Synthesis and coordination chemistry of aminophosphine derivatives of adenine

Article abstract of DOI:10.1039/b303715k

Two aminophosphine derivatives of adenine N⁹-(N^2′-diphenylphosphinoaminoethyl)adenine L1 and N⁹-(N^2′-diphenylphosphino-N^2′-n-pro pylaminoethyl)adenine L2 were synthesized. Oxidation of L1 and L2 with

Full text of DOI:10.1039/b303715k

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis and coordination chemistry of aminophosphine
derivatives of adenine†
Qingzhi Zhang, Guoxiong Hua, Pravat Bhattacharyya, Alexandra M. Z. Slawin and
J. Derek Woollins*
Department of Chemistry, University of St Andrews, Fife, Scotland, UK KY16 9ST.
E-mail: jdw3@st-and.ac.uk; Fax: 01334 463384; Tel: 01334 463861
Received 3rd April 2003, Accepted 19th June 2003
First published as an Advance Article on the web 7th July 2003
Two aminophosphine derivatives of adenine N 9-(N 2Ј-diphenylphosphinoaminoethyl)adenine L1 and N 9-
(N 2Ј-diphenylphosphino-N 2Ј-n-propylaminoethyl)adenine L2 were synthesized. Oxidation of L1 and L2 with H2O2,
elemental sulfur or elemental selenium led to the corresponding oxidized products 5–10. Both L1 and L2 behave
as monodentate ligands towards late transition metals. Reaction of compound L1 or L2 with [AuCl(tht)], [{RhCl-
(µ-Cl)(η5-C5Me5)}2], [{IrCl(µ-Cl)(η5-C5Me5)}2], [{Rh(µ-Cl)(cod)}2], [{RuCl(µ-Cl)(η3:η3-C10H16)}2] and [{RuCl(µ-Cl)-
i
(p-MeC6H4 Pr)}2] gave the corresponding “dangling” monodentate complexes 11–20, leaving the adenine ring free for
complementary hydrogen bonding. Interaction of L1 and L2 with [MX2(cod)] (M = Pt; X = Cl, Me) in 2 : 1 molar
ratio also gave monodentate complexes 21 and 22. All compounds have been fully characterized by microanalysis, IR,
31P-{1H} NMR, 1H NMR and EI/CI/FAB MS spectroscopies. 1H-{31P} NMR, 1H–1H-COSY or 1H–13C correlation
experiments were used to confirm the spectral assignments. Four compounds were structurally characterized by
crystallographic X-ray analysis.
powdered calcium hydride and distilled under nitrogen. Diethyl
Introduction
ether and tetrahydrofuran were purified by reflux over sodium/
Owing to their resemblance to the structure of adenosine
benzophenone and distillation under dinitrogen. Ligand pre-
and their broad-spectrum of antiviral or anticancer activity,
parations were performed under an oxygen-free nitrogen
atmosphere using standard Schlenk techniques. Coordination
adenine derivatives substituted at the N 9-position constitute
an important class of pharmacologically active compounds.1
reactions and work-up were performed in dry solvents.
Among them (N 9-[2-(phosphonomethoxy)ethyl]adenine and
[MX2(cod)] (M = Pd, Pt; X = Cl; cod = cycloocta-1,5-diene)16
its analogues have been extensively studied.2,3 In bioinorganic
chemistry, metal complexes capable of forming complementary
hydrogen bonds occupy an ever-increasing important position
and [{RuCl(µ-Cl)(η3:η3-C10H16)}2]17 were prepared using
literature procedures. Preparation of all the compounds apart
from the representative L1 and 11 are available as electronic
supplementary information (ESI†).
in the development of biochemically active molecules. Houlton
et al.4 have prepared some interesting bifunctional complexes
Infrared spectra were recorded (KBr discs) on a Perkin-
which combine the covalent bond-forming capabilities of the
metal ion and a ligand surface capable of recognizing nucleo-
tide bases by means of hydrogen bonding. The same group
developed the concept of directed metalation and reported a
1
Elmer system 2000 spectrometer, H NMR spectra (300 MHz)
on a Varian Gemini 2000 spectrometer, 31P-{1H} NMR spectra
at 121.4 MHz with δ referenced to external 85% H3PO4,
13C-{1H} NMR spectra at 67.9 MHz on a JEOL GSX 270
series of nucleoside analogues in which the ribose group is
1
1
spectrometer, 2D-NMR (COSY, and H–13C or H–13P hetero-
nuclear correlation) on a Bruker Advance 300. Microanalyses
were performed by the University Service within this Depart-
ment and fast atom bombardment (FAB) or chemical ioniz-
ation (CI) mass spectra by the EPSRC Mass Spectrometer
Service (Swansea, UK). Precious metal salts were provided on
loan by Johnson Matthey PLC.
replaced by a dimethylene/trimethylene tethered ethylene-
5–8
diamine
or 1,2-dithioethane.9 Interaction of such chelate-
tethered nucleoside analogues with metal ions gave interesting
A-N 3-bound or A-C8-bound mono- or poly-nuclear complexes.
In an approach which combines the antitumour activities of
diphosphines and their gold() complexes10,11 and our experi-
ence in the synthesis and coordination chemistry of P–N com-
pounds12–15 we have incorporated the aminophosphine unit into
adenine through an aminodimethylene linkage at N 9-position.
The adenine analogues prepared in this way possess two func-
tions: excellent coordination tendency towards transition
metals and the capacity for base-pair or complementary hydro-
gen bonding interactions. The new combination of the bio-
active adenine and aminophosphine as well as the corre-
sponding complexes may lead to some enhanced biological
activities.
N 9-(N 2Ј-Diphenylphosphinoaminoethyl)adenine (L1)
N 9-(2Ј-Aminoethyl)adenine (2.48 g, 13.91 mmol) was dissolved
in hot CH3CN. To this solution was added Et3N (2.20 cm3,
14.38 mmol) and Ph2PCl (2.52 cm3, 14.04 mmol) in CH3CN (20
cm3). The reaction mixture was refluxed for 1 h and cooled to
room temperature. The solvent was removed in vacuo and H2O
(40 cm3) was added to remove the salt. The mixture was filtered
washing with H2O (2 × 20 cm3) and EtOH (2 × 20 cm3) and
Et2O (2 × 20 cm3) successively to give 3.40 g of crude product.
