Topotecan dynamics, tautomerism and reactivity - 1H/ 13C NMR and ESI MS study

Article abstract of DOI:10.1002/mrc.2625

Topotecan (TPT) is in clinical use as an antitumor agent, hycamtin. Because of this, it requires both biologically and chemically useful information to be available. TPT acts by binding to the covalent complex formed by nicked DNA and topoisomerase I. This has a poisonous effect since inserted into the single-strand nick and TPT inhibits its religation. We used NMR to trace TPT dynamics, tautomerism and solvolysis products in various solvents and conditions. Chemical stability was assessed in methanol and DMSO as compared to water, and the regioselectivity of the N- and O-methylation was studied using various alkylating agents. The reaction products of quaternization of the nitrogen atom and methylation of the oxygen atom were characterized by means of ESI MS, ¹H/¹³C-HMBC and -HSQCAD NMR. We have focused on the NMR characterization of TPT with an anticipation that its aggregation, tumbling properties and the intramolecular dipolar interactions will be a common feature for other compounds described in this article. These features can also be useful in tracing the interactions of this class of topoisomerase I (TopoI) poisons with DNA. Moreover, the results explained shed light on the recently disclosed problem of lack of stability of TPT in the heart tissue homogenate samples using the analytical assays developed for this class of compounds carried out in the presence ofmethanol. Copyright

Full text of DOI:10.1002/mrc.2625

Research Article
Received: 18 March 2010
Revised: 6 May 2010
Accepted: 6 May 2010
Published online in Wiley Interscience: 8 June 2010
(www.interscience.com) DOI 10.1002/mrc.2625
Topotecan dynamics, tautomerism
and reactivity – 1H/13C NMR and ESI MS study
a
a,b
c
a,c
˙
Karolina Hyz, Robert Kawe¸cki, Elzbieta Bednarek, Wojciech Bocian,
Jerzy Sitkowskia,c and Lech Kozerskia,c∗
Topotecan (TPT) is in clinical use as an antitumor agent, hycamtin. Because of this, it requires both biologically and chemically
useful information to be available. TPT acts by binding to the covalent complex formed by nicked DNA and topoisomerase
I. This has a poisonous effect since inserted into the single-strand nick and TPT inhibits its religation. We used NMR to trace
TPT dynamics, tautomerism and solvolysis products in various solvents and conditions. Chemical stability was assessed in
methanol and DMSO as compared to water, and the regioselectivity of the N- and O-methylation was studied using various
alkylating agents. The reaction products of quaternization of the nitrogen atom and methylation of the oxygen atom were
characterized by means of ESI MS, 1H/13C-HMBC and -HSQCAD NMR. We have focused on the NMR characterization of TPT
with an anticipation that its aggregation, tumbling properties and the intramolecular dipolar interactions will be a common
feature for other compounds described in this article. These features can also be useful in tracing the interactions of this class of
topoisomerase I (TopoI) poisons with DNA. Moreover, the results explained shed light on the recently disclosed problem of lack
of stability of TPT in the heart tissue homogenate samples using the analytical assays developed for this class of compounds
c
carried out in the presence of methanol. Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Supporting information may be found in the online version of this article.
Keywords: NMR; ESI MS; topotecan dynamics; chemical transformations
Introduction
the aggregation state and dynamics of interacting molecules in
solution. Our recent studies[17,18] and the present results indicate
clearly that TPT is in a form of aggregates in a buffered water
solution and tumbles slowly. All protons are shifted to lower
frequencies in the self-association state, consistent with the
expected stacked structure. The intermolecular NOEs, observed
for the distant protons (see Table 2, for example from 23-CH2 to
19-CH3, etc.) suggest head-to-tail stacking of the TPT molecules in
aggregates.
The self-association binding constant of the TPT, calculated
from the present DNMR results, ca 3.3 1 mM−1 (see Supporting
Information) is consistent with our earlier results obtained from
dilution studies and earlier published UV experiments.[19] We have
also found that the carboxylate form binds ca 40 times weaker
than the lactone form to the nicked DNA decamer, in agreement
with the biological assays.[18]
A continuing interest is observed in the new class of compounds
in pursue to find a lead and better formulate an anticancer
drug pharmacophore.[1–5] Topoisomerases I, II poisons attract
much attention in the biomedical research in this direction as
they constitute important chemotherapeutic ways in combating
cancer.
The cytotoxic plant alkaloid camptothecin (CPT) and the
synthetic derivative topotecan (TPT) have emerged as clinically
usefulanticancerdrugs[6] andareweaklyDNAbindingcompounds
difficult to study by NMR. They are poorly soluble in water at
physiological pH, and have strongly pH-dependent 1H and 13C
NMR chemical shifts. A lactone-carboxylate equilibrium shown in
Scheme 1 is slow on the NMR timescale at room temperature.
They act as inhibitors of TopoI[7–14] via the formation of stable
ternary complexes composed of the drug, DNA and TopoI.[7,8,15]
The enzyme forms a clamp around a nicked B-form of DNA.
The TPT intercalates, stacking with its long axis parallel to the
GC(+1).[16] An important feature of the binding interaction is the
lactone functionality of the TPT, which hydrolyzes to the ring-
opened carboxylate form depending on pH (Scheme 1) and is
>80% ring-opened at pH 7.4.[10,11,13] Despite this, most biological
assays indicate that the carboxylate form of TPT is not active as a
TopoI inhibitor; on the contrary, the lactone form is stabilized in
the presence of DNA.
In Scheme 1, the two most important constituents of the
equilibrium are shown, which are observed in the NMR spectrum
in buffered water solutions at a wide pH range of 5–8. We
characterized these two forms of TPT in solution and have
∗
Correspondence to: Lech Kozerski, Institute of Organic Chemistry, Polish
Academy of Sciences, 01-224 Warsaw, Kasprzaka 44, Poland.
