Transformations of the natural cytokinin N6-isopentenyladenine in aqueous acidic media: Structural aspects

Article abstract of DOI:10.1039/c1ob05649b

N6-Isopentenyladenine (L1) was subjected to variously acidic media in 0.1 M, 1 M and 2 M HCl. In dependence on the acidity of the medium, the formation of three main acid hydrolysis products, involving the N6-isopentenyladeninium (HL1) (1), 7,8,9,10-tetrahydro-7,7-dimethyl-3H-pyrimido[2,1-i]purin-6-ium (HL2) (2) or 5-amino-4-(4,4-dimethyl-3,4,5,6-tetrahydropyrimidin-2-yl)-imidazolium (H₂L3) (3-5) cations, were determined and characterized by multinuclear solution-state NMR spectroscopy and in the solid state by single crystal X-ray analysis. The coordination abilities of these transformation products have been also investigated. The compounds of the compositions [Zn(HL1)Cl₃]·H₂O (1), [Zn₃(HL2) ₂Cl₈] (2), (H₂L3)[CuCl₄] (4) and (H₂L3)[ZnCl₄] (5) have been prepared in dependence on the acidity of the medium used by the reactions of L1 with ZnCl₂· 1.5H₂O or CuCl₂·2H₂O. Based on the NMR spectroscopic and X-ray crystallographic results, the mechanism of transformation of L1 in the acidic medium, involving the protonation, cyclization and ring fission, has been suggested. The Royal Society of Chemistry 2011.

Full text of DOI:10.1039/c1ob05649b

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 5703
www.rsc.org/obc
PAPER
Transformations of the natural cytokinin N6-isopentenyladenine in aqueous
acidic media: structural aspects†
a
a
b
a
b
ˇ
Zdeneˇk Tra´vn´ıcˇek,* Radka Novotna´, Jarom´ır Marek, Igor Popa and Michal Sipl
Received 26th April 2011, Accepted 16th May 2011
DOI: 10.1039/c1ob05649b
N6-Isopentenyladenine (L1) was subjected to variously acidic media in 0.1 M, 1 M and 2 M HCl. In
dependence on the acidity of the medium, the formation of three main acid hydrolysis products,
involving the N6-isopentenyladeninium (HL1) (1), 7,8,9,10-tetrahydro-7,7-dimethyl-3H-
pyrimido[2,1-i]purin-6-ium (HL2) (2) or 5-amino-4-(4,4-dimethyl-3,4,5,6-tetrahydropyrimidin-
2-yl)-imidazolium (H2L3) (3–5) cations, were determined and characterized by multinuclear
solution-state NMR spectroscopy and in the solid state by single crystal X-ray analysis. The
coordination abilities of these transformation products have been also investigated. The compounds of
the compositions [Zn(HL1)Cl3]·H2O (1), [Zn3(HL2)2Cl8] (2), (H2L3)[CuCl4] (4) and (H2L3)[ZnCl4] (5)
have been prepared in dependence on the acidity of the medium used by the reactions of L1 with
ZnCl2·1.5H2O or CuCl2·2H2O. Based on the NMR spectroscopic and X-ray crystallographic results,
the mechanism of transformation of L1 in the acidic medium, involving the protonation, cyclization
and ring fission, has been suggested.
composition and structure (generally C2, N6, N9 trisubstituted
Introduction
derivatives of adenine) proved to act as more potent and selective
CDKI. The most promising representative of the adenine-based
CDKI, roscovitine, 2-(1-ethyl-2-hydroxyethylamino)-N6-benzyl-
9-isopropyladenine, has successfully passed in vivo and preclinical
testing and it is currently being tested (named as Seliciclib or
CYC202) in the 2b-phase of clinical trials on patients with non-
small cell lung cancer.7,8
N6-isopentenyladenine (L1) belongs to a group of plant growth
regulators (hormones) known as cytokinins whose main phys-
iological role is to induce plant cell division.1 This compound
was among the first substances identified as nonspecific cyclin-
dependent kinase inhibitors (CDKI).2 Cyclin-dependent kinases
(CDK) represent a group of serine-threonine kinase enzymes,
which control the progression of cell cycle in proliferating eu-
karyotic cells.3,4 CDK become active after association with their
regulatory partners, i.e. coenzymes called cyclins. Inappropriate
expression and/or mutations of CDK, cyclins, and natural cyclin-
dependent kinase inhibitors (CDKI) have been found in various
cancers.5 For these reasons, CDK activity has been targeted for
drug development and a number of small molecules have now been
identified as pharmaceutical CDK inhibitors, which can be used
for the reduction of cell division in various tumours.6 However, N6-
isopentenyladenine as CDKI was of limited interest because of its
poor selectivity.2 Later, it was found that the compounds of similar
We have focused on the investigation of the reaction course
and reaction products between cytokinins, CDKI or derived
compounds and simple inorganic salts of selected transition metals
in dependence on different pH and molar ratio in our systematic
study of these systems. The findings of our laboratory have
proved that the resulting biological activity (cytotoxicity, SOD-
like activity) can be substantially increased by coordination of
the organic substance to a suitable transition metal centre.9,10 The
question of stability of adenine-based cytokinins in low pH has
to be considered, because acidic treatments are usually performed
during their isolation and purification from plant tissues. From the
literature sources, we have found that the acid hydrolysis of N6-
isopentenyladenine has been investigated and two products have
been isolated: the hydrated derivative N6-(3-hydroxy-3-methylbut-
2-enyl)adenine and, under more vigorously acidic conditions, tri-
cyclic 7,8,9-trihydro-7,7-dimethyl-3H-pyrimido[2,1-i]purine. The
aRegional Centre of Advanced Technologies and Materials, Department of
Inorganic Chemistry, Faculty of Science, Palacky´ University, 17. listopadu
12, CZ-771 46, Olomouc, Czech Republic. E-mail: zdenek.travnicek@
upol.cz; Fax: +420585 634 954; Tel: +420 585 634 352
bDepartment of Inorganic Chemistry, Faculty of Science, Palacky´ University,
17. listopadu 12, CZ-771 46, Olomouc, Czech Republic
1
products were characterised by elemental analysis, UV and H
NMR spectroscopy and mass spectrometry.11 Ring-fission was
not observed under the used conditions. The study of stability of
a structurally similar compound, zeatin (2-methyl-4-(7H-purin-
6-ylamino)but-2-en-1-ol), another natural cytokinin, has been
† Electronic supplementary information (ESI) available: Additional data
regarding the hydrogen bonding for compounds 1–5 (Table S1–S5) and
crystal structures (Fig. S1–S5) of the presented compounds. CCDC
reference numbers 809163–809167. For ESI and crystallographic data in
CIF or other electronic format see DOI: 10.1039/c1ob05649b
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5703–5713 | 5703
©

View Article Online
performed.12 The transformation of zeatin in 1 M HCl at 100 ◦C
was studied; a mixture of three products was acquired and charac-
terized by elemental analysis, IR, UV and 1H NMR spectroscopy,
EI MS spectrometry and HPLC analysis. These compounds were
identified as a diol resulting from the hydration of the double
bond of the aliphatic chain, and as two cyclized products, a
hydroxypyrrolidine and dihydropyrrole. The cyclization of the
aliphatic chain to purine forming a tricyclic compound was not
proved and no ring-fission was detected under the used conditions.
In general, ring-opening of different adenine derivatives has
been extensively studied in various pH and temperatures, by 15N
isotope marking or with different influence of substituents.13–19
Also, the ring-opening mechanism occurs in varied synthetic
procedures.20–23 For example, recently, the ring-opening reactions
of 3-b-D-ribofuranosyl-3,7,8,9-tetrahydropyrimido[1,2-i]purin-8-
ol (prepared by cyclization from adenosine and epichlorohydrin)
in basic pH and of its aglycon in acidic conditions, have been
reported. Both reactions afforded 2-(5-amino-1-R-imidazol-4-yl)-
1,4,5,6-tetrahydropyrimidin-5-ol (R = b-D-ribofuranosyl or H).24
This compound was reacted with CS2 or HNO2 leading to
ring-closure which resulted in a thione, and triazine product,
respectively. The physical study of the compounds was performed
by HPLC, MS and NMR methods.
