Investigations into the DNA-binding mode of doxorubicinone

Article abstract of DOI:10.1039/c8ob02344a

Cancer treatment is one of the major challenges facing the modern biomedical profession. Development of new small-molecule chemotherapeutics requires an understanding of the mechanism of action for these treatments, as well as the structure-activity relat

Full text of DOI:10.1039/c8ob02344a

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Investigations into the DNA-binding mode of
doxorubicinone†
Cite this: DOI: 10.1039/c8ob02344a
Samuel Steucek Tartakoff,
Jennifer M. Finan, Ellis J. Curtis, Haley M. Anchukaitis,
Danielle J. Couture and Samantha Glazier
*
Cancer treatment is one of the major challenges facing the modern biomedical profession. Development
of new small-molecule chemotherapeutics requires an understanding of the mechanism of action for
these treatments, as well as the structure–activity relationship. Study of the well-known DNA-intercalating
agent, doxorubicin, and its aglycone, doxorubicinone, was undertaken using a variety of spectroscopic
and calorimetric techniques. It was found that, despite conservation of the planar, aromatic portion of
doxorubicin, the agylcone does not intercalate; it instead likely binds to the DNA minor-groove.
Received 21st September 2018,
Accepted 31st October 2018
DOI: 10.1039/c8ob02344a
rsc.li/obc
anthracycline-family of natural products, was first isolated in
Background and introduction
1969 from Streptomyces peucetius6 and has been used to treat
Despite an enormous amount of research that has been various forms of cancer since 1974. DOX is known to interrupt
focused on understanding and treating cancer, the United DNA transcription by intercalation between DNA base pairs,
States National Cancer Institute anticipates over 1.7 million which inhibits topoisomerase II.
new cases of will be diagnosed in the US alone in 2018, with
However, the mechanism by which intercalation occurs has
600 000 deaths attributed to cancer and related compli- been vigorously debated.7 While it is generally accepted that
cations.1 While many advances have been made, cancer DOX first pre-coordinates weakly to the DNA minor-groove
remains one of the major challenges for the biomedical field,2 before intercalation, both computational and experimental
in part due to the multifaceted nature of the disease, which in efforts have put forward a number of more complex mecha-
turn has given rise to multifaceted approaches to treatment. nisms and have measured binding constants that, when nor-
One such approach is the use of small-molecule ligands, malized for salt concentrations, vary by several orders of
which interact with DNA by a variety of mechanisms and magnitude.
prevent transcription, ultimately leading to cell death. Since
the discovery of cisplatin, one of the most widely-known DNA-
binding chemotherapeutic agents,3 a number of other natural
and synthetic drug-molecules have been discovered that bind
to DNA.4 However, the complexity of drug–DNA interactions
has made it challenging to understand fully how this binding
occurs, which, in turn, has hampered efforts to develop more
selective small-molecule chemotherapeutics.
For this reason, we became interested in studying the
details of DNA–ligand interactions and how those interactions
change with subtle modifications to the molecular structure of
known small-molecule ligands. Our previous work5 with doxo-
rubicin (DOX, Fig. 1) had provided insight into the entropic
and enthalpic influences for DNA-binding but also left a
number of questions unanswered. DOX, a member of the
St. Lawrence University, Department of Chemistry, 23 Romoda Drive, Canton,
New York, 13617, USA. E-mail: sglazier@stlawu.edu
†Electronic supplementary information (ESI) available: 1H NMR spectrum and
HRMS data for DOXY, as well as DNA melting curves for CT DNA with DOX and
DOXY. See DOI: 10.1039/c8ob02344a
Fig. 1 The structures of naturally occurring doxorubicin (DOX) and
several synthetic analogues; hydroxyrubicin (HDX), annamycin (ANN),
and doxorubicinone (DOXY).
