Improving cytotoxic properties of ferrocenes by incorporation of saturated N-heterocycles

Article abstract of DOI:10.1016/j.jorganchem.2017.06.005

A family of ferrocene derivatives of the general formula [Fe(η⁵-C₅H₄CH₂(p-C₆H₄)CH₂(N-het))₂] bearing saturated six- and five-membered N-heterocycles (N-het) was prepar

Full text of DOI:10.1016/j.jorganchem.2017.06.005

Accepted Manuscript
Improving cytotoxic properties of ferrocenes by incorporation of saturated N-
heterocycles
Tomáš Hodík, Martin Lamač, Lucie Červenková Šťastná, Petra Cuřínová, Jindřich
Karban, Hana Skoupilová, Roman Hrstka, Ivana Císařová, Róbert Gyepes, Jiří Pinkas
PII:
S0022-328X(17)30400-X
10.1016/j.jorganchem.2017.06.005
JOM 19992
DOI:
Reference:
To appear in: Journal of Organometallic Chemistry
Received Date: 17 March 2017
Revised Date: 5 June 2017
Accepted Date: 7 June 2017
Please cite this article as: Tomáš. Hodík, M. Lamač, L. Červenková Šťastná, P. Cuřínová, Jindř. Karban,
H. Skoupilová, R. Hrstka, I. Císařová, Ró. Gyepes, Jiří. Pinkas, Improving cytotoxic properties of
ferrocenes by incorporation of saturated N-heterocycles, Journal of Organometallic Chemistry (2017),
doi: 10.1016/j.jorganchem.2017.06.005.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Improving cytotoxic properties of ferrocenes by incorporation of saturated N-heterocycles
Tomáš Hodík,a,ǁ Martin Lamač,a Lucie Červenková Šťastná,b Petra Cuřínová,b Jindřich
Karban,b Hana Skoupilová,c Roman Hrstka,c Ivana Císařová,d Róbert Gyepes,a,d Jiří Pinkasa,*
a J. Heyrovský Institute of Physical Chemistry, Academy of Sciences of the Czech Republic,
v.v.i., Dolejškova 2155/3, 182 23 Prague 8, Czech Republic
b Institute of Chemical Process Fundamentals, Academy of Sciences of the Czech Republic,
v.v.i., Rozvojová 135, 165 02 Prague 6, Czech Republic
c Regional Centre for Applied and Molecular Oncology, Masaryk Memorial Cancer Institute,
Žlutý kopec 7, 65653 Brno, Czech Republic
d Department of Inorganic Chemistry, Charles University, Hlavova 2030, 128 43, Prague 2,
Czech Republic
* Corresponding Author: tel. (+420) 266053735; e-mail, jiri.pinkas@jh-inst.cas.cz.
ǁ Present address: Institut für Organische Chemie, Universität Leipzig, Johannisallee 29 D-
04103 Leipzig, Deutschland
† Electronic supplementary information (ESI) available: experimental conditions for
preparation of commercially unavailable starting materials, NMR spectra of new prepared
compounds, crystallographic data for structures 4a and 4a·2HCl.
Abstract
A family of ferrocene derivatives of the general formula
[Fe(η5-C5H4CH2(p-C6H4)CH2(N-het))2] bearing saturated six- and five-membered
N-heterocycles (N-het) was prepared. Reactions of the selected complexes with acids (HCl,
acetic acid) afforded either the corresponding hydrochlorides or led to deprotection of the
functionalized pendant N-heterocycles. The reaction of [{Ru(η6-p-cymene)Cl2}2] with the
corresponding cyclopentadienide derivatives afforded cationic ruthenium complexes
[Ru(η6-p-cymene)(η5-C5H4CH2(p-C6H4)CH2(N-het))]Cl while ruthenocenes
[Ru(η5-C5H4CH2(p-C6H4)CH2(N-het))2] were formed as minor byproducts. The prepared
complexes (20 examples) were characterized by elemental analysis, melting point, NMR and
ESI-MS and the molecular structures of selected ferrocene derivatives were determined by X-
ray diffraction analysis. The ferrocene derivatives and the ruthenium complexes were tested in
vitro for their cytotoxic properties against three cell lines derived from ovarian cancer
(A2780, A2780cis, and SK-OV-3) and against non-tumour embryonic cell line HEK293

ACCEPTED MANUSCRIPT
(human kidney cells). The most active ferrocene derivatives displayed cytotoxicity in
submicromolar and low micromolar concentration against both cisplatin (CisPt) sensitive and
resistant cells. The results showed a significant effect of the pendant N-heterocycle on the
ferrocene derivative toxicity and selectivity against cancer cells. Ultimately, ferrocene
derivatives bearing either piperidine or morpholine groups were proposed to be the most
promising substitutes for platinum drugs, as they exhibited comparable or even higher activity
(in comparison to CisPt) against cancer cells, whereas these compounds were found to exhibit
lower toxicity against embryonic HEK293 cells.
Keywords: Metallodrugs; Ferrocene derivatives; N-heterocycles; Cytotoxicity; Ovarian
cancer.
Introduction
Cancer is a family of diseases characterized by aggressive, uncontrolled growth of
cells, representing one of the greatest causes of death, leading to an estimated 8 million of
deaths occurring in 2012 worldwide.[1] Ovarian cancer is the seventh most common
malignancy among women, and it represents the eighth leading cause of cancer-related death
among women, being the deadliest among gynaecologic tumours.[2] It is a heterogeneous
disease and even though chemotherapy and surgery have improved the outcome for patients
with gynaecologic malignancies over the last 20 years, "women’s cancers" still account
annually for over 10% of cancer related deaths. The current treatment protocol consists in
delivering platinum-drug chemotherapy (e.g. cisplatin(CisPt), oxaliplatin, carboplatin) which
is however frequently accompanied by a number of side-effects. In addition, the therapy is
often hampered by acquired resistance of cancer cells towards platinum-drugs.[3-4]
Therefore, the introduction of new types of active metallodrugs with a different mode of
action in comparison with platinum drugs is strongly desirable.
Among non-platinum coordination (organometallic) compounds, ferrocene derivatives
play a prominent role due to their unique properties, such as thermal and hydrolytic stability,
inclusion of the reversibly redox active Fe2+/3+ center, their simple preparation and
straightforward modification. These properties render ferrocene derivatives very popular for
biological applications, especially as potential cytostatic drugs in cancer treatment, which has
been documented by several reviews.[5-8]

ACCEPTED MANUSCRIPT
In our previous work, we described the new ferrocene derivative containing
1-piperidinyl moiety [Fe(η5-C5H4CH2(p-C6H4)CH2(N(CH2)2CH2))2]·2HCl (4a·2HCl, Scheme
2) which possessed superior cytotoxicity in comparison with CisPt against ovarian (A2780
and A2780cis) and breast (MCF7) cancer cell lines in 24 h assay.[9] In addition, a newly
developed electrochemical measurement allowed us to determine an efficient intracellular
transport of the complex, and to establish that its accumulation occurs predominantly in
cytoplasm.
In this article, we report the continuation of investigation of the promising concept of
ferrocenes bearing saturated N-heterocycles at their periphery. A series of ferrocene
derivatives bearing either substituted or unsubstituted saturated five- and six-membered
nitrogen-containing heterocycles was synthesized. In addition, two nitrogen-free ferrocenes
were prepared to compare their efficacy. We also investigated the possibility of preparing
cationic ruthenium cyclopentadienyl-arene complexes accommodating a saturated
N-heterocycle, while ruthenocene complexes were formed as byproducts. To understand the
role of the particular N-heterocycle on the organometallic complex activity towards tumour
cells and to compare the anticancer activity of our compounds with cisplatin representing the
treatment of choice for ovarian cancer patients, the cytotoxic effects of these newly prepared
complexes as well as of two structurally related organic species against three human ovarian
cancer cell lines (A2780, A2780cis, and SK-OV-3) were investigated. Human embryonic
kidney cell line HEK293 was included in the study to determine tumour selectivity of our
compounds and to exclude incidental nephrotoxicity, which is frequently observed in patients
treated with cisplatin.
Results and discussion
Synthesis of 6-substituted fulvenes 3a − 3h
Fulvenes 3a − 3h were prepared according to the general synthetic Scheme 1 adapted
from literature.[9-10] As starting secondary amines 1 we used simple saturated 5- and
6-membered nitrogen heterocycles: piperidine (1a), morpholine (1b), thiomorpholine (1c) and
pyrrolidine (1e). To probe the influence of hydroxyl substituents, we also employed the
readily available O-protected iminoalditols: O-benzyl- (OBn) and O-methoxymethyl-
(OMOM) protected (3S,4S)-pyrrolidine-3,4-diol (1,4-dideoxy-1,4-iminothreitol) 1f and 1g,
1,4-dideoxy-2,3:5,6-di-O-isopropylidene-1,4-imino-D-talitol 1h,[11] and 2,3,4,6-tetra-O-

ACCEPTED MANUSCRIPT
benzyl-1,5-dideoxy-1,5-imino-D-glucitol (O-benzylated deoxynojirimycin) 1d.[12-13] In
choosing iminoalditols for ferrocene functionalization we were also inspired by a recent
discovery that some iminocyclitol-ferrocene conjugates show cytotoxic properties.[14-15]
R1R2N
Br
R1R2N
O
R1R2NH
O
pyrrolidine, MeOH
N
K2CO3, MeCN
1a - 1h
3a - 3h
2a - 2h
O
O
BnO
N
N
OMOM
OMOM
N
N
N
N
OBn
OBn
R1R2N =
N
H
O
O
b
S
c
BnO
BnO
OBn
e
g
d
h
a
f
O
Scheme 1. Synthesis of 6-substituted fulvenes 3a − 3h
With the set of the required heterocycles in hand, we subjected them to reaction with
4-(bromomethyl)benzaldehyde[16] in acetonitrile in the presence of potassium carbonate to
obtain N-(4-formylphenyl)methyl derivatives 2a − 2h. The pyrrolidine-promoted
condensation with freshly cracked cyclopentadiene according to the procedures by Stone[17]
or Erden[18] afforded 6-substituted fulvenes 3a − 3h. Detailed experimental procedures for
the new fulvenes 3c, 3d, and 3f − 3h are given in Supporting Information.
A cytostatic effect of 6-aryl fulvenes against a wide family of cancer cells was recently
published.[19] Therefore, we prepared a fulvene hydrochloride 3a·HCl by the reaction of 3a
with HCl in THF. In comparison with 3a the hydrochloride showed an increased solubility
both in DMSO and in water, which was advantageous for biological evaluations. We also
prepared the corresponding cyclopentadienyl hydrochloride 3aa·HCl (as a mixture of
isomers) by the reaction of 3a with LiBEt3H, followed by protolysis with an excess of HCl.
Formally, 3aa·HCl represents a free ligand released from 4a·2HCl and the concept was used
for determination of the ligand cytotoxicity and its potential contribution to the cytotoxicity of
ferrocene 4a·2HCl.
Syntheses and characterization of ferrocenes 4a − 4j