Recrystallization from CH3CN gave the pure product as a white
solid. Yield: 2.254, 40%. Microanalysis (%): Found (calc.) for
C19H19N6P: C, 62.23 (62.98); H, 5.02 (5.28); N, 23.49 (23.19). IR
(KBr disc, cmϪ1): 3384m, 3328m, 3293m, 3142m, 3098w, 3066w,
2934w, 2891w, 2847w, 1655vs, 1598vs, 1575s, 1488m, 1432m,
1416s, 1390w, 1375w, 1359m, 1325m, 1307s, 1238s, 1205w,
1165m, 1125s, 1094m, 1082w, 1057m, 1025w, 1010w, 958w,
941w, 911s, 854w, 797s, 753s, 739s, 695vs, 644m, 588m, 556m,
519s, 482m. CIMS (m/z): 363 [M ϩ H]ϩ. EIMS (m/z): 362 [M]ϩ.
Experimental
All solvents and reagents were purchased from Aldrich and
Lancaster. Dichloromethane was heated to reflux over
† Electronic supplementary information (ESI) available: Preparation of
all compounds apart from L1 and complex 11; Tables S1–S8: NMR
spectral data of all the compounds. See http://www.rsc.org/suppdata/dt/
b3/b303715k
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
3250

View Article Online
Table 1 Details of the X-ray data collections and refinements for compounds 5, 11, 15 and 17
Compound
5ؒ1.125CHCl3
11ؒ1/2H2O
15ؒ1/4C4H4O
17ؒ1/2CHCl3
Empirical formula
M
C20.125H20.125Cl3.375N6OP
512.7
C19H20AuClN6O0.50
603.8
P
C30H35Cl2N6O0.25PRh
688.4
C29.50H34.50Cl3.50IrN6P
820.4
Crystal colour, habit
Crystal dimensions/mm
Crystal system
Space group
Colorless, block
0.1 × 0.08 × 0.08
Monoclinic
P21
Colorless, block
0.2 × 0.1 × 0.03
Monoclinic
P2/c
Deep red, prism
0.18 × 0.1 × 0.06
Monoclinic
P21/c
Orange, prism
0.05 × 0.05 × 0.12
Monoclinic
P21/n
a/Å
b/Å
c/Å
15.8751(18)
8.4370(10)
20.658(2)
111.474(2)
2575
18.2826(3)
11.9140(2)
12.7693(2)
106.965(1)
2660
16.3100(2)
9.0989(1)
22.2647(2)
96.689(1)
3282
14.8735(11)
10.7643(8)
20.7773(15)
98.120(2)
3293
β/Њ
U/Å3
Z
4
4
4
4
Dc/g cmϪ3
1.322
0.480
1.508
5.706
1.393
0.726
1.655
4.417
µ/mmϪ1
F(000)
1053
12999
7069 (0.0840)
0.0780, 0.1722
1164
12619
3727 (0.0503)
0.0378, 0.0836
1412
13837
4659 (0.0274)
0.0369, 0.0995
1620
15660
4575 (0.1231)
0.0486, 0.0695
Measured reflections
Independent reflections (Rint
Final R1, wR2 [I > 2σ(I )]
)
[Au(Cl)(L1)] (11)
To L1 (57 mg, 157 µmol) in dichloromethane (30 cm3) was
added [AuCl(tht)] (50 mg, 156 µmol). The mixture was stirred
at room temperature for 30 min and then vacuumed to
ca. 0.5 cm3. Et2O was added and the solid was filtered off
and washed with Et2O (3 × 3 cm3). Yield: 90 mg, 97%. Micro-
analysis (%): Found (calc.) for C19H19AuClN6P: C, 38.71
(38.37); H, 3.02 (3.22); N, 13.80 (14.13). IR (KBr disc,
cmϪ1): 3327m, 3162m, 2916w, 2864w, 1649vs, 1599s, 1575m,
1481m, 1436s, 1416s, 1359w, 1327w, 1308m, 1242m, 1202w,
1164w, 1108s, 1056m, 998w, 922w, 901w, 798m, 746m, 710w,
694s, 649m, 546m, 491m, 449m. FABϩ (m/z): 595 [M ϩ H]ϩ, 594
[M]ϩ, 559 [M Ϫ Cl]ϩ.
X-Ray crystallography
Table 1 lists details of data collections and refinements for 5, 11,
15 and 17. Data were collected at room temperature using
Mo-Kα radiation with a SMART system. Intensities were
corrected for Lorentz-polarisation and for absorption. The
structures were solved by the heavy atom method or by direct
methods. The positions of the hydrogen atoms were idealised.
Refinements were by full-matrix least squares based on F 2 using
SHELXTL.18 In 5 there is a total of 1.125CHCl3 solvate mole-
cules. The half occupancy molecule is disordered, its carbon
atom was refined isotropically and its hydrogen atom was not
included in the final refinement. The NH protons in 5 were
refined in idealised positions.
Scheme 1 Synthesis of aminophosphine analogues of adenosine.
Reagents and conditions: (i) anhydrous DMF, NaH (60% in mineral
oil), 2 h; (ii) ClCH2CH2Br, rt, 36 h; (iii) NaN3, DMSO, 80 ЊC, 24 h;
(iv) MeOH, H2, 1 atm, 10% Pd/C; (v) CH3CH2CH2NH2, rt, 4 d:
(vi) Ph2PCl, Et3N, CH3CN, reflux, 12 h: (vii) THF, H2O2 (30%), 0 ЊC; or
THF, S8/Se, reflux, 2 h.
CCDC reference numbers 207487–207490.
See http://www.rsc.org/suppdata/dt/b3/b303715k/ for crystal-
lographic data in CIF or other electronic format.