E-mail: lkoz@icho.edu.pl
a
b
c
Institute of Organic Chemistry, Polish Academy of Sciences, 01-224 Warszawa,
Kasprzaka 44, Poland
A thorough understanding of the action mechanism of the drug
requires detailed knowledge of its physicochemical properties in
the physiological medium. With respect to that, the TPT structure
in water at biological pH has not been studied in detail. In addition,
with regard to the interaction with DNA, it is important to know
Institute of Chemistry, University of Podlasie, ul. 3 Maja 54, 08-110 Siedlce,
Poland
National Medicines Institute, 00-725 Warszawa, Chełmska 30/34, Poland
c
Magn. Reson. Chem. 2010, 48, 575–584
Copyright ꢀ 2010 John Wiley & Sons, Ltd.

K. Hyz et al.
Scheme 1. R1 = H, R2 = H, R3 = H [camptothecin]; R1 = OH, R2
R2 = H, R3 = C2H5 [irinotecan].
=
23CH2N(22CH3)2, R3 = H [topotecan]; R1 = 1-[4ꢁ-pipridynyl-piperidine] carboxy,
1H-T1 relaxation times and NOE enhancements
presented their dynamics in water, in conditions close to
physiological. In addition, we gave an evidence for a tautomer, an
enolformshowninScheme 3, which, incontrasttothelactoneand
carboxylate forms, is observed in water and methanol solutions as
intermediate species. We also determined the molecular tumbling
parametersofaleadcompoundinthisclassoftheTopoIinhibitors,
TPT in water. Moreover, the intramolecular NOEs were established
as a reference data set useful in tracing the intermolecular
contacts with DNA as targets. We focused on the TPT with an
anticipation that its tumbling properties and the intramolecular
dipolar interactions will be a common feature for the other
compounds described in this article. The alkylation products of the
nitrogen or oxygen sites, formed under various conditions, were
characterized by ESI MS, 1H/13C-HMBC and -HSQCAD NMR.
Theseresultsshedlightontherecentlydisclosedproblemoflack
of the TPT stability in the heart tissue homogenate samples[20,21]
using methanol in the analytical assays developed for this class of
compounds.
In order to evaluate the correlation time for the TPT and
intramolecular NOEs in buffered water solution at pH 5, T1 and
NOE measurements were performed at room temperature and at
13 and 3 ◦C. As a reference, a TPT solution in DMSO-d6/CDCl3 (2 : 4,
v/v ratio) was also studied (Table S2, Supporting Information).
The change of T1 values with the temperature variation is a
sensitive gauge of the molecular tumbling and, indirectly, of the
state of aggregation. We do not intend to quantitatively discuss
the absolute T1 values of aromatic versus aliphatic protons. This is
because different relaxation mechanisms can dominate in both,
i.e. chemical shift anisotropy for the protons attached to the
extended aromatic plane and spin rotation for the side methyl
groups. The data cited in Table 1 clearly show the characteristic
tendencies for large molecules outside the extreme narrowing
regime. With decreasing temperature, the T1 of the ring protons
7-H, 12-H, 11-H and 18-CH2 increases whereas for the 19-CH3
protons, it is essentially constant and shows the opposite trend
for the dimethylamino group. These results indicate that the
main skeleton protons are characterized by the ωτc ꢂ 1, i.e.
fall into regime of large molecules which exhibit the negative
NOEs. The side group has apparently a much shorter correlation
time and its rotational correlation time is characterized by the
condition ωτc ꢃ 1, observed in the region of extreme narrowing
and positive NOEs. Lowering the temperature is accompanied
by the increase in the solvent viscosity. This effect should result
in shortening of all relaxation times, likewise the aggregation,
whereas opposite trends are still clearly visible. Therefore, the
Results and Discussion
Structure of TPT × HCl in water
The structure in water, as being close to physiological medium,
is of primary importance. At pH 4.5, after dissolving, TPT × HCl
is ca 95% in the lactone form [see Scheme 2), chemical shifts are
pH-sensitive (compare data in Table 3 at pH 6.0)]. We studied 1H-T1
relaxation times and NOE to address the aggregation state and
the molecular tumbling in this medium.
Scheme 2. 1-H chemical shifts of TPT × HCl in D2O, pH 4.5.
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 575–584

Topotecan NMR and chemistry
The above experiments suggest the aggregation of TPT in
buffered water solution at ambient temperatures. The slow
molecular tumbling of highly solvated monomeric molecule,
although possible theoretically, is less probable in the case of low
molecular weight structure (MW = 421). The observed long-range
effects, for example, from 11-H to 18-CH2, (Table 2) interpreted
here as intermolecular effects, can, in principle, be seen as spin
diffusion artefacts. Again, this is highly unlikely due to the fact
that both groups are very distant in a molecule and there is no
intervening proton between them which might ease the spin
diffusion.
Table 1. Comparison of the T1 relaxation times in TPT at different
temperatures in buffered D2O solutiona
T1 (s)
13 ◦C
Proton/
Temperature
30 ◦C
3 ◦C
7-H
0.72 0.10
0.89 0.03
0.97 0.03
12-H
1.93, 2.01 0.54 2.40, 2.45 0.11 2.37, 2.34 0.09
1.76, 1.89 0.44 2.34, 2.41 0.11 2.37,2.47 0.11
11-H
14-H
2.91 0.65
3.08 0.13
2.70 0.11
Nme2
18-CH2
19-CH3
0.51, 0.52 0.01 0.44, 0.44 0.01 0.41, 0.41 0.01
0.34, 0.35 0.03 0.43, 0.43 0.01 0.49, 0.50 0.01
With regard to the problem of aggregation, an earlier paper
citestheself-associationbinding constantforirinotecan2.41×104
[M−1], forming dimers at the level of 2 mM in water, as obtained
from UV spectra.[25] In a recent paper, we have established
0.72 0.10
0.89 0.03
0.97 0.03
a The solution of 0.3 mg of TPT in 600 µl D2O buffer; 50 mM in NaCl,
50 mM in K3PO4, pH 5. Signals close to water; 5, 23, 17-CH2 are not cited.
