6.74 (bs), 6.90 (bs), 7.07 (broad
multiplet), 7.28 (s), 7.41 (bs), 7.55 (broad multiplet)
ppm. Fluorescence emission spectrum: broad peak cen-
tered at 488 nm. UV spectrum: lmax=255 and 343 nm,
anhyd DMF. The reaction vessel was very slowly shaken
for 20 h and the Kaiser test was negative.
The resin was then thoroughly rinsed with DMF, DCM
and MeOH. To cleave the resin and concurrently
deprotect the arginine, TFA was added to the reaction
vessel along with 2 drops of water and 2 drops of TIS.
The reaction vessel was then very slowly shaken for 18 h
and the resin beads then ®ltered from the product/TFA
mixture to give a yellow, mildly ¯uorescent solution.
Addition of 8 volumes of ether led to a white precipitate
which was separated from the solution by centrifugation.
The crude solid was then redissolved in water followed by
removal of any undissolved impurities by centrifugation
and then lyopholized back to a dry solid. Puri®cation by
preparative HPLC and determination of product quan-
tity (through NMR spectroscopy) was achieved in the
same manner as for FMGT-1. Product purity was
l
min=289, E343=2.2Â104 at [8.95Â10 6]. LRMS (FAB):
639 (M+H)+1. Analytical HPLC (reverse-phase, C18)
0.1% aqueous tri¯uoroacetic acid (TFA) with gradient
of 0±100% acetonitrile in 30 min, product eluted at
11.8 min.
6
shown by analytical HPLC (yellow solid, 2.1Â10
moles, yield 11%). 1H NMR (D2O) d 1.4±1.8 (over-
lapping signals, 12H,-C-CH2-CH2-C-N-, arginine), 2.131
(m, 2H, Ar-O-C-CH2-), 2.526 (broad triplet, J=7.0 Hz,
2H,-CH2-CONR), 2.9±3.3 (overlapping signals, H3C-N-
(CH2-)2+-CH2-N-C-(N)2), 3.680, 3.708, 3.829, 3.861
(4H,-(CH2)2N-Ar, piperazine), 4.026 (bs, 2H, Ar-O-CH2-),
4.15±4.28 (overlapping signals, 3H,-N-CHR-(CˆO)-,
arginine), integration of the aromatic region between 7.0
and 8.1 accounts for all 10 aromatic, Ar H, protons, 7.028
(bs), 7.190 (s), 7.261, 7.284 (overlapping singlets), 7.391
(bs), 7.476 (broad triplet), 7.554, 7.572 (overlapping sing-
lets), 7.620, 7.642 (overlapping singlets), 7.817 (bs), 8.027
(bs) ppm. Fluorescence emission spectrum: broad peak
centered at 478 nm. UV spectrum: lmax=256 and 343,
FMGT-2. meta-(4-Hydroxybutyric acid)-Hoechst ace-
tate salt (4) (10 mg, 0.016 mmol) was dissolved in 1 mL
dilute HCl (aq) and lyopholized to attain the HCl salt.
The acid was then dissolved in 1 mL DMF with 1.4 equiv
PyBOP1 (11.4 mg, 0.022 mmol) and 10 equiv spermine
(32 mg, 0.16 mmol). TLC (silica, 100% methanol)
showed about 50% completion of the reaction, product
Rf ꢂ 0. The solvent was evaporated under vacuum and
the crude reaction mixture redissolved in water. Product
was puri®ed by HPLC in the same manner as FMGT-1
and collected as the HCl salt. Product purity was shown
by analytical HPLC. Quanti®cation of product quantity
was achieved by NMR spectroscopy as for FMGT-1 and
showed puri®cation of 6.1Â10 6 moles of product (yellow
solid, yield 38%). 1H NMR (D2O+D3COD) d 1.85 (bs,-
N-C-CH2-CH2-C-N-), 1.966 (m, Ar-O-C-CH2-), 2.141
(m, 4H,-N-C-CH2-C-N-), 2.513 (broad triplet, J=
7.42 Hz, 2H,-CH2-CONHR), 3.03±3.367 (overlapping
signals, protons in b position relative to nonaromatic
amines+OˆC-N-CH2-), 3.708 (d, 2H,-(CH2)2N-Ar,
piperazine), 3.906 (d, 2H,-(CH2)2N-Ar, piperazine),
4.076 (bs, 2H, Ar-O-CH2-), integration of the aromatic
region between 6.9 and 8.2 accounts for all 10 aromatic,
Ar H, protons, 7.009 (m), 7.246 (s), 7.291 (m), 7.445 (t,
J=7.98 Hz), 7.533 (bs), 7.597±7.670 (overlapping signals),
8.114 (s), 8.155 (s) ppm. Fluorescence emission spectrum:
broad peak centered at 478 nm. UV spectrum: lmax=256
and 343, lmin=289 nm, E343=3.5Â104 at [7.56Â10 6].