HCl for acid hydrolysis. Additionally, in our study centred on the
influence of the presence of the metal in the reaction mixture on the
reaction course, we used simple inorganic salts ZnCl2·1.5H2O or
CuCl2·2H2O. Therefore, using HCl in these syntheses was rational
in order not to introduce a different anion (i.e. from another acid)
into the reaction system. The course of acid hydrolysis of L1 can
be observed on the prepared compounds 1, 2 and 3–5. Simple
protonation at the N1 atom of the purine moiety occurs in 0.1
M HCl. When the concentration is raised to 1 M, the cyclization
of the isopentenyl chain to the purine skeleton is observed, which
affords the formation of the 7,8,9,10-tetrahydro-7,7-dimethyl-3H-
pyrimido[2,1-i]purin-6-ium cation (HL2). Finally, when L1 is
held in 2 M HCl and 100 ◦C, the ring-opening occurs within
the pyrimidine moiety which is accompanied by the elimina-
tion of the C2 atom, thus giving the 5-amino-4-(4,4-dimethyl-
3,4,5,6-tetrahydropyrimidin-2-yl)-imidazolium dication (H2L3)
(Scheme 1).
Although various adenine derivatives (as mentioned above, Ref.
13–24) have been studied for their behaviour in acidic conditions,
their hydrolysis products have not been characterized by single
crystal X-ray analysis yet. Moreover, to our best knowledge, no
ring-opening products have been confirmed and/or described for
adenine-type cytokinins or cytokinin-derived compounds so far.
Herein, we describe the first concurrent NMR spectroscopic and
X-ray crystallographic study of the transformation products of
N6-isopentenyladenine (L1), in aqueous HCl in the presence of
Zn(II) and Cu(II). The aims of this study included determination
of the reaction conditions (especially, pH and reaction time)
leading to L1 acid hydrolysis to the point of ring-opening and
unambiguous elucidation of the structures of the hydrolysis
products, and moreover, we also studied the coordination abilities
of the L1 acid hydrolysis products to the metal ions [Cu(II) and
Zn(II)].
Scheme 1 A schematic representation showing the reaction condi-
tions leading to individual acid hydrolysis transformation products of
N6-isopentenyladenine.
Results and discussion
Five compounds [Zn(HL1)Cl3]·H2O (1), [Zn3(HL2)2Cl8] (2),
(H2L3)Cl2·2H2O (3), (H2L3)[CuCl4] (4) and (H2L3)[ZnCl4] (5)
have been obtained employing N6-isopentenyladenine (L1) and an
appropriate zinc(II) or copper(II) salt as the starting compounds
subjected to 0.1 M HCl (in the case of 1), 1 M HCl (in the case
of 2) and 2 M HCl (in the cases of 3–5). These variously acidic
reaction conditions were used in this study because, in general,
the knowledge of the stability of cytokinins and cytokinin-derived
compounds is essential for determining the suitable conditions for
their isolation and purification from plant tissues12,25,26 where wide
ranges of pH are commonly used (e.g. pH 2.9 in the process of
extraction of zeatin from barley27). Moreover, acidic conditions
are also often used during the preparation of their transition
metal complexes.28 Aqueous hydrochloric acid was used for the
presented syntheses based on the information from the literature
research and mainly from the studies11,12 focused on structurally
similar compounds where the authors used variously concentrated
The previously reported studies on acid hydrolysis of the title
compound, cytokinin L1,11 and the structurally related zeatin,12
also a cytokinin, attracted our attention to the problem of acid
hydrolysis in cytokinins. The authors Martin and Reese11 described
as early as in 1968 the course of acid hydrolysis of L1 up to the
tricyclic compound HL2 and they suggested the ring-opening in
the system, however, no direct proof of the assumption about the
formation of any product resulting from pyrimidine ring fission
was presented. The isolated compounds were all characterized
by means of physical-chemical methods analyzing the samples in
solution. Our motivation was to supply the most unambiguous
evidence about the hydrolysis product structures in the solid state,
which can be unequivocally accomplished only by obtaining single
crystals of the compounds and determining their molecular and
crystal structures by X-ray analysis. Additionally, we wanted to
go one step further in the acid hydrolysis of L1 to find the
5704 | Org. Biomol. Chem., 2011, 9, 5703–5713
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
reaction conditions for successful isolation of the ring-fission
hydrolysis product and confirm the structural assumptions from
the literature. Metal ions Zn(II) and Cu(II) were introduced in the
system because no similar study focused on the metal presence
influence on acid hydrolysis has been performed up to now. We
wanted to find out if there would be any differences in the reaction
courses with and without the metal in the system. In our efforts
to elucidate the X-ray structures of the acid hydrolysis products,
we were unable to prepare single crystals of most of the organic
compounds, which is why we tried adding a metal in the system
and see if the resulting coordination compounds might more
easily form single crystals. We wanted to analyze if any and what
type of coordination compounds would form in the process of
acid hydrolyses. We were confident that L1 would coordinate
to the transition metals based on our previous rich experience
in coordination chemistry of cytokinins and cytokinin-derived
compounds.9,10,28 No data have been found in the literature about
the formation of metal complexes with the compounds formed
during acid hydrolyses. The isolation of the coordination com-
pounds of the hydrolysis products might, moreover, be interesting
for future studies of biological activity (e.g. cytotoxicity) of these
compounds. The starting title compound L1 exhibits intrinsic bi-
ological activity (see Introduction). Our experience from previous
years confirms that coordination to a suitable transition metal ion
can lead to significant increase in biological activity.9,10,28 It would
therefore be interesting to look at the differences in biological
activity with the change of structural parameters caused by acid
hydrolysis.
In determining the reaction times for isolating the hydrolysis
products, the aliquots were checked by tlc. The desired termination
of reactions was, when no mixture of compounds was found by
tlc, so that we could identify the final hydrolysis product under
the used conditions. Moreover, in our efforts to obtain single
crystals of the hydrolysis products, the probability of single crystal
formation is higher, if a single chemical individuum is present in
the system. Slow evaporation of the solution solvent then might
lead to the successful formation of single crystals suitable for
X-ray analysis. This approach was, however, less successful in
the case of the acid hydrolysis performed without the presence
of the metals. When L1 was subjected to 0.1 M HCl for 30
min, slow evaporation of the solvent, after the termination of
the reaction, afforded a white powder, whose composition was
after NMR spectroscopic analysis suggested to be a mixture of
three protonated L1 molecules with the proton attached to the
nitrogens N7, N3 and N1 in the approximate ratio of 6 : 3 : 1. The
compounds were inseparable and have almost the same Rf on
tlc. On the other hand, when the same reaction was performed
in the presence of ZnCl2·1.5H2O, single crystals were obtained,
and NMR in the solution and X-ray analysis in the solid state
proved the formation of [Zn(HL1)Cl3]·H2O (1). The coordination
of the organic molecule to Zn(II) via N7 caused the redistribution
of electron density and the preferred site for protonation was
therefore N1. A similar situation occurred in the case of acid
hydrolysis in 1 M HCl. We were not able to find the reaction
time that would lead to the formation of a single product in
the system; NMR showed the tricyclic compound HL2 to be the
prevailing component but no single crystals were obtained. This
is in accordance with the findings of Martin and Reese,11 who
concluded that a quantitative conversion of L1 to the tricyclic
HL2 occurred only under more vigorously acidic conditions, e.g.
in aqueous alcoholic fluoroboric acid at 80◦ (the product isolated
as a crystalline fluoroborate) or in trifluoroacetic acid at 60◦. On
the other hand, when ZnCl2·1.5H2O was present, single crystals of
[Zn3(HL2)2Cl8] (2) involving the 7,8,9,10-tetrahydro-7,7-dimethyl-
3H-pyrimido[2,1-i]purin-6-ium cation coordinated also via N7,
were obtained and the molecular and crystal structures were
determined. The formation of the trimetallic structure bridged
by the chlorido ligands might be possibly connected to the higher
concentration of the chloride anions in the reaction mixture and
their high donor ability. The ring-opening did not occur under
these conditions no matter what reaction time was used. More
vigorously acidic conditions were needed. The only case, where
we were able to isolate single crystals of the purely organic acid
hydrolysis product, was from a 2-day reaction at 100 ◦C in 2
M HCl. The compound was identified to be the in previous
studies assumed product of ring fission (H2L3)Cl2·2H2O (3), in-
volving the 5-amino-4-(4,4-dimethyl-3,4,5,6-tetrahydropyrimidin-
2-yl)-imidazolium dication. Still, as compared to the reaction
involving ZnCl2·1.5H2O or CuCl2·2H2O, the acid hydrolysis
without the presence of the metal took twice as long reaction
time (24 h with, and 2 days without the metal ion). When we
analysed the purely organic product after the reaction time of
24 h, the NMR study identified a mixture of two predominant
compounds, HL2 and H2L3. Concerning the coordination ability,
the H2L3 organic cation was not shown to coordinate to the metal
atom, even if copper, with a greater variability in the coordination
number than zinc, was used as the metal atom. Only the formation
of ionic compounds with the complex anion [MCl4]2- was observed
(M = Zn, Cu). This fact might be again connected with the even
higher concentration of the Cl- anions than in the case of 2 and
their high donor ability.
