Tripodal trisamides based on nicotinic and picolinic acid derivatives

Article abstract of DOI:10.1139/v98-030

A number of polydentate arylamide ligands have been prepared by coupling various acyclic tripodal or linear polyamines with derivatives of nicotinic and picolinic acids. Two synthetic procedures were utilized; tris{[(2-hydroxynicotinyl)carbonyl]-2-aminoethyl}amine (H₃NICTREN) was prepared by Method A, the HOSu/DCC method, and the other arylamides in this study were prepared by Method B, the CDI method. Method A involved the reaction of N-hydroxysuccinimide with 2-hydroxynicotinic acid (in the presence of dicyclohexylcarbodiimide (DCC) as a dehydrative coupling reagent) to form the succinimide ester, followed by reaction with TREN to yield H₃NICTREN. Method B involved reaction of a carboxylic acid (2-hydroxynicotinic, 3-hydroxypicolinic, nicotinic, or picolinic acids) with carbonyldiimidazole (CDI) to form the N-acylimidazolide, followed by reaction with the amine (TREN, TAME, spermidine, or TRPN) to yield the desired arylamide. The X-ray structure of 1,1,1-tris([(3-hydroxypicolinyl)carbonyl]-2-aminomethyl)ethane (H₃PICTAME) was determined; crystals of H₃PICTAME are monoclinic, a = 10.257(2), b = 15.572(3), c = 15.208(2) A, β = 96.124(15)°, Z = 4, space group P2₁/a. The structure was solved by direct methods and refined by full-matrix least-squares procedures to R = 0.041 and R_w = 0.038 for 2506 reflections with I ≥gt; 3σ(I). In the solid state, H₃PICTAME contains an extensive hydrogen-bonding network, with eight intra- and one intermolecular H-bonds per molecule; the ligand is partially preorganized for metal ion chelation. The acid dissociation constants of H₃NICTREN and those of 1,1,1-tris{[(2-hydroxynicotinyl)carbonyl]-2-aminomethyl}ethane (H₃NICTAME) have been determined; pK_a1 = 11.2 (10.68), pK_a2 = 10.7 (10.58), pK_a3 = 10.0 (9.71), and pK_a4 = 6.25 for H₃NICTREN (H₃NICTAME); the high phenolic pK_a's are consistent with the hydrogen bonding observed in the solid state.

Full text of DOI:10.1139/v98-030

414
Tripodal trisamides based on nicotinic and
picolinic acid derivatives
H.R. Hoveyda, Veranja Karunaratne, Christopher J. Nichols, Steven J. Rettig,
Ashley K.W. Stephens, and Chris Orvig
Abstract: A number of polydentate arylamide ligands have been prepared by coupling various acyclic tripodal or linear
polyamines with derivatives of nicotinic and picolinic acids. Two synthetic procedures were utilized;
tris{[(2-hydroxynicotinyl)carbonyl]-2-aminoethyl}amine (H3NICTREN) was prepared by Method A, the HOSu/DCC
method, and the other arylamides in this study were prepared by Method B, the CDI method. Method A involved the reaction
of N-hydroxysuccinimide with 2-hydroxynicotinic acid (in the presence of dicyclohexylcarbodiimide (DCC) as a dehydrative
coupling reagent) to form the succinimide ester, followed by reaction with TREN to yield H3NICTREN. Method B involved
reaction of a carboxylic acid (2-hydroxynicotinic, 3-hydroxypicolinic, nicotinic, or picolinic acids) with carbonyldiimidazole
(CDI) to form the N-acylimidazolide, followed by reaction with the amine (TREN, TAME, spermidine, or TRPN) to yield the
desired arylamide. The X-ray structure of 1,1,1-tris{[(3-hydroxypicolinyl)carbonyl]-2-aminomethyl}ethane (H3PICTAME)
was determined; crystals of H3PICTAME are monoclinic, a = 10.257(2), b = 15.572(3), c = 15.208(2) Å, β = 96.124(15)°, Z =
4, space group P21/a. The structure was solved by direct methods and refined by full-matrix least-squares procedures to R =
0.041 and Rw = 0.038 for 2506 reflections with I ≥ 3σ(I). In the solid state, H3PICTAME contains an extensive
hydrogen-bonding network, with eight intra- and one intermolecular H-bonds per molecule; the ligand is partially
preorganized for metal ion chelation. The acid dissociation constants of H3NICTREN and those of
1,1,1-tris{[(2-hydroxynicotinyl)carbonyl]-2-aminomethyl}ethane (H3NICTAME) have been determined; pKa1 = 11.2 (10.68),
pKa2 = 10.7 (10.58), pKa3 = 10.0 (9.71), and pKa4 = 6.25 for H3NICTREN (H3NICTAME); the high phenolic pKa’s are
consistent with the hydrogen bonding observed in the solid state.