The T1 relaxation times are cited for each component of multiplet.
TPT
that the self-association constant Ka
= 3.4 × 103 [M−1] from
the concentration dependence of TPT chemical shifts.[17] In the
Supporting Information, the temperature dependence of TPT
chemical shifts is given, which were obtained from the DNMR
observed effects cannot be attributed solely to the increase of
solvent viscosity and/or aggregation but also, as expected, to
different correlation times of the skeleton and side groups in
the molecule. This is evident from comparison of the T1 values
at 30 and 13 ◦C whereas the comparison of the values at 13
and 3 ◦C suggests that both contributions compete effectively.
While we do not attempt to quantify individual shares of
each effect, we only state qualitatively that these experiments
suggest that at ambient temperatures, and in a buffered water
solution, the TPT molecules form molecular clusters. These clusters
tumble slowly in solution which should give rise to nulled
NOEs.
experiment performed in this work and yield the self-aggregation
TPT
constant of ca Ka
= 3.0 × 103 [M−1]. The intermolecular NOEs
seen in Table 2 suggest head-to-tail orientation of TPT molecules
in aggregate. A more detailed analysis of the aggregate geometry
was not attempted because the observed NOEs are very small and,
in addition, they are derived from conformational ensemble in fast
exchange.
We present this information which can have application in
biological studies, DNA binding experiments and pharmaceutical
tests for medicine bioavailability.
In Table 2, the steady state NOEs for D2O solution of TPT at 30,
13 and 3 ◦C are listed. DPFGSE NOE[22] and ROE data are cited
at 30 ◦C for comparison. As expected, Table 2 shows that in D2O
solution at 30 ◦C, the intramolecular effects are nearly nulled and
at the same time intermolecular NOEs appear for all parts of the
TPT molecule. The selective DPFGSE ROE in a few cases show
larger positive intramolecular effects, in contrast to DPFGSE NOE
performed with the same timing. With decreasing temperature,
both intramolecular (11-H vs 12-H) and intermolecular effects
tend to be larger and negative. This is a clear evidence that self-
association of TPT in buffered D2O solution occurs even at very
low concentrations. The change of NOE from a small positive
value at 30 ◦C to a negative one at 13 ◦C evidences that the
molecular tumbling of TPT cluster approaches the correlation
time characteristic for biomolecules. This also confirms that
these molecular clusters have the rotational correlation time,
at ambient temperatures and at 500 MHz, fulfilling the condition
ωoτc ≈ 1.12 giving the correlation time, τc, of the order of
3.55 × 10−10 s.
TPT structure in CD3OD solution
The information on the medicinal behavior in different solvents is
important for biological assays and pharmaceutical formulations.
Methanol is used in the analytical assays for the heart tissue
homogenate samples developed for this class of compounds. For
this reason, we present the TPT reactivity in this solvent in the
following.
Two chemical processes are disclosed in CD3OD solution of
TPT × HCl which have very different time dependencies, i.e.
substitution of the –NHMe2+ Cl− group yielding 23-CH2 –OCD3,
Fig. 1, and deuteration of 5-CH2 group which undergoes through
intermediate enol species formation as shown in Scheme 3
and Fig. 2. Figures 1 and 2 show that the NHMe2+ Cl− group
substitution is completed within 10 days and is displayed by
the spectral process evidence presented in Fig. 1. The most
diagnostic feature of the process is the disappearance of the
signal at δ = 4.78 ppm (23-CH2 –N) and the rise of the signal at
δ = 5.01 ppm (23-CH2 –O). Another clue is the disappearance of
the signal due to Me2ND+ Cl−+ group and the rise of the signal
at 2.72 ppm, due to Me2ND2 Cl− group. This process is also
readily confirmed by high-frequency shift of both nuclei in the
23-13C1H2 group with respect to their positions in parent TPT ×
HCl (Table 3). It can also be deduced from the data (Table 3) that
the ether is in the lactone form. This is done by comparing its
21-13C = O chemical shift and the lack of proton non-equivalence
in 18-CH2 group which is characteristic for carboxylate but not for
the lactone form in all solvents used.
The reference sample in CDCl3/DMSO-d6, 4 : 2 v/v shows no
steady state NOE effects at room temperature whereas DPFGSE
ROE (Table S2) shows expected intramolecular effects but no
intermolecular effects. Theoretical intramolecular NOEs are given
for comparison. They were calculated using the geometry and
interproton distances as obtained using protocol described
previously.[23,24] Table S3 shows the comparison of the spectral
parameters obtained after the experiments were carried out in
water and DMSO-d6/CDCl3 solutions of TPT. There are negligible
differences in chemical shifts despite the difference of dielectric
constants of both solvents. Although it is unexpected, the TPT
chemical shifts are insensitive to the solvent and concentration
changes.
Sanna et al.[26,27] have postulated the amide–enaminoenol tau-
tomerism on rings C/D junction at pH < 4 on the basis of
theoretical calculations, but it was not documented experimen-
tally. Wehaveobservedthatafterprolongedstanding ofTPT×HCl
c
Magn. Reson. Chem. 2010, 48, 575–584
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

K. Hyz et al.
Table 2. Comparison of the NOE enhancements in TPT in buffered water solution at different temperaturesa
Proton irrad.