Based on the literature research and also results in this work, the
course of acid hydrolysis of L1 in 2 M HCl can be described in four
steps. The first step involves hydration of the isopentenyl double
bond of the aliphatic chain following the Markovnikov’s rule, thus
affording the intermediate I (Scheme 2). We were not able to isolate
and identify this intermediate I, but it is described in Ref. 11, where
the authors, after a reaction of N6-isopentenyladenosine in HCl
at 95 ◦C for 35 min, obtained a mixture of unmodified L1 and its
1
hydrated analogue (I). This compound was characterized by H
NMR, MS, UV-Vis and MS spectroscopy. This step of hydration is
followed by protonation, and consequently by the dehydration and
cyclization of the aliphatic chain to the purine skeleton, which re-
sults in the formation of HL2 with a stable six-membered ring. Up
tothispoint, theacidhydrolysisproductsweredescribedbyMartin
and Reese,11 but still the structures of the transformation products
were elucidated indirectly from a combination of solution-state
physical-chemical methods (NMR, UV-Vis, MS). Subsequently,
this cationic compound HL2 undergoes hydrolysis followed by
the Dimroth rearrangement, when the C10 atom shifts from the
position N1 to the exocyclic N6 atom. The N1–C2 bond undergoes
cleavage, which results in ring-fission, thus forming the keto-enolic
tautomeric dication (II, III). The formamine group at the former
C4 position of purine is formed. The existence of the intermediate
III was suggested by Reese et al., however, in their work, it was
supposedly formed during alkaline hydrolysis of L1. Neither in
their, nor in our work, the intermediate was isolated, which might
be explained by the fact that this intermediate is probably not very
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5703–5713 | 5705
©

View Article Online
Scheme 2 A schematic representation of the suggested hydrolysis pathway of L1 in 2 M HCl.
1
stable and easily transforms into H2L3. The whole process of acid
hydrolysis is finalized by the elimination of formic acid and the
formation of H2L3 (see Scheme 2). The detailed characterizations
by NMR spectroscopy and X-ray analysis of the isolated products
are given below.
the conclusions from the H NMR spectrum. Due to the ligand
coordination to Zn(II) via the N7 atom, the signals of C5 (Dd =
-6.75ppm) andC8(Dd = 4.87 ppm) were most significantly shifted.
Moreover, because of N1 being the protonation site, the signal of
C2 was shifted upfield by 5.69 ppm and C6 upfield by 3.37 ppm.
Additionally, significant coordination shifts were detected for C9
(3.93 ppm), C10 (-3.47 ppm) and C11 (3.17 ppm), which can
be caused by a different orientation of the aliphatic chain in
the structure of 1 and also the presence of a network of non-
covalent interactions (proved in the solid state by X-ray analysis,
see below) also influencing the chemical shifts of the carbon
resonances.
NMR spectroscopic study
An extensive NMR spectroscopic study was performed in order to
elucidate the course of N6-isopentenyladenine (L1) acid hydrolysis
and the composition of the hydrolysis products in the solution
phase. Therefore, the reference spectra of L1 were compared
with those of the isolated products of L1 acid hydrolysis, namely
[Zn(HL1)Cl3]·H2O (1), [Zn3(HL2)2Cl8] (2), (H2L3)Cl2·2H2O (3)
and (H2L3)[ZnCl4] (5). The chemical shifts of the individual proton
and carbon signals were assigned by the one-dimensional 1H and
13C as well as two-dimensional correlation experiments 1H–1H gs-
COSY, 1H–13C gs-HMQC and 1H–13C gs-HMBC.
The proton and carbon spectra of [Zn(HL1)Cl3]·H2O (1)
were qualitatively similar to those of N6-isopentenyladenine (L1)
because the organic part of complex 1 maintains the L1 structure.
Only N1-protonation and N7-coordination to the zinc(II) atom
influence the chemical shifts in the spectra of 1. The changes in the
chemical shifts of the resonances can be discussed as coordination
shifts, Dd = dcomplex - dligand . The analysis of coordination shifts then
gives crucial information leading to elucidation of protonation and
coordination sites in 1. Therefore, in the 1H NMR spectrum of 1,
the most significant coordination shifts were found for the signals
belonging to C2H (0.42 ppm) and C8H (0.67 ppm) due to N1-
protonation, and coordination via N7, respectively. The greatest
coordination shift was, however, found for the signal N6H, which
was shifted downfield by 2.47 ppm. This can be explained by
the incorporation of the N6H group in the intramolecular N–
H ◊ ◊ ◊ Cl hydrogen bonds (as proved by X-ray analysis, see below).
The results following from the 13C NMR spectrum supported
The organic part of [Zn3(HL2)2Cl8] (2) is represented by
the 7,8,9,10-tetrahydro-7,7-dimethyl-3H-pyrimido[2,1-i]purin-6-
ium cation (HL2), a tricyclic acid hydrolysis product of L1 in
1 M HCl. As expected, the proton and carbon spectra of 2
differed significantly from the spectra of L1. The signals exhibited
significant changes in chemical shifts. The cyclization of the
aliphatic chain results in hindrance to free rotation, therefore,
the two C9-attached hydrogens appeared as two distinct signals
mutually shifted by 0.15 ppm, contrary to the L1 1H NMR
spectrum where the two hydrogens exhibited the equivalent
chemical shifts. Moreover, one new signal of C-attached hydrogen
was observed in the spectrum of 2 because there are two hydrogens
bonded to the C10 atom (there is no C10 C11 double bond, as
in L1). The change of C10 from a methynyl to methenyl carbon
results in a shift of the hydrogen signals towards lower frequencies.
The C10Ha and C10Hb resonances were detected to be mutually
shifted by 0.19 ppm. This change of the C10 carbon environment
was even more significant in the 13C NMR spectrum. It was clearly
distinguishable from the APT (attached proton test) experiment,
where the peak assignable to C10 was positive (= CH2 group)
contrary to the C10 signal in L1, which was inverted (= CH group).
Moreover, the signal in the spectrum of 2 was shifted upfield by
90.23 ppm. A similarly high change in chemical shift as compared
5706 | Org. Biomol. Chem., 2011, 9, 5703–5713
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
to L1 was observed for the C11 signal (in L1 at 133.76 ppm and
in 2 at 60.08 ppm), as this carbon is a part of a cycle and bonded
to nitrogen in the structure of 2.
products, we decided to prepare single crystals of all the above-
discussed products for X-ray characterizations in the solid state.
The crystals of the compounds 1–5 were obtained by slow
evaporation of a solvent from the reaction solution at room
temperature. The detailed description of molecular and crystal
structures is given below.
The compounds (H2L3)Cl2·2H2O (3) and (H2L3)[ZnCl4] (5)
contain the 5-amino-4-(4,4-dimethyl-tetrahydropyrimidin-2-yl)-
imidazolium dication (H2L3) which formed after L1 hydrolysis in
2 M HCl. Therefore, the NMR spectra of the two compounds
are analogous apart from several chemical shift differences
caused most likely by diverse network of hydrogen-bonds in
both structures (confirmed also in the solid state by X-ray
analysis, see below). The ring-fission results in two single-bonded
heterocyclic rings, the tetrahydropyrimidine and imidazole ones.
The C-attached protons on the tetrahydropyrimidine ring became
equivalent and their resonances appeared at the same values of
chemical shift. Also the signals of all the six methyl hydrogens
The structure of [Zn(HL1)Cl3]·H2O (1)
The crystal structure of 1 consists of one molecule of [Zn(HL1)Cl3]
and one crystal water molecule (Fig. 2). The Zn(II) atom is
four-coordinated by three terminal chlorido ligands and single
protonated L1 molecule (HL1) forming a ZnNCl3 donor set. HL1
is represented by the N1-protonated N9-tautomer coordinated to
Zn via the N7 atom of the purine skeleton. The geometry around
the Zn(II) centre can be described as a distorted tetrahedron (see
angles in the Fig. 2 legend). The distortion can be quantified using
a four-coordinate geometry index developed by Houser et al.29
t4 = [360◦-(a + b)]/141◦ (a and b being the largest angles in the
four-coordinate species); for 1 the value equals 0.89. The selected
bond lengths and angles are listed in the legend to Fig. 2 and the
hydrogen bond parameters are summarized in Table S1 (ESI†).