Key words: arylamide, hydrogen bonding, preorganization.
Résumé : On a préparé une série de ligands arylamides polydentates en procédant au couplage de diverses polyamines
linéaires ou acycliques tripodaux avec des dérivés des acides nicotinique et picolinique. On a fait appel à deux voies de
synthèse : on a préparé la tris{[(2-hydroxynicotinyl)carbonyl]-2-aminoéthyl}amine (H3NICTREN) par la méthode A, la
méthode au HOSu/DCC, alors que les autres arylamides examinés dans cette étude ont été préparés par la méthode B, la
méthode CDI. La méthode A implique la réaction du N-hydroxysuccinimide avec l’acide 2-hydroxynicotinique (en présence
de dicyclohexylcarbodiimide (DCC) comme agent de couplage déshydratant) qui conduit à l’ester succinimide, suivie d’une
réaction avec du « TREN » pour obtenir le H3NICTREN. La méthode B implique la réaction d’un acide carboxylique (acides
2-hydroxynicotinique, 3-hydroxypicolinique, nicotinique ou picolinique) avec le carbonyldiimidazole (CDI) pour former le
N-acylimidazolide, suivie d’une réaction avec l’amine (« TREN », « TAME », spermidine ou « TRPN ») conduisant à
l’arylamide désiré. On a déterminé la structure cristalline du 1,1,1-tris{[(3-hydroxypicolinyl)carbonyl]-2-aminométhyl}éthane
(H3PICTAME). Les cristaux sont monocliniques, groupe d’espace P21/a, avec a = 10,257(2), b = 15,572(3) et c = 15,208(2)
D, β = 96,124(15)° et Z = 4. On a résolu la structure par des méthodes directes et on l’a affinée par la méthode des moindres
carrés (matrice entière) jusqu’à des valeurs de R = 0,041 et Rw = 0,038 pour 2506 réflexions avec I ≥ 3σ(I). À l’état solide, le
H3PICTAME comporte un important réseau de liaisons hydrogènes, avec huit liaisons intra- et une liaison intermoléculaire
par molécule; le ligand est partiellement préorganisé pour donner une chélation avec le métal. On a déterminé les constantes
de dissociation acide du H3NICTREN et du 1,1,1-tris{[(3-hydroxypicolinyl)carbonyl]-2-aminométhyl}éthane (H3PICTAME)
qui sont respectivement : pKa1 = 11,2 (10,68), pKa2 = 10,7 (10,58), pKa3 = 10,0 (9,71) et pKa4 = 6,25; les valeurs élevées du
pKa des phénols sont en accord avec les liaisons hydrogènes observées à l’état solide.
Mots clés : arylamide, liaison hydrogène, préorganisation.
[Traduit par la rédaction]
Received August 25, 1997.
H.R. Hoveyda, V. Karunaratne, C.J. Nichols, S.J. Rettig, A.K.W. Stephens, and C. Orvig.1 Medicinal Inorganic Chemistry Group,
Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, BC V6T 1Z1, Canada.