Results and discussion
Ligand preparation and oxidation
Scheme 1 shows the synthesis of the ligands L1 and L2 and the
corresponding oxidised compounds. The precursors 119,20 221
and 322 were prepared according to literature methods with
slight modification. Compound 4 was easily obtained from 1
and much excess neat n-propylamine at room temperature in a
similar method to the preparation of ethylenediamine-N 9-
ethyladenine hydrochloride.5,7,23 However, according to micro-
analysis results, we obtained 4 as a free amine rather than a
hydrochloride salt.
Due to the low solubility of 3 in common organic solvents at
room temperature, its reaction with Ph2PCl was carried out
in CH3CN at reflux in the presence of Et3N. Monitoring by
31P-{1H} NMR indicated that the reaction of 3 and Ph2PCl in
1 : 1 molar ratio occurred mainly at the dangling alkyl primary
Scheme 2 Prdoduct mixture when 3 and Ph2PCl were reacted in 1 : 2
molar ratio.
amino group to give L1 though a trace amount of A (Scheme 2)
was found in the reaction mixture. Work-up by removing the
salt from the reaction mixture with water and washing with
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
3251

View Article Online
EtOH and Et2O and recrystallization from CH3CN gave an air-
and moisture-stable white solid L1 in modest yield. The selectiv-
ity of P–N bond formation at the dangling amino group rather
than the aryl C6-NH2 at the purine ring can be attributed to the
difference in basicities of the two primary amino groups. The
alkyl group at the N 9-dangling chain is more electron-rich than
the amino group at the 6-position whose lone pair is highly
delocalized in the electron-withdrawing purine ring. Attempts
to get N 9-(N 2Ј,N 2Ј-bis(diphenylphosphino)aminoethyladenine
B proved to be unsuccessful. Reaction of 3 and Ph2PCl in 1 : 2
molar ratio gave a mixture of compounds shown in Scheme 2
which are difficult to separate. It is obvious that in the second-
ary substitution, the favorable basicity at the amino nitrogen of
dangling chain is compensated by the unfavorable steric effect
from the diphenylphosphine group. Therefore no selectivity
between the secondary alkyl amino group and the primary aryl
amino group at C6-position was observed in the second
substitution.
tion of C᎐N bond of the purine ring were observed at around
᎐
1655 and 1598 cmϪ1 as two very strong bands. Compound 3
shows the typical strong νN᎐N band for azide compounds at
2100 cmϪ1. Compounds 5 and 6 displayed strong νP᎐O at ca. 1180
and 1183 cmϪ1. Since the absorption intensities᎐ of νP᎐S and
᎐
νP᎐Se tend to be medium to weak, it is difficult to assign them
unambiguously.
In the 31P-{1H} NMR spectrum (Table S2 in ESI†), L1 shows
a similar resonance to other aminophosphines derived from
equimolar of alkyl primary amines and chlorodiphenyl-
phosphine27 with δP at ca. 43.8. L2 (Table S3 in ESI†) displays a
higher frequency phosphorus signal at δP 62.8, similar chemical
shifts were observed for other aminophosphines28,29 from
secondary alkyl amines and Ph2PCl. Their oxygen derivatives 5
and 6 show lower frequency shifted signals at δP 25.0 and 32.4
compared with the precursors. The chemical shifts of their sul-
fide compounds 7 and 8 fall in higher frequencies at ca. δP 60.7
and 70.5. The selenium compounds 9 and 10 also show singlets
at higher frequencies than the L1 and L2 at δP 57.0 and 69.3
with the selenium satellites. The coupling constants 1JPSe = 755
As a consequence of the n-propyl group at the N 9-position,
compound 4 is more soluble in organic solvent than 3. In di-
chloromethane the reaction of 4 with Ph2PCl went smoothly at
room temperature in the presence of triethylamine to give L2 in
61% yield. Again the basicities of the secondary alkyl amino
group prevailed and the substitution occurred almost exclus-
ively at the “dangling” amino position to give L2 as an air- and
moisture-stable white solid.
1
and JPSe = 753 Hz are typical for one bond P–Se couplings30
(Tables S2 and S3 in ESI†).
1H NMR spectra of 1,31 32 and 333 and the 13C NMR spec-
2
trum of 220 have been reported in DMSO or DMSO-d6–D2O.
1
To compare the H NMR spectra of the precursors and the
aminophosphines L1 and L2, the same solvent should be
employed. However catalytic oxidation (with either catalyst
or irradiation) of P() compounds by DMSO have been
reported,34–36 and we did observe the gradual oxidation of
aminophosphine compounds in DMSO-d6 during spectral
acquisition, therefore employing DMSO-d6 for L1 and L2 was
precluded. Fortunately, L1 and L2 are readily soluble in CDCl3
and CD2Cl2.
Oxidation of L1 and L2 with elemental sulfur and selenium
proceeded smoothly in THF at reflux within 2 h. However oxid-
ation of these ligands with aqueous H2O2 (30%) or H2O2–
NH2CONH2 adduct needs care. The oxidizing agent can only
be used stoichiometrically or slight excess, otherwise the oxid-
ation at N 1 position of the purine ring occurs.24,25 Usually, the
urea-hydrogen peroxide adduct is a very convenient reagent for
oxidation of P().26 Suspension of the excess H2O2–NH2-
CONH2 adduct in the CH2Cl2 solution of P() compound
leads to the stoichiometric conversion of P() to P(), simple
work-up with filtration removes most of the excess adduct
and the resulting urea. The filtrate can be washed with water
to guarantee the absolute removal of the trace urea or adduct
and dried over anhydrous magnesium sulfate to give the
product. However, in our cases, due to the high polarity of the
product and the complementary hydrogen bonding capacity of
the adenine moiety, the urea was found to be difficult to remove
just by filtration. Washing with water is not efficient, especially
with 5 which is more soluble in water than in CH2Cl2. The
involvement of H2O (formed or from aqueous H2O2 or from
washing) and the tendency of hydrogen bonding interactions
between H2O and the adenine moiety make it very hard to get
satisfactory microanalysis results with the oxygen oxidized
compounds 5 and 6. Oxidation of L1 and L2 by air proved too
slow to be useful as a preparative technique. Stirring the solu-
tion of L1 or L2 in THF in air at room temperature for 4 days
only gave ca. 30% oxidation of the starting material.