ACCEPTED MANUSCRIPT
R
R
i, ii
2 eqiv.
Fe
R
3a - 3j
4a - 4j
OBn
OBn
OBn
O
O
O
O
N
N
a
N
N
(46%)
g
ref.9
BnO
d
(77%)
(77%)
O
N
N
O b
(67%)
N
O
H
e
O
c
S
(48%)
(44%)
OBn
O
(66%)
h
N
CH3
H
(63%)
(62%)
i
j
OBn
f
(64%)
Scheme 2. Preparation of ferrocene derivatives 4a − 4j. Reagents and conditions: (i) LiBEt3H,
THF, rt; (ii) FeCl2, THF, rt. Values in parentheses correspond to isolated yields of 4a − 4j.
Ferrocenes 4b − 4j were prepared by a procedure similar as reported for 4a.[9] In addition to
fulvenes 3b − 3h, we also used heterocycle-free fulvenes 3i and 3j as starting materials. The
reaction of fulvenes 3b − 3j with LiBEt3H in THF afforded the corresponding lithium
cyclopentadienide. This was reacted in situ with solid anhydrous FeCl2 in the same solvent for
10 − 24 h at room temperature as schematically depicted in Scheme 2. The crude products
were purified by chromatography on silica gel and in selected examples recrystallized from an
appropriate solvent (alcohols, THF) or a mixture of solvents (THF/alcohol). Pure ferrocene
derivatives 4b − 4j were obtained in moderate yields (44 − 77%) as yellow to orange solids or
orange waxes depending on the nature of the substituent. It should be noted, that two
alternative routes to ferrocene 4j were already published: either by reaction of thallium
benzylcyclopentadienide with FeCl2, or by reduction of 1,1´-dibenzoylferrocene with borane-
dimethyl sulfide complex.[20-21]
Ferrocenes 4a − 4i showed good solubility in low polar organic solvents (toluene, CHCl3,
THF) and low solubility in organic polar solvents such as DMSO (as an exception, complex
4c was totally insoluble in DMSO, which hampered its biological evaluation), and in organic
protic solvents (EtOH). The solubility of all complexes in water was negligible.
All new prepared ferrocene complexes 4b − 4i were characterized by 1H and 13C
NMR, ESI-MS spectrometry and elemental analysis. In addition, melting points of solid
samples were determined.

ACCEPTED MANUSCRIPT
1H NMR spectra of ferrocenes 4b − 4i show the characteristic presence of two
doublets (3JHH ~ 8 Hz) for aromatic phenylene C6H4 protons positioned in the region 7.1−7.3
ppm. Cyclopentadienide protons C5H4 appeared as a singlet in the region 4.03−4.07 ppm
(with an exception of 4g where two multiplets centered at 4.03 and 4.07 ppm were found). A
signal for methylene group C5H4CH2C6H4 connecting cyclopentadienyl and phenylene
appeared as a singlet in a narrow region for both 1H and 13C NMR spectra (δH/δC = 3.59 −
3.68 ppm/34.82 − 35.73 ppm). A methylene group connecting the phenylene group with a
nitrogen atom displayed as a singlet signal (in the region 3.45 − 3.57 ppm) in 1H NMR spectra
only for achiral ferrocenes 4b, 4c, and 4e. On the other hand, the presence of chiral centers in
ferrocenes 4d and 4f − 4h led to splitting of the methylene group signal into two doublets
(2JHH = 12.9 − 13.8 Hz) centered at 3.37 and 4.01 for 4d, 3.52 and 3.60 ppm for 4f, 3.43 and
3.51 ppm for 4g, and 3.55 and 4.18 ppm for 4h. ESI-MS spectrometry proved a molecular
composition of all prepared compounds with [M + H]+ and [M + Na]+ ions as the most
abundant within the spectra of 4b − 4h. Complex 4i lacking any electronegative atom (such as
O, N) represented the only exception showing just the molecular ion M+ at m/z = 394 as the
only peak of relevance.
Reaction of selected ferrocenes with acids (quaternization, deprotection)
4x + excess HCl → 4x·2HCl (where 4x = 4b, 4c, 4e, 4f) eq. 1.
Reaction (equation 1) of selected ferrocene amines 4b, 4c, 4e, 4f with a slight excess
of anhydrous HCl solution in ether (either 3M in methylcyclopentyl ether or 1M in Et2O)
gave the corresponding amine hydrochlorides 4b·2HCl, 4c·2HCl, 4e·2HCl, 4f·2HCl in
medium to excellent yields (55 − 97%), similarly as reported for 4a·2HCl.[9] The reaction
was advantageously conducted in ether solvents (THF, Et2O), where the starting materials
showed high solubility, whereas the products were almost insoluble and could be simply
isolated by filtration. The amine hydrochlorides were well soluble in polar organic solvents
(DMSO, EtOH, MeOH) and in water. They remained intact in aqueous environment for
several days as was proved by 1H NMR. The formation of hydrochlorides 4·2HCl was
supported by 1H NMR spectroscopy, where the signal of the ammonium proton appeared as a
broad singlet (at 11.37 ppm for 4b·2HCl, 10.93 ppm for 4c·2HCl, 11.27 ppm for 4e·2HCl,
and 11.67 ppm for 4f·2HCl).

ACCEPTED MANUSCRIPT
In the case of 4g, we observed a different reactivity with HCl due to lability of the
protecting methoxymethyl group under acidic conditions. The reaction of 4g with HCl (either
anhydrous or aqueous) gave ferrocene bishydrochloride 5 (Scheme 3) having four free
hydroxyl groups in excellent yields (89 and 95% in two independent experiments). Both
ammonium protons in 5 were easily removed by treatment with anion-exchange resin
maintained in -OH− cycle to give neutral ferrocene alcohol 6 in 63% yield (Scheme 3). As
expected, 5 and 6 were well soluble in alcoholic solvents, while the former one exhibited a
considerably higher solubility in water in comparison with the latter one.
O
O
O
O
N
Fe
N
O
O
O
O
4g
i
H
N+
Cl-
OH
HO
HO
Fe
N+
H
Cl-
OH
5 (95%)
ii
OH
HO
N
Fe
N
OH
HO
6 (63 %)
Scheme 3. Reaction of 4g with an excess of HCl and formation of 5 and 6. Reagents and
conditions: (i) HCl, THF, rt; (ii) anion exchange resin AG-1-X8 (in OH− cycle), H2O, rt.
1H NMR spectra displayed signals attributable to hydroxyl groups as a broad singlet
positioned at 5.8 ppm and 4.8 ppm for 5 and 6, respectively. In IR spectra, valence O-H
vibration was found at 3286 and 3337 cm−1 for 5 and 6, respectively. The signal of
ammonium protons in 5 appeared at 11.05 ppm in 1H NMR spectrum.
A selective removal of the exocyclic isopropylidene groups in 4h was conducted
similarly as described in literature.[11] A reaction of 4h with aqueous 80% acetic acid
resulted in selective hydrolysis of both exocyclic isopropylidene groups and formation of 7
(Scheme 4) in satisfactory yield (56%). 1H and 13C NMR spectra of the product were in
accordance with the proposed structure of 7, although the hydroxyl protons could not be
localized in 1H NMR spectra. However, the presence of hydroxyl groups was supported by
the presence of a vibration band at 3364 cm−1 in IR spectra. The attempted total deprotection

ACCEPTED MANUSCRIPT
of all hydroxyls in 4h with aqueous trifluoroacetic acid led to the decomposition of the
ferrocene scaffold and no product was isolated.
O
O
H
O
O
N
Fe
N
O
O
O
H
O
4h
i
OH
N
HO
H
O
O
Fe
N
O
O
H
HO
7 (56%)
OH
Scheme 4. Selective removal of two isopropylidene groups resulting in formation of 7.
Reagents and conditions: (i) 80% AcOH, 50 °C.
Preparation of cationic ruthenium complexes 8 and ruthenocenes 9
The high cytotoxic potency of 4a led us to the idea to prepare analogous ruthenium
species. A reaction of an excess of lithium cyclopentadienides generated in situ from 3a or
3b, with [{Ru(η6-p-cymene)Cl2}2] in THF (Scheme 5) gave cationic arene-cyclopentadienyl
complexes 8a (yield 63%) or 8b (yield 33%), respectively. Formation of the corresponding
ruthenocenes 9a and 9b as by-products in crude reaction mixtures was observed by NMR
spectroscopy. Indeed, the similar formation of ruthenocene from a reaction of [{Ru(η6-p-
cymene)Cl2}2] with sodium cyclopentadienide was mentioned in the literature.[22] The
complex 9a was generated only in 2% yield (as calculated from 1H NMR of crude reaction
mixture) and was not isolated. A high amount of 3b and increased reaction temperature led to
a slightly higher yield (7%) of 9b in comparison to 9a. A different polarity of 8b and 9b
allows their efficient separation by column chromatography on neutral alumina. In an
independent experiment an excess of lithium cyclopentadienide derivative (prepared from 3b
and LiBEt3H) reacted with 8b in boiling toluene for 63h. After workup, 9b was isolated in
87% yield. We propose that a reaction of [{Ru(η6-p-cymene)Cl2}2] with lithium
cyclopentadienide Cp1Li (equation 2.) in THF at room temperature, followed by reaction of
generated (and eventually purified) [(η6-p-cymene)(η5-Cp1)Ru]Cl with Cp2Li (equation 3.) in