Proton observed NOE (%)
14-H
7-H
12-H
11-H
23-CH2
18-CH2
–
19-CH3
–
7-H
−1.0
−0.9
−1.5
−1.0
0.7
(0; 16.9)
–
∗
–
−1.3
−2.5
−0.6
–
−1.0
−0.8
−1.5
−2.9
−2.0
3.2
−1.6
−19.6
12-H
11-H
14-H
−0.6
−1.2
−3.2
2.9
−1.1(−1.0; 10.8)
−18.4
−3.2
−3.2
−4.5
−4.9
−4.0
−5.6
−2.4
−3.7
−3.1
4.1
∗
−33.7
−8.2
1.2(0.8; 7.0)
−16.8
−29.3
1.4
∗
−0.5
−0.7
2.9
–
0.5
2.0
−1.1
2.7
–
2.8
∗
2.1
0.4
−0.8
−5.9
–
0.6
−1.4
–
−5.9
0.4
–
0.9
b
23-CH2
–
–
–
–
∗
–
–
–
–
–
−54.9
0.5
−7.3
0.2
−8.9
0.9
−3.2
3.5
−3.0
−2.4
5.7(3.8; 7.0)
0.4
18-CH2
19-CH3
∗
−1.6
−2.1
1.1
−2.3
−6.7
3.9
−1.4
1.2
−6.7
1.1
−4.9
4.0(2.3; 3.5)
−1.8
∗
−2.7
−3.3
–
−1.9
−3.1
−3.4
−6.8
−2.1
a The DPFGSE NOE and ROE are cited at 30 ◦C for comparison (round brackets). The solution of 0.3 mg TPT in 600 µl D2O buffer; 50 mM in NaCl, 50 mM
in K3PO4, pH 5. The values are cited from the top to the bottom at 30, 13 and 3 ◦C, respectively. Positions not cited have less than 0.4 absolute
value effect. 23-CH2 resonance was not irradiated at 30 and 13 ◦C due to the overlap with residual HOD signal. Although the precision of the NOE
enhancements is less than a tenth, we present them here in this precision in order to have better comparison of temperature effect on NOEs. The data
are not used for structural assessments.
Figure 1. Time dependence of the –NDMe2+ Cl− substitution in TPT × HCl in CD3OD. The presence of the OCD3 group is evidenced in the HMBC
spectrum.
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 575–584

Topotecan NMR and chemistry
Scheme 3. ETHER form as a product of deuteration in methanol-d4 and an ENOL form as intermediate species in deuteration of TPT × HCl.
comprising 5-CH2, 7-H, 11-H and 12-H protons is shown in
Figure S1.
Reaction of TPT on silica in ethyl acetate/methanol
TPT × HCl cannot be purified using silica gel. The reaction
described above in methanol-d4 proceeds instantaneously on
developing the silica plates with ethyl acetate/methanol (4 : 1,
vol%). Silica catalyzes the elimination reaction which results in TPT
derivatization yielding an ether 23-CH2OCH3. The compounds are
characterized in Table 3 by 1H and 13C NMR; Table 3 comprises the
data for the various TPT forms in the studied media.
Quaternization reactions of TPT by various alkylating agents
Following the clinical application of the Hycamtin and
Camptosar (TPT and irinotecan, respectively, Scheme 1) several
Figure 2. The 5 ppm region of the 1H NMR of spectrum of TPT × HCl in
groups became interested in structural modifications of CPT[3,28,29]
CD3OD after a few weeks.
with the goal to improve its water solubility and cytotoxicity ver-
sus tumor cells. We have attempted to synthesize the quaternary
inmethanol-d4 solution, the5-CH2 signaldisappearsgraduallyand
after few months, the process of deuteration is nearly complete
(Fig. 2).
This can be interpreted as a result of the presence of the enol-
type tautomer shown in Scheme 3. This tautomer is present only
as an intermediate species, most probably due to its higher energy
than that of the parent amide. The dynamics of this process is
very slow. This confirms the theoretical calculations according to
which there is an energy difference of 100 kJ/mol in favor of the
keto form.[27]
trimethylammonium salt using CH3SO2OCH3 and CF3SO2OCH3 as
alkylating agents. It turned out that, in case of these reagents,
the reaction was completed, yielding ca 10–50% of trimethy-
lammonium salts. In addition, the reaction yielded at least two
other types of TPT derivatives, i.e. expected N-methylation and
double methylation, 23-N and 10-O, or 20-O doubly methylated
products (m/z 450). More importantly, N-methylated product,
characterized after workup of reaction mixture by MS (m/z
436.1) and 1H/13C-HSQCAD, -HMBC NMR, turned out to be un-
stable. For this reason, we did not attempt to separate the
individual compounds but assigned their structures using ESI
MS, 1H NMR, and HSQCAD and 1H/13C-GHMBC NMR in DMSO-
d6 solution. Figure 3 and Scheme 4 show the 1H NMR spectra
which were used in spectral assignment and the structures with
assigned proton/carbon resonances, respectively. We have la-
beled compounds in the reaction mixture according to product
structure.
Each structure in Scheme 4 is confirmed by a few cross-peaks
which characterize the polarization transfer in HMBC spectrum
used to connect given 23-CH2 group to the carbon atom in
ring A, 10-C. The assignment of the structures is accomplished
with the aid of a spectrum of TPT × HCl in DMSO-d6 which
shows the alcohol after standing several months. The other
parts of the spectra are shown in Fig. S1 (3.5–5.5 ppm region)
Interestingly, both processes are not observed instantaneously
in D2O solution at pH ∼6.
The bottom spectrum in Fig. 1 shows also the second set of
signals which is tentatively assigned to the carboxylate form
which cannot be characterized due to its very small population.
Figure 2 shows the spectral result of the second process, i.e.
5-CH2 deuteration proceeding via intermediate enol (Scheme 3).