L1, L2 and compounds 1–10 were fully characterized by mul-
tiple NMR spectrometry (Tables S1–S3 in ESI†), EI/CI/FAB
mass spectrometry, infrared spectroscopy and microanalyses.
All the compounds gave reasonable microanalysis results apart
from 5 and 6, which showed slightly lower carbon percentage
than the calculated value due to the involvement of H2O men-
tioned above. Their mass spectra show the molecular ions and
the expected fragmentation pattern with appropriate isotope
distributions. The sulfur and selenium species 7–10 showed the
[M Ϫ S]ϩ and [M Ϫ Se]ϩ fragment ions in their EIMS spectra.
The corresponding [M Ϫ O]ϩ were not observed in 5 and 6,
For comparative purposes, in addition to the spectra in
1
DMSO-d6, H NMR data of 1–4 in CDCl3 containing a few
drops of DMSO-d6 were also collected and are listed (Table S1
in ESI†).. It was found that in the 1H NMR spectra the chemical
shifts of the C6-NH2 group and one of the CH groups of the
purine ring (C2 or C8) varies with solvent (CDCl3, DMSO-d6 or
CDCl3–DMSO-d6). On the other hand, in the 13C-{1H} NMR
spectra, the resonance frequencies of the carbons on the purine
ring of compounds 1–4 showed little variation in different sol-
vents, these data are in accord with the 13C-{1H} NMR spectra
of other analogues of adenine.37 To assign the C2-H and C8-H
unambiguously, 1H–13C correlation experiments (1H–13C
HMQC) were performed. It turned out that the C8-H appeared
in lower frequency at around δH 7.8 in CDCl3 or CDCl3–
DMSO-d6, while in DMSO-d6, it appeared close to the C2-H
resonance at around δH 8.2. The C6–NH2 resonance in CDCl3
or CDCl3–DMSO-d6 was observed at around δH 5.6, in DMSO-
d6 at around δH 7.2.
Based on the above observation, we assign the higher fre-
quency singlet at ca. δH 8.2 to C2–H and the lower frequency
sinlget at ca. δH 7.8 to C8 in the 1H NMR spectra of L1, L2 and
5–10 (Tables S2 and S3 in ESI†). The assignments were con-
firmed by 1H–13C correlation experiments. In the 1H NMR
spectra, the broad singlet of C6-NH2 appeared at around δH
6
and can be confirmed by H/D exchange experiment. Again in
the 13C-{1H} NMR spectra, these compounds show closely
similar carbon signals of the purine ring (Tables S2 and S3 in
ESI†).
Compounds L1 and 5, 7 and 9 (Table S2 in ESI†) show a
triplet at ca. δH 4.3 and a multiplet at ca. δH 3.3. The multiplet
was attributed to the multiple spin–spin couplings between
CHCH, CHNH and PNCH and thus was assigned to
AdeCH2CH2, the triplets to the AdeCH2CH2. In addition, the
NH resonances of L1, 5, 7 and 9 were observed as slightly
broad pseudo-quartet or doublet of triplet due to the multiple
implying that the P᎐S and P᎐Se bond are not as stable as the
᎐
᎐
P᎐O bond. In their IR spectra, all compounds display broad
᎐
medium intensity νN–H vibrations, usually two, in some cases
several, depending on the different extent of hydrogen bonding
and dryness, in the range 3349–3106 cmϪ1. The N–H bending
absorption of NH2 and NH groups and the stretching absorp-
3
2
coupling JCHNH and JNHP and were confirmed by H/D
exchange experiment. The fact that in the oxidized compounds
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
3252

View Article Online
the NH signal fell to higher frequency at δH 4.4–4.6 than the
NH at δH 2.2 in the P() compound L1 is in agreement with our
former observation.12
AdeCH2CH2 in 2 was found to be contrary to the assignment in
the reference.20
As for compound 4, H–1H COSY and H–13C correlation
spectra reveal a similar tendency in chemical shifts to its phos-
phorus species, with the triplets of protons in the order of
AdeCH2CH2 (δH 4.30) > AdeCH2CH2 (δH 3.08) > NCH2-
CH2CH3 (δH 2.60), while the carbon singlets in an reversed
order of AdeCH2CH2 (δC 44.35) < AdeCH2CH2 (δC 49.04) <
NCH2CH2CH3 (δC 51.64) (Table S1 in ESI†).
1
1
In the 1H NMR spectra of L2 and 6, 8, 10 (Table S3 in ESI†),
the triplet at ca. δH 4.3 can be easily assigned to AdeCH2CH2,
the sextet at ca. δH 1.5 to NHCH2CH2CH3, the triplet at ca. δH
0.7 to NHCH2CH2CH3 respectively. However, it is not straight
forward to assign the doublets of triplets at ca. δH 3.3 and the
1
multiplet at ca. δH 3.0. H–1H COSY spectra reveals the rel-
ationship of the signals and the doublets of triplets δH 3.3 was
assigned to AdeCH2CH2 and the multiplets at ca. δH 3.0 to
NCH2CH2CH3. It is still not clear why the latter appeared as
multiplets rather than doublets of triplets as AdeCH2CH2.
Maybe one of the phenyl rings of the Ph2P group at the N
atom blocks the free rotation of the N–C bond and makes
the two protons of the NCH2CH2CH3 group inequivalent. The
1H-{31P} NMR spectra were collected and the NCH2CH2CH3
signal was simplified as a pseudo-triplet, however, no more
information was obtained.