ACCEPTED MANUSCRIPT
boiling toluene would be a feasible general pathway for generation of ruthenocenes [(η5-
Cp1)(η5-Cp2)Ru] bearing differently substituted cyclopentadienyl rings.
[{Ru(η6-p-cymene)Cl2}2] + 2 Cp1Li → 2 [(η6-p-cymene)(η5-Cp1)Ru]Cl eq. 2.
[(η6-p-cymene)(η5-Cp1)Ru]Cl + Cp2Li → [(η5-Cp1)(η5-Cp2)Ru]
eq. 3.
Cationic complexes 8a and 8b were well soluble in polar solvents and in water, while the
considerably less polar 9b showed a reasonable solubility only in less polar organic solvents,
such as chloroform. 1H and 13C NMR spectra of 8a, 8b and 9b were in accordance with their
proposed structures. The signals of cyclopentadienyl protons in 8 (pseudotriplets centered at
5.31 and 5.43 ppm for 8a; 5.31 and 5.42 ppm for 8b) were considerably downfield shifted in
comparison to 9 (pseudotriplets centered at 4.38 and 4.44 ppm for 9a; multiplet 4.38 − 4.40
ppm for 9b) in 1H NMR spectra as a result of decreased electron density in the former species.
ESI-MS spectra of 8a and 8b showed a peak corresponding to arene cyclopentadienyl
ruthenium cation at m/z 488 and 490, respectively. The molecular composition of 9b was
supported by the presence of [M + H]+ ion at m/z 611.
[N]
1. i
2. ii
Ru Cl-
+
excess
3a
8a (63%)
8b (33%)
+
([N] = piperidin-1-yl)
[N]
3b
([N] = morpholin-1-yl)
[N]
Ru
[N]
9a (2%, not isolated)
9b (7%)
Scheme 5. Formation of cyclopentadienyl arene ruthenium complexes 8 and ruthenocenes 9.
Reagents and conditions: (i) LiBEt3H, THF, rt; (ii) [{Ru(η6-p-cymene)Cl2}2], THF, rt or
reflux.
Molecular structure determination

ACCEPTED MANUSCRIPT
Fig. 1 An ORTEP view of molecular structure of 4a at 50% probability level. Hydrogen
atoms are omitted for clarity. Selected distances [Å] and angles [°]: Fe1-Cg1 1.6510(17), C-
C(ring C1-C5) 1.410(6) − 1.423(5), C-C(ring C7-C12) 1.378(6) − 1.399(5), C-C(C14-C18) 1.511(5) −
1.526(5), C1-C6 1.504(5), C6-C7 1.525(5), C10-C13 1.509(5), N1-C(C13, C14, C18) 1.462(4) −
1.467(5), C1-C6-C7 112.8(3), N1-C13-C10 112.8(3), Fe1-Cg1-C6-C7 178.11.
Crystals of 4a suitable for X-ray analysis were grown by diffusion of methanol into a solution
of 4a in THF. 4a crystallizes in a monoclinic lattice with space group P21/c. The molecule
view and the most important structural parameters are given in Fig. 1. Crystallographic
parameters for 4a and its molecular packing view are given in Supporting Information.
The molecule of 4a possessed an ideal staggered conformation of its cyclopentadienyl rings
with the substituents oriented anti-periplanarly within the ferrocene scaffold. The phenylene
rings are positioned outwards from the ferrocene core (torsion angle Fe1-Cg1-C6-C7 178.11°)
and their orientation towards the cyclopentadienyl rings is close to perpendicular (dihedral
angle 110.26°). The piperidine rings adopted chair conformations (θ = 4.8(4)°).
Crystals of 4a·2HCl suitable for X-ray analysis were obtained from saturated aqueous
solution upon storing several days at room temperature. The molecule view as well as
structural parameters are given in Supporting Information.
Cytotoxicity studies
The cytotoxicity of complexes 4 − 9 was tested in vitro against selected ovarian cancer
lines: A2780 (sensitive to cisPt), A2780cis (acquired resistance to cisPt), and SK-OV-3
(intrinsic resistance to cisPt) using MTT test. As nephrotoxicity remains the main side effect
in cisPt treatment, the cytotoxicity of selected (most active) complexes against human
embryonic kidney cells (HEK293) was evaluated as well. Results (IC50) for 24h and 72h
treatments are collected in Table 1.
In comparison to CisPt (IC50 = 13±2 and 1.7±0.3 µM for 24h and 72h assay,
respectively), several complexes possessed similar activity in both 24 h assay and 72 h assay

ACCEPTED MANUSCRIPT
against A2780 cells. The high activities below 10 µM were achieved with ferrocene
derivatives 4a, 4b, 4e, 4a·2HCl, 4c·2HCl, 4e·2HCl, 5, 6, 7 and ruthenocene 9b in both 24
and 72 h treatment. The moderate activities below 50 µM were found for ferrocenes 4g, 4h
and organic hydrochlorides 3a·HCl and 3aa·HCl. Benzyloxy-substituted ferrocenes 4d, 4f,
ferrocenes lacking N-heterocycle 4i, 4j and cationic ruthenium complexes 8 were found
inactive (for all studied cell lines) and will not be discussed further.
Interestingly, exposure of cisplatin resistant cell line A2780cis to CisPt revealed ca
five-fold decrease in cytotoxic activity (IC50 = 50±9 and 10±2 µM for 24h and 72h assay,
respectively), while the most of the active ferrocene derivatives (4b, 4g, 4b·2HCl, 4c·2HCl,
5, 6, 7) and ruthenocene 9b retained their potency or dropped in their cytotoxicity only 2-3
times (4a, 4e).
Similarly SK-OV-3 cell line showed low response to CisPt treatment (IC50 > 100 and
5.6±1.0 µM for 24h and 72 h assay, respectively). However, ferrocenes 4a, 4b, 4g, 4b·2HCl
and ruthenocene 9b maintained cytotoxicity below 50 µM, while ferrocenes 4e·2HCl, (the
most potent one showed IC50 = 1.76±0.08 µM in 72h assay), 4e, 4a·2HCl, 4c·2HCl, 5, and 6
possessed activity below 10 µM in 72h assay. These results together with the results obtained
for A2780 imply a different mechanism of action of ferrocene and ruthenocene derivatives in
comparison to CisPt.
Unfortunately, most of the derivatives (4c·2HCl, 4e·2HCl, 4e, 4g, 5, 6, 7) having high
cytotoxic activity against all three cancer cell lines displayed a similar or even higher toxicity
against HEK293 cell line and therefore lack of selectivity necessary for application in cancer
treatment. Nonetheless, another potent ferrocene derivatives 4a, 4b, 4a·2HCl, and 4b·2HCl
bearing the six-membered piperidine or morpholine rings displayed an enhanced selectivity
against cancer cell lines with selectivity factor Sf in the range from 2 to 6 (where Sf =
IC50(HEK293) / IC50(cancer cell line); for details see Table 1).
In addition to transition metal complexes, hydrochlorides of fulvene 3a·HCl and
cyclopentadiene 3aa·HCl (a presumed product of ferrocene 4a·2HCl degradation) were
evaluated for their cytotoxic properties. Both species revealed a quite remarkable cytotoxicity
in micromolar range against all cell lines, although their potency did not reach the level of
4a·2HCl and lacked its high selectivity. Interestingly, the cytotoxicity of fulvene 3a·HCl was
found ca twice higher in comparison to 3aa·HCl for all studied lines, which is in accordance
with known cytotoxicity of different 6-arylfulvenes in low to medium micromolar range.[19]
All the above given results let us draw several conclusions concerning this type of
compounds: (i) The studied complexes showed a reasonable stability in biological

ACCEPTED MANUSCRIPT
environment as the prolonged assay usually led to increased efficiency against the studied
human cell lines. (ii) The presence of a saturated aza-heterocycle is essential for achieving
acceptable activity of the complexes, as the nitrogen-free ferrocenes (4i and 4j) displayed no
activity. (iii) The nature of the substituents attached to the N-heterocycle also plays a
significant role. The complexes bearing benzyloxy groups (4d, 4f and even the polar and well
soluble 4f·2HCl) were completely inactive in comparison to their unsubstituted, highly active
parent compounds (4a and 4e). However, the presence of hydroxyl substituents protected by
easily hydrolysable methoxymethyl acetals in 4g resulted in a species exerting an activity
comparable to analogous unprotected ferrocenes 5 and 6 bearing (3S,4S)-pyrrolidine-3,4-diol,
or even to 4e bearing pendant unsubstituted pyrrolidine. Similarly, ca 5-fold increase in
activity was also observed after hydrolysis of the exocyclic isopropylidene acetal in 4h to 7.
(iv) The cationic ruthenium complexes 8 showed little promise as anticancer agents despite
their high stability and solubility in aqueous environment. This is in accordance with the
known inactivity of a range of cationic cyclopentadienyl-arene ruthenium complexes recently
reviewed by Morais et al.[23] (v) The cytotoxicity of complexes roughly relates to the
toxicity of the pendant heterocycle as could be seen when comparing of the cytotoxicity of
ferrocenes (following the order 4a ~ 4b < 4e) with the toxicity of the corresponding cyclic
amines which increases in the order morpholine << piperidine < pyrrolidine (in other words,
six-membered heterocycles are less toxic than five-membered ones).[24] However, ferrocenes
with six-membered aza-heterocycles (particularly 4a, 4b, 4a·2HCl, and 4b·2HCl) seem to be
more promising as potential drugs in comparison to ferrocenes with a five-membered
heterocycle (e.g. 4e bearing pyrrolidine functionality) as they showed higher selectivity
against cancer cell lines compared to embryonic kidney cells HEK293.
Present, mostly preliminary results indicate a potential mechanism of action in an
increased formation of reactive oxygen species (ROS) by selected ferrocene derivatives. An
ability of ferrocene/ferrocenium pair to induce ROS production, which subsequently induces
DNA damage is well known.[25] However, a detailed mechanism of action of the most active
complexes remains to be elucidated. Predominantly cytostatic activity in terms of cell cycle
blockage without killing the cells or clonogenic survival assay to test long-term activity of our
complexes should accompany determination of cytotoxic activity that should also be extended
to determination of the mechanism of cell death.