The coordinated organic ligand contains two aromatic rings,
i.e. pyrimidine and imidazole. These rings were found to be nearly
coplanar, because they form the dihedral angle of 0.81(15)◦. One
methyl group (C13 atom) of the isopentenyl chain of HL1 has been
found disordered over two positions with the occupancy factors of
0.819(8) and 0.181(8). The molecular structure of uncoordinated
N6-isopentenyladenine (L1) was previously reported.30 Significant
differences have been found after comparing the interatomic
parameters of free L1 and HL1 coordinated in 1. The angle
C8–N7–C5 is significantly enlarged in 1 [104.5(3)◦] as compared
to free L1 [101.83(5)◦] due to coordination to Zn(II). The same
1
were in the H NMR spectra of 3 and 5 observed as one singlet.
Additionally, the C2H signal disappeared from the 1H spectra as
well as the C2 signal (in the original purine ring numbering) from
the carbon spectra. This is due to deformylation in the final phase
of L1 acid hydrolysis (see Scheme 2). This change in the number of
carbon atoms is well observable on the comparison of the 1H–13C
gs-HMQC spectra of 2 and 5 which are displayed in Fig. 1.
The values of chemical shifts from the NMR spectra of all
the studied compounds are listed in the Experimental section.
The atom numbering used in the description of NMR spectra
is consistent with the numbering in the section X-ray structures
(below, Fig. 2–6). The carbon atom numbering of the organic parts
in 2 and 3 is also shown in Fig. 1.
X-ray structures
In an effort to independently validate the results following from
the solution-state NMR spectra regarding the L1 acid hydrolysis
Fig. 1 The 1H–13C gs-HMQC NMR spectra of 2 (left) and 5 (right) demonstrating the change in the number of carbon atoms after ring fission (right);
inset: the cationic parts of the compounds with the carbon atom numbering.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5703–5713 | 5707
©

View Article Online
- y, 0.5 + z; (v) 1 - x, 1 - y, -z]. The figure showing a part of the
crystal structure of 1 (Fig. S1) is presented in ESI.†
The structure of [Zn3(HL2)2Cl8] (2)
The molecular structure of 2 (Fig. 3) is comprised of one trimetallic
Zn(II) complex [Zn3(HL2)2Cl8] with two bridging and six terminal
chlorido ligands, and two terminal N-donor organic ligands, i.e.
the 7,8,9,10-tetrahydro-7,7-dimethyl-3H-pyrimido[2,1-i]purin-6-
ium cations (HL2). All the three Zn(II) atoms are tetrahedrally
coordinated, the central Zn(II) centre in a ZnCl4 donor set,
while the other two terminal Zn(II) atoms in ZnNCl3 donor
˚
sets. The Zn(II) atoms are mutually separated by 3.780(4) A
˚
[Zn1 ◊ ◊ ◊ Zn2] and 3.885(4) A [Zn2 ◊ ◊ ◊ Zn3], while the terminal ones
˚
are distanced by 7.285(5) A [Zn1 ◊ ◊ ◊ Zn3]. The four-coordinate
geometry indexes29 are equal to 0.88, 0.93, and 0.91 for the
geometry around Zn1, Zn2 and Zn3, respectively. The selected
bond lengths and angles are given in the legend to Fig. 3 and the
hydrogen bond parameters are summarized in Table S2 (ESI†).
Fig. 2 The molecular structure of [Zn(HL1)Cl3]·H2O (1). One methyl
group on the isopentenyl chain has been found disordered over two
positions with the occupancy factors of 0.819(8) and 0.181(8). The
atoms with the occupancy factor 0.181(8) are displayed in light grey
colour for clarity reasons. The non-hydrogen atoms are displayed as
displacement ellipsoids at the 50% probability level. Selected bond
◦
˚
lengths (A) and angles ( ): Zn–N7, 2.062(3); Zn–Cl1, 2.2557(11); Zn–Cl2,
2.2778(11); Zn–Cl3, 2.2223(11); N1–C2, 1.368(5); N3–C4, 1.350(6);
N7–C8, 1.327(5); N9–C8, 1.345(6); N7–Zn–Cl1, 108.28(10); N7–Zn–Cl2,
101.63(10); N7–Zn–Cl3, 104.85(10); Cl1–Zn–Cl2, 105.60(4); Cl1–Zn–Cl3,
118.35(4); Cl2–Zn–Cl3, 116.62(5); C8–N7–Zn, 120.7(3); C8–N7–C5,
104.5(3); C8–N9–C4, 107.4(4); C2–N1–C6, 124.1(4); C2–N3–C4, 112.0(4).
result arises from protonation at N1; the angle C2–N1–C6 equals
124.1(4)◦ in 1 and 119.05(6)◦ in L1. Moreover, due to hydrogen
bonding, in which protons attached to N9 and N6 are involved,
the angles at these sites decrease in comparison to free L1 [C8–
N9–C4 equals 107.4(4)◦ in 1 and 103.69(5)◦ in L1; and C6–N6–C9
is 122.6(3)◦ in 1 and 123.6(5)◦ in L1]. Additionally, the aliphatic
chain bonded at N6 has been found to be differently oriented in
both discussed compounds, because the torsion angles C6–N6–
C9–C10 in 1 and L1 differ significantly; 172.73(35)◦ as compared
to 94.08(71)◦ in 1, and L1, respectively.
Fig. 3 The molecular structure of [Zn3(HL2)2Cl8] (2). A part of the
3,4,5,6-tetrahydropyrimidine ring in HL2A has been found disordered
over two positions with the occupancy factors 0.823(5) and 0.177(5). The
atoms with the occupancy factor 0.177(5) are displayed in light grey colour
for clarity reasons. The non-hydrogen atoms are displayed as displacement
˚
ellipsoids at the 50% probability level. Selected bond lengths (A) and
angles (◦): Zn1–N7, 2.040(2); Zn1–Cl1, 2.1975(6); Zn1–Cl2, 2.2660(6);
Zn1–Cl3, 2.3059(6); Zn2–Cl4, 2.2018(7); Zn2–Cl5, 2.2278(7); Zn2–Cl6,
2.3399(7); Zn2–Cl3, 2.3576(6); Zn3–N7A, 2.044(2); Zn3–Cl7, 2.2137(7);
Zn3–Cl8, 2.2316(7); Zn3–Cl6, 2.3227(7); N7–Zn1–Cl1, 108.60(6);
N7–Zn1–Cl3, 104.90(6); Cl1–Zn1–Cl2, 120.95(2); Cl4–Zn2–Cl5,
12 molecular structures have been found in the Cambridge
Structural Database (CSD),31 which contain an N7-coordinated
adenine derivative coordinated to Zn(II) in the ZnCl3N donor set.
The average Zn–Cl and Zn–N7 bond lengths in the compounds
113.49(3);
Cl4–Zn2–Cl6,
111.02(3);
Cl4–Zn2–Cl3,
114.13(3);
˚
˚
were calculated to be 2.246(19) A and 2.064(23) A, respectively.
These dimensions are quite comparable with those found in 1,
where the Zn–Cl bond lengths are equal to 2.2223(11)–2.2778(11)
Cl6–Zn2–Cl3, 94.48(2); N7A–Zn3–Cl7, 112.70(6); N7A–Zn3–Cl8,
105.75(6); Cl7–Zn3–Cl8, 118.86(3); Cl7–Zn3–Cl6, 111.38(3).
˚
˚
A and the Zn–N7 bond length to 2.062(3) A.