Compound 5 was also characterized by X-ray crystallo-
graphy (Fig. 1, Table S5 in ESI†). As shown in Fig. 1(b), hydro-
gen bonding plays a predominant role in the stabilization of the
structure. Various types of intermolecular hydrogen bonds
occur between adjacent molecules, including the Hoogsteen
type between one N(6)–H and N(7) [N(6)–H(6A) ؒ ؒ ؒ N(37A):
d(D ؒ ؒ ؒ A) = 3.088(14) Å, d(H ؒ ؒ ؒ A) = 2.14 Å, Є(DHA)
161.5Њ; N(36A)–H(36A) ؒ ؒ ؒ N(7): d(D ؒ ؒ ؒ A) = 3.094(13) Å,
d(H ؒ ؒ ؒ A) = 2.12 Å, Є(DHA) = 171.8Њ ], the Watson–Crick
type [N(6)–H(6B) ؒ ؒ ؒ N(31B): d(D ؒ ؒ ؒ A) = 3.045(16) Å,
In the 13C-{1H} NMR spectra of L1 and 5, 7 and 9 (Table S2
in ESI†) the signals for the two methylene carbons could not be
assigned directly. The 1H–13C correlation experiment reveals
that the doublet at higher frequency between δC 46.2–43.9
belongs to AdeCH2CH2, the lower frequency doublet or
singlet between δC 45.9–40.4 to the AdeCH2CH2. This is in
d(H ؒ ؒ ؒ A)
=
2.09 Å, Є(DHA)
=
164.3Њ; N(36)–
H(36B) ؒ ؒ ؒ N(1B): d(D ؒ ؒ ؒ A) = 3.093(15) Å, d(H ؒ ؒ ؒ A) =
2.14 Å, Є(DHA) = 162.5Њ] in the adenine moiety and that
between the N–H and the P᎐O groups (and vice versa) of the
᎐
dangling chain [N(12)–H(12A) ؒ ؒ ؒ O(31): d(D ؒ ؒ ؒ A)
=
2.822(12) Å, d(H ؒ ؒ ؒ A) = 2.14 Å, Є(DHA) = 125.1Њ; N(42)–
H(42) ؒ ؒ ؒ O(1): d(D ؒ ؒ ؒ A) = 2.811(12) Å, d(H ؒ ؒ ؒ A) = 2.20
Å, Є(DHA) = 119.3Њ].
1
accord with the proton signals order in the H NMR spectra,
the AdeCH2CH2 appearing at higher frequency and AdeCH2-
CH2 at a lower frequency. Interestingly, though in L1 the two-
bond P–C coupling 2JPC = 17 Hz is bigger than the three-bond
C–P coupling 3JPC = 6 Hz, no 2JPC coupling was observed in its
oxidized compounds 5, 7 and 9.
Coordination chemistry of L1 and L2
Attempts to get chelate complexes D shown in Scheme 3 from
L1 and L2 were not very successful. Reaction of [PdX2(cod)]
(X= Cl, Br) and L1 or L2 (in both 1 : 1 and 1 : 2 molar ratio) led
to the immediate precipitation of a yellow solid which did not
dissolve in any solvent (not even in DMSO and DMF), the
insolubility makes the characterization very difficult. Inter-
action of L1 and [Pt(CH3)2(cod)] in 1 : 1 molar ratio gave a
However, in the 13C-{1H} NMR spectra of L2, 6, 8 and 10
(Table S3 in ESI†), as revealed by the 1H-13C correlation experi-
ments, the carbon signals in the alkyl carbon region for
AdeCH2CH2NCH2CH2CH3 are not in the same order as that of
proton signals. The chemical shifts δC appeared as the following
order: NCH2CH2CH3 > AdeCH2CH2 > AdeCH2CH2. Herein
2
3
1
both JCP and JCP were observed for AdeCH2CH2 and
NCH2CH2CH3.
mixture of 20 (DMSO-d6, δP 60.6, JPtP = 2160 Hz) and the
1
proposed eight-membered chelate complex (δP 61.18, JPtP
=
The phenyl protons in the phosphorus species L1, L2, 5–10
2184 Hz) in ca. 3 : 4 ratio based on integration of 31P-{1H}
NMR. H NMR spectrum of this mixture also displayed two
1
appeared as complicated multiplets due to the H–P coupling in
1
the H NMR spectra. The phenyl carbons, however, display
sets of signals, in which one set was in agreement with that of
20. Further characterization of the chelate complex was impos-
sible because of the failure in separation. However, L1 and L2
proved to be excellent monodentate ligand towards transition
metals. Reaction of L1 and L2 with [Au(tht)Cl] and Rh(),
Rh(), Ir(), Ru() and Ru() gave monodentate complexes
11–20. Two molar equivalents of L1 or L2 reacted with [PtX2-
(cod)] (X = CH3 or Cl) also gave monodentate complexes 21
and 22. The complexes gave reasonable microanalysis results.
Like ligands L1 and L2, the IR spectra of the complexes display
the typical medium νNH stretching bands at around 3300 and
well-resolved signals in their 13C-{1H} NMR spectra (Tables S2
and S3 in ESI†). In the P() compounds L1 and L2, the coup-
ling between the ipso-carbon and the phosphorus is unobserv-
able or relatively small, 1JPC = 14 Hz, the coupling between the
phosphorus and the ortho-carbons or meso-carbons is relatively
large, 2JPC = 20 Hz, 3JPC = 25 Hz, and the coupling between the
para-carbon and the phosphorus is not observed. In contrast,
the coupling between the ipso-carbon and phosphorus in the
1
oxidized P() compounds is huge, JPC ≈ 128, 100 and 90 Hz,
respectively, corresponding to O, S, Se oxidized species, the
4
small coupling JCP = 3 Hz between the phosphorus and the
3150 cmϪ1, the νNH bending and C᎐N stretching were observed
᎐
2
para-carbon is also observed. The P–Co coupling JPC ≈ 10 Hz
as very strong bands at ca. 1640 and 1590 cmϪ1. Their FAB MS
spectra show the molecular ions and the expected fragmen-
tation ions with appropriate isotope distributions.
3
and P–Cm coupling JPC ≈ 12 Hz were smaller than that in the
P() compounds.
Comparing the proton signals and carbon signals of the two
methylene groups of AdeCH2CH2 moiety between compounds
L1, L2 and 5–10 (Tables S2 and S3 in ESI†), we found that the
proton signal of AdeCH2CH2 always appeared at higher fre-
quency at around δH 4.3 and that of AdeCH2CH2 always at the
lower frequency around δH 3.5. However the carbon signals do
not necessarily follow this order. This made us reconsider the
assignment of proton and carbon resonances of AdeCH2CH2
of 1, 2 and 3.20,31–33 In terms of the general tendency of the L1,
L2 and 5–10, we assign the higher frequency proton signal at ca.