1
Author to whom correspondence may be addressed. Telephone: (604) 822-4449. Fax: (604) 822-2847. E-mail: orvig@chem.ubc.ca
Can. J. Chem. 76: 414–425 (1998)
© 1998 NRC Canada

415
Fig. 1. Catecholate (a) and salicylate (b) modes of bonding for the
pH-dependent coordination environments of Fe(III) bound to
catechoylamide ligands. Prototypical “arylamide” ligand in
protonated form (c).
Hoveyda et al.
Introduction
In searching for new ligands to be used in the sequestration of
trivalent metal ions (e.g., group 13 and lanthanides), much
insight can be gained from an examination of naturally occur-
ring chelating agents for trivalent iron. The physical and
chemical similarities between Fe(III) and group 13 metals, es-
pecially Ga(III), are rather pronounced; they share the same
charge and similar ionic radii, e.g., 0.645 Å for Fe(III) com-
pared to 0.620 Å for Ga(III) (1). In addition, the electronic
configuration of (aqueous) ferric ion is high-spin d5, while the
group 13 metal ions are diamagnetic (Al(III), d0; Ga(III) and
In(III), d10), and, accordingly, none of these metal ions or the
lanthanides are perturbed by crystal field effects (2). The kin-
ship between Ga(III) and Fe(III) is nicely reflected in the ki-
netics of ligand exchange, for instance, the first-order water
exchange rates for Ga(III) and Fe(III) (at 298 K) are 4.0 × 102
and 1.6 × 102 s–1, respectively (3). Also, microorganisms and
plants have evolved a host of low-molecular-weight ligands;
the siderophores (for a general overview, see ref. 4), which
show both a high affinity and a high specificity for Fe(III)
(formation constants ranging from 1025 to 1049) (5). Therefore,
in a biomimetic approach, the knowledge gained from ferric
ion transport can be applied, for example, to the development
of 67Ga radiopharmaceuticals (6, 7). Our approach in design-
ing the ligands prepared in this paper has been biomimetic, as
is outlined below.
Enterobactin forms one of the most stable Fe(III) com-
plexes known (8), with an overall stability constant of 1049 (9).
The backbone of this macrocyclic ligand is derived from three
L-serine units, and it utilizes three catechol moieties to seques-
ter Fe(III) in a hexadentate coordination sphere as in Fig. 1a
(10–12). However, since the effective ligand is the conjugate
base form (the catecholate), the coordination environment of
the metal ion is a function of pH. In a series of studies (13–15),
Raymond and co-workers have unequivocally established that
sequential protonation of the ferric enterobactin complex
(three one-proton steps) results in a concomitant shift from a
“catechol binding mode” to a “salicylate binding mode” (see
Fig. 1). The “arylamides” that possess the structural features
depicted in Fig. 1c may thus be considered as “salicylate type”
OH
OH
H+
O
Fe
O
O
O
Fe
N
HN
O
HN
O
H
a
b
c
ligands, potentially capable of metal ligation via the phenolate
oxygen and the amide carbonyl oxygen.
It was our interest in maltol as an effective bidentate chelat-
ing ligand for group 13 metal ions (16, 17) that instigated the
investigation of its structural isomer, isomaltol, as yet another
naturally occurring bidentate chelator (18). However, once it
was established that isomaltol was not a good ligand for the
formation of other hexadentate ligands (i.e., its denticity can-
not be increased via an imine linkage), due to nucleophilic
attack being favored at the C-5 carbon (19), attention was
turned to using the amide linkage both as a metal binding
group, as well as the “arm” to increase the denticity of the
ligand. As a consequence, a bidentate analog Hahm was syn-
thesized starting from the appropriate ethyl ester Hehm (19).
In spite of the fact that Al and Ga tris(ligand) complexes could
be isolated with Hahm, it was discovered that in H2O, Hahm
undergoes decomposition that is time, concentration, and
pH dependent (19).
Given the fact that, in general, the half-life for hydrolysis
of an amide bond (in neutral aqueous solution) is ca. 7 years
(20), it is clear that replacement of the furan with another aro-
matic ring system would do much to alleviate the instability
of these ligands, not to mention the reduction in the number of
steps required for the ligand synthesis. Therefore, we have
used derivatives of nicotinic and picolinic acid as (commer-
cially available) starting materials for the formation of new
arylamide ligands.