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Even less clear is the importance of the daunosamine sugar ring. A 1 M aqueous HCl solution was added (0.6 mL, resulting
for DOX–DNA binding. This sugar, which is protonated under in 0.2 M overall HCl concentration) and heating was continued
physiological conditions, presumably interacts with the nega- for 90 minutes. Upon complete cleavage of the sugar, as
tively-charged phosphate backbone of DNA. What is not clear judged by TLC analysis, the reaction volume was reduced
is whether this electrostatic interaction is necessary for the in vacuo to approximately 25% of the original volume. H2O
final intercalated state of DOX. Hydroxyrubicin (HDX), which (0.5 mL) was added, after which the reaction mixture was
lacks the 3′ amino-moiety, still inhibits topoisomerase II and it extracted with CH2Cl2 (3 × 1 mL). The organic extracts were
has been assumed that the activity of HDX is also due to DNA dried with anhydrous MgSO4, filtered, and concentrated
intercalation, albeit to a lesser degree than DOX based on DNA in vacuo, and the resulting dark-red solid was purified by
unwinding assays.8,9 Similarly, annamycin (ANN) actually exhi- column chromatography (SiO2, 19 : 1 CHCl3/CH3OH, Rf = 0.13,
bits enhanced topoisomerase activity, despite significant UV) to afford doxorubicinone as a bright-orange solid (24 mg,
modification to the sugar.10 Considering the importance of 0.058 mmol, 67% yield) that matched reported spectroscopic
DOX as a chemotherapeutic and the uncertainty about the role values by 1H NMR analysis.12 This characterization was con-
that the sugar plays in DNA-binding, it is puzzling that the firmed by HRMS data.
doxorubicin-aglycone (doxorubicinone, adriamycinone, DOXY)
has received very little scholarly attention.7a
DNA melting curves. Samples contained 30 μM calf thymus
DNA (CT DNA) and 2 μM of DOX or DOXY in a low
Given the proposed pre-coordination of DOX to the minor- ionic strength phosphate buffer (3 mM phosphate buffer +
groove, the absence of the amino-sugar on DOXY would be 2.0 mM NaCl). They were heated at 5 °C increments and the
expected to lead to at least one of the following outcomes: (1) absorbance of CT DNA was monitored at 258 nm. Tm values
slower DNA–ligand binding kinetics but the same final bound were determined from the first-derivative of the melting
state, (2) a similar final intercalative state, albeit with a weaker profile.
binding constant, or (3) an entirely different final binding
mode (i.e. DOXY could remain bound in the minor-groove or recorded on a Jasco spectrometer with the following para-
only intercalate slightly). meters: 1000 mdeg sensitivity, 0.1 nm resolution, 1.0 nm
Circular dichroism (CD) spectra. The CD spectra were
To differentiate between those three potential outcomes, we bandwidth, 1.0 s response, 50 nm min−1 scan speed, 1.0 cm
undertook a side-by-side comparison study of DOX and DOXY, path-length cell at 25 °C. A fixed concentration of CT DNA
looking specifically at conformational changes to DNA and the (10 μM BP in 10 mM phosphate buffer + 50 mM NaCl + 1 mM
relative strengths of the DNA–ligand binding. Our methods cacodylate) was titrated with DOX and DOXY solutions to
focused on DNA melting temperature analysis, circular dichro- provide [drug] : [DNA] ratios typically in the range of 0.1 to 10.
ism (CD), and fluorescence titration, although additional
methods were also examined.
Fluorescence titrations and the osmotic stress method.
Fluorescence spectroscopy was used to measure the binding
constants (Kb) of DOX and DOXY with CT DNA. A fixed concen-
tration of 2 μM DOX or DOXY was titrated with DNA solution
in order to provide DNA/drug ratios typically in the range of
zero to 18. The titration solutions were mixed well and were
allowed to equilibrate to room temperature. The fluorescence
Methods and experimental
General experimental
The mono-HCl salt of doxorubicin and CT DNA was purchased spectrum of each titration solution was collected using a
from Sigma-Aldrich. Silica gel used for chromatographic separ- Horiba Jobin Yvon Fluoromax-3 fluorimeter with the following
ation (60 Å, 230–400 mesh) was purchased from Sigma- parameters: λexcitation = 480 nm and λemission = 592 nm, incre-
Aldrich. Thin-layer chromatographic (TLC) analysis was con- ment of 0.5 nm, integration time of 0.1 s, excitation slit of
ducted using glass-backed EMD/Millipore silica-gel plates 1.00 nm, emission slit of 3.00 nm, and room temperature.