LRMS (FAB): 697 (M+H)+1. Analytical HPLC
(reverse-phase, C18) 0.1% aq TFA with gradient of 0±
100% acetonitrile in 20 min, product eluted at 9.4 min.
l
min=289 nm, E343=3.6Â104 at [6.25Â10 6]. LRMS
(ESI): 978 (M+H)+1, 488 (M+2H)+2. Analytical HPLC
(reverse-phase, C18) 0.1% tri¯uoroacetic acid (TFA)
aqueous with gradient of 0±100% acetonitrile in 20 min,
product eluted at 9.9 min.
FMGT-5. The resin-(Arg)5-NH2 was synthesized in the
same manner as for FMGT-3 using MBHA resin
(0.02 mmol loading sites, 50 mg). Coupling of the depro-
tected terminal amine of the resin with the Hoechst moi-
ety was also accomplished as for FMGT-3 beginning
with meta-(4-hydroxybutyric acid)-Hoechst acetate salt
(2 equiv, 20 mg, 0.04 mmol). Puri®cation by preparative
HPLC and determination of product quantity (through
NMR spectroscopy) was achieved in the same manner as
for FMGT-1. Product purity was shown by analytical
6
HPLC (yellow solid, 5.1Â10
mol, yield 26%). 1H
NMR (D2O) d 1.48±1.84 (overlapping signals, 20H,-C-
CH2-CH2-C-N-, arginine), 2.135 (m, 2H, Ar-O-C-CH2-),
2.538 (broad triplet, 2H,-CH2-CONR), 3.001±3.245 (over-
lapping signals, H3C-N-(CH2-)2+-CH2-N-C-(N)2), 3.679,
3.707, 3.826, 3.860 (4H,-(CH2)2N-Ar, piperazine), 4.051
(bs, 2H, Ar-O-CH2-), 4.233-4.307 (overlapping signals,
5H,-N-CHR-(CˆO)-, arginine), integration of the aro-
matic region between 7.3 and 8.1 accounts for all 10 aro-
matic, Ar H, protons, 7.038 (bs), 7.188 (s), 7.259, 7.283
(d), 7.415 (bs), 7.493 (bs), 7.563, 7.622, 7.644 (overlapping
signals), 7.836 (s), 8.037 (bs) ppm. Fluorescence emission
spectrum: broad peak centered at 477 nm. UV spectrum:
FMGT-3. The resin-(Arg)3-NH2 was synthesized using
manual FMOC solid-phase techniques with MBHA
resin (0.02 mmol loading sites, 50 mg) and 2 equiv of the
coupling reagent PyBOP1 for each amino acid addi-
tion. meta-(4-Hydroxybutyric acid)-Hoechst acetate salt
(2 equiv, 20 mg, 0.04 mmol) was dissolved in 1 mL dilute
HCl (aq) and lyopholized to attain the HCl salt. The 2
equiv of meta-(4-hydroxybutyric acid)-Hoechst HCl salt
were then added to the deprotected resin with 2 equiv
PyBOP1 and 4 equiv DIPEA in approximately 2 mL of
l
max=256 and 343, lmin=290 nm, E343=1.74Â104 at

1880
A. L. Satz, T. C. Bruice, / Bioorg. Med. Chem. 8 (2000) 1871±1880
[6.60Â10 6]. LRMS (ESI): small signal at 1291 (M+H)+1
,
23. Oakley, M. G.; Mrksich, M.; Dervan, P. B. Biochemistry
1992, 31, 10969.