ACCEPTED MANUSCRIPT
Table 1. Cytotoxicitya of ferrocenes 4 − 7, cationic ruthenium complexes 8, ruthenocene 9b, and organic hydrochlorides (3a·HCl, 3aa·HCl)
against selected ovarian cancer cell lines (A2780, A2780cis, SK-OV-3) and human embryonic kidney cells (HEK293)
cmd.
Cell line (treatment time)
A2780
(24h)
A2780
(72h)
A2780cis A2780cis SK-OV-3 SK-OV-3 HEK293
HEK293
(72h)
(24h)
(72h)
(24h)
(72h)
(24h)
6.9±0.9
3.4±0.1
18±7
6.6±1.5
91±11
37±1
>100
19±3
4a
4b
b
b
Sf = 6
Sf = 3
6.6±0.9
5.4±2.3
9±4
5.5±0.2
>100
32±10
>100
9.4±0.1
b
b
Sf = 2
Sf = 2
n.d.
6.2±1.6
n.d.
n.d.
2.2±0.6
n.d.
>100
3.7±0.5
>100
>100
2.8±0.4
>100
>100
8.9±1.3
>100
>100
4.6±1.9
>100
>100
6±3
>100
2.2±0.2
>100
4d
4e
>100
>100
>100
>100
>100
7.8±1.0
4f
30±5
7±2
23±2
8±2
34±11
95±18
>100
>100
n.d.
7±2
4g
29±9
37±7
53±29
>100
33±20
>100
n.d.
4h
>100
>100
n.d.
n.d.
4i
>100
>100
>100
>100
>100
n.d.
n.d.
4j
2.9±0.3
0.7±0.6
5.4±0.6
5.9±0.8
6.2±0.3
29±9
4.3±0.8
4a·2HCl
b
Sf = 6
60±5
11±3
65±12
6.5±1.4
>100
20.1±0.3
>100
23±4
4b·2HCl
b
b
Sf = 2
Sf = 4

ACCEPTED MANUSCRIPT
6.8±1.6
4.7±0.2
>100
5.3±1.4
2.7±0.3
>100
6±3
10.5±1.0
>100
7.4±0.6
3.3±0.4
>100
15±2
6.2±0.6
>100
9.4±1.0
1.76±0.08
>100
3.0±0.3
5.4±0.5
n.d.
2.8±0.7
1.46±0.04
n.d.
4c·2HCl
4e·2HCl
4f·2HCl
5
12±5
6±4
8.4±0.5
12±1
5.4±1.1
5.3±0.1
7.3±1.6
37±8
19±1
9±3
12±4
10±1
20±5
n.d.
3.6±0.9
3.6±0.3
5.7±0.5
n.d.
9±3
4.3±0.8
5.6±0.5
39±21
28±4
8.1±0.3
55±11
>100
4.8±0.4
24±8
6
5.9 ±0.3
>100
12±2
7
>100
>100
8a
58±10
10±2
39±3
37±14
10±4
>100
>100
n.d.
n.d.
8b
9±5
50±15
44±10
62±8
>100
49±20
22±7
n.d.
n.d.
9b
31 ± 5
53 ± 3
13±2
12±2
14±4
38±3
>100
n.d.
12±3
3a·HCl
3aa·HCl
CisPt
29 ± 3
1.7±0.3
34±6
>100
43±5
n.d.
50±9
10±2
>100
5.6±1.0
35±5
3.8±0.5
b
Sf = 2
a Cytotoxicity was expressed as IC50 [µM] i.e. concentration leading to 50% inhibition of cells; given as the mean of at least three
experiments±standard deviation (SD); SD = [1/N(Σ(xi−xav)2]½ (N is number of measurements, xav is average).
b Selectivity factor Sf is defined for a given compound as the ratio between its activity (IC50) in normal cells compared to cancer cells.[26] For
our purposes Sf = IC50(HEK293) / IC50(cancer cell line). Only values where Sf > 1 are displayed.
n.d. = not determined

ACCEPTED MANUSCRIPT
Conclusions
A series of 8 ferrocene derivatives 4a − 4h of the general formula
[Fe(η5-C5H4CH2(p-C6H4)CH2(N-het))2] were prepared by a straightforward way starting from
N-heterocycle substituted 6-phenylenefulvenes, LiBEt3H and FeCl2. The obtained ferrocenes
bearing a pendant saturated N-heterocycle were further modified either by nitrogen
quaternization or by deprotection. The reaction of [{Ru(η6-p-cymene)Cl2}2] with the
corresponding cyclopentadienide derivatives afforded cationic ruthenium complexes
[Ru(η6-p-cymene)(η5-C5H4CH2(p-C6H4)CH2(N-het))]Cl 8, while ruthenocenes
[Ru(η5-C5H4CH2(p-C6H4)CH2(N-het))2] 9 were found as minor byproducts.
An evaluation of the cytotoxic properties of the prepared complexes against 3 ovarian
cancer cell lines showed that the presence of a pendant saturated N-heterocycle linked to the
ferrocene moiety is an essential feature for high cytotoxicity. A careful choice of the pendant
N-heterocycle allowed achieving not only a high activity, but also a sufficient selectivity of a
particular ferrocene. Hydrochloride 4a·2HCl (bearing pendant piperidinium cation) with a
cytotoxicity in submicromolar region and high selectivity (Sf ~ 6) would be suggested as the
most promising CisPt substitute for the treatment of ovarian cancer. Interestingly, ferrocenes
4a, 4b, and 4b·2HCl (bearing pendant piperidine or morpholine groups) also showed
remarkable cytotoxicity against the cisplatin resistant cell line A2780cis (ca twice more
efficient than CisPt) while still maintaining selectivity against the cancer line (Sf = 2 − 4)
compared to human embryonic HEK293 cell line. The ferrocene 4e and the corresponding
hydrochloride 4e·2HCl were found to be the most cytotoxic species towards CisPt
intrinsically resistant SK-OV-3 cell line, however they lack any selectivity and possessed a
comparable or even higher cytotoxicity against embryonic HEK293 cells as well.
Experimental
All manipulations with air sensitive compounds were carried out under argon atmosphere
using standard Schlenk techniques. NMR spectra were measured either on a Varian Mercury
300 (1H at 300 MHz; 13C at 75 MHz), Varian Inova 500 (1H at 500MHz, 13C at 125 MHz) or
Bruker Avance 400 (1H at 400 MHz; 13C at 100 MHz) at 25 ºC. 1H and 13C chemical shifts
(δ/ppm) are given relative to solvent signals (δH/δC: CDCl3 7.26/77.16; DMSO-d6 2.50/39.52).
Electrospray mass spectra (ESI-MS) were measured with a Bruker Esquire 3000 instrument
on dichloromethane/acetonitrile solutions. IR spectra were measured in Nujol suspensions on

ACCEPTED MANUSCRIPT
a Nicolet Avatar FTIR spectrometer in the range of 400−4000 cm−1. Melting points were
determined on a Kofler block and were uncorrected. Elemental analyses were carried out on a
FLASH EA1112 CHN-O Automatic Elemental Analyzer (Thermo Scientific).
Chemicals. Solvents were appropriately dried by refluxing over either Na/benzophenone
(THF, diethyl ether, toluene), or CaH2 (dichloromethane), distilled and stored over 4Å
molecular sieves. LiBEt3H (1.0 M solution in THF), HCl (1M solution in Et2O, 3M in
methylcyclopentyl ether), [{Ru(η6-p-cymene)Cl2}2], were purchased from Aldrich and used
as received. 6-(4-(piperid-1-ylmethyl)-phenyl)fulvene (3a), 6-(4-(morpholin-4-ylmethyl)-
phenyl)fulvene (3b),[10] 6-(4-(pyrrolidin-1-yl)methyl)-phenyl)fulvene (3e),[10] 6-(4-methyl-
phenyl)fulvene (3i),[18], 6-phenylfulvene (3j);[17] 4a, 4a·2HCl[9] were prepared as
described in literature.
Anion exchange resin AG-1-X8 (Bio-Rad) was converted into OH− cycle by washing with
1M NaOH solution followed by water washing to neutral reaction.
Preparation of 1-[4-(cyclopenta-2,4-dien-1-ylidenemethyl)benzyl]piperidinium chloride
(3a·HCl)
A solution of HCl in Et2O (1.00 ml, 1.00 mmol, 1M) was added dropwise into a
solution of 3a (0.112 g, 0.45 mmol) in THF (5 ml). The mixture was stirred for 15 min, a
formed precipitate was isolated, washed with THF (3 × 3 ml), Et2O (3 × 2 ml) and dried in
vacuum to obtain 3a·HCl as an orange powder. Yield 0.087 g (67%).
M.p. 160 ºC (decomp). 1H NMR (300 MHz, DMSO-d6): 1.24 − 1.43 (m, 1H, NCH2CH2CH2);
1.62 − 1.93 (m, 5H, NCH2CH2CH2); 2.76 − 2.93 (m, 2H, NCH2CH2CH2); 3.27 (d, 2JHH = 12.0
Hz, 2H, NCH2CH2CH2); 4.28 (d, 3JHH = 5.1 Hz, 2H, NCH2C6H4); 6.35 − 6.39 (m, 1H, C5H4);
6.48 − 6.53 (m, 1H, C5H4); 6.67 − 6.75 (m, 2H, C5H4); 7.37 (s, 1H, =CH); 7.72 (s, 4H, C6H4);
11.00 (bs, 1H, NH). 13C{1H} NMR (300 MHz, DMSO-d6): 21.41 (NCH2CH2CH2); 22.07
(NCH2CH2CH2); 51.55 (NCH2CH2CH2); 58.30 (NCH2C6H4); 119.85 (C5H4, CH); 127.53
(C5H4, CH); 130.67 (C6H4, CH); 130.81 (C5H4, CH); 130.96 (C6H4, CqCH2N); 132.00 (C6H4,
CH); 136.06 (C5H4, CH); 137.00 (C6H4, CqCHC5H4); 137.48 (=CH); 145.22 (C=CH). ESI-
MS, m/z (ESI+): 539 ([2M − Cl]+), 503 ([2M − Cl − HCl]+), 252 ([M − Cl]+), 167. Elemental
analysis for C18H22ClN, calculated C, 75.11; H, 7.70; N, 4.87%, found C, 75.26; H 7.75; N,
4.81%.

ACCEPTED MANUSCRIPT
Preparation of 1-[4-(cyclopentadienylmethyl)benzyl]piperidinium chloride (a mixture of
isomers) (3aa·HCl)
A slight excess of LiBEt3H solution in THF (0.70 ml, 0.70 mmol, 1M) was added
slowly dropwise into an orange solution of 3a (0.163 g, 0.65 mmol) in THF (6 ml), which
caused a gradual decolorization of the mixture. The mixture was stirred for 4h and then a
solution of HCl in Et2O (2.1 ml, 2.1 mmol, 1M) was added in several portions. The mixture
was stirred for 30 min and volatiles were evaporated in vacuum. A solid residue was extracted
in chloroform and filtered. Solvent evaporation in vacuum gave 3aa·HCl as a mixture of
regioisomers. Yield 0.138 g (73%).
1H NMR (300 MHz, DMSO-d6): 1.20 − 2.00 (m, overall 6H, NCH2CH2CH2); 2.45 − 3.25 (m,
overall 4H, NCH2CH2CH2); 3.50, 3.67, 3.72 (3 × s, overall 2H, C5H4CH2C6H4); 4.15 − 4.25
(m, overall 2H, NCH2C6H4); 6.90 − 6.86 (m, overall 5H, C5H5); 7.10 − 7.70 (m, overall 4H,
C6H4); 11.38 (bs, 1H, NH). ESI-MS, m/z (ESI+): 543 ([2M − Cl]+), 507 ([2M − Cl − HCl]+),
254 ([M − Cl]+), 169. Elemental analysis for C18H24ClN, calculated C, 74.59; H, 8.35; N,
4.83%, found C, 74.71; H 8.38; N, 4.80%.
General procedure for preparation of ferrocenes 4
A 1M solution of LiBEt3H (1.00 ml, 1.00 mmol) in THF was slowly added dropwise
into a solution of the fulvene 3 (1.00 mmol) in the same solvent (10 ml). As the reaction
proceeded the reaction mixture changed colour from intense yellow-orange to almost
colourless within minutes. The reaction mixture was stirred for 4 h, a solid FeCl2 (0.064 g,
0.50 mmol) was added at once and the reaction mixture was stirred for 15 h. Solvents were
evaporated in vacuum and the solid residue was purified as described below for individual
compounds (either by column chromatography on silica or by crystallization).
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where N-het = morpholin-1-yl (4b)
Prepared by the general procedure mentioned above from 3b (0.840 g, 3.32 mmol),
LiBEt3H (3.40 ml, 1 M solution in THF, 3.40 mmol) and FeCl2 (0.212 g, 1.67 mmol). Purified
by recrystallization from hot EtOH and obtained as a yellow powder. Yield 0.633 g (67%).
M.p. 115 ºC. 1H NMR (300 MHz, CDCl3): 2.38 − 2.45 (m, 8H, NCH2CH2O); 3.45 (s, 4H,
NCH2C6H4); 3.64 (s, 4H, C5H4CH2C6H4); 3.66 − 3.72 (m, 8H, NCH2CH2O); 4.04 (s, 8H,