(60 Å, 230–400 mesh). All other reagents were purchased from After generating the binding curve in the absence of osmolyte,
commercial vendors and used as received, without further osmolyte solution was added to each titration solution,
1
purification. H NMR spectra were collected using a 300 MHz allowed to equilibrate, and a new binding curve was generated.
JEOL-Eclipse NMR spectrometer (75 MHz on the 13C channel) Typically, five additional binding curves were collected, with
and referenced using the residual CHCl3 solvent peak osmolalities ranging from approximately 0–3.3 osm.
(7.26 ppm for 1H, 77.16 ppm for 13C).
Triethylene glycol (TEG) was used as the osmolyte. Origin©
Hydrolysis of doxorubicin. Using a modification of the con- data analysis and graphing software was used to determine the
ditions reported by Menna et al.,11 doxorubicin-mono-HCl salt binding constant at each osmolality. The raw fluorescence data
(50 mg, 0.086 mmol) was dissolved in methanol (3.0 mL), were fit with an independent, non-site-specific binding
topped with a water condenser, then heated to 80 °C with stir- function:13
ꢂ
ꢃ
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢀ
ꢁ
ꢀ
ꢁ
2
2
Kb½Sꢀ0 þ Kb½Dꢀ0 þ 1 ꢁ
Kb½Sꢀ0 þ Kb½Dꢀ0 þ 1 ꢁ4Kb ½Sꢀ0½Dꢀ0 ½Ib ꢁ If ꢀ
2Kb½Dꢀ0
ð1Þ
I0 ¼
þ If :
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
In eqn (1), I0 is the fluorescence intensity at a given to eqn (6), where t is time, y is fluorescence intensity, and y0 is
DNA/drug ratio, If is the fluorescence intensity of the drug in the baseline fluorescence intensity.
the absence of DNA, Ib is the fluorescence intensity of the drug
y ¼ y0 þ A1eꢁt=τ þ A2eꢁt=τ ð6Þ
ð6Þ
1
2
in its bound state, [D]0 is the total concentration of drug in
solution, and [S]0 is the total concentration of DNA in
solution (in units of moles of base pairs per liter of solution).
The absolute fluorescence response, I0, was plotted as a
function of total DNA concentration, [S]0. The total drug
concentration, [D]0, was treated as a constant, and Kb treated
as a parameter.
Results and discussion
After some optimization, acidic hydrolysis of the sugar
After calculating the binding constant at each osmolality in moiety on DOX, followed by chromatographic purification,
a titration series, ln(K) was plotted against solution osmolality provided DOXY in high purity and sufficient quantities for our
to obtain a linear relationship. The number of water molecules studies.
exchanged during the binding event, Δnw, was calculated
As the first of our three study questions dealt with the kine-
using the slope of this line and eqn (2). The binding constants tics of DNA-binding and how the removal of the amino-sugar
were then plotted against solution osmolality. A form of the would impact binding-rate, we used a stopped-flow fluo-
Gibbs–Duhem equation14 tells us that the slope of this line is rescence experiment with each drug and CT DNA. From this
directly proportional to Δnw, the net uptake or release of water method, the binding event was fit using a second order expo-
molecules that accompanies the binding event:
nential, which yielded two lifetimes for DOXY, τ1a = 620 100
s (A1a = 0.11) and τ2a = 6700 500 s (A2a = 0.89). This means
that there are at least two steps involved for binding. Similarly,
@ ln Kb ꢁΔnw
¼
:
ð2Þ
@½Osmꢀ
55:5
fits of DOX-DNA binding data also yielded two lifetimes τ1a
=
0.019 0.002 s (A1a = 0.98) and τ2a = 0.87 0.08 s (A2a = 0.02).
The difference in association binding rates was significant,
though not entirely surprising considering the uncharged
DOXY could not benefit from a strong electrostatic attraction
to the DNA backbone. Dissociation kinetics measurements for
A van’ t Hoff analysis (eqn (3)–(5)) was used to calculate
enthalpy (ΔH) and entropy (ΔS) values. Data was collected over
a temperature range (T) of 15 °C to 60 °C in 5 °C increments.