24. Adnet, F.; Liquier, J.; Taillandier, E.; Singh, M. P.; Rao,
K. E.; Lown, J. W. J. Biomol. Struct. Dyn. 1992, 10, 565.
25. He, G. X.; Browne, K. A.; Blasko, A.; Bruice, T. C. J.
Am. Chem. Soc. 1994, 116, 3716.
strong signals at 646 and 431 for (M+2H)+2 and
(M+3H)+3, respectively. Analytical HPLC (reverse-
phase, C18) 0.1% aqueous tri¯uoroacetic acid with gra-
dient of 0±100% acetonitrile in 20 min, product eluted
at 10.2 min.
26. Browne, K. A.; He, G. X.; Bruice, T. C. J. Am. Chem.
Soc. 1993, 115, 7072.
27. Bostock-Smith, C. E.; Searle, M. S. Nucl. Acids Res. 1999,
27, 1619.
Acknowledgement
28. Steinmetzer, K.; Reinert, K. E. J. Biomol. Struct. Dyn.
1998, 15, 779.
29. Haq, I.; Ladbury, J. E.; Chowdhry, B. Z.; Jenkins, T. C.;
Chaires, J. B. J. Mol. Biol. 1997, 271, 244.
This work was supported by a grant from the National
Institute of Health (5R37DK09171-36).
30. Loontiens, F. G.; Regenfuss, P.; Zechel, A.; Dumortier,
L.; Clegg, R. M. Biochemistry 1990, 29, 9029.
31. Harshman, K. D.; Dervan, P. B. Nucl. Acids Res. 1985,
13, 4825.
32. Loontiens, F. G.; McLaughlin, L. W.; Diekmann, S.;
Clegg, R. M. Biochemistry 1991, 30, 182.
33. Eftink, M. R. Flouorescence Methods for Studying Equili-
brium Macromolecule±Ligand Interactions; Brand, L., Johnson,
M. L., Eds.; Academic: San Diego, 1998; Vol. 278, pp 221±257.
34. Clark, G. R.; Squire, C. J.; Gray, E. J.; Leupin, W.; Nei-
dle, S. Nucl. Acids Res. 1996, 24, 4882.
35. For a recent investigation of the interactions between
DNA and polyarginine, see Mascotti, D. P.; Lohman, T. M.
Biochemistry 1997, 36, 7272.
References and Notes
1. Aymami, J.; Nunn, C. M.; Neidle, S. Nucl. Acids Res. 1999,
27, 2691.
2. Singh, A. K.; Lown, J. W. Synth. Commun. 1998, 28, 4059.
3. Bostock-Smith, G. E.; Laughton, C. A.; Searle, M. S. Nucl.
Acids Res. 1998, 26, 1660.
4. Pilch, D. S.; Poklar, N.; Baird, E. E.; Dervan, P. B.; Bre-
slauer, K. J. Biochemistry 1999, 38, 2143.
5. Pilch, D. S.; Yu, C.; Makhey, D.; LaVoie, E. J.; Srinivasan,
A. R.; Olson, W. K.; Sauers, R. R.; Breslauer, K. J.; Gea-
cintov, N. E.; Liu, L. F. Biochemistry 1997, 36, 12542.
6. Guan, L. L.; Zhao, R. L.; Lown, J. W. Biochem. Biophys.
Res. Commun. 1997, 231, 94.
36. Mitsunobu, O. Synthesis, 1981, 1.
37. Dardoize, F.; Goasdoue, C.; Goasdoue, N.; Laborit, H.
M.; Topall, G. Tetrahedron 1989, 45, 7783.
7. Clark, G. R.; Gray, E. J.; Neidle, S.; Li, Y. H.; Leupin, W.
Biochemistry 1996, 35, 13745±13752.