ACCEPTED MANUSCRIPT
C5H4); 7.12, 7.22 (2 × d, 3JHH = 8.2 Hz, 2 × 4H, C6H4). 13C{1H} NMR (75 MHz, CDCl3):
35.65 (C5H4CH2C6H4); 53.71 (NCH2CH2O); 63.29 (NCH2C6H4); 67.11 (NCH2CH2O); 68.42,
69.44 (C5H4, CH); 88.11 (C5H4, Cipso); 128.33, 129.27 (C6H4, CH); 135.36, 140.70 (C6H4, Cq).
ESI-MS, m/z (ESI+): 587 [M + Na]+); 565 ([M + H]+); 490. Elemental analysis for
C34H40FeN2O2, calculated C, 72.33; H, 7.14; N, 4.96%, found C, 72. 53; H 7.18; N, 4.91%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where N-het = thiomorpholin-1-yl
(4c)
Prepared by the general procedure mentioned above from 3c (1.000 g, 3.71 mmol),
LiBEt3H (3.70 ml, 1 M solution in THF, 3.70 mmol) and FeCl2 (0.232 g, 1.83 mmol). Purified
by recrystallization from hot THF and obtained as yellow microcrystals. Yield 0.485 g (44%).
M.p. 177 ºC. 1H NMR (300 MHz, CDCl3): 2.53 − 2.58 (m, 16H, NCH2CH2S); 3.46 (s, 4H,
NCH2C6H4); 3.63 (s, 4H, C5H4CH2C6H4); 4.04 (s, 8H, C5H4); 7.11, 7.19 (2 × d, 3JHH = 8.0 Hz,
2 × 4H, C6H4). 13C{1H} NMR (75 MHz, CDCl3): 28.14 (NCH2CH2S); 35.67 (C5H4CH2C6H4);
55.01 (NCH2CH2S); 63.54 (NCH2C6H4); 68.45, 69.47 (C5H4, CH); 88.13 (C5H4, Cipso);
128.34, 129.14 (C6H4, CH); 135.66, 140.68 (C6H4, Cq). ESI-MS, m/z (ESI+): 597 ([M + H]+).
Elemental analysis for C34H40FeN2S2, calculated C, 68.44; H, 6.76; N, 4.70%, found C, 68.48;
H 6.73; N, 4.71%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where
N-het = 2,3,4,6-tetra-O-benzyl-1,5-dideoxy-1,5-imino-D-glucitol-N-yl (4d)
Prepared by the general procedure mentioned above from 3d (0.550 g, 797 µmol),
LiBEt3H (0.80 ml, 1 M solution in THF, 800 µmol) and FeCl2 (0.050 g, 398 µmol). Purified
by chromatography on silica using heptane/ethyl acetate (5/1, v/v) and obtained as an orange
wax. Yield 0.438 g (77%).
Rf = 0.2 heptane/ethyl acetate (5/1, v/v). 1H NMR (300 MHz, CDCl3): 1.99 (t, 2JHH ~ 3JHH
11.2 Hz, 2H, C(1)H2); 2.37 (d, 3JHH = 9.8 Hz, 2H, C(5)H); 3.02 (dd, 2JHH = 11.2 Hz, 3JHH
~
=
4.6 Hz, 2H, C(1)H2); 3.37 (d, 2JHH = 13.7 Hz, 2H, NCH2C6H4); 3.41 − 3.66 (m, 6H, C(2)H,
C(3)H, C(4)H); 3.68 (s, 4H, C5H4CH2C6H4); 3.71 (dd, 2JHH = 10.6 Hz, 3JHH = 2.6 Hz, 2H,
C(6)H2); 3.79 (dd, 2JHH = 10.6 Hz, 3JHH = 1.2 Hz, 2H, C(6)H2); 4.01 (d, 2JHH = 13.7 Hz, 2H,
NCH2C6H4); 4.03 − 4.07 (m, 8H, C5H4); 4.40 − 4.61 (m, 9H, CH2, CH2Ph); 4.75 − 4.97 (m,
7H, CH2, CH2Ph); 7.01 − 7.40 (m, 48H, C6H4 and Ph). 13C{1H} NMR (75 MHz, CDCl3):

ACCEPTED MANUSCRIPT
35.73 (C5H4CH2C6H4); 54.31 (C(1)); 56.35 (NCH2C6H4); 65.01 (C(5)); 66.67 (C(6)); 68.55,
69.50 (C5H4, CH); 72.65, 73.42, 75.36(2C) (CH2Ph); 78.38 (C(2)); 78.84 (C(4)); 87.43 (C(3));
88.14 (C5H4, Cipso); 127.52 − 128.50 (Ph and C6H4, CH); 129.12 (C6H4, CH); 135.78 (C6H4,
Cq); 138.12, 138.59, 138.68, 139.17 (Ph, Cipso); 140.54 (C6H4, Cq). ESI-MS, m/z (ESI+): 1486,
1460 [M + Na]+); 1444 ([M + Li]+). Elemental analysis for C94H96FeN2O8, calculated C,
78.53; H, 6.73; N, 1.95% found C, 78.56; H 6.90; N, 1.93 %.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where N-het = pyrrolidin-1-yl (4e)
Prepared by the general procedure mentioned above from 3e (0.833 g, 3.51 mmol),
LiBEt3H (3.5 ml, 1 M solution in THF, 3.5 mmol) and FeCl2 (0.222 g, 1.75 mmol). Purified
by chromatography on silica using THF/hexane (1/1, v/v). Recrystallization from methanol
gave product as a yellow solid. Yield 0.446 g (48%).
M.p. 75 − 78 ºC. 1H (300 MHz, CDCl3): 1.73-1.80 (m, 8H, NCH2CH2); 2.46-2.52 (m, 8H,
NCH2CH2); 3.57 (s, 4H, NCH2C6H4); 3.64 (s, 4H, C5H4CH2C6H4); 4.03 (s, 8H, C5H4); 7.11,
7.22 (2 × d, 2 × 3JHH = 8.1 Hz, 2 × 4H, C6H4). 13C{1H} (75 MHz, CDCl3): 23.57 (NCH2CH2);
35.67 (C5H4CH2C6H4); 54.28 (NCH2CH2); 60.56 (NCH2C6H4); 68.41, 69.47 (C5H4, CH);
88.23 (C5H4, Cipso); 128.35, 128.98 (C6H4, CH); 137.01 (C6H4, CqCH2N); 140.39 (C6H4,
CqCH2C5H4). ESI-MS, m/z (ESI+): 533 ([M + H]+). Elemental analysis for C34H40FeN2,
calculated C, 76.68; H, 7.57; N, 5.26%, found C, 76.42; H 7.61; N, 5.32%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where
N-het = (3S,4S)-3,4-bis(benzyloxy)pyrrolidin-1-yl (4f)
Prepared by the general procedure mentioned above from 3f (1.142 g, 2.54 mmol),
LiBEt3H (2.50 ml, 1 M solution in THF, 2.50 mmol) and FeCl2 (0.157 g, 1.24 mmol). Purified
by chromatography on silica using hexane/THF (2/1, v/v) and obtained as a yellow-orange oil.
Yield 0.760 g (64%).
Rf = 0.4 hexane/THF (2/1, v/v). 1H NMR (400 MHz, CDCl3): 2.57 (dd, 2JHH = 10.0 Hz, 3JHH
=
4.1 Hz, 4H, NCH2CH); 2.88 (dd, 2JHH = 10.0 Hz, 3JHH = 6.3 Hz, 4H, NCH2CH); 3.52, 3.60 (2
× d, 2 × 2JHH = 12.9 Hz, 2 × 2H, NCH2C6H4); 3.64 (s, 4H, C5H4CH2C6H4); 4.04 (bs, 12H,
CHOBn, C5H4); ); 4.45, 4.49 (2 × d, 2 × 2JHH = 11.9 Hz, 2 × 4H, CH2Ph); 7.11 (d, 3JHH = 7.9
Hz, 4H, C6H4); 7.21 (d, 3JHH = 7.9 Hz, 4H, C6H4); 7.31 (m, 20H, Ph). 13C{1H} NMR (100
MHz, CDCl3): 35.67 (C5H4CH2C6H4); 58.48 (NCH2CH); 60.15 (NCH2C6H4); 68.45, 69.47