The two coordinated organic cations (further denoted as HL2
and HL2A) contain three heterocyclic rings, i.e. pyrimidine (A),
imidazole (B) and 3,4,5,6-tetrahydropyrimidine (C). The rings A
and B are almost coplanar forming dihedral angles of 2.59(8)◦
and 2.81(8)◦ in HL2, and HL2A, respectively. A part of the ring
C in HL2A (C9, C10, C11, C12, C13 and the attached hydrogen
atoms) has been found disordered over two positions with the
occupancy factors of 0.823(5) and 0.177(5). The only molecular
The crystal structure of 1 has been found to be stabilized
by hydrogen bonds of the types N–H ◊ ◊ ◊ Cl, N–H ◊ ◊ ◊ O and O–
H ◊ ◊ ◊ Cl (Fig. S1). Also, the non-covalent interactions C–H ◊ ◊ ◊ Cl,
C–H ◊ ◊ ◊ O and C ◊ ◊ ◊ Cl contribute to stabilization of the crystal
ii
iii
˚
˚
structure [C9 ◊ ◊ ◊ Cl1 = 3.841(4) A, C8 ◊ ◊ ◊ Cl3 = 3.515(5) A,
C13A ◊ ◊ ◊ O1iv = 3.550(7) A, C6 ◊ ◊ ◊ Cl1 = 3.372(4) A; symmetry
v
˚
˚
codes: (ii) 1 - x, -0.5 + y, 0.5 - z; (iii) -x, 1 - y, -z; (iv) 1 + x, 0.5
5708 | Org. Biomol. Chem., 2011, 9, 5703–5713
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
structure deposited in CSD31 containing the same heterocyclic
rings, yet differently substituted and protonated on a different
nitrogen, is 7,8,9,10-tetrahydro-7,8-dihydroxy-7,9-dimethyl-3H-
pyrimido[2,1-i]purine-9-carboxylate dihydrate.32
The Zn(II) ions in the structure of 2 are connected by chlorido
bridges. Chloride, and halide in general, anions often act as
bridging ligands due to their high donor abilities connected to the
presence of four lone pairs in the valence shell. For this reason, they
can connect up to four metal centres forming polynuclear species.
The [Zn2Cl6]2- anion can be named as a simple representative
of a zinc compound involving chlorido bridges. There are two
Cl- anions bridging the two metal centres and four terminal
chlorido ligands. The search in CSD31 showed that 32 X-ray
structures have already been deposited in the database containing
this complex anion. Moreover, the CSD database also contains
2 X-ray structures33,34 in which three Zn(II) centres are bridged
by one chlorido ligand and 3 X-ray structures35–37 where one Cl-
anion bridges four Zn(II) atoms. The specific search in CSD for a
trimetallic moiety involving three Zn(II) centres one-dimensionally
bridged by the chlorido ligands, as present in 2, showed that no
molecular structures of such trinuclear Zn(II) complexes have been
determined up to now. Therefore, the molecular structure of 2
presented herein represents the first complex of this kind. Only
11 polymeric and polynuclear Zn(II) complexes involving at least
three metal centres with chlorido bridges have been deposited
in the database. The calculated average Zn–Cl(bridging) bond
Fig. 4 The molecular structure of (H2L3)Cl2·2H2O (3). The non-hy-
drogen atoms are displayed as displacement ellipsoids at the 50%
◦
˚
probability level. Selected bond lengths (A) and angles ( ): N1–C2,
1.3123(19); N3–C4, 1.3804(18); N8–C7, 1.3151(19); N12–C11, 1.4704(18);
C2–N1–C5, 109.03(12); C2–N3–C4, 109.66(12); C7–N8–C9, 121.54(12);
C7–N12–C11, 123.43(12).
tetrahydropyrimidin-2-yl)-imidazole moiety has been deposited in
this database.
An extensive network of hydrogen bonds was found to be
present in the crystal structure of 3 (Fig. S3). Additionally, the
˚
length was found to be 2.348(59) A. The Zn–Cl(bridging) bond
˚
lengths in 2 equal 2.3059(6)–2.3576(6) A and thus are quite
comparable with the average bond length. In general, it can be
said that the Zn–Cl(bridging) are significantly longer than the
non-covalent contacts also stabilize the crystal structure of 3, i.e.
v
C–H ◊ ◊ ◊ Cl [C14 ◊ ◊ ◊ Cl2iv = 3.704(2) A], C–H ◊ ◊ ◊ O [C2 ◊ ◊ ◊ O2 =
˚
˚
Zn–Cl(terminal) ones, which equal 2.1975(6)–2.2660(6) A and are
i
i
˚
˚
˚
in good agreement with those in 1 [2.2218(12)–2.2777(12) A].
3.440(2) A] and C ◊ ◊ ◊ Cl [C4 ◊ ◊ ◊ Cl2 = 3.386(14) A, C5 ◊ ◊ ◊ Cl2 =
˚
N–H ◊ ◊ ◊ Cl and N–H ◊ ◊ ◊ N hydrogen bonds stabilize the crystal
structure of 2 (Fig. S2). Furthermore, the non-covalent contacts
of the types C–H ◊ ◊ ◊ Cl and C ◊ ◊ ◊ Cl are present within the
3.258(14) A; symmetry codes: (i) 1 + x, y, z; (iv) 0.5 + x, 1.5 - y,
-0.5 + z, (v) -0.5 + x, 1.5 - y, -0.5 + z]. The figure showing a part
of the crystal structure of 3 (Fig. S3) is presented in ESI.†
i
iv
˚
crystal structure of 2 [C12A ◊ ◊ ◊ Cl5 = 3.730(4) A, C10 ◊ ◊ ◊ Cl4
=
iv
v
˚
˚
3.746(3) A, C13 ◊ ◊ ◊ Cl4 = 3.720(3) A, C8 ◊ ◊ ◊ Cl1 = 3.543(3)
The structure of (H2L3)[CuCl4] (4)
vi
iii
˚
˚
˚
A, C13A ◊ ◊ ◊ Cl6 = 3.682(4) A; C8A ◊ ◊ ◊ Cl7 = 3.416(3) A,
C5 ◊ ◊ ◊ Cl2vii = 3.375(2) A, C6 ◊ ◊ ◊ Cl2 = 3.255(2) A; symmetry
codes: (i) -1 + x, y, z; (iii) 1 - x, 1 - y, 1 - z; (iv) x, -1 + y,
z; (v) 1 - x, 1 - y, -z; (vi) -1 + x, -1 + y, z; (vii) 1 - x, -y, -z].
The figure displaying a part of the crystal structure of 2 (Fig. S2)
is presented in ESI.†
The molecular structure (Fig. 5) of 4 is comprised of the
doubly protonated 5-amino-4-(4,4-dimethyl-3,4,5,6-tetrahydro-
pyrimidin-2-yl)-imidazolium cation and its positive charge is
compensated by the presence of the [CuCl4]2- complex anion. The
Cu(II) atom is four coordinated by four terminal chlorido ligands.
The geometry around Cu(II) is distorted tetrahedral. All the angles
around the Cu(II) centre are very significantly different from the
angle in “ideal tetrahedron”, i.e. 109◦47¢ (see Fig. 5 legend). The
compound 4 has a t4 value of 0.52. The hydrogen bond parameters
are listed in Table S4 (ESI†).
The differences of the bond lengths and angles in H2L3 in 3
and 4 are caused by diverse types of hydrogen bonds involved
in the two structures. The greatest difference was found for
the angle at N2 [N8 in 3] which equals 124.9(2)◦ in 4 and
121.54(12)◦ in 3. It is caused by the presence of the N8–H8 ◊ ◊ ◊ Cl2
hydrogen bond in the structure of 3. The two heterocyclic rings,
i.e. imidazole and 3,4,5,6-tetrahydropyrimidine, were found to be
differently mutually oriented than in 3. The dihedral angle between
the planes through the ring systems was determined to equal
40.25(10)◦. As for the [CuCl4]2- anion, 580 compounds containing
the CuCl4 moiety deposited in CSD have been found. The average
vii
˚
˚
The structure of (H2L3)Cl2·2H2O (3)
The asymmetric unit of 3 consists of one doubly proto-
nated 5-amino-4-(4,4-dimethyl-3,4,5,6-tetrahydropyrimidin-2-yl)-
imidazolium cation (H2L3), two chloride anions and two water
molecules of crystallization (Fig. 4). Selected bond lengths and
angles are summarized in the legend to Fig. 4 and the hydrogen
bond parameters are given in Table S3 (ESI†).