8. Ebrahimi, S. E. S.; Bibby, M. C.; Fox, K. R.; Douglas, K.
T. Anti-Cancer Drug Des 1995, 10, 463.
9. Parkinson, J. A.; Ebrahimi, S. E.; McKie, J. H.; Douglas,
K. T. Biochemistry 1994, 33, 8442.
10. Bailly, C.; Chaires, J. B. Bioconjugate Chem. 1998, 9, 513.
11. Geierstanger, B. H.; Wemmer, D. E. Annu. Rev. Biophys.
Biomol. Struct. 1995, 24, 463.
12. Zimmer, C.; Wahnert, U. Prog. Biophys. Mol. Biol. 1986,
47, 31.
38. Yadagiri, B.; Lown, J. W. Synth. Commun. 1990, 20, 955.
39. Coste, J.; Lenguyen, D.; Castro, B. Tetrahedron Lett.
1990, 31, 205.
40. Kubota, Y. Bull. Chem. Soc. Jpn 1990, 63, 758.
41. Kubota, Y.; Nakamura, H. Chem. Lett., 745±748.
42. Bailly, C.; Colson, P.; Houssier, C.; Wang, H. Y.; Bathini,
Y.; Lown, J. W. J. Biomol. Struct. Dyn. 1994, 12, 173.
43. Bathini, Y.; Rao, K. E.; Shea, R. G.; Lown, J. W. Chem.
Res. Toxicol. 1990, 3, 268.
13. Bruice, T. C.; Sengupta, D.; Blasko, A.; Chiang, S. Y.;
Beerman, T. A. Bioorg. Med. Chem. 1997, 5, 685.
14. Bremer, R. E.; Baird, E. E.; Dervan, P. B. Chem. Biol.
1998, 5, 119.
44. Robles, J.; McLaughlin, L. W. J. Am. Chem. Soc. 1997,
119, 6014.
45. Blasko, A.; Bruice, T. C. Proc. Natl. Acad. Sci. USA 1993,
90, 10018.
15. Chiang, S. Y.; Bruice, T. C.; Azizkhan, J. C.; Gawron, L.;
Beerman, T. A. Proc. Natl. Acad. Sci. USA 1997, 94, 2811.
16. Sengupta, D.; Blasko, A.; Bruice, T. C. Bioorg. Med.
Chem. 1996, 4, 803.
17. Bellorini, M.; Moncollin, V.; Dincalci, M.; Mongelli, N.;
Mantovani, R. Nucl. Acids Res. 1995, 23, 1657.
18. Welch, J. J.; Rauscher, F. J.; Beerman, T. A. J. Biol.
Chem. 1994, 269, 31051.
46. The equilibrium association constant and stoichiometry
for the complexation of dsDNA 12 by Hoechst 33258 has been
previously reported by Loontiens et al. (ref 32) to be
1
Log(K)=8.5 M and 1:1.
47. The equilibrium association constant for the binding of
1
Hoechst 33258 to 12b was determined to be Log(K)=8.5 M
(ref 29).
48. de Clairac, R. P. L.; Seel, C. J.; Geierstanger, B. H.;
Mrksich, M.; Baird, E. E.; Dervan, P. B.; Wemmer, D. E. J.
Am. Chem. Soc. 1999, 121, 2956.
19. Chiang, S. Y.; Welch, J.; Rauscher, F. J.; Beerman, T. A.
Biochemistry 1994, 33, 7033.
20. Dorn, A.; A€olter, M.; Muller, M.; Gehring, W. J.; Leu-
pin, W. EMBO J. 1992, 11, 279.
49. Herman, D. M.; Turner, J. M.; Baird, E. E.; Dervan, P. B.
J. Am. Chem. Soc. 1999, 121, 1121.
21. Dickinson, L. A.; Gulizia, R. J.; Trauger, J. W.; Baird, E.
E.; Mosier, D. E.; Gottesfeld, J. M.; Dervan, P. B. Proc. Natl.
Acad. Sci. USA 1998, 95, 12890.
50. Swalley, S. E.; Baird, E. E.; Dervan, P. B. J. Am. Chem.
Soc. 1999, 121, 1113.
51. Kopka, M. L.; Goodsell, D. S.; Han, G. W.; Chiu, T. K.;
Lown, J. W.; Dickerson, R. E. Structure 1997, 5, 1033.
22. Neidle, S. Biopolymers 1997, 44, 105.