ACCEPTED MANUSCRIPT
(C5H4, CH); 71.52 (CH2Ph); 83.73 (CHOBn); 88.13 (C5H4, Cipso); 127.75, 127.93 (Ph, CH);
128.30, (C6H4, CH); 128.47 (Ph, CH); 128.93 (C6H4, CH); 135.94 (C6H4, CqCH2N,), 138.93
(Ph, Cipso); 140.57 (C6H4, CqCH2C5H4). ESI-MS (m/z; ESI+): 957 ([M + H]+), 674. Elemental
analysis for C62H64FeN2O4, calculated C, 77.81; H, 6.74; N, 2.93%, found C, 78.03; H 6.77;
N, 2.90%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where
N-het = (3S,4S)-3,4-bis(methoxymethyl)pyrrolidin-1-yl (4g)
Prepared by the general procedure mentioned above from 3g (1.170 g, 3.26 mmol),
LiBEt3H (3.20 ml, 1 M solution in THF, 3.20 mmol) and FeCl2 (0.203 g, 1.60 mmol). Purified
by chromatography on silica using hexane/THF (3/1, v/v) and obtained as a yellow-orange oil.
Yield 0.957 g (77%).
Rf = 0.2 hexane/THF (3/1, v/v). 1H NMR (300 MHz, DMSO-d6): 2.39 (dd, 2JHH = 10.2 Hz,
3JHH = 4.5 Hz, 4H, NCH2CH); 2.76 (dd, 2JHH = 10.2 Hz, 3JHH = 6.0 Hz, 4H, NCH2CH); 3.21
(s, 12H, OMe); 3.43, 3.51 (2 × d, 2 × 2JHH = 12.9 Hz, 2 × 2H, C6H4CH2N); 3.59 (s, 4H,
C5H4CH2C6H4); 3.98 (pseudo t, 3JHH = 6.0 Hz, 3JHH = 4.5 Hz, 4H, NCH2CH); 4.01-4.04, 4.05-
4.08 (2 × m, 2 × 4H, C5H4); 4.54, 4.58 (2 × d, 2 × 2JHH = 6.9 Hz, 2 × 4H, OCH2O); 7.12, 7.17
(2 × d, 2 × 3JHH = 8.4 Hz, 2 × 4H, C6H4). 13C{1H} NMR (75 MHz, DMSO-d6): 34.82
(C5H4CH2C6H4); 54.80 (OMe); 58.16 (NCH2CH); 59.07 (C6H4CH2N); 67.93, 68.97 (C5H4,
CH); 81.13 (NCH2CH); 88.16 (C5H4, Cipso); 94.98 (OCH2O); 128.07, 128.33 (C6H4, CH);
135.89, 140.42 (C6H4, Cq). ESI-MS, m/z (ESI+): 773 [M + H]+. Elemental analysis for
C42H56FeN2O8, calculated C, 65.28; H, 7.30; N, 3.63%, found C, 65.49; H 7.35; N, 3.60%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where
N-het = 1,4-dideoxy-2,3:5,6-di-O-isopropylidene-1,4-imino-D-talitol-N-yl (4h)
Prepared by the general procedure mentioned above from 3h (0.336 g, 0.82 mmol),
LiBEt3H (0.82 ml, 1 M solution in THF, 0.82 mmol) and FeCl2 (0.051 g, 0.40 mmol). Purified
by chromatography on silica using hexane/ethyl acetate (5/1 → 1/1, v/v) and obtained as a
yellow powder. Yield 0.233 g (66%).
M.p. 121 ºC. 1H NMR (300 MHz, CDCl3): 1.30, 1.34, 1.43, 1.51 (4 × s, 4 × 6H, CMe2); 2.49
(dd, 2JHH = 9.9 Hz, 3JHH = 3.9 Hz, 2H, C(1)H2); 2.84 − 2.99 (m, 2H, C(4)H); 3.17 (dd, 2JHH
9.9 Hz, 3JHH = 6.0 Hz, 2H, C(1)H2); 3.55 (d, 2JHH = 13.8 Hz, 2H, NCH2C6H4); 3.64 (s, 4H,
=

ACCEPTED MANUSCRIPT
C5H4CH2C6H4); 3.83 (t, 2JHH = 3JHH = 7.8 Hz, 2H, C(6)H2); 3.96 − 4.06 partly overlapped (m,
2H, C(6)H2); 4.03 (s, 8H, C5H4); 4.11 − 4.24 (m, 4H, NCH2C6H4 and C(5)H); 4.33 − 4.44 (m,
2H, C(3)H); 4.51 − 4.63 (m, 2H, C(2)H); 7.10, 7.19 (2 × d, 2 × 3JHH = 7.9 Hz, 2 × 4H, C6H4).
13C{1H} NMR (75 MHz, CDCl3): 25.11, 25.46, 26.69, 27.44 (CMe2); 35.69 (C5H4CH2C6H4);
58.94 (C(1)); 59.29 (NCH2C6H4); 66.54 (C(6)); 68.50, 69.45 (C5H4, CH); 70.71 (C(5)); 77.32
(C(4)); 78.87 (C(2)); 82.38 (C(3)); 88.19 (C5H4, Cipso); 109.61, 113.03 (CMe2); 128.38,
128.87 (C6H4, CH); 136.78 (C6H4, CqCH2N,); 140.56 (C6H4, CqCH2C5H4). ESI-MS, m/z
(ESI+): 915 [M + K]+, 899 [M + Na]+, 883 [M + Li]+, 877 [M + H]+. Elemental analysis for
C50H64FeN2O8, calculated C, 68.48; H, 7.36; N, 3.20%, found C, 68.11; H 7.31; N, 3.15%.
Preparation of [{η5-C5H4(4-methylbenzyl)}2Fe] (4i)
Prepared by the general procedure mentioned above from 3i (2.76 g, 16.4 mmol),
LiBEt3H (16.4 ml, 1 M solution in THF, 16.4 mmol) and FeCl2 (1.04 g, 8.2 mmol). Crude
product was purified by filtration through a short silica column and crystallization from EtOH
to give 2i as a yellow powder. Yield 2.04 g (63%).
M.p. 113 ºC. 1H NMR (300 MHz, CDCl3): 2.34 (s, 6H, C6H4Me); 3.66 (s, 4H, CH2); 4.07 (s,
8H, C5H4); 7.10 (s, 8H, C6H4). 13C{1H} NMR (75 MHz, CDCl3): 21.13 (C6H4Me); 35.55
(C5H4CH2C6H4); 68.40, 69.41 (C5H4, CH); 88.38 (C5H4, Cipso); 128.37, 129.05 (C6H4, CH);
135.42, 138.80 (C6H4, Cq). ESI-MS, m/z (ESI+): 394 ([M]+). Elemental analysis for C26H26Fe,
calculated C, 79.19; H, 6.65%, found C, 79.15; H 6.68%.
Preparation of [{η5-C5H4(benzyl)}2Fe] (4j)
Prepared by the general procedure mentioned above from 6-phenylfulvene (0.100 g,
0.65 mmol), LiBEt3H (0.65 ml, 1 M solution in THF, 0.65 mmol) and FeCl2 (0.041 g, 0.32
mmol). Purified by chromatography on silica using Et2O as an eluent and obtained as a yellow
solid. Yield 0.074 g (62%). NMR spectra were consistent with literature.[20-21]
1H NMR (300 MHz, CDCl3): 3.69 (s, 4H, CH2), 4.08 (s, 8H, C5H4), 7.20 − 7.33 (m, 10H, Ph).
13C{1H} NMR (75 MHz, CDCl3): 35.98 (CH2), 68.47, 69.47 (C5H4, CH), 88.11 (C5H4, Cipso),
126.02, 128.37, 128.49 (Ph, CH), 141.77 (Ph, Cipso).
Preparation of 4b·2HCl

ACCEPTED MANUSCRIPT
A HCl solution in methylcyclopropylether (0.75 ml, 2.25 mmol, 3M solution) was
added dropwise into solution of 4b (0.244 g, 0.43 mmol) in THF (5 ml), which caused an
immediate precipitate formation. The mixture was stirred for 30 min and the precipitate was
collected on frit, washed with THF (3 × 1 ml) and dried in vacuum. The product was obtained
as a yellow-orange solid. Yield 0.155 g (56%).
M.p. 165 ºC (decomp.). 1H NMR (400 MHz, DMSO-d6): 3.01 − 3.17 (m, 8H, NCH2CH2);
3.67 (bs, 4H, C5H4CH2C6H4); 3.82 − 3.86 (m, 8H, CH2O); 4.06, 4.10 (2 x bs, 2 x 4H, C5H4);
4.24 (bs, 4H, NCH2C6H4); 7.25 (bd, 3JHH = 6.7 Hz, 4H, C6H4), 7.52 (bs, 4H, C6H4), 11.37 (bs,
2H, NH). 13C{1H} NMR (100 MHz, DMSO-d6): 34.88 (C5H4CH2C6H4); 50.46 (NCH2CH2);
58.63 (NCH2C6H4); 62.95 (CH2O); 68.09, 69.02 (C5H4, CH); 87.67 (C5H4, Cipso); 126.66
(C6H4, CqCH2N); 128.50, 131.39 (C6H4, CH); 143.27 (C6H4, CqCH2C5H4).
Elemental analysis for C34H42Cl2FeN2O2, calculated C, 64.06; H, 6.64; N, 4.40%, found C,
64.13; H 6.72; N, 4.35%.
Preparation of 4c·2HCl
A HCl solution in methylcyclopropylether (0.70 ml, 2.10 mmol, 3M solution) was
added dropwise into solution of 4c (0.311 g, 0.52 mmol) in a THF/CH2Cl2 (10 ml/17 ml)
solvent mixture. The mixture was stirred for 1h, while formation of precipitate was gradually
observed. The volume of the mixture was reduced to a half and the precipitate was isolated.
The precipitate was washed with THF (3 × 4 ml), Et2O (3 × 4 ml) and dried in vacuum to give
a product as a yellow powder. Yield 0.190 g (55%).
M.p. 176 ºC (decomp.). 1H NMR (500 MHz, DMSO-d6): 2.76 (bd, 2JHH = 14.3 Hz, 4H,
CH2S); 3.03 (bq, 2JHH ~ 3JHH ~ 12.0 Hz, 4H, NCH2CH2); 3.15 (bt, 2JHH ~ 3JHH ~ 12.7 Hz, 4H,
CH2S); 3.52 (bd, 4H, 2JHH = 12.3 Hz, NCH2CH2); 3.66 (bs, 4H, C5H4CH2C6H4); 4.06, 4.12 (2
× bs, 2 × 4H, C5H4); 4.27 (bs, 4H, NCH2C6H4); 7.27, 7.48 (2 × d, 3JHH = 7.7 Hz, 2 × 4H,
C6H4), 10.93 (bs, 2H, NH). 13C{1H} NMR (100 MHz, DMSO-d6): 23.70 (CH2S); 34.90
(C5H4CH2C6H4); 52.41 (NCH2CH2S); 59.01 (NCH2C6H4); 68.14, 69.05 (C5H4, CH); 87.66
(C5H4, Cipso); 126.64 (C6H4, CqCH2N); 128.59, 131.47 (C6H4, CH); 143.39 (C6H4,
CqCH2C5H4). Elemental analysis for C34H42Cl2FeN2S2, calculated C, 60.99; H, 6.32; N,
4.18%, found C, 61.20; H 6.38; N, 4.22%.
Preparation of 4e·2HCl