The H2L3 cation contains two heterocyclic ring systems, i.e.
imidazole and 3,4,5,6-tetrahydropyrimidine. The imidazole ring
only slightly deviates from planarity with the maximum deviation
˚
from the mean planes being 0.005(2) A for the atom C2. The
dihedral angle between the planes created through the two ring
systems was found to equal 35.44(5)◦. Based on the search
in CSD,31 no molecular structure containing the 4-(3,4,5,6-
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5703–5713 | 5709
©

View Article Online
Fig. 5 The molecular structure of (H2L3)[CuCl4] (4). The non-hydrogen
atoms are displayed as displacement ellipsoids◦at the 50% probability
˚
level. Selected bond lengths (A) and angles ( ): Cu1–Cl1, 2.2733(8);
Cu1–Cl2, 2.2121(8); Cu1–Cl3, 2.2741(8); Cu1–Cl4, 2.2382(7); N2–C3,
1.485(4); N6–C5, 1.455(4); N9–C8, 1.376(4); N11–C7, 1.398(3); N14–C8,
1.338(4); Cl1–Cu1–Cl3, 138.13(3); Cl1–Cu1–Cl2, 96.37(3); Cl2–Cu1–Cl4,
149.59(3); Cl3–Cu1–Cl4, 94.70(3); C1–N2–C3, 124.9(2); C1–N6–C5,
121.3(2); C8–N9–C10, 109.8(2); C7–N11–C10, 109.2(2).
˚
Cu–Cl bond length was calculated to be 2.38(20) A. This value
Fig. 6 The molecular structure of (H2L3)[ZnCl4] (5), showing the two
crystallographically independent molecules in the unit cell. The non-hydro-
gen atoms are displayed as displacement ellip◦soids at the 50% probability
is significantly longer than those found for 4 [2.2121(8)–2.2741(8)
A]. This could be caused by the fact, that greater variety of Cu–
Cl bond types is involved in the average value, i.e. compounds
containing the CuCl4 moiety in other than tetrahedral geometry
(octahedron etc.).
The crystal structure is stabilized by hydrogen bonds of the N–
H ◊ ◊ ◊ Cl type. Moreover, there are non-bonding interactions of the
type C–H ◊ ◊ ◊ Cl [C10 ◊ ◊ ◊ Cl2 = 3.552(3) A, C10 ◊ ◊ ◊ Cl3 = 3.515(3)
˚
˚
level. Selected bond lengths (A) and angles ( ) for the molecule contain-
ing Zn2: Zn2–Cl4, 2.2881(8); Zn2–Cl6, 2.2270(7); Zn2–Cl8, 2.2700(8);
Zn2–Cl9, 2.3349(7); N1–C2, 1.310(4); N6–C4, 1.354(4); N8–C7, 1.327(4);
N12–C11, 1.488(4); Cl6–Zn2–Cl9, 109.93(3); Cl6–Zn2–Cl8, 114.86(3);
Cl8–Zn2–Cl9, 103.38(3); Cl4–Zn2–Cl6, 112.91(3); C2–N1–C5, 109.7(2);
C2–N3–C4, 109.9(2); C7–N8–C9, 122.7(3); C7–N12–C11, 123.6(3).
iii
iii
˚
iv
˚
˚
A, C13 ◊ ◊ ◊ Cl2 = 3.876(4) A; symmetry codes: (iii) x, -1 + y, z;
(iv) 1 - x, 2 - y, 1 - z]. The figure showing a part of the crystal
structure of 4 (Fig. S4) is given in ESI.†
dihedral angle in 3 [35.44(5)◦]. The search in CSD showed that
479 compounds have been deposited therein, which contain the
ZnCl4 moiety. The average calculated Zn–Cl bond length equals
˚
2.271(32) A. This value correlates well with the bond lengths found
The structure of (H2L3)[ZnCl4] (5)
˚
in 5, 2.2360(8)–2.3202(7) A.
A great number of hydrogen bonds of the type N–H ◊ ◊ ◊ Cl
have been found to contribute to stabilization of the crystal
structure of 5 (Fig. S5). Furthermore, the non-covalent contacts
The asymmetric unit of 5 consists of the 5-amino-4-(4,4-dimethyl-
tetrahydropyrimidin-2-yl)-imidazolium dication and the [ZnCl4]2-
complex anion (Fig. 6). There are two crystallographically inde-
pendent molecules of 5 in the unit cell. The geometry around
the Zn(II) centre can be described as a distorted tetrahedron. The
distortion from the regular tetrahedron is far smaller than that of
the [CuCl4]2- anion in the structure of 4. The compound 5 has a t4
value of 0.95 for both crystallographically independent molecules.
The selected interatomic parameters are listed in the legend to Fig.
6 and the hydrogen bond parameters are in Table S5 (ESI†).
If we compare the interatomic parameters of H2L3 in 3 and
5, we find significant differences which are caused by different
hydrogen bonding system. The biggest difference, like in the case of
4, was found for the C7–N12–C11 angle [121.54(12)◦ and 123.6(3)◦
in 3 and 5]. The heterocyclic rings, i.e. imidazole and 3,4,5,6-
tetrahydropyrimidine, present in the two crystallographically
independent molecules, have been found to be differently mutually
oriented, they form dihedral angles of 28.69(10)◦ in H2L3 and
39.70(10)◦ in H2L3¢. These values also differ significantly from the
iii
of the types C–H ◊ ◊ ◊ Cl [C9¢ ◊ ◊ ◊ Cl6iii = 3.699(4) A, C9¢ ◊ ◊ ◊ Cl10
=
˚
iii
v
˚
˚
˚
3.611(4) A, C14 ◊ ◊ ◊ Cl10 = 3.634(3) A, C10¢ ◊ ◊ ◊ Cl6 = 3.814(3) A,
vi
ii
˚
˚
C13 ◊ ◊ ◊ Cl6 = 3.891(4) A] and C ◊ ◊ ◊ Cl [C7¢ ◊ ◊ ◊ Cl7 = 3.427(3) A;
symmetry codes: (ii) -x, 1 - y, 1 - z; (iii) 1 - x, 1 - y, 1 - z; (v) x,
y, -1 + z; (vi) -x, 2 - y, 1 - z]. The figure showing a part of the
crystal structure of 5 (Fig. S5) is presented in ESI.†
Experimental
Materials and general methods
ZnCl2·1.5H2O, CuCl2·2H2O, N6-isopentenyladenine and solvents
were purchased from commercial sources (Sigma–Aldrich Co.,
Lachema Co.) and were used as received. Elemental analyses (C,
H, N) were performed on a Fisons EA-1108 CHNS-O Elemental
Analyzer (Thermo Scientific).
5710 | Org. Biomol. Chem., 2011, 9, 5703–5713
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
NMR spectroscopy
[Zn3(HL2)2Cl8] (2). The preparation of 2 followed the
same synthetic procedure as described for 1 employing N6-
isopentenyladenine and ZnCl2·1.5H2O as the starting compounds.
However, 1 M HCl was used as the medium and the reaction time
was 5 h in this case. Colourless crystals of 2 were obtained by slow
evaporation of the reaction solution. Found: C, 26.8; H, 3.2; N,
15.7. Calc. for C20H14N10Cl8Zn3: C, 27.0; H, 3.2; N, 15.8%. Yield:
60%.
NMR data: 1H-NMR (DMF-d7, TMS, ppm): dH 10.37 (1H, s,
N6H), 9.16 (1H, bs, N9H), 8.94 (1H, s, C2H), 8.56 (1H, s, C8H),
3.79 (2H, t, J 6.0, C9H), 2.37 (2H, t, J 6.0, C10H), 1.88 (6H,
s, C12,C13H). 13C-NMR (DMF-d7, TMS, ppm): dC 147.44 (C6),
146.92 (C4), 143.81 (C2), 143.32 (C8), 117.84 (C5), 60.16 (C11),
35.66 (C9), 32.13 (C10), 27.51 (C12,C13).
1H and 13C NMR spectra were recorded on a 400 MHz Varian
spectrometer at 300 K in deuterated N,N¢-dimethylformamide
1
(DMF-d7). The internal reference standard used for H and 13C
NMR measurements was tetramethylsilane (TMS). J values are
1
given in Hz. The individual H and 13C signals were assigned by
2D correlation experiments of 1H–1H gs-COSY, 1H–13C gs-HMQC
and 1H–13C gs-HMBC. The samples for the NMR measurements
were prepared by dissolution of the compounds L1 (obtained from
commercial sources) and 1, 2, 3 and 5 (isolated from the reactions
of L1, and L1 with ZnCl2·1.5H2O in variously acidic media) in
DMF-d7.
NMR data for N6-isopentenyladenine (L1): 1H-NMR (DMF-
d7; TMS; ppm): dH 12.96 (1H, bs, N9H), 8.25 (1H, s, C2H), 8.17
(1H, s, C8H), 7.50 (1H, bs, N6H), 5.42 (1H, tt, J 6.9 and 1.4,
C10H), 4.23 (2H, bs, C9H), 1.76 (3H, s, C12H), 1.71 (3H, s, C13].