The triplet at δ = 5.3 ppm is a residual di-hydrogenated 5-CH2
group characterized by 2Jgem = −19 Hz, ꢀυgem = 8 Hz, and long-
range couplings to H-7, 0.9 Hz. The two broad, low-frequency
signals are due to stereoselectively monodeuterated 5-CHD
2
groups; each signal is characterized by reduced Jgem/6.52 Hz
geminal coupling, spin coupling to deuterium and long-range
coupling to 7-H proton. The simulation of the full spin system
c
Magn. Reson. Chem. 2010, 48, 575–584
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

K. Hyz et al.
Figure 3. The aromatic part of the 1H NMR spectrum of the product mixture (DMSO-d6 solution) in the quaternization reaction of neutral TPT with
CF3SO2OCH3, spectrum A. The product composition after 1 month reaction, spectrum B. The TPT × HCl spectrum in DMSO-d6, a month after dissolving,
is shown for comparison, spectrum C. The signals are labeled with the acronyms characterizing the type of changes in ring A.
and Fig. S2 (0.3–3.3 ppm region). The structural evidence of Conclusion
the [Me3N+ CF3SO3−] salt and [10-OMe, Me3N+ CF3SO3−] salt
The tautomeric structure and stability in solution, in different
is backed by the presence of signal at 3.12 ppm belonging to
media CH3OH, DMSO and water, has been established for the
Me3N+ group (Fig. S2), and 1H and 13C NMR chemical shifts
TopoI inhibitor in clinical use, hycamtin. We have shown in the
present work that TPT tumbles in water solution as aggregate
(Scheme 4).ThechemicalprocesswhichisevidencedinFig. 1isthe
substitution in the [Me3N+ CF3SO3−] salt at 23-C position, yielding
with the correlation time characteristic for large molecules at
the alcohol 23-CH2OH and the [Me3HN+ CF3SO3−] salt (Fig. S2,
decreased temperature, ca 5 ◦C, but gives nulled NOEs, both in
−
doublet at 2.73 ppm). Interestingly, in the [Me2HN+ CF3SO3
]
water or DMSO-d6/CDCl3 solutions at room temperature. The TPT
× HCl is unstable in CH3OH solution and two processes were
disclosed, i.e. substitution of the (CH3)2HN+ –group yielding the
ether 23-CH2 –OCH3 and the presence of enamine–enol form
(Scheme 2) at ring C/D junction. This was evidenced by the
deuteration of the 5-CH2 group. This is the first experimental
evidenceoftheexistenceofthistypeoftautomer,whichwasearlier
predicted theoretically.[25] The alkylation of neutral TPT using
CF3 –SO2 –OCH3 yields 10-O-, 20-O- and 22-N-alkylated products,
and N,O-doubly methylated products with the predominance of
the latter, i.e. 22-N-alkylated product. The simultaneous presence
of both N-dimetylammonium- and N-trimetylammonium-triflates
in DMSO-d6 solution allows the observation that the latter is
much more susceptible to substitution. Present results on the
TPT instability in methanol give an explanation for the observed
degradation of TPT in the heart tissue homogenates on storage at
room temperature in methanol for an extended period of time.[21]
Further studies on the chemistry of derivatives of the CPT family
are in progress.
salt, this substitution in the [Me2HN+ CF3SO3−] group does not
proceed so easily, as confirmed by the presence of negligible
amount of alcohol in spectrum C representing the process of
[Me2HN+ Cl−] salt incubation in DMSO-d6 for few months.
In Fig. 4, the 1H/13C-HMBC evidence for the structure of the
[Me3HN+ CF3SO3−]saltofTPTinsolutionispresented. The1H/13C-
HMBC evidence for the structure of the [Me3N+ CH3SO3−] salt of
TPT in solution is presented in Fig. 5.
Figure 5 shows the long-range correlation from methyl groups
in (13CH3)3N+ CH3SO3 ion pair (δ = 3.11/δ = 53.1 ppm and
−
δ = 3.05/δ = 45.5 ppm for cation and ion, respectively) to 23-
C1H2 methylene group (δ = 53.9 ppm) identified in the 1H trace at
δ = 4.93 ppm.Inadditiontothat,themethylgroup,(C1H3)314N+ –,
in trimethylammonium salt, at δ 3.11 ppm, is easily identified by
small coupling, 3J(1H, 14N) ca 2 Hz to 14N nucleus, often observed
in nitrogen quaternary salts due to symmetric environment of
nitrogen atom.
With respect to the above evidence of structure, we were not
able to reconcile it with the result which cites the chemical
−
shift of (CH3)3N+ –group at δ = 2.94 ppm and CH3SO3 at
Experimental
δ = 2.78 ppm for the same compound but in methanol/CDCl3
solution (vol. ratio not given) with no information on 13C.[30] Our
experience is that in methanol solution the quaternary salt is not
stable.
TPT hydrochloride was purchased from Alexis Biochemicals and
was used directly without applying any additional purification
steps. To obtain free base TPT, TPT hydrochloride (2 mg,
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 575–584

Topotecan NMR and chemistry
Table 3. 1H and 13C chemical shifts, ppm, of various TPT tautomers and derivatives in D2O, methanol-d4 and DMSO-d6 solutions
TPT × HCla
Lactone
D2O, pH 6
TPT × HClb
Carboxylate
D2O, pH 10
TPTd
Ether
23-CH2OCH3
TPTe
Lactone
DMSO-d6
TPTe
Carboxylate
DMSO-d6
TPTf
Alcohol
DMSO-d6
TPTc
Free base
2-C
150.8
147.9
148.5
146.9
148.9
148.0
147.3
148.4
150.2
146.1
149.6
146.6
149.5
146.4
3-C
5-CH2
6-C
5.20; 53.4
133.0
5.04; 53.4
n.o.
5.25; 51.2
n.o.
5.23; 51.3
n.o.
5.28, 50.6
131.2
5.26, 50.7
130.3
5.25; 50.8
130.1
7-CH
8.65; 128.5
131.5
8.25; 126.0
132.4
8.61; 126.4
130.6
8.52; 127.0
132.3
8.91, 126.8
n.o.
8.75, 127.7
129.5
8.75; 127.9
129.6
8-C
9-C
110.8
109.9
113.3
115.8
109.1
109.4
119.6
10-C
160.5
169.4
162.4
165.1
n.o.
n.o.