ACCEPTED MANUSCRIPT
A HCl solution in methylcyclopropylether (0.90 ml, 2.70 mmol, 3M solution) was
added dropwise into solution of 4e (0.231 g, 0.43 mmol) in THF (8 ml) and the mixture was
stirred overnight. A formed solid was isolated, washed with Et2O (2 × 5 ml) and dried in
vacuum to give a product as a yellow powder. Yield 0.206 g (79%).
M.p. 155 ºC (decomp.). 1H (400 MHz, DMSO-d6): 1.89 (bs, 8H, NCH2CH2); 2.87-3.37 (m, 8,
NCH2CH2); 3.65 (s, 4H, C5H4CH2C6H4); 4.05, 4.10 (2 × pseudo t, 2 × 3JHH = 1.8 Hz, 2 × 4H,
C5H4); 4.22 (s, 4H, NCH2C6H4); 7.24, 7.50 (2 × d, 2 × 3JHH = 7.9 Hz, 2 × 4H, C6H4); 11.27
(bs, 2H, NH). 13C{1H} (100 MHz, DMSO-d6): 22.55 (NCH2CH2); 34.85 (C5H4CH2C6H4);
52.35 (NCH2CH2); 56.33 (NCH2C6H4); 68.09, 69.03 (C5H4, CH); 87.76 (C5H4, Cipso); 128.50
(C6H4, CH); 129.31 (C6H4, CqCH2N); 130.34 (C6H4, CH); 142.77 (C6H4, CqCH2C5H4).
Elemental analysis for C34H42Cl2N2FeS2, calculated C, 60.99; H, 6.32; N, 4.18%, found C,
61.20; H 6.38; N, 4.22%.
Preparation of 4f·2HCl
The ferrocene 4f (0.088 g, 0.09 mmol) was dissolved in Et2O (3 ml) and an excess of
solution of HCl in Et2O (0.25 ml, 1 M solution, 0.25 mmol) was added gradually dropwise,
which caused an immediate formation of yellow precipitate. The mixture was stirred for 2 min
and the precipitate was isolated, washed with Et2O (3 × 2 ml) and dried in vacuum. The
product was obtained as a yellow powder. Yield 0.093 g (97%).
M.p. 72 ºC. 1H (300 MHz, DMSO-d6): 3.20−3.30 (m, 2H, NCH2CH); 3.32−3.48 (m, 4H,
NCH2CH); 3.58−3.67 partly overlapped (m, 2H, NCH2CH); 3.66 (s, 4H, C5H4CH2C6H4);
4.06, 4.12 (2 × pseudo t, 2 × 4H, C5H4); 4.20−4.38 (m, 8H, NCH2C6H4 and CHOBn); 4.48,
4.55 partly overlapped (2 × d, 2 × 2JHH = 11.8 Hz, 2 × 2H, CH2Ph); 4.57 (s, 4H, CH2Ph); 7.25
(d, 3JHH = 7.9 Hz, 4H, C6H4); 7.27−7.41 (m, 20H, Ph); 7.51 (d, 3JHH = 7.9 Hz, 4H, C6H4);
11.67 (bs, 2H, NH). 13C{1H} (75 MHz, DMSO-d6): 34.89 (C5H4CH2C6H4); 55.38, 55.76
(NCH2CH); 58.85 (NCH2C6H4); 68.14, 69.06 (C5H4, CH); 70.42, 70.82 (CH2Ph); 79.30,
79.92 (CHOBn); 87.68 (C5H4, Cipso); 127.67−128.36 (Ph, CH); 128.59, 130.79 (C6H4, CH);
137.45 (Ph, Cipso); 137.47 (C6H4, CqCH2N); 143.17 (C6H4, CqCH2C5H4). ESI-MS (m/z; ESI+):
957 ([M − HCl − Cl]+). Elemental analysis for C62H66Cl2FeN2O4, calculated C, 72.30; H,
6.46; N, 2.72%, found C, 72.12; H 6.44; N, 2.78%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where
N-het = (3S,4S)-3,4-dihydroxypyrrolidin-1-yl hydrochloride (5)

ACCEPTED MANUSCRIPT
The ferrocene 4g (0.950 g, 1.23 mmol) was dissolved in THF (20 ml) and a slight
excess of concentrated hydrochloric acid (0.4 ml, 4.8 mmol) was added in one portion. The
mixture was stirred for 4h, which led to separation of a brown wax on the flask bottom.
Solvents were decanted off and the wax was washed with THF (3 × 15 ml), acetone (10 ml),
Et2O (2 × 10 ml) and dried in vacuum. Dissolving of the residue in MeOH and evaporation of
the formed solution in vacuum gave 5 as a brown solid. Yield 0.783 g (95%).
Note: The same product was obtained when anhydrous HCl (0.54 ml, 0.54 mmol, 1M) in Et2O
was used for reaction with 4g (0.125 g, 0.16 mmol) in THF (5 ml). Yield of 5 was 0.96 g
(89%).
M.p. 80 ºC (decomp.). 1H (300 MHz, DMSO-d6): 2.96−3.17 (m, 2H, NCH2CH); 3.32−3.62
partly overlapped (m, 6H, NCH2CH); 3.66 (s, 4H, C5H4CH2C6H4); 4.05 (s, 4H, CHOH); 4.10
(s, 8H, C5H4); 4.25, 4.33 (2 × d, 2 × 2JHH = 12.8 Hz, 2 × 2H, NCH2C6H4); 5.8 (bs, 4H,
CHOH); 7.25, 7.51 (2 × d, 2 × 3JHH = 7.9 Hz, 8H, C6H4); 11.05 (bs, 2H, NH). 13C{1H} (75
MHz, DMSO-d6): 34.82 (C5H4CH2C6H4); 58.35 (NCH2CH); 59.35 (NCH2C6H4); 68.09, 69.02
(C5H4, CH); 74.41 (CHOH); 87.68 (C5H4, Cipso); 128.54 (C6H4, CH); 128.75 (C6H4, CqCH2N);
130.56 (C6H4, CH); 142.98 (C6H4, CqCH2C5H4). ESI-MS (m/z; ESI−): 631 ([M − HCl − H]−).
IR (nujol, cm−1): 3286 (br, νO-H). Elemental analysis for C34H42Cl2FeN2O4, calculated C,
61.00; H, 6.32; N, 4.19%, found C, 61.43; H 6.42; N, 4.11%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where
N-het = (3S,4S)-3,4-dihydroxypyrrolidin-1-yl (6)
Ferrocene bishydrochloride 5 (0.259 g, 0.39 mmol) solution in water (2 ml) was
transferred to the top of a small column (2 × 8 cm) packed with anion-exchange resin AG-1-
X8 in OH− form. The column was washed with 200 ml of water and the crude product was
eluted with acetone (70 ml) as an orange band. The acetone solution was adsorbed on the top
of a silica column and the column was washed with an additional acetone (total amount 300
ml). A product was eluted with ethanol (total amount 250 ml) as a yellow band. Evaporation
of the solvent gave 6 as a yellow-orange wax, which spontaneously solidified at room
temperature. Yield 0.146 g (63%).
M.p. 128 ºC. 1H NMR (500 MHz, DMSO-d6): 2.27 (dd, 2JHH = 9.6 Hz, 3JHH = 4.2 Hz, 4H,
NCH2CH); 2.71 (dd, 2JHH = 9.6 Hz, 3JHH = 6.0 Hz, 4H, NCH2CH); 3.40 (d, 2JHH = 13.0 Hz,
2H, NCH2C6H4); 3.51 (d, 2JHH = 13.0 Hz, 2H, NCH2C6H4); 3.60 (s, 4H, C5H4CH2C6H4); 3.82

ACCEPTED MANUSCRIPT
(dd, 3JHH = 6.0 Hz, 3JHH = 4.2 Hz, 4H, CHOH); 4.02, 4.07 (2 × pseudo t, 2 × 3JHH = 1.8 Hz, 2
× 4H, C5H4); 4.80 (bs, 4H, OH); 7.11, 7.16 (2 × d, 3JHH = 8.1 Hz, 2 × 4H, C6H4). 13C{1H}
NMR (125 MHz, DMSO-d6): 34.86 (C5H4CH2C6H4); 59.64 (NCH2C6H4); 60.74 (NCH2CH);
67.95, 68.98 (C5H4, CH); 77.60 (CHOH); 88.18 (C5H4, Cipso); 127.94, 128.35 (C6H4, CH);
136.40 (C6H4, CqCH2N); 140.26 (C6H4, CqCH2C5H4). ESI-MS (m/z; ESI+): 597 ([M + H]+).
IR (nujol, cm−1): 3337 (br, νO-H). Elemental analysis for C34H40FeN2O4, calculated C, 68.45;
H, 6.76; N, 4.70%, found C, 68.44; H 6.79; N, 4.70%.
Preparation of [{η5-C5H4CH2(p-C6H4)CH2(N-het)}2Fe] where
N-het = 1,4-dideoxy-2,3-O-isopropylidene-1,4-imino-D-talitol-N-yl (7)
A ferrocene 4h (0.248 g, 0.28 mmol) was dissolved in 80% acetic acid (5 ml) and the
mixture was saturated with argon (stripping with argon for 5 min). The mixture was stirred at
50 °C until the starting material no longer remained in the mixture (ca 40 h) as shown by
TLC. Volatiles were evaporated in vacuum under slight heating (up to 50 °C) leaving the
crude product as an orange oil. The crude product was purified by chromatography on silica
using dichloromethane/methanol mixture (30/1 → 10/1, v/v) as an eluent. After evaporation
of solvents and trituration in heptane, pure 7 was obtained as an orange solid. Yield 0.125 g
(56%).
M.p. 61 ºC. 1H NMR (300 MHz, DMSO-d6): 1.22, 1.43 (2 × s, 2 × 6H, CMe2); 2.48 partly
overlapped by solvent signal (dd, 2JHH = 11.4 Hz, 3JHH = 3.2 Hz, 2H, C(1)H2); 2.95 (dd, 2JHH
= 11.4 Hz, 3JHH = 5.3 Hz, 2H, C(1)H2); 3.00 (dd, 2JHH = 4.8 Hz, 3JHH = 1.5 Hz, 2H, C(6)H2);
3.32 − 3.40 (m, 2H, C(5)H); 3.44 − 3.56 (m, 4H, C(4)H and C(6)H2); 3.59 (s, 4H,
C5H4CH2C6H4); 3.72, 4.00 (2 × d, 2 × 2JHH = 13.3 Hz, 2 × 2H, NCH2C6H4); 4.02, 4.07 (2 ×
pseudo t, 2 × 4H, C5H4); 4.49 − 4.56 (m, 2H, C(2)H); 4.60 (dd, 3JHH = 6.2 Hz, 3JHH = 1.5 Hz,
2H, C(3)H); 7.12, 7.17 (2 × d, 2 × 3JHH = 8.2 Hz, 2 × 4H, C6H4). 13C NMR (75 MHz, DMSO-
d6): 24.39, 27.15 (CMe2); 34.87 (C5H4CH2C6H4); 58.20 (C(1)); 59.52 (NCH2C6H4); 63.19
(C(6)); 67.96, 69.00 (C5H4, CH); 71.04 (C(4)); 71.55 (C(5)); 79.94 (C(3)); 82.98 (C(2)); 88.23
(C5H4, Cipso); 110.81 (CMe2); 128.08, 128.35 (C6H4, CH); 137.26 (C6H4, CqCH2N,); 140.29
(C6H4, CqCH2C5H4). ESI-MS (m/z; ESI+): 819 ([M + Na]+), 803 ([M + Li]+), 797 ([M + H]+).
IR (nujol, cm−1): 3364 (br, νO-H). Elemental analysis for C44H56FeN2O8, calculated C, 66.32;
H, 7.08; N, 3.52%, found C, 66.45; H 7.12; N, 3.47%.
Reaction of [{Ru(η6-p-cymene)Cl2}2] with a lithium cyclopentadienide generated from 3a