ΔG ¼ ꢁRT ln K
ð3Þ
ð4Þ
DOX were also measured and yielded lifetime values of τ1d
=
ꢄ
ꢅ
0.43 0.005 s (A1d = 0.57) and τ2d = 1.7 0.008 s (A2d = 0.02).
The dissociation lifetime of DOXY appears to be <100 ns and,
therefore, beyond the detection limits of stopped-flow. Again,
the kinetic behavior of DOXY differs significantly from that of
DOX. These kinetics are summarized in Fig. 2. Because the
rate of association was so much slower than known DNA-inter-
ꢁΔH
1
1
ln K1 ꢁ ln K2
¼
ꢁ
R
T1 T2
ΔH ꢁ ΔG
ΔS ¼
:
ð5Þ
T
Stopped flow kinetics. A Hi-Tech Scientific SFA-20 Rapid calating agents,15 the question about whether a different
Kinetics stopped flow accessory was used to mix solutions for binding mode might be operative was reinforced.
kinetic experiments. Fluorescence parameters were: λexcitation
=
We therefore proceeded to examine the drug–DNA complex
480 nm and λemission = 592 nm, integration times 0.005 s for for structural changes that might point towards either interca-
DOX and 5.0 s for DOXY, and T = 30 °C. Samples for the lation or an alternate binding mode. Measurement of DNA-
association kinetics were prepared with DNA in excess to melting temperature (Tm) is a widely used method for recog-
assure complete binding: [DOX] = 4 μM + [CT DNA] = 160 μM nizing structural changes in DNA.16 Specifically, DNA dena-
and [DOX] = 1 μM + [CT DNA] = 10 μM. For the dissociation turation has been used to extensively characterize whether or
kinetics measurements, one syringe contained CT DNA and not anthracycline small-molecules have specificity for certain
DOX or DOXY at the same concentrations used for the associ- DNA sequences.17 In addition, it has been shown that interca-
ation experiments. Samples were left overnight to assure lators stabilize the DNA double-helix through pi-stacking inter-
binding equilibrium. The second syringe contained 2% actions, raising the Tm,18 while minor-groove binding has little
sodium dodecyl sulphate (SDS), which sequesters bound DOX effect on the melting point.19 It should be noted that neomy-
or DOXY upon mixing in the cuvette. Data was analysed using cin, a known major-groove binder, is reported to have no effect
exponentials to obtain florescence lifetimes, τn, and weighting on the Tm,20 although there is a paucity of melting data about
coefficients. In addition to goodness of fit, R2, several criteria other major-groove-bound DNA–drug complexes.
were used to determine the order of the best fit exponential.
The results of our DNA melting-point study are shown in
The weighting coefficients, An, had to be physically significant Table 1. In agreement with previously published studies,
(>1%), the difference between lifetime values had to be at least known intercalators (ethidium bromide and DOX) increase Tm
an order of magnitude, and fits were robust, allowing for a for CT DNA significantly (9.1 °C and 11.4 °C, respectively.)
range of initial parameters. Using these criteria, binding by Hoechst, a known minor-groove binder, had little effect on Tm,
DOX and DOXY were determined to be second order according lowering the value by 0.8 °C. DOXY raised the temperature but
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
above 50 µM in a solution with ionic strength of 2.5 mM.7b,24
Furthermore, DOX may form higher-order complexes with
DNA at high concentrations, further complicating analysis.
DOXY, which lacks the hydrophilic sugar-moiety, could be
expected to be even less soluble and, indeed, self-aggregation
was observed at concentrations of 15–20 µM. Thus, the range
of concentrations that we were able to analyze for DOXY fell
outside the ranges of many previously published studies on
DOX.