13C-NMR (DMF-d7, TMS, ppm): dC 154.96 (C6), 152.82 (C2),
150.21 (C4), 139.87 (C8), 133.76 (C11), 122.56 (C10), 119.52 (C5),
38.19 (C9), 25.21 (C13), 17.46 (C12).
Crystal data and structure refinement for complex 2:
C20H28Cl8N10Zn3, M = 888.23, triclinic, a = 8.3518(3), b =
3
˚
˚
10.0642(5), c = 21.5006(10) A, V = 1646.72(13) A , T = 120(2)K,
¯
space group P1, Z = 2, reflections collected/unique: 11581/5752
(Rint = 0.0082), final R indices: R1 = 0.0214, wR2 = 0.0695, R indices
(all data): R1 = 0.0223, wR2 = 0.0702.
(H2L3)Cl2·2H2O
(3).
A
solution
containing
N6-
X-ray crystallography
isopentenyladenine (0.20 g, 1 mmol) in 2 M HCl (20 ml)
◦
was refluxed at 100 C for 2 days. The reaction solution was left
Single-crystal data were collected with an Xcalibur2 (Oxford
Diffraction) for 1, 2 and 5, and with four-circle kappa-axis KUMA
KM-4 diffractometer (KUMA Diffraction, Wroclaw) for 3 and 4,
at 120 K using graphite-monochromated MoKa radiation and
CCD detector. CrysAlis software package was used for data
collection and reduction. The structures were solved by direct
methods [SHELXS-97] and refined on F2 using full-matrix least-
squares procedure [SHELXL-97]38 All H-atoms were found from
difference Fourier maps and were refined using a riding model,
expect for those belonging to the crystal water molecule in 1 and
3, whose positions were refined freely. Molecular graphics were
made and additional structural parameters were interpreted by
means of DIAMOND.39
standing at room temperature and crystals of 3 suitable for single
crystal X-ray analysis formed after a few days. Found: C, 35.5; H,
7.0; N, 22.9. Calc. for C9H21N5O2Cl2: C, 35.8; H, 7.0; N, 23.2%.
Yield: 45%.
NMR data: 1H-NMR (DMF-d7, TMS, ppm): dH N1,N8H – not
detected, 10.36 (1H, s, N12H), 10.06 (1H, s, N3H), 9.54 (2H, bs,
N6H), 8.91 (1H, s, C2H), 3.61 (2H, m, C11H), 1.93 (2H, t, J 6.1,
C10H), 1.49 (6H, s, C13,C14). 13C-NMR (DMF-d7, TMS, ppm):
dC 148.42 (C7), 143.48 (C4), 131.80 (C2), 99.97 (C5), 51.35 (C9),
36.46 (C11), 31.76 (C10), 27.90 (C13,14).
Crystal data and structure refinement for 3: C9H21Cl2N5O2, M =
302.21, monoclinic, a = 7.0591(4), b = 15.6552(7), c = 13.2731(6)
A, V = 1451.69(12) A , T = 120(2)K, space group P 21/n, Z =
4, reflections collected/unique: 7984/2562 (Rint = 0.0354), final R
indices: R1 = 0.0261, wR2 = 0.0674, R indices (all data): R1 =
0.0278, wR2 = 0.0684.
3
˚
˚
Syntheses
[Zn(HL1)Cl3]·H2O (1). To a stirred warm (50 ◦C) solution of
0.10 g (0.5 mmol) of N6-isopentenyladenine in 35 ml of 0.1 M HCl,
2 mmol of ZnCl2·1.5H2O was added. The resulting mixture was
kept under stirring at 50 ◦C for 30 min and then was allowed to stay
at room temperature. After a week, crystals of 1 suitable for single
crystal X-ray analysis formed by slow evaporation of the solvent.
Found: C, 30.3; H, 4.0; N, 17.9. Calc. for C10H16N5OCl3Zn: C,
30.5; H, 4.1; N, 17.8%. Yield: 72%.
NMR data: 1H-NMR (DMF-d7, TMS, ppm): dH 1H, HN9- not
detected, 9.97 (1H, t, J 6.7, N6H), 8.84 (1H, s, C8H), 8.67 (1H, s,
C2H), 5.43 (1H, t, J 6.7, C10H), 4.32 (2H, t, J 5.7, C9H), 1.77 (3H,
s, C12H), 1.73 (3H, s, C13H). 13C-NMR (DMF-d7, TMS, ppm): dC
151.59 (C6], 147.35 (C4), 147.13 (C2), 143.74 (C8), 136.93 (C11),
119.09 (C10), 112.77 (C5), 42.12 (C9), 25.08 (C13), 17.54 (C12).
Crystal data and structure refinement for complex 1: C10H16Cl3-
(H2L3)[CuCl4] (4). N6-isopentenyladenine (0.20 g, 1 mmol)
was dissolved in 10 ml of 2 M HCl. Then, a solution (10 ml) of
CuCl2·1.5H2O (0.16 g, 1 mmol) in the same solvent was added.
The resulting mixture was refluxed at 100 ◦C for 24 h. By slow
evaporation of the solvent at room temperature, single crystals
formed after a period of 5–6 weeks. Found: C, 27.1; H, 4.2; N,
17.4. Calc. for C9H17N5Cl4Cu: C, 27.0; H, 4.3; N, 17.5%. Yield:
58%.
Crystal data and structure refinement for 4: C9H17N5Cl4Cu, M =
˚
400.62, triclinic, a = 7.6673(4), b = 7.8804(5), c = 13.4349(7) A, V =
3
¯
˚
794.67(8) A , T = 120(2)K, space group P1, Z = 2, reflections
collected/unique: 4809/2778 (Rint = 0.0448), final R indices: R1 =
0.0352, wR2 = 0.0944, R indices (all data): R1 = 0.0366, wR2 =
0.0955.
N5OZn, M = 394.00, monoclinic, a = 11.4767(3), b = 9.9382(3), c =
3
˚
˚
14.0372(5) A, V = 1599.34(9) A , T = 120(2)K, space group P 21/c,
Z = 4, reflections collected/unique: 9458/2811 (Rint = 0.0144), final
R indices: R1 = 0.0413, wR2 = 0.0904, R indices (all data): R1 =
0.0418, wR2 = 0.0905.
(H2L3)[ZnCl4] (5). The same synthetic procedure as described
above for 4 was applied, only ZnCl2·1.5H2O was used as the
inorganic salt. During the reaction time, the colour of the reaction
solution changed from colourless to pale yellow. Pale yellow single
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5703–5713 | 5711
©

View Article Online
crystals formed by slow evaporation of the solvent. Found: C,
27.0; H, 4.1; N, 17.1. Calc. for C18H34N10Cl8Zn2: C, 26.9; H, 4.3;
N, 17.4%. Yield: 65%.
is alkylated {the case of 7,8,9,10-tetrahydro-7,7-dimethyl-3H-
pyrimido[2,1-i]purin-6-ium cation}, the electrophilic N1–C2–N3
is the preferred amidine-group for the hydrolysis and the opening
of the pyrimidine ring is therefore rendered possible. The compo-
sition of the presented compounds 1–5 in solution and solid state
was elucidated by heteronuclear NMR spectroscopy, and X-ray
analysis, respectively. The molecular and crystal structures of the
acidic hydrolysis transformation products of an adenine derivative
in general are herein reported for the first time.
1
NMR data: H-NMR (DMF-d7, TMS, ppm): dH N1,N2H –
not detected, 10.20 (2H, bs, N6H), 9.72 (1H, bs, N8H), 9.60 (1H,
bs, NH), 8.62 (1H, s, C2H), 3.64 (2H, m, C9H), 1.96 (2H, t, J 5.9,
C10H), 1.46 (6H, s, C13,C14). 13C-NMR (DMF-d7, TMS, ppm):
dC 149.89 (C7), 143.33 (C4), 132.59 (C2), 100.37 (C5), 51.11 (C11),
36.75 (C9), 31.85 (C10), 27.79 (C13,C14).
Crystal data and structure refinement for 5: C9H17N5Cl4Zn, M =
˚
402.45, triclinic, a = 7.8356(4), b = 13.1475(6), c = 15.8482(9) A,
Acknowledgements
3
¯
˚
V = 1568.96(14) A , T = 120(2)K, space group P1, Z = 4, reflections
collected/unique: 9645/5487 (Rint = 0.0639), final R indices: R1 =
0.0403, wR2 = 0.1079, R indices (all data): R1 = 0.0432, wR2 =
0.1100.