ACCEPTED MANUSCRIPT
3a (0.308 g, 1.23 mmol) was dissolved in THF (15 ml) and LiBEt3H (1.20 ml, 1.20
mmol, 1M solution in THF) was added dropwise. The mixture was stirred for 4h and then
solid [{Ru(η6-p-cymene)Cl2}2] (0.355 g, 0.58 mmol) was added. The resulting orange mixture
was stirred for 7 days and the volatiles were evaporated in vacuum. 1H NMR spectrum of the
mixture showed 8a as a main product (molar ratio 96%), while the corresponding ruthenocene
9a was detected as a minor product (molar ratio 4%, total yield ca 2%). The crude 8a was
purified by chromatography on alumina (neutral, Brockmann II). The column was first eluted
with CH2Cl2 (giving a fraction enriched with 9a). Then elution of the column with ethanol,
followed by solvent evaporation gave a brown waxy solid. The solid was recrystallized from
acetone/heptane mixture to give 8a as a light grey wax. Yield 0.385 g (63%).
[{η6-p-cymene}{η5-C5H4CH2(p-C6H4)CH2(piperidin-1-yl)}Ru]Cl (8a)
1H NMR (300 MHz, DMSO-d6): 1.16 (d, 6H, 3JHH = 6.9 Hz, CHMe2); 1.31 − 1.40 (m, 2H,
NCH2CH2CH2); 1.40 − 1.51 (m, 4H, NCH2CH2CH2); 2.24 (s, 3H, Me, p-cymene); 2.22 − 2.32
(m, 4H, NCH2CH2CH2); 2.63 (septuplet, 1H, 3JHH = 6.9 Hz, CHMe2); 3.36 (s, 2H,
NCH2C6H4); 3.54 (s, 2H, C5H4CH2C6H4); 5.31, 5.43 (2 × pseudo t, 2 × 2H, C5H4); 6.13, 6.16
(2 × d, 2 × 3JHH = 6.8 Hz, 2 × 2H, C6H4, p-cymene); 7.22 (s, 4H, C6H4). 13C{1H} NMR (75
MHz, DMSO-d6): 19.12 (Me, p-cymene); 23.07 (CHMe2); 23.98 (NCH2CH2CH2); 25.52
(NCH2CH2CH2); 31.09 (CHMe2); 32.65 (C5H4CH2C6H4); 53.83 (NCH2CH2CH2); 62.47
(NCH2C6H4); 79.81, 80.80 (C5H4, CH); 84.37, 86.62 (p-cymene, CH), 100.74 (p-cymene, Cq);
102.32 (C5H4, Cipso); 111.46 (p-cymene, Cq); 128.19, 128.96 (C6H4, CH); 137.01, 138.16
(C6H4, Cq). ESI-MS (m/z; ESI+): 488 ([M − Cl]+). Elemental analysis for C28H36ClNRu,
calculated C, 64.29; H, 6.94; N, 2.68%, found C, 64.35; H 6.96; N, 2.65%.
Reaction of [{Ru(η6-p-cymene)Cl2}2] with a lithium cyclopentadienide generated from 3b
An excess of 3b (0.498 g, 1.97 mmol) was dissolved in THF (20 ml) and LiBEt3H
(2.00 ml, 2.00 mmol, 1M solution in THF) was added dropwise. The mixture was stirred for
4h and then solid [{Ru(η6-p-cymene)Cl2}2] (0.294 g, 0.48 mmol) was added. The resulting
red mixture was stirred overnight at room temperature and then refluxed for 8 h. Volatiles
were evaporated in vacuum, resulting red solid was redissolved in toluene (20 ml) and
refluxed for additional 7h. Evaporation of toluene gave dark brown solid which was purified

ACCEPTED MANUSCRIPT
by chromatography on alumina (neutral, Brockmann II). The first elution with
dichloromethane gave crude 9b (0.182 g), second elution with THF was discarded and final
elution with ethanol gave pure 8b (0.173 g) as a brownish wax.
[{η6-p-cymene}{η5-C5H4CH2(p-C6H4)CH2(morpholin-1-yl)}Ru]Cl (8b)
Yield 0.173 g (33%). 1H NMR (300 MHz, DMSO-d6): 1.17 (d, 6H, 3JHH = 6.9 Hz, CHMe2);
2.24 (s, 3H, Me, p-cymene); 2.28 − 2.34 (m, 4H, NCH2CH2); 2.63 (septuplet, 1H, 3JHH = 6.9
Hz, CHMe2); 3.42 (s, 2H, NCH2C6H4); 3.53 (s, 2H, C5H4CH2C6H4); 3.51 − 3.58 partly
overlapped (m, 4H, NCH2CH2); 5.31, 5.42 (2 × pseudo t, 2 × 2H, C5H4); 6.14 (s, 4H, C6H4, p-
cymene); 7.22, 7.26 (2 × d, 2 × 3JHH = 8.2 Hz, 2 × 2H, C6H4,). 13C{1H} NMR (75 MHz,
DMSO-d6): 19.15 (Me, p-cymene); 23.07 (CHMe2); 31.13 (CHMe2); 32.71 (C5H4CH2C6H4);
53.11 (NCH2CH2); 62.05 (NCH2C6H4); 66.17 (NCH2CH2); 79.84, 80.78 (C5H4, CH); 84.36,
86.60 (p-cymene, CH), 100.76 (p-cymene, Cq); 102.18 (C5H4, Cipso); 111.50 (p-cymene, Cq);
128.24, 129.19 (C6H4, CH); 136.20, 138.36 (C6H4, Cq). ESI-MS (m/z; ESI+): 490 ([M − Cl]+).
Elemental analysis for C27H34ClNORu, calculated C, 61.76; H, 6.53; N, 2.67%, found C,
61.68; H 6.55; N, 2.70%.
[{η5-C5H4CH2(p-C6H4)CH2(morpholin-1-yl)}2Ru] (9b)
Crude 9b was purified by trituration with ethanol. Pure 9b was obtained as a beige
powder. Yield 0.041 g (7%).
1H NMR (300 MHz, CDCl3): 2.33 − 2.53 (m, 8H, NCH2CH2O); 3.46 (s, 8H, NCH2C6H4 and
C5H4CH2C6H4); 3.62 − 3.79 (m, 8H, NCH2CH2O); 4.38 − 4.40 (m, 4H, C5H4); 4.42 − 4.45
(m, 4H, C5H4); 7.12, 7.22 (2 × d, 2 × 3JHH = 8.0 Hz, 2 × 4H, C6H4). 13C{1H} NMR (75 MHz,
CDCl3): 35.48 (C5H4CH2C6H4); 53.73 (NCH2CH2O); 63.31 (NCH2C6H4); 67.12
(NCH2CH2O); 70.49; 71.96 (C5H4, CH); 91.70 (C5H4, Cipso); 128.47, 129.20 (C6H4, CH);
135.41, 140.79 (C6H4, Cq). ESI-MS, m/z (ESI+): 611 ([M + H]+), 490. Elemental analysis for
C34H40N2O2Ru, calculated C, 66.97; H, 6.61; N, 4.60%, found C, 67.12; H 6.67; N, 4.58%.
An alternative preparation of 9b could be performed by reaction of 8b with excess of
3a in boiling toluene as follows. A suspension of an excess of lithium cyclopentadienide
(0.066 g, 0.25 mmol; prepared from LiBEt3H and 3b) and 8b (0.052 g, 0.08 mmol) was
stirred in boiling toluene for 63 h. After cooling to room temperature, the suspension was
passed through short column (2 × 10 cm) of neutral alumina (Brockmann II). The column was

ACCEPTED MANUSCRIPT
washed with toluene (total volume 30ml) and the product was eluted with dichloromethane.
Evaporation of solvents from the eluent gave 9b as an almost white solid. Yield 0.045 g
(87%). NMR spectra of the product were identical as for the above prepared complex.
X-ray Crystallography
Single-crystal X-ray diffraction data for 4a and 4a·2HCl were obtained using Nonius
KappaCCD difractometer equipped with Bruker ApexII detector. Data were collected using
monochromatized MoKα radiation (λ=0.71073 Å) at 150(2)K. The phase problem was solved
by direct methods (SHELXS)[27] and refined by full-matrix least squares based on F2
(SHELXL97).[28] The hydrogen atoms were fixed into idealized positions (riding model) and
assigned temperature factors Hiso(H) = 1.5 Ueq(pivot atom) for the methyl groups and Hiso(H)
= 1.2 Ueq(pivot atom) for the rest. The graphic depiction of molecular structure was carried
out with the PLATON program.[29] Selected crystal data for 4a and 4a·2HCl are presented in
Supporting information. Crystallographic data for the structure of 4a and 4a·2HCl have been
deposited with the Cambridge Crystallographic Data Centre with CCDC numbers 1508819
and 1508820. Copies of the data can be obtained, free of charge, on application to Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK, (fax: +44-(0)1223-
336033 or e-mail: deposit@ccdc.cam.ac.Uk).
Cytotoxicity Tests
Due to the growth rate and size, the A2780, A2780cis cells were seeded in density
10000 cells/well for 24h and 5000 cells/well for 72h treatment. SK-OV-3 cells were seeded in
density 5000 cells/well for 24h and 2000 cells/well for 72h treatment. The next day the cells
were exposed to all tested compounds diluted in DMSO in concentration range 0 − 100 µM
(each in pentaplicates) for mentioned time points. The cell viability was measured using
colorimetric MTT assay as described previously.[30] All experiments were made
independently in triplicates. Data from cytotoxicity assay were analysed in GraphPadPrism
software and expressed as IC50 values (compound concentrations that produce 50% of cell
metabolic inhibition). Errors were calculated as standard deviations (SD) and confidence
interval is expressed as 1SD. A cytotoxicity of complex 4c was not tested due to its
insolubility in DMSO as well as in biological medium.