Therefore, to achieve the desired DOXY/DNA ratios (ranging
from 0 to 10), with DOX/DNA ratios in the same regime, we
found it expedient to fix the CT DNA concentration at 10 µM,
while adding variable amounts of DOX or DOXY. The results
can be seen in Fig. 3. The spectrum of CT DNA shows a nega-
tive band at 245 nm and a positive band at 275 nm, which are
measures of right-handed helicity and ππ* stacking, respect-
ively.25 It had previously been reported by Garcia and co-
workers that intercalation of DOX with CT DNA could be
observed in the DOX CD spectrum by a red-shift in the spectral
band at 300 nm, accompanied by a decrease in intensity. They
observed a maximal shift at a DOX/DNA ratio of 0.35, which
Fig. 2 Association kinetics of DOXY (top) and DOX (bottom) with CT
DNA. SDS induced dissociation kinetics of DOX from CT DNA (inset).
Solid lines are the best fit curves used to calculate fluorescence life-
times. Dissociation kinetics of DOXY was too fast to be measured using
the same method. While association kinetics is second order for both,
DOXY binds several orders of magnitude more slowly than DOX.
Table 1 The difference in melting points of DNA binding molecules
compared to the melting temperature of CT DNA alone
Ethidium bromide
+9.1
DOX
Hoechst
DOXY
+2.5
ΔTm (°C)
+11.4
–0.8
only by 2.5 °C, suggesting that if intercalation is occurring, it
is only occurring to a small extent.
To clarify the results of the melting point study, we next
attempted viscosity titrations using methods published in pre-
vious work.21 Despite the general reliability of this method to
determine DNA binding mode,22 the low solubility of DOXY
prevented us from obtaining meaningful results. Because CD
is such a widely used tool for analyzing structural changes in
DNA, particularly in the context of DOX binding to DNA,7c,23 it
seemed natural to undertake a similar analysis with DOXY.
When considering our experimental conditions, there were
several potential challenges, of which we needed to be cogni-
Fig. 3 CD spectra of CT DNA, DOX, and DOX/CT DNA ratios between 0
and 2.5 (top). CD spectra of CT DNA, DOXY, and DOXY/CT DNA ratios
between 0 and 10 (bottom). DOX is a known intercalator and induces
band shifts at 195 nm, 225 nm, and 300 nm. These shifts are not
zant. DOX exhibits significant aggregation at concentrations observed in the spectra for DOXY bound to CT DNA.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
agrees with the accepted binding of DOX at roughly every third
DNA base pair. We also observed this shift in the spectrum for
DOX/DNA ratios between 0.25–1.0.
Garcia also pointed out the positive spectral band at
293 nm and assigned that as a shift of the DOX band from
250 nm. However, CT DNA has a broad positive band centered
at 275 nm and it seems likely that in the regime examined by
Garcia (with a 5-fold excess of CT DNA relative to DOX), the
DOX band at 250 nm was simply buried under the larger CT
DNA band. We did observe a blue-shift in the CT DNA spectral
band, from 225 nm to 215 nm, suggesting structural changes
in the DNA helix. This same shift was noted by Giustini and
was attributed to DOX binding to poly-GC DNA sequences.7c
This shift was observed with as little as 0.1 DOX/DNA and
became more pronounced up to a 1 : 1 ratio, after which the
DOX band at 234 nm started to predominate.
The final noteworthy spectral shift was observed in the CD
spectrum of CT DNA at 195 nm. The positive band was
remained largely unchanged up to a DOX/DNA ratio of 0.50.
However, at and above concentrations of 10 μM DOX, this
band was completely replaced by a large, negative band at
197 nm. The fact that DOX has a positive spectral feature cen-
tered at 200 nm suggests that this shift represents a binding
event between DNA and DOX; possibly through a secondary
binding mode, as postulated by Garcia.
Fig. 4 Binding isotherms for the interactions of CT DNA with DOX and
DOXY. Solid lines are the best fit curves used to calculate equilibrium
constants: KDOX = 1.3 × 107 0.2 and KDOXY = 2.9 × 104 0.2. The fluo-
rescence signal is quenched to a greater degree at low concentrations
of DNA for DOX, which is indicates more efficient binding by DOX.
104 0.2 (Fig. 4). When corrected for the difference in ionic
strength, the relative magnitudes match well with Chaires’
values. If, as we think, DOXY is binding to the minor-groove of
DNA, it is additionally worth comparing with Garcia’s value for
pre-intercalative binding of DOX, which was measured as K =
1 × 104 when I = 100 mM. After correcting for ionic strength,
that number is almost identical to our DOXY binding constant.