OH
OH
O
N
CO2H
R
CO2H
R
OH
Me
O
O
Me
O
O
N
O
Me
R
Maltol
Isomaltol
R = OEt (Hehm)
R = N(H)Me (Hahm)
R = H nicotinic acid R = H picolinic acid
OH 2-hydroxy-
nicotinic acid
(H2nic)
OH 3-hydroxy-
picolinic acid
(H2pic)
Experimental
Na benzophenone ketyl under Ar prior to use; 1,4-dioxane was
predried over activated 4Å molecular sieves before use. Water
was deionized (Barnstead D8902 and D8904 cartridges) and
distilled (Corning MP-1 Megapure still). Plates used for thin-
layer chromatography (TLC) were aluminum-supported
0.2 mm silica gel 60 F254 (EM Science). Column chromatog-
raphy was carried out using 230–400 mesh size silica gel
Materials and methods
All chemicals were reagent grade and were used as received
from commercial vendors (Aldrich and Sigma) unless other-
wise stated. 1,1,1-Tris(azidomethyl)ethane was prepared ac-
cording to a literature procedure (21). THF was distilled from
© 1998 NRC Canada

416
Can. J. Chem. Vol. 76, 1998
(E. Merck, Silica Gel 60) and the “flash” technique (22), or by
using lipophilic Sephadex LH-20™. All reactions were carried
out under N2 unless otherwise noted. The yields reported refer
to isolated yields.
Tris(aminopropyl)amine (TRPN)
TRPN was made by reduction of nitrilotrispropionamide with
LAH in THF with some modifications to the reported method
(23). The intermediate trisamide was synthesized as follows.
To a suspension of acrylamide (43.0 g, 0.60 mol) in H2O (20
mL) was added 26 g of 30% NH4OH (0.22 mol). The solution
became clear in ca. 10 min, and the mixture was left stirring at
20°C for 24 h. The precipitate thus formed was filtered,
washed with H2O (3 × 10 mL), and dried in vacuo, yielding
nitrilotrispropionamide (25 g, 57%). 1H NMR and IR spectra
of the product were identical with those previously reported.
To a suspension of the dry trisamide thus obtained (10.0 g,
0.04 mol) in 1000 mL of THF at 0°C, LAH (10.0 g, 0.26 mol)
was added in small portions over 30 min. The resulting white-
grey slurry was stirred at 0°C for 10 min, allowed to warm up
to 20°C, and then stirred for 18 h. The mixture was then cooled
to 0°C and NaOH (50 mL, 8 M) was added slowly to quench
the excess LAH. To the resulting off-white slurry was added
Et2O (500 mL) and the reaction mixture was stirred for 30 min.
The off-white precipitate was filtered and washed with
Et2O. The organic layer was separated, dried (MgSO4), and
evaporated to afford a yellow oil. The crude product was vac-
uum distilled (0.2 Torr; 1 Torr = 133.3 Pa) to remove the
bisamine impurity (ca. 10% by 1H NMR; 65°C) from the prod-
Instrumentation
NMR spectra were recorded on Bruker WH-400 (1H NMR,
400 MHz), or Bruker AC-200E (1H NMR, 200 MHz; 13C
NMR, 50 MHz) instruments. 1H NMR data are reported as δ
downfield from TMS (external reference) at 400 MHz in
DMSO-d6 unless otherwise noted. 13C NMR data are reported
as δ downfield from TMS (external reference) at 50 MHz in
DMSO-d6 unless otherwise noted. Infrared (IR) spectra were
recorded as KBr disks in the range 4000–400 cm–1 on a Perkin
Elmer PE 783 spectrophotometer and were referenced to poly-
styrene unless stated otherwise. Mass spectra (MS) were ob-
tained with a Kratos MS 50 (electron-impact ionization, EI),
an AEI MS 9, or a Kratos Concept IIH (fast-atom-bombard-
ment ionization, FAB). Melting points were measured on a
Mel-Temp apparatus and are uncorrected. Analyses were per-
formed on a Carlo Erba elemental analyzer by Mr Peter Borda
of this department.