154.2
11-CH
12-CH
13-C
7.47; 125.4
7.94; 134.6
145.3
7.25; 131.1
7.85; 133.3
143.7
7.36; 12 6.4
7.95; 131.0
144.6
7.36; 128.9
7.87; 130.8
143.9
7.64, 122.5
8.19, 133.5
144.0
7.49, 123.0
7.97, 130.5
144.2
7.48; 123.2
7.96; 130.6
144.1
14-CH
15-C
7.37; 101.2
153.4
7.50; 103.5
159.7
7.60; 98.7
153.0
7.59; 98.3
153.0
7.29, 96.6
150.7
7.25, 96.2
n.o.
7.25; 965
150.6
16-C
120.9
129.0
119.2
118.5
119.2
119.0
118.6
16a-C(O)
17-CH2
18-CH2
19-CH3
C-20
n.o.
165.2
159.5
159.5
157.5
n.o.
157.3
5.37, 5.50; 68.4
1.96, 33.5
0.96; 9.6
76.2
4.84; 58.7
2.20, 2.38; 34.2
1.07; 10.2
82.2
5.37, 5.57; 66.7
1.96; 32.6
1.01; 8.15
74.3
5.37, 5.57; 66.7
1.96; 32.6
1.01; 8.15
74.3
5.29, 5.42; 65.6
1.86; 30.7
0.87, 8.2
73.0
5.40, n.o.
1.38, 26.3
0.82, n.o.
n.o.
5.39; 65.7
1.84;30.7
0.86; 8.3
72.9
21-C(O)
22-CH3(N)
23-CH2
Other
177.4
181.1
175.1
175.1
173.1
n.o.
173.0
2.91; 45.1
4.68; n.o.
–
2.85; 44.5
4.51; 56.7
–
2.49; 44.5
4.19; 56.7
–
2.84, 43.0
4.72, 51.0
n.o.
–
4.98; 66.3
OCH3 3.44; 57.8
4.95, 53.9
4.94; 54.2
a D2O solution pH = 6.
b D2O solution pH = 10.
c Obtained from TPT × HCl, see Experimental section, in DMSO-d6 solution.
d Product of the reaction on gel using ethyl acetate: CH3OH eluent, CD3OD solvent.
e DMSO-d6.
f Product of quaternization reaction, see Scheme 4, DMSO-d6.
4.4 × 10−3 mmol) was dissolved in H2O (0.5 ml) and pH was
set to 7 using NaHCO3 aqueous solution. Free TPT was extracted
with CH2Cl2 and dried under vacuum.
1HNMR (CD3OD), just after dissolving: δ = 1.01 (t, 3H, J = 7 Hz),
1.96 (q, 2H, J = 7 Hz), 2.90 (s, 6H), 4.74 (s, 2H), 5.35 (bs, 2H), 5.38 (d,
1H, J = 16 Hz), 5.59 (d, 1H, J = 16 Hz), 7.57 (d, 1H, J = 9 Hz), 7.64
(s, 1H), 8.21 (d, 1H, J = 9 Hz), 8.81 (s, 1H).
1H NMR (CD3OD) after 10 days: δ = 1.01 (t, 3H, J = 7 Hz), 1.96
(q, 2H, J = 7 Hz), 2.68 (s, 6H = (CH3)2N × HCl), 5.00 (s, 2H), 5.31
(bs, 0.5 H, partly deuterated), 5.38 (d, 1H, J = 16 Hz), 5.58 (d, 1H,
J = 16 Hz), 7.47 (d, 1H, J = 9 Hz), 7.63 (s, 1H), 8.03 (d, 1H, J = 9 Hz),
8.74 (s, 1H).
Reaction procedures
General procedure of TPT reaction on silica gel
TPThydrochloride(2 mg,4.4×10−3 mmol)wasdissolvedinCH3OH
(1 ml), applied to a PLC glass plate (silica gel, 2 mm) and dipped
into a mixture of CH3OH and CH3CO2Et (1 : 9, v/v). Two fractions
were collected and eluted from gel with CH3OH. Fraction with
Rf = 0.7 was recognized as impurities while the one with Rf = 0.1
as a methyl ether, 23-CH2OCH3, C22H21N2O6.
The aromatic part of the spectrum differs from that obtained
after the reaction of TPT on silica, most probably due to the
deuteration of a methyl group and/or due to the differences in pH
and concentration of the sample.
HR-MS (ESI): Calculated for C22H19D3N2O6, [M + H]+ 412.1582,
1H NMR (CD3OD): δ = 1.01 (t, 3H, J = 7 Hz), 1.96 (q, 2H,
J = 7 Hz), 3.44 (s, 3H), 4.98 (s, 2H), 5.22 (residual signal because
of deuteration), 5.37 (d, 1H, J = 16 Hz), 5.55 (d, 1H, J = 16 Hz),
7.36 (d, 1H, J = 9 Hz), 7.59 (s, 1H), 7.87 (d, 1H, J = 9 Hz), 8.54
(s, 1H).
found: 412.1664.
General procedures for alkylation of TPT
CF3SO3CH3 (methyl trifluoromethanesulfonate)
HR-MS (ESI): Calculated for C22H20N2O6, [M − H]+ 407.1237,
Methyl trifluoromethanesulfonate (0.1 ml, 0.9 mmol) was dis-
solvedinCH2Cl2 (3 ml)andthensmallamountofK2CO3 wasadded.
The mixture was stirred for 30 min to neutralize possible formation
of trifluoromethanesulfonic acid (CF3SO3H). The solution was fil-
tered from K2CO3, added to dry free TPT (1.8 mg, 4.4×10−3 mmol)
and stirred for 24 h at room temperature. The mixture was evapo-
ratedtodrynessundervacuumatroomtemperatureanddissolved
found: 407.1231.