In previous work in our laboratory, we found that the
number of water molecules exchanged during the binding
process of small molecules with DNA depended on the
binding mode. In the case of DOXY, the water exchange value
was +66, which is a large increase in the number of co-
ordinated waters, consistent with other groove binding drugs
(Fig. 5). For example, netropsin binding to DNA takes up
+50–60 water molecules, while Hoechst uptakes +74 water
By contrast, the CD spectrum for DOXY with CT DNA exhi-
bits no such shifts, even when DOXY was added in great
excess. The lack of a strong signal from the DOXY is not sur-
prising, considering the cleavage of the amino-sugar signifi-
cantly decreases (but does not entirely negate) the chiral char-
acter of DOXY. However, the complete lack of change to the CT
DNA bands, particularly the band at 225 nm, supports the
theory that DOXY is not intercalating.
Considering the mounting evidence that DOXY does not
intercalate, but instead stays in the DNA minor-groove, by
analogy with the DOX pre-intercalative state, DOXY should
also bind less strongly to DNA than does DOX. Granted, the
lack of the protonated amino-sugar on DOXY removes the
electrostatic attraction to the phosphate backbone, which
would automatically decrease the binding affinity. Fortunately,
Chaires et al.7a had already shown that, by using salt back-titra-
tion26 and applying the Manning–Record equation,27 the
electrostatic component of the binding constant can be iso-
lated. They found that HDX, which possess the sugar-moiety
but lacks the charged ammonium, binds to CT DNA 100-times
less strongly than DOX, corresponding to a ΔΔG of 2.5
kcal mol−1, of which 1.9 kcal mol−1 were electrostatic. Taken
together, the non-electrostatic ΔΔG between DOX and HDX
was only about 0.6 kcal mol−1. By contrast, after accounting for
electrostatic differences, the ΔΔG between DOX and DOXY was
2.0 kcal mol−1, which makes up over 1/5th of the total binding
free energy. He attributes this difference in binding energy to
the importance of the sugar rather than a shift in DNA
binding mode.
Fig. 5 Eqn (2) and the equilibrium constants obtained from titrations of
DOXY and CT DNA in the presence of various concentrations of TEG
were used to make a water exchange plot. From the linear regressions
of the data, a net uptake of water, ΔnDOXY = +66 waters (R2 = 1.00), was
Under our titration conditions ([Na+] = 63 mM), DOX had a
binding constant of 1.3 × 107 0.2, while for DOXY, K = 2.9 × calculated.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
groove based on the precedence of other non-intercalating,
DNA-binding small-molecules. This is supported by the evi-
dence obtained from CD, DNA melting temperature, the shift
in binding constant, and water uptake studies. The possibility
of quasi-intercalation cannot be discounted, however, based
on the unusual entropy and enthalpy values for DNA binding.
There is no obvious structural reason that DOXY could not
intercalate and this, therefore, invites a more careful study of
all new proposed intercalating agents, rather than basing
assumptions on structural analogy.
It is, furthermore, clear from this study that the protonated
amino-sugar moiety plays a role, not only in rate of pre-coordi-
nation between the drug-molecule and DNA, but also in deter-
mining the ultimate bound state of the small-molecule ligand.
The implications of these results for chemotherapeutics
Fig. 6 Van’t Hoff plot to determine the enthalpy of binding of DOXY
with CT DNA. The equilibrium constants were measured at five different research warrants further study, as understanding the specifics
temperatures. From the linear regression of the data, an exothermic
enthalpy, ΔHDOXY = −52 kJ mol−1 (R2 = 1.00), was calculated.
of molecular binding modes may prove useful in determining
other cellular outcomes, which would have implications for
cell physiology and eventual patient outcomes.
molecules.28 Intercalating molecules typically uptake a signifi-
cantly smaller number of waters (0 to +30),14a as exemplified
Conflicts of interest
by DOX (+13 waters). It is interesting to note that, in this
experiment, DOXY more closely resembles the charged, struc-
turally dissimilar groove binders than it does the anthra-
cycline, DOX.
There are no conflicts to declare.