Syntheses
Caution! The handling of polyazides in large quantities may
be hazardous. DCC is a severe skin irritant!
uct (97–100°C), yielding 2.6 g (35%) of TRPN. H and 13C
NMR spectra of TRPN thus synthesized were identical with
those reported in the literature (23).
1
1,1,1-Tris(aminomethyl)ethane (TAME)
TAME was prepared with some modifications to the reported
method (21). To a solution of 1,1,1-tris(azidomethyl)ethane
(16.8 g, 0.086 mol) in THF (200 mL) at 0°C, lithium alu-
minium hydride (LAH) (14.0 g, 0.364 mol) was added in small
portions over 30 min. The resulting white-grey slurry was
stirred at 0°C for 10 min, allowed to warm up to 20°C, and then
refluxed for 24 h. The mixture was cooled to 0°C and NaOH
(50 mL, 8 M) was added slowly to quench the excess LAH. To
the resulting off-white slurry was added Et2O (500 mL) and the
reaction mixture was stirred for 30 min. The off-white precipi-
tate was filtered off and washed with Et2O. The combined fil-
trate and washings were dried (MgSO4), and the solvent was
removed on a rotary evaporator to afford TAME (8.6 g, 86%)
as a pale yellow oil. 1H NMR and IR spectra of the product thus
obtained were identical with those previously reported (21).
This product was used without any further purification.
Method A: Succinimido 2-hydroxynicotinate (HOSuNic)
To a mixture of 2-hydroxynicotinic acid (1.370 g, 9.85 mmol)
in THF (100 mL) was added N-hydroxysuccinimide (1.148 g,
9.98 mmol) and DCC (2.275 g, 11.03 mmol). The mixture was
then refluxed for 20 h, whereupon it was allowed to cool to
20°C. The white precipitate was filtered, washed with CH2Cl2
(3 × 50 mL), and dried in vacuo. The crude product thus ob-
tained was still contaminated with dicyclohexylurea (DCU).
The product was purified via Soxhlet extraction (48 h) using
CH2Cl2 to yield 1.02 g (42%) of HOSuNic. 1H NMR: 2.98 (s,
4H1,2), 6.49 (t, 1Hβ, J = 7 Hz), 7.99 (dd, 1Hγ, J = 7, 2 Hz), 8.44
(dd, 1Hα, J = 7, 2 Hz), 12.56 (br s, OH). MS (EI), m/z: 236
(M+).
2
4
1
2
2
O
O
N
N
5
C CH3
1
1
2
HN
O
HN
N
O
N
O
O
O
OH
N
OH
N
OH
N
3
3
OH
N
HImNic
H3NICTAME
H3NICTREN
HOSuNic
Tris{[((2-hydroxynicotinyl)carbonyl]-2-aminoethyl}amine
(H3NICTREN)
To a mixture of HOSuNic (1.17 g, 4.95 mmol) in THF (200
mL) was added TREN (0.240 g, 1.64 mmol) in one portion.
The resultant mixture was refluxed for 18 h, whereupon it was
left to cool to 20°C. The white precipitate was filtered, washed
with CH2Cl2 and Et2O, and dried in vacuo to afford H3NIC-
TREN (0.73 g, 87%), mp 242°C (dec.). IR: 3640–2600 (br,
νOH), 3280 (m, νNH), 1680 (s, νCTO), 1610 (m, νCTC), 1560 (s,
1
amide II (24)), 1473 (m, νCTC), 1328, 1255, 1060, 980. H
NMR: 2.68 (t, 6H1, J = 6 Hz), 3.10–3.60 (partially obscured m,
6H2), 6.43 (t, 3Hβ, J = 6 Hz), 7.64 (dd, 3Hγ, J = 6, 4 Hz), 8.28
© 1998 NRC Canada

417
Hoveyda et al.