Reaction in methanol
TPT hydrochloride (0.5 mg, 1.1 × 10−3 mmol) was dissolved in
CD3OD (0.7 ml) and incubated at 4 ◦C for 10 days. Product of this
reaction was identified as an ether C22H18D3N2O6.
c
Magn. Reson. Chem. 2010, 48, 575–584
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

K. Hyz et al.
Scheme 4. 1H and 13C chemical shifts of species characterized in situ after alkylation of TPT × HCl.
1
in deuterated DMSO-d6. In the H NMR spectrum, five products
are observed: N-methylated TPT, N,O-dimethylated TPT, proto-
nated TPT [TPTH+ CF3SO3−], alcohol 23-CH2OH. N-methylated
TPT constitutes up to 50% of the mixture.
13C, and 15N nuclei, respectively. An amount of 0.1–0.3 mg of
the compound was dissolved in 0.7 ml of H2O/D2O (9 : 1) and
transferred to the 5-mm NMR tube. Chemical shifts, δ, given
in ppm, were referenced against internal reference TSPA. The
spectrometer was equipped with indirect detection 5 mm Nalorac
probe with an actively shielded z-gradient coil (ID-PFG).
1H NMR spectra were run using standard method of presatura-
tion of residual water implemented in Varian software.
DPFGSE NOE experiments were run using a pulse sequence
published by Stott et al.[22] using shaped selective π soft pulses
generated by a standard Varian program. The rsnob π pulse was
calibrated for each individual multiplet.
LR-MS (ESI): m/z 149 [CF3SO3]−, 422 [M + H]+, 436 [M]+,
450 [M]+.
CH3SO3CH3 (methyl methanesulfonate)
Methyl methanesulfonate (0.1 ml, 1.2 mmol) was dissolved in
CH2Cl2 (3 ml) and then a small amount of K2CO3 was added.
The mixture was stirred for 30 min to neutralize methanesulfonic
acid (CH3SO3H), if present. The solution was filtrated from K2CO3
and added to dry TPT (1.8 mg, 4.4 × 10−3 mmol). The mixture
was stirred for 24 h at room temperature. Then, the mixture was
evaporated to dryness under vacuum (at 50 ◦C) and dissolved in
deuterated DMSO. In the 1H NMR spectrum, four products are
observed: N-methylated TPT, N,O-dimethylated TPT, protonated
TPT [TPTH+ CH3SO3−], and alcohol, 23-CH2OH. N-methylated TPT
constitutes less than 20% of the mixture.
DPFGSE ROE experiments were run using a pulse program
obtained from W. Koz´min´ski (Warsaw University) using shaped
selective π soft pulses (rsnob shape, 50 Hz selectivity, ca 30 ms
pulse duration), 3 kHz spin lock field, in cw mode, of 500 ms
duration (mixing time) was centered on water resonance.
1D NOE spectra of TPT solutions were measured on 500 MHz
Varian INOVA and were acquired in steady state mode using 15 s
preirradiation time of each multiplet, 6 kHz spectral window,
300 transients, and 2 s acquisition time. Absorption mode
spectra were analyzed by comparing with dummy irradiation
spectrum and considering the saturation factor ranging from 0.6
to 0.9.
LR-MS (ESI): m/z 422 [M + H]+, 450 [Mꢁꢁ]+; 95 [CH3SO3]−.
NMR experiments
The NMR spectra were recorded at 303 K on Varian INOVA 500
spectrometer operating at 499.8, 125.7 and 50.51 MHz for 1H,
1H-T1 measurements were performed using the standard
inversion-recovery protocol implemented in Varian software.
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 575–584

Topotecan NMR and chemistry
Figure 4. The overlay of the aliphatic part of the 1H/13C-HSQCAD-HMBC spectrum (colored/black) of the one of the product mixtures (DMSO-d6 solution)
in the quaternization reaction of neutral TPT with CF3SO2OCH3. The colored HSQCAD signals are linked with the black HMBC peaks evidencing the
attachment of respective methyl groups to 23-CH2 carbon atom. The HSQCAD signal in purple has no correlation and is assigned to a salt.
−
Figure 5. Overlay of the 500 MHz phase-sensitive 1H/13C-HMBC trace (blue–red) and absolute value 1H/13C-HSQCAD trace (green) of [Me3N+ CH3SO3
salt of TPT in DMSO-d6 solution. CH3 –SO2 –OCH3 was used for the quaternization reaction.
]
1H/13C-HSQCAD spectra:[31] The echo–antiecho phase-sensitive
gradient-selected 1H–13C HSQCAD (heteronuclear single quan-
tum coherence adiabatic version) NMR spectra were obtained
with a spectral width of 5000 Hz, 2048 points in the 1H dimension
and 8000 Hz, 800 × 2 increments in the 13C dimension; 128 tran-
sients per t1 increment, with a relaxation delay of 1 s and 1J(C,H) =
135 Hz. The data were linearly predicted to 1600 points and zero
filled to 4096 points in F1 before Fourier transformation.
1H/13C-HMBC: The gradient-selected 1H–13C HMBC spectra
were performed with an acquisition time of 0.2 s, 1H – 90◦ pulse
width of 7.8 µs, 13C – 90◦ pulse width of 11.5 µs, a spectral width
of 5000 Hz, 2048 data points in the 1H dimension and 25 000 Hz,
1024 increments in the 13C dimension, and a relaxation delay of
1.2 s. The data were acquired as absolute value mode, with 64
transients per t1 increment. The experiment was optimized for
nJ(C,H) = 8 Hz and low pass filter for 1J(C,H) = 140 Hz was used.
The data were linear predicted to 2048 points and zero filled to
4096 points in F1 prior to Fourier transformation.
Acknowledgements
The authors acknowledge the financial support during this
research from Ministry of Science and Higher Education Grant
No. NN204 155736. The authors thank Prof. S. Szyman´ski (Institute
of Organic Chemistry, Polish Academy of Sciences) for the detailed
spin simulation (Fig. S1). The authors acknowledge the help of Dr
K. Zdanowski in preparing the manuscript.
c
Magn. Reson. Chem. 2010, 48, 575–584
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

K. Hyz et al.
[15] L. P. Rivory, J. Robert, Bull. Cancer 1995, 82, 265.