The thermodynamic data from the van’t Hoff analysis in
Acknowledgements
Fig. 6, also supports a groove-bound state for DOXY. While DOX
displays an entropic increase for binding (ΔS = +21 J K−1 mol−1),
We thank St Lawrence University for research funds and for
fellowship support to J. M. Finan, H. M. Anchukaitis,
E. J. Curtis, and D. J. Couture. We thank the McNair Scholars
Program for funding support to E. J. Curtis. We are also grate-
indicative of a more disorder in the intercalated state, DOXY has
a negative entropy of binding (ΔS = −88 J K−1 mol−1). The
large uptake of water molecules by DOXY likely contributes to
the large, negative entropic cost of binding. Similarly, the
ful to Dr J. Schmeisser and K. Park for their contributions to
increase in hydrogen-bonding from a more extended water
this project. We are indebted to Hamilton College and
network is consistent with a large negative enthalpy for DOXY
University of Toronto for allowing us to use their CD
spectrometers.
(ΔH = −52 kJ mol−1). DOX (ΔH = −31 kJ mol−1), by compari-
son, is significantly less exothermic than DOXY. Chaires
et al.29 reviewed thermodynamic data in the literature and
found entropy changes for various DNA-binding small-
Notes and references
molecules to range from −40 to +220 J K−1 mol−1, while
enthalpy changes were between −38 to +18 kJ mol−1. The ΔS
and ΔH values calculated for DOXY both extend beyond the
ranges for either typical minor-groove binders or intercalators,
indicating either another binding mode (such as quasi-interca-
lation)30 or a unique thermodynamic profile due to the neutral
charge (the thermodynamic data surveyed by Chaires all came
from cationic DNA-binding molecules.)
1 National Institutes of Health - National Cancer Institute,
Cancer Statistics. Statistics at a Glance: The Burden
of Cancer in the United States. Retrieved from https://
www.cancer.gov/about-cancer/understanding/statistics, 9-10-
2018.
2 A. Von Eschenbach, Med. Prog. Bull., 2005, 1, 1.
3 S. Dasari and P. B. Tchounwou, Eur. J. Pharmacol., 2015, 5,
364.
4 For a representative review, see: C.-H. Leung, D. S.-H. Chan,
V. P.-Y. Ma and D.-L. Ma, Med. Res. Rev., 2012, 1.
5 R. M. Kenney, K. E. Buxton and S. Glazier, Biophys. Chem.,
2016, 216, 9.
6 N. Lomovskaya, S. L. Otten, Y. Doi-Katayama, L. Fonstein,
X.-C. Liu, T. Takatsu, A. Inventi-Solari, S. Filippini, F. Torti,
A. L. Colombo and C. R. Hutchinson, J. Bacteriol., 1999,
181, 305.
Conclusions
Despite conservation of the entire planar, aromatic tricyclic
portion of DOX – in short, the entire portion of the molecule
that intercalates between DNA base pairs – DOXY has a com-
pletely different DNA-binding mode with CT DNA. It is pre-
sumed that the mode involves binding to the DNA minor-
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Organic & Biomolecular Chemistry
Paper
7 For several of these studies, see: (a) J. B. Chaires,
S. Satyanarayana, D. Suh, I. Fokt, T. Przewloka and
K. C. Majundar and T. K. Das, Biochem. Anal. Biochem.,
2016, 5, 278.
W. Priebe, Biochemistry, 1996, 35, 2047; (b) C. Pérez-Arnaiz, 19 P. L. Hamilton and D. P. Arya, Nat. Prod. Rep., 2012, 29,
N. Busto, J. M. Leal and B. García, J. Phys. Chem. B, 2014, 134.
118, 1288; (c) M. Airoldi, G. Barone, G. Gennaro, 20 D. P. Arya, R. L. Coffee and L. Xue, Bioorg. Med. Chem. Lett.,
A. M. Giuliani and M. Giustini, Biochemistry, 2014, 53,
2197.
8 B. Jollès, A. Laigle, W. Priebe and A. Garnier-Suillerot,
Chem.-Biol. Interact., 1996, 100, 165.
2004, 14, 4643.