δH 4.3 to AdeCH2CH2, the lower frequency signal at around
δH 3.5 to AdeCH2CH2. The carbon resonances were assigned
through 1H–13C correlation spectra as listed in Table S4 in ESI.†
The assignment of the carbon signals for AdeCH2CH2 and
In the 31P-{1H} NMR spectra (Tables S6 and S7 in ESI†),
apart from complexes 17, 18 and 22 all the complexes show the
chemical shifts at higher frequencies than L1 and L2. The rho-
dium species exhibited the expected doublets with appropriate
coupling constants (1JRhP = 149–161 Hz). The platinum com-
plex 21 shows a singlet at δP 60.6 with platinum satellites 1JPtP
=
2106 Hz. The coupling constant is typical for a phosphine lig-
and trans to the methyl group. The chemical shift of complex
22, like other aminophosphine complexes from [PtCl2(cod)],
falls in a lower frequency at δP 59.9 than the free ligand.27 The
1
magnitude of the JPtP = 3983 Hz is typical for a phosphine
ligand trans to chloride and thus in agreement with the cis
geometry. The lower frequency-shift of the iridium complexes
17 (δP 34.7) and 18 (δP 49.0) compared with L1 (δP 43.8) and L2
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
3253

View Article Online
Fig. 1 (a) The crystal structure of compound 5. (b) The crystal structure of compound 5 showing part of hydrogen bonding.
Scheme 3 Coordination chemistry of L1 and L2. Reagents and conditions: (i) [AuCl(tht)], CH2Cl2, rt, 2 h; (ii) [{Rh(µ-Cl)(cod)}2], THF, rt, 2 h;
i
(iii) 0.5 [{RhCl(µ-Cl)(η5-C5Me5)}2], [{IrCl(µ-Cl)(η5-C5Me5)}2], 0.5 [{RuCl(µ-Cl)(η3:η3-C10H16)}2] or [{RuCl(µ-Cl)(p-MeC6H4 Pr)}2]; (iv) 0.5 [PtCl2-
(cod)] or Pt(cod)Me2; (v) [PdX2(cod)] (X = Cl, Br, Me), CH2Cl2.
(δP 62.8) was also observed in other iridium complex of
aminophosphine.38
CDCl3 or CDCl3–DMSO-d6 solution of the complexes, if
necessary, 1H-{31P} NMR spectra were also collected. Interest-
ingly, while the C6-NH2 signal on the purine ring at around
δH 5.7 disappeared immediately after shaking the solution
vigorously for all the compounds, the signal for NH on the
dangling chain lingered for 0.5–2 h before vanishing in L1, 5, 7
and 9. For the complexes in Table S6 in ESI,† it is very difficult
to get spectra free of the NH signal. Standing the sample for
two days with D2O added to CDCl3 solution with shaking or
The 1H NMR spectra of the complexes are closely similar to
the free ligand, in spite of the fact that in some cases the signals
of AdeCH2CH2 and NCH2CH2CH3 are broadened. Com-
parison of the 1H NMR data of L1 (Table S2 in ESI†) with its
complexes in Table S6 in ESI† shows that the NH signals of the
dangling chain are shifted to higher frequency on coordination,
from δH 2.2 to δH 3.6–3.8. In the case of 11, even further higher
frequency shift (δH 4.86) was observed because of the involve-
ment of DMSO-d6. In order to identify the NH signal, to sim-
plify the coupling system of AdeCH2CH2 and to measure the
1
sonication from time to time led to H NMR spectra with the
NH signal decreased and AdeCH2CH2 signal simplified to
some extent. The sluggish H/D exchange may be mainly attri-
buted to the steric effect and the hydrophobic properties of
3
2
coupling constants JCHNH and JPNH, D2O was added to the
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
3254

View Article Online
both the diphenylphosphine group and other moiety intro-
duced together with the metal.
1
The H NMR spectrum of 12 (Table S7 in ESI†) shows a
closely similar spectrum to L2, with the NCH2CH2CH3 signal
appearing as complicated multiplets, which is simplified as a
pseudo-triplet in the 1H-{31P} NMR spectrum. The complexity
of NCH2CH2CH3 of other complexes of L2 was less clear
because of the slight broadness of the signal. Surprisingly,
though complex 18 gave a sharp singlet in the 31P-{1H} NMR
spectrum, the 1H NMR signals are extremely broad except that
of C2-H, C8-H, C6-NH2 on the purine ring and the CH3 of
the pentylmethylcyclopentadienyl ring. The broadening of
the signals of the dangling moiety at N 9 including the phenyl
protons suggests that there may be some conformational
exchange around the nitrogen atom. The conformations may
arise from the interaction of the large C5Me5 group and the
purine ring and thus may block the free conversion of the chiral
N center. Low-temperature spectra obtained at Ϫ30 and
Ϫ55 ЊC look even more complicated with the signals splitting
into two sets of broad peaks, suggesting that the exchange
between the two conformations slow down to some extent, but
not slow enough to give two sets of sharp signals. A spectrum
collected at 50 ЊC exhibited slightly narrower and more intensi-
fied peaks, however, only the NCH2CH2CH3 resonance looked
like a triplet, all other peaks remained as broad singlets. The
reaction of L2 and [{RhCl(µ-Cl)(η5-C5Me5)}2] in 1 : 1 molar
ratio was also tested to give 16. The 31P-{1H} NMR spectrum
1
showed a broad doublet at δP 80.3 with JRhP = 152 Hz. The
1H NMR at room temperature is as broad and complicated
(appearing like two sets of signals) as that of 18 at low temper-
ature, implying that there are two conformations for this
compound in the CDCl3 solution.
Among the above complexes, single crystals suitable for
X-ray analysis of compound 11, 15 and 17 were obtained.
These crystals show two common features: the co-crystallis-
ation of some solvate molecules (Table 1) and the presence of a
variety of hydrogen bonds. For clarity the solvate molecules are
omitted in the structures (Fig. 2(a)–(c), Fig. 3(a) and (b), Fig.