The authors would like to thank the Ministry of Education,
Youth and Sports of the Czech Republic (MSM6198959218),
and the Operational Program Research and Development for
Innovations – European Social Fund (CZ.1.05/2.1.00/03.0058)
for the financial support.
Conclusions
The behaviour of N6-isopentenyladenine (L1) in variously acidic
media and in the presence of Zn(II) or Cu(II) has been studied
by solution-state NMR spectroscopy and single crystal X-ray
analysis. In general, three transformation products have been ob-
served, N1-protonated N6-isopentenyladenine (HL1), the tricyclic
cation 7,8,9,10-tetrahydro-7,7-dimethyl-3H-pyrimido[2,1-i]purin-
6-ium (HL2) originating from cyclization of the aliphatic chain
to purine, and the 5-amino-4-(4,4-dimethyl-tetrahydropyrimidin-
2-yl)-imidazolium dication (H2L3), which is formed after ring-
fission. In most cases, we were unable to obtain single crystals of
the organic transformation products, which was one of the aims of
this study. We therefore added the metal ions into the system and
this study showed that the presence of a metal ion [in our case Zn(II)
or Cu(II)] in the reaction mixture had significant influence on the
course of the whole reaction process. The reaction speed of acid
hydrolysis with the inorganic chlorides present was significantly
increased and pure chemical species in the forms of single crystals
suitable for X-ray analysis could be obtained. The results following
from this approach involve structural description of the organic
hydrolysis products and at the same time, elucidation of their
interactions with the metal centres. Therefore, apart from the L1
acid hydrolysis product formation, the coordination abilities of
the corresponding intermediates to the mentioned metal ions have
been investigated. As can be seen from the composition of the
[Zn(HL1)Cl3]·H2O (1), [Zn3(HL2)2Cl8] (2), (H2L3)[CuCl4] (4) and
(H2L3)[ZnCl4] (5) complexes, the HL1 and HL2 cations directly
coordinate to the metal ion through the N7 atoms, while H2L3
does not coordinate and remains in its cationic form compensating
the charge of the complex anion [MCl4]2-. This observation might
be explained by a strong donor ability of chloride anions and
their increasing concentration across the row 1, 2, 4 and 5.
Therefore, generally, the proton and Cl- concentration are most
likely indirectly responsible for coordination or non-coordination
of the organic hydrolysis products to the metal sites. The proton
concentration influences the acid transformation of the organic
compounds and the chloride concentration affects coordination
to the metal or the formation of the complex anion [MCl4]2-.
The first step of the ring opening of L1 follows the Dimroth
rearrangement. As was previously reported,13,18 in the non-
substituted purine skeleton of adenine, the amidine group N7–
C8–N9 is the hydrolysis target. However, if either N1 or N3
References
1 P. J. Davies, Plant Hormones, third ed., Springer, Dordrecht, 1997.
2 L. Meijer and E. Raymond, Acc. Chem. Res., 2003, 36, 417–425.
3 C. E. Arris, F. T. Boyle, A. H. Calvert, N. J. Curtin, J. A. Endicott,
E. F. Garman, A. E. Gibson, B. T. Golding, S. Grant, R. J. Griffin,
P. Jewsbury, L. N. Johnson, A. M. Lawrie, D. R. Newell, M. E. M.
Noble, E. A. Sausville, R. Schultz and W. Yu, J. Med. Chem., 2000, 43,
2797–2804.
4 P. Imbach, H. G. Capraro, P. Furet, H. Mett, T. Meyer and J.
Zimmermann, Bioorg. Med. Chem. Lett., 1999, 9, 91–96.
5 M. Hall and G. Peters, Adv. Cancer Res., 1996, 68, 67–108.
6 P. M. Fischer and D. P. Lane, Curr. Med. Chem., 2000, 7, 1213–1245.
7 L. Meijer, A. Borgne, O. Mulner, J. P. J. Chong, J. J. Blow, N. Inagaki,
M. Inagaki, J. G. Delcros and J. P. Moulinoux, Eur. J. Biochem., 1997,
243, 527–536.
8 C. Benson, S. Kaye, P. Workman, M. Garret, M. Walton and J. de
Bono, Br. J. Cancer, 2005, 92, 7–12.
ˇ
9 P. Starha, Z. Tra´vn´ıcˇek and I. Popa, J. Inorg. Biochem., 2010, 104,
639–647.
ˇ
10 Z. Dvorˇa´k, R. Vrzal, P. Starha, A. Klanicova´ and Z. Tra´vn´ıcˇek, Toxicol.
in Vitro, 2010, 24, 425–429.
11 D. M. G. Martin and C. B. Reese, J. Chem. Soc. C, 1968, 1731–1738.
12 M. Haidoune, C. Pethe, M. Laloue and R. Mornet, J. Chem. Soc.,
Perkin Trans. 1, 1994, 3009–3012.
13 P. Lehikoinen, J. Martinem and H. Lonnberg, J. Org. Chem., 1986, 51,
3819–3823.
14 N. J. Kos and H. C. van der Plas, J. Org. Chem., 1983, 48, 1207–1210.
15 N. J. Kos, H. C. van der Plas and W. J. F. Blees, J. Org. Chem., 1983, 48,
850–855.
16 N. J. Leopard and T. R. Henderson, J. Am. Chem. Soc., 1975, 97,
4990–4999.
17 J. D. Engel, Biochem. Biophys. Res. Commun., 1975, 64, 581–586.
18 E. R. Garrett and P. J. Mehta, J. Am. Chem. Soc., 1972, 94, 8542–
8547.
19 T. Fujii, T. Saito and T. Nakasaka, Chem. Pharm. Bull., 1989, 37,
3245–3246.
20 J. A. Montgomery, A. G. Laseter, A. T. Shortnacy, S. J. Clayton and H.
J. Thomas, J. Med. Chem., 1975, 18, 564–567.
21 P. D. Sattsangi, J. R. Bardko and N. J. Leonard, J. Am. Chem. Soc.,
1980, 102, 770–774.
22 J. Ma¨ki, R. Sjo¨holm and L. Kronberg, J. Chem. Soc., Perkin Trans. 1,
2000, 4445–4450.
23 I. Nowak, J. F. Canon and M. J. Robins, Org. Lett., 2006, 8, 4565–4568.
24 P. Sund and L. Kronberg, Nucleosides, Nucleotides Nucleic Acids, 2008,
27, 1215–1226.
25 W. A. Stirk, V. Ordog and J. van Staden, J. Phycol., 1999, 35, 89–92.
26 N. J. Leonard and T. R. Henderson, J. Am. Chem. Soc., 1975, 97,
4990–4999.
27 H. A. van Onckelen and R. Verbeek, Phytochemistry, 1972, 11, 1677–
1680.
5712 | Org. Biomol. Chem., 2011, 9, 5703–5713
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
ˇ
34 P. B. Hitchcock, M. F. Lappert and Xue-Hong Wei, Dalton Trans.,
2006, 1181–1187.
28 Z. Tra´vn´ıcˇek, V. Krysˇtof and M. Sipl, J. Inorg. Biochem., 2006, 100,
214–225.
35 I. G. Dance, J. Chem. Soc., Chem. Commun., 1980, 818–820.
36 I. G. Dance, Aust. J. Chem., 1985, 38, 1391–1411.
37 M. Molon, T. Cadenbach, T. Bollermann, C. Gemel and R. A. Fischer,
Chem. Commun., 2010, 46, 5677–5679.
38 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008,
64, 112–122.
29 L. Yang, D. R. Powell and R. P. Houser, Dalton Trans., 2007, 955–964.
30 J. N. Low and R. A. Howie, Private Communication, 2003 (CSD
Refcode XACJIV; CCDC 207389).
31 F. A. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 380–388.
32 C. Routaboul, L. Dumas, I. Gautier-Luneau, J. Vergne, M.-C. Maurel
and J.-L. Decout, Chem. Commun., 2002, 1114–1115.
39 K. Brandenburg, DIAMOND, Release 3.1c, Crystal Impact GbR:
Bonn, Germany, 2006.
33 R. Dobrovetsky, D. Bravo-Zhivotovskii, B. Tumanskii, M. Botoshan-
sky and Y. Apeloig, Angew. Chem., Int. Ed., 2010, 49, 7086–7088.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5703–5713 | 5713
©