(dd, 3Hα, J = 6, 4 Hz), 9.79 (t, 3H, C(O)NH, J = 6 Hz), 12.20
(br s, 3H, OH). 13C NMR (DMSO-d6): 37.42 (C1), 53.46 (C2),
106.14 (Cb), 120.63 (Cγ), 139.30 (Cα), 143.77 (Cδ), 162.32
(C-OH), 163.36 (CTO). Anal. calcd. (found) for
C24H27N7O6 0.5H2O: C 55.60 (55.42), H 5.21 (5.33), N 18.90
(18.83).
product (150 mg, 73% recovery), equivalent to a 51% overall
yield; mp 115–117°C. TLC (eluant: 50% MeOH in EtOAc) Rf
= 0.3. IR: 3600–2600 (br,νOH), 3260 (m, νNH), 1675 (s, νCTO),
1605 (m, νCTC), 1560 (s, amide II), 1230, 775. 1H NMR (δ at
200 MHz in DMSO-d6): 1.3–1.9 (three overlapping m, 6
H3,7,8), 2.9–3.6 (partially obscured overlapping m, 8H2,4,6,9),
6.03–6.22 (m, 1Hβ), 6.40–6.51 (m, 2Hβ′,β′′), 7.24–7.45 (two
overlapping m, 2Hα,γ), 7.60–7.75 (m, 2Hγ′,γ′′), 8.22–8.38 (m,
2Hα′,α′′), 9.60–9.90 (m, 2H, C(O)NH), 11.85 (br s, 1H, OH),
12.45 (br s, 2H, OH). MS (FAB), m/z: 510 ± 1 (M+H)+. Anal.
calcd. (found) for C25H28N6O6 H2O: C 57.03 (57.13), H 5.74
(5.80), N 15.96 (15.82).
Method B: N-(2-Hydroxynicotinyl)-imidazole (HImNic)
To a mixture of 2-hydroxynicotinic acid (5.34 g, 38.4 mmol)
in THF (200 mL) was added CDI (6.43 g, 39.4 mmol) in one
portion. The mixture was refluxed for 19 h. After cooling the
reaction mixture to 20°C, the white precipitate was filtered,
washed with Et2O, and dried in vacuo to afford HImNic
(6.65 g, 98%). 1H NMR: 6.38 (t, 1Hβ, J = 6.6 Hz), 7.05 (dd,
1H4, J = 1.6, 0.8 Hz), 7.61 (t, 1H5, J = 1.6 Hz), 7.74 (dd, 1Hγ,
J = 6.3, 2.2 Hz), 7.93 (dd, 1Hα, J = 6.3, 2.2 Hz), 8.16 (t, 1H2, J
= 0.8 Hz), 12.32 (br s, OH). MS (EI), m/z: 189 (M+). Anal.
calcd. (found) for C9H7N3O2: C 57.14 (56.95), H 3.73 (3.81),
N 22.21 (22.15).