[16] M. R. Redinbo, L. Stewart, P. Kuhn, J. J. Champoux, W. G. Hol, Science
Supporting information
1998, 279, 1504.
[17] W. Bocian, R. Kawe¸cki, E. Bednarek, J. Sitkowski, A. Pietrzyk, M. P.
Williamson, P. E. Hansen, L. Kozerski, Chem. Eur. J. 2004, 10, 5776.
[18] W. Bocian, R. Kawe¸cki, E. Bednarek, J. Sitkowski, M. P. Williamson,
P. E. Hansen, L. Kozerski, Chem. Eur. J. 2008, 14, 2788.
[19] S. Streltsov, V. Oleinikov, M. Ermishov, K. Mochalov, A. Sukhanova,
Y. Nechipurenko, S. Grokhovsky, A. Zhuze, M. Pluot, I. Nabiev,
Biopolymers 2003, 72, 442.
Supporting information may be found in the online version of this
article.
References
[20] N. A. de Vries, M. Ouwehand, T. Buckle, J. H. Beijnen, O. van
[1] P. B. Arimondo, G. S. Laco, C. J. Thomas, L. Halby, D. Pez, P. Schmitt,
A. Boutorine,T. Garestier,Y. Pommier,S. M. Hecht,J. S. Sun,C. Bailly,
Biochemistry 2005, 44, 4171.
Tellingen, Biomed. Chromatogr. 2007, 21, 1191.
[21] N. R. Srinivas, Biomed. Chromatogr. 2009, 23, 447.
[22] K. Stott, J. Stonehouse, J. Keeler, T.-L. Hwang, A. J. Shaka, J. Am.
Chem. Soc. 1995, 117, 4199.
[2] K. Cheng, N. J. Rahier, B. M. Eisenhauer, R. Gao, S. J. Thomas,
S. M. Hecht, J. Am. Chem. Soc. 2005, 127, 838.
[23] T. Thanh Danh, W. Bocian, L. Kozerski, P. Szczukiewicz, J. Frelek,
M. Chmielewski, Eur. J. Org. Chem 2005, 429.
[3] N. J. Rahier, B. M. Eisenhauer, R. Gao, S. H. Jones, S. M. Hecht, Org.
Lett. 2004, 6, 321.
[24] M. P. Williamson, Mag. Reson. Chem. 1987, 25, 356.
[25] R.Aiyama,H.Nagai,S.Sawada,T.Yokokura,H.Itokawa,M. Nakanishi,
Chem. Pharm. Bull. 1992, 40, 2810.
[26] N. Sanna, G. Chillemi, A. Grandi, S. Castelli, A. Desideri, V. Barone,
J. Am. Chem. Soc. 2005, 127, 15429.
[27] N. Sanna, G. Chillemi, L. Gontrani, A. Grandi, G. Mancini, S. Castelli,
G. Zagotto, C. Zazza, V. Barone, A. Desideri, J. Phys. Chem. B 2009,
113, 5369.
[4] A. Cagir, S. H. Jones, R. Gao, B. M. Eisenhauer, S. M. Hecht, J. Am.
Chem. Soc. 2003, 125, 13628.
[5] P. Pourquier, Y. Pommier, Adv. Cancer Res. 2001, 80, 189.
[6] F. Zunino, G. Pratesi, Expert Opin. Investig. Drugs 2004, 13, 269.
[7] L. F. Liu, P. Duann, C. T. Lin, P. D’Arpa, J. Wu, Ann. NY Acad. Sci. 1996,
803, 44.
[8] C. H. Takimoto, J. Wright, S. G. Arbuck, Biochim. Biophys. Acta 1998,
1400, 107.
[28] S. Sawada, S. Okajima, R. Aiyama, K. Nokata, T. Furuta, T. Yokokura,
E. Sugino, K. Yamaguchi, T. Miyasaka, Chem. Pharm. Bull. 1991, 39,
1446.
[29] S. Dallavalle, T. Delsoldato, A. Ferrari, L. Merlini, S. Penco, N. Carenini,
P. Perego, M. De Cesare, G. Pratesi, F. Zunino, J. Med. Chem. 2000,
43, 3963.
[9] W. J. Slichenmyer, E. K. Rowinsky, R. C. Donehower, S. H. Kaufmann,
J. Natl. Cancer Inst. 1993, 85, 271.
[10] M. C. Wani, A. W. Nicholas, M. E. Wall, J. Med. Chem. 1987, 30, 2317.
[11] C. Jaxel, K. W. Kohn, M. C. Wani, M. E. Wall, Y. Pommier, Cancer Res.
1989, 49, 1465.
[12] R. P. Hertzberg, M. J. Caranfa, S. M. Hecht, Biochemistry 1989, 28,
4629.
[30] W. D. Kingsbury, J. C. Boehm, D. R. Jakas, K. G. Holden, S. M. Hecht,
G. Gallagher,
M. J. Caranfa,
F. L. McCabe,
L. F. Faucette,
[13] R. P. Hertzberg, M. J. Caranfa, K. G. Holden, D. R. Jakas,
G. Gallagher, M. R. Mattern, S. M. Mong, J. O. Bartus, R. K. Johnson,
W. D. Kingsbury, J. Med. Chem. 1989, 32, 715.
[14] Y. Pommier, G. Kohlhagen, K. W. Kohn, F. Leteurtre, M. C. Wani,
M. E. Wall, Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 8861.
R. K. Johnson, R. P. Hertzberg, J. Med. Chem. 1991, 34, 98.
[31] M. F. Summers, L. G. Marzilli, A. Bax, J. Am. Chem. Soc. 1986, 108,
4285.
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 575–584