21 S. R. Dalton, S. Glazier, B. Leung, S. Win, C. Megatulski
and S. J. N. Burgmayer, JBIC, J. Biol. Inorg. Chem., 2008, 13,
1133.
9 E. Solary, V.-H. Ling, R. Perez-Soler, W. Priebe and 22 (a) D. Suh and J. B. Chaires, Bioorg. Med. Chem., 1995, 3,
Y. Pommier, Int. J. Cancer, 1994, 58, 85.
723; (b) S. Satyanarayana, J. Dabrowiak and J. B. Chaires,
Biochemistry, 1992, 31, 9319.
10 A. V. Trevino, B. A. Woynarowska, T. S. Herman, W. Priebe
and J. M. Woynarowski, Mol. Cancer Ther., 2004, 3, 1403.
11 P. Menna, E. Salvatorelli, R. Giampietro, G. Liberi,
G. Teodori, A. M. Calafiore and G. Minotti, Chem. Res.
Toxicol., 2002, 15, 1179.
12 T. H. Smith, A. Fujiwara, W. W. Lee, H. Y. Wu and
D. W. Henry, J. Org. Chem., 1977, 42, 3653.
13 M. J. Waring, J. Mol. Biol., 1965, 13, 269.
23 (a) K. K. Patel, E. A. Plummer, M. Darwish, A. Rodger and
M. J. Hannon, J. Inorg. Biochem., 2002, 91, 220;
(b) C. B. Spillane, J. A. Smith, J. L. Morgan and F. R. Keene,
J. Biol. Inorg. Chem., 2007, 12, 819; (c) A. Ghosh, P. Das,
M. R. Gill, P. Kar, M. G. Walker, J. A. Thomas and A. Das,
Chem. – Eur. J., 2011, 17, 2089; (d) L. Gallois, M. Fiallo and
A. Garnier-Suillerot, Biochim. Biophys. Acta, Biomembr.,
1998, 1370, 31.
14 (a) X. Qu and J. B. Chaires, J. Am. Chem. Soc., 2001, 123, 1;
(b) C. Stanley and D. C. Rau, Biophys. J., 2006, 91, 912.
15 (a) H. J. Li and D. M. Crothers, J. Mol. Biol., 1969, 39, 461;
24 A. Walter, H. Schutz and E. Stutter, Int. J. Biol. Macromol.,
1983, 5, 351.
(b) M. M. Islam, S. Fujii, S. Sato, T. Okauchi and 25 (a) R. Gomathi, A. Ramu and A. Murugan, Bioinorg. Chem.
S. Takenaka, Bioorg. Med. Chem., 2015, 23, 4769;
(c) V. Rizzo, N. Sacchi and M. Menozzi, Biochemistry, 1989,
Appl., 2014, 2014, 1; (b) R. M. Pagni, J. Chem. Educ., 1998,
75, 1095.
28, 274; (d) W. Förster and E. Stutter, Int. J. Biol. Macromol., 26 T. M. Lohman and D. M. Mascotti, Methods Enzymol., 1992,
1984, 6, 114. 212, 400.
16 K. J. Breslauer, R. Frank, H. Blocker and L. A. Marky, Proc. 27 M. T. Record Jr., T. M. Lohman and H. P. De, J. Mol. Biol.,
Natl. Acad. Sci. U. S. A., 1986, 83, 3746. 1976, 107, 145.
17 G. W. Lehman and J. P. Mctague, J. Chem. Phys., 1968, 49, 28 (a) N. Y. Sidorova and D. C. Rau, Biopolymers, 1995, 35, 377;
3170.
(b) Anuradha, M. S. Alam and N. K. Chaudhury, Chem.
Pharm. Bull., 2010, 58, 1447.
18 (a) P. U. Maheswari, V. Rajendiran, M. Palaniandavar,
R. Parthasarathi and V. Subramanian, J. Inorg. Biochem., 29 J. B. Chaires, Arch. Biochem. Biophys., 2006, 453, 26.
2006, 100, 3; (b) S. Roy, S. Ganai, R. K. Nandi, 30 P. Lincoln and B. Norden, J. Phys. Chem. B, 1998, 102, 9583.
This journal is © The Royal Society of Chemistry 2018
Org. Biomol. Chem.