1
3
b
''
''
2
''
HN
O
O
OH
H
8
2
4
6
5
1
N
9
3
7
N
N
N
N
N
H
'
'
OH O
OH
O
'
N
OH
HpropylPic
H3NICSPERM
Tris{[(2-hydroxynicotinyl)carbonyl]-2-aminoethyl}amine
(H3NICTREN)
[(3-Hydroxypicolinyl)carbonyl]propylamine (HpropylPic)
To a mixture of HImNic (6.30 g, 33.3 mmol) in 1,4-dioxan
(400 mL) was added TREN (1.56 g, 10.67 mmol) in one por-
tion. The resulting mixture was refluxed for 18 h. After cooling
the reaction mixture to 20°C, the precipitate was filtered,
washed with THF and Et2O, and dried in vacuo to afford
To a mixture of 3-hydroxypicolinic acid (1.48 g, 10.6 mmol) in
THF (200 mL) was added CDI (1.92 g, 11.8 mmol) in one por-
tion. The mixture was refluxed for 2 h whereupon n-propylamine
(10.1 g, 0.12 mol) was added and the resultant mixture was
refluxed for a further 19 h. After cooling the reaction mixture
to 20°C, the volatiles were removed under reduced pressure
and the resultant oil was column chromatographed on silica gel
(eluant: 50% pentane in Et2O, Rf = 0.84) to afford Hpropylpic
(1.40 g, 74%) as a yellow-brown liquid. IR (neat on KBr
plates): 3600–2600 (br, νOH), 3380 (m, νN-H), 1650 (s, νCTO),
1580 (m, νCTC), 1535 (s, amide II), 1450 (m, νCTC), 1340,
1295, 1255, 1195, 810, 795. 1H NMR: 0.85 (t, 3H1, J = 7 Hz),
1.54 (m, 2H2), 3.25 (dt, 2H3, J = 8, 7 Hz), 7.36 (dd, 1Hγ, J = 8,
0.8 Hz), 7.48 (dd, 1Hβ, J = 8, 4 Hz), 8.12 (dd, 1Hα, J = 4,
1
H3NICTREN as a white solid (5.20 g, 95%). H NMR, 13C
NMR, IR, and MS spectra were identical to those reported for
the sample obtained by Method A. Anal. calcd. (found) for
C24H27N7O6 0.25H2O: C 56.08 (56.15), H 5.39 (5.57), N 19.07
(19.07).
1,1,1-Tris{[(2-hydroxynicotinyl)carbonyl]-2-aminomethyl}eth
ane (H3NICTAME)
To a mixture of HImNic (6.05 g, 32.0 mmol) in 1,4-dioxan
(400 mL) was added TAME (1.10 g, 9.39 mmol) in one por-
tion. The resulting mixture was refluxed for 19 h. After cooling
the reaction mixture to 20°C, the precipitate was filtered,
washed with THF and Et2O, and dried in vacuo to afford
H3NICTAME as an off-white solid (4.22 g, 89 %), mp 270°C
(dec.). IR: 3600–2600 (br, νOH), 3280 (m, νNH), 1665 (s,
νCTO), 1600 (m, νCTC), 1562 (s, amide II), 1480 (m, νCTC),
0.8 Hz), 9.14 (t, 1H, C(O)NH, J = 6 Hz), 12.70 (br s, OH). 13
C
NMR: 11.32 (C1), 22.27 (C2), 40.05 (C3), 125.84 (Cγ), 128.95
(Cβ), 131.49 (Cα), 139.66 (CC(O)NH), 157.37 (C-OH),
168.70 (CTO). Exact Mass calcd. (found) for C9H12N2O2:
180.0904 (180.0898). Anal. calcd. (found) for C9H12N2O2: C
59.99 (59.77), H 6.71 (6.78), N 15.55 (15.50).
1
1430, 1320, 1235, 1150, 790, 770. H NMR: 0.91 (s, 3H1),
3.14–3.47 (partially obscured d, 6H2, J = 6 Hz), 6.46 (t, 3Hβ J
= 6 Hz), 7.69 (dd, 3Hγ, J = 8, 1.2 Hz), 8.30 (dd, 3Hα, J = 8,
1.2 Hz), 10.01 (t, 3H, C(O)NH, J = 6 Hz), 12.38 (br s, 3H,
OH). MS (FAB), m/z: 481 (M+H)+. Anal. calcd. (found) for
C23H24N6O6 H2O: C 55.42 (55.68), H 5.22 (5.28), N 16.87
(16.55).
1,1,1-Tris{[(3-hydroxypicolinyl)carbonyl]-2-aminomethyl}-
ethane (H3PICTAME)
To a mixture of 3-hydroxypicolinic acid (5.34 g, 38.39 mmol)
in THF (150 mL) was added CDI (6.43 g, 39.65 mmol) in one
portion. The mixture was refluxed for 5 h whereupon TAME
(1.50 g, 12.80 mmol) was added and the solution was refluxed