4(a)) except the solvent involved in hydrogen bonding (Fig.
4(b)). The comparative bond lengths and angles are listed in
Table S8 in ESI.†
Complex 11 (Fig. 2(a), (b)) co-crystallised with one half
molecule of water. The coordination around the Au(1) adopts
the typical linear geometry [P(1)–Au(1)–Cl(1) 176.62(8)Њ] for
gold() complexes. There are three types of intermolecular
hydrogen bonding observed in this structure, including the
typical common and reversed Watson–Crick type N(6)–
H(6) ؒ ؒ ؒ N(1A) and N(6A)–H(6) ؒ ؒ ؒ N(1) [d(D ؒ ؒ ؒ A)
=
2.997(8) Å, d(H ؒ ؒ ؒ A) = 2.02 Å, Є(DHA) = 172.0Њ], the weak
and less common H-bond N(6)–H(6) ؒ ؒ ؒ N(3B) or
N(3) ؒ ؒ ؒ H(6C)–N(6C) [d(D ؒ ؒ ؒ A) = 3.124(8) Å, d(H ؒ ؒ ؒ A) =
2.40 Å, Є(DHA) = 130.6Њ] and the Hoogsteen type between
N(7) and the dangling N–H of an adjacent molecule [N(12)–
H(12) ؒ ؒ ؒ N(7C): d(D ؒ ؒ ؒ A) = 2.989(7) Å, d(H ؒ ؒ ؒ A) = 2.08
Å, Є(DHA) = 154(6)Њ]. Different from 5, 15 and 17, the struc-
ture of 11 displays intramolecular π–π-stacking interaction
(Fig. 2(c)). One of the phenyl ring of the phosphine moiety
stacks approximately parallel (α = 12Њ) with the six-membered
ring of the adenine. The inter-planar centroid–centroid separ-
ation between the stacked rings is of 3.7 Å. No intermolecular
π–π-stacking was observed.
Single molecules of 15 and 17 are similar to each other
(Fig. 3(a), Fig. 4(a)). The geometry around the Rh(1) or Ir(1)
can be viewed as a three-legged ‘piano stool’ with the two chlor-
ides and the phosphorus P(1) supporting the pentamethyl-
cyclopentadienyl top. The bond angles between the ‘legs’ are
approximately 90Њ. An intramolecular five-membered-ring
hydrogen bond between one of the chloride atoms and the NH
group of the aminophosphine moiety is observed in both
compounds [N(12)–H(12) ؒ ؒ ؒ Cl(1): d(D ؒ ؒ ؒ A) = 3.149(3) Å,
Fig. 2 (a) The crystal structure of complex 11. (b) The crystal
structure of complex 11 showing hydrogen bonding. (c) The crystal
structure of complex 11 showing π–π-stacking effect.
d(H ؒ ؒ ؒ A) = 2.49(4) Å, Є(DHA) = 124(3)Њ for 15, and corre-
spondingly 3.396(2), 2.73(2) Å, 125.8(1)Њ for 17]. However the
two compounds co-crystallise with various solvent molecules, a
quarter of THF for 15 and a half of CHCl3 for 17. The struc-
tures of 15 and 17 also adopt different types of intermolecular
hydrogen bonds. Two molecules in 15 pair with each other by
Hoogsteen type hydrogen bonds between the N(7) and one
N(6)–H [N(6)–H(6A) ؒ ؒ ؒ N(7A) and N(6A)–H(6AA) ؒ ؒ ؒ
N(7): d(H ؒ ؒ ؒ A) = 2.13(2) Å, Є(DHA) = 157(4)Њ] and one
additional intermolecular hydrogen bond between the other
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
3255

View Article Online
In conclusion, we describe two aminophosphine analogues
of adenine L1 an L2 and their chalcogenide derivatives. L1 and
L2 proved to be good monodentate ligands towards late transi-
tion metals to give a series of monodentate complexes. All the
compounds retain a free adenine moiety for complementary
hydrogen bonding. Further studies on the interaction of these
compounds with DNA and bioactivities are undergoing.
Acknowledgements
We wish to thank the JREI for equipment grants and Johnson
Matthey plc for loan of precious metals and Dr Andrew
Houlton for his inspiration of this work. Qingzhi Zhang
is indebted to St Andrews University for financial support.
We also appreciate Dr Stephen M Aucott and Mrs. Joanna
Wheatley in this lab for the kind donation of some of the metal
precursors and Dr Petr Kilian for help in low-temperature
NMR.
References
1 J. Bourguignon, L. Desaubry, P. Raboisson, C-G. Wermuth and
C. Lugnier, J. Med. Chem., 1997, 40, 1768.
2 L. M. Schultze, H. H. Chapman, N. J. P. Dubree, R. J. Jones,
K. M. Kent, T. T. Lee, M. S. Louie, M. J. Postich, E. J. Prisbe,
J. C. Rohloff and R. H. Yu, Tetrahedron Lett., 1998, 39, 1853.
3 A. Holy, J. Gunter, H. Dvorakova, M. Masojidkova, G. Andrei,
T. Snoeck, J. Balzarini and E. De Clercq, J. Med. Chem., 1999, 42,
2064.
Fig. 3 (a) The crystal structure of complex 15. (b) The crystal
structure of complex 15 showing the hydrogen bonding.
4 A. Houlton, D. M. P. Mingos and D. J. Williams, J. Chem. Soc.,
Chem., Commun., 1994, 503.
5 C. Price, M. R. J. Elsegood, W. Clegg, N. H. Rees and A. Houlton,
Angew. Chem., Int. Ed. Engl., 1997, 36, 1762.
6 M. A. Shipman, C. Price, A. E. Gibson, M. R. J. Elsegood, W. Clegg
and A. Houlton, Chem. Eur. J., 2000, 6, 4317.
7 M. A. Shipman, C. Price, M. R. J. Elsegood, W. Clegg and
A. Houlton, Angew. Chem., Int. Ed., 2000, 39, 2360.
8 C. Price, M. A. Shipman, N. H. Rees, M. R. J. Elsegood, Andrew J.
Edwards, W. Clegg and A. Houlton, Chem. Eur. J., 2001, 7, 1194.
9 C. Price, B. R. Horrocks, A. Mayeux, M. R. J. Elsegood, W. Clegg
and A. Houlton, Angew. Chem., Int. Ed., 2002, 41, 1047.
10 D. T. Hill, G. R. Girard, F. L. McCabe, R. K. Johnson, P. D. Stupik,
J. H. Zhang, W. M. Reiff and D. S. Eggleston, Inorg. Chem., 1989,
28, 3529.
11 C. K. Mirabelli, D. T. Hill, L. F. Faucette, F. L. McCabe,
G. R. Girard, D. B. Bryan, B. M. Sutton, J. O. Bartus, S. T. Crooke
and R. K. Johnson, J. Med. Chem., 1987, 30, 2181.
12 A. M. Z. Slawin, J. D. Woolins and Q. Zhang, J. Chem. Soc. Dalton
Trans., 2001, 621.
13 S. M. Aucott, A. M. Z. Slawin and J. D. Woollins, J. Chem. Soc.,
Dalton Trans., 2000, 2559.
14 T. Q. Ly, A. M. Z. Slawin and J. D. Woollins, J. Chem. Soc.,
Dalton Trans., 1997, 1611.
15 P. Bhattacharyya, A. M. Z. Slawin, M. B. Smith and J. D. Woollins,
J. Chem. Soc., Dalton Trans., 1996, 3647.
16 J. X. McDermott, J. F. White and G. M. Whitesides, J. Am. Chem.
Soc., 1976, 98, 6521.
17 J. G. Toerien and P. H. van Rooyen, J. Chem. Soc., Dalton Trans.,
1991, 1563.
18 G. M. Sheldrick, SHELXTL, Brucker AXS, Madison, WI, 1999.
19 P. H. C. Mundill, R. W. Fries, W. Richard, C. Woenckhaus and
B. V. Plapp, J. Med. Chem., 1981, 24, 474.
20 O. F. Schall and G. W. Gokel, J. Am. Chem. Soc., 1994, 116, 6089.
21 M. El Masri and K. D. Berlin, Org. Prep. Proc. Int., 1995, 27, 161.
22 M. Ho, B. A. Wilson and I. Katampe, Bioorg. Med. Chem. Lett.,
1996, 6, 899.
23 C. Price, M. R. J. Elsegood, W. Clegg and A. Houlton, J. Chem. Soc.,
Chem. Commun., 1995, 2285.
24 T. Saito, H. Hayashibara, Y. Kumazawa, T. Fujisawa and T. Fujii,
Heterocycles, 1990, 31, 1503.
25 T. Fujii, T. Saito, K. Kizu, Y. Hayashibara, S. Nakajuma and
T. Fujisawa, Chem. Pharm. Bull., 1991, 39, 301.
26 Y. Lu, M. Freytag, P. G. Jones and R. Schmutzle, Z. Anorg. Allg.
Chem., 2002, 628, 274.
27 A. D. Burrows, M. F. Mahon and M. T. Palmer, J. Chem. Soc.,
Dalton Trans., 2000, 1669.
28 A. D. Burrows, M. F. Mahon and M. T. Palmer, J. Chem. Soc.,
Dalton Trans., 2000, 3515.
Fig. 4 (a) The crystal structure of complex 17. (b) The crystal
structure of complex 17 showing hydrogen bonding.
N(6)–H and chloride atom Cl(2) on the metal [N(6)–H(6B) ؒ ؒ ؒ
Cl(2) and N(6A)–H(6AB) ؒ ؒ ؒ Cl(2B): d(D ؒ ؒ ؒ A) = 3.348(4) Å,
d(H ؒ ؒ ؒ A) = 2.52(3) Å, Є(DHA) = 143(3)Њ]. In compound 17,
however, two molecules couple with each other only through
the Watson–Crick type of hydrogen bondings between N(6)–H
and N(1) [N(6)–H(6A) ؒ ؒ ؒ N(1A) and N(6A)–H(6AA) ؒ ؒ ؒ
N(1): d(D ؒ ؒ ؒ A) = 3.073(7) Å, d(H ؒ ؒ ؒ A) = 2.09 Å, Є(DHA)
= 175.7Њ]. The other N(6)–H proton forms an H-bond with
one of the chlorides of the solvate CHCl3. Since there are
two more hydrogen bonds between two molecules, the struc-
ture of 15 is more compact than that of the long spread-
eagled 17.
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
3256

View Article Online
29 M. S. Balakrishna, R. M. Abhyhyankar and J. T. Magae, J. Chem.
Soc., Dalton Trans., 1999, 1407.
30 S. W. Carr and R. Colton, Aust. J. Chem, 1981, 34, 35.
31 M. D’Auria and A. Vantaggi, J. Heterocycl. Chem., 1994, 31, 375.
32 L. Soulere, J.-C. Sturm, L. J. Nunez-Vergara, P. Hoffmann and
J. PerPie, Tetrahedron, 2001, 57, 7173.
35 J. M. Berg and R. H. Holm, J. Am. Chem. Soc., 1985, 107,
925.
36 A. Baceiredo, G. Bertrand and G. Sicard, J. Am. Chem. Soc., 1985,
107, 4781.
37 S. V. Calenbergh, J. K. von F. D. Kunzel, N. M. Blaton, O. M.
Peeters, J. Rozenski, A. V. Aerschot, A. De Bruyn, D. De
Keukeleire, A. P. Uzerman and P. Herdewijn, J. Med. Chem., 1997,
40, 3765.
33 S. V. Calenbergh, C. L. M. J. Verlinde, J. Soenens, A. D. Bruyn and
M. Callens, J. Med., Chem., 1995, 38, 3838.
34 S. Bhattacharjee and R. Bhattacharyya, J. Chem. Soc.,
Dalton Trans., 1992, 1357.
38 S. M. Aucott, A. M. Z. Slawin and J. D. Woollins, J. Organmet.
Chem., 1999, 582, 83.
D a l t o n T r a n s . , 2 0 0 3 , 3 2 5 0 – 3 2 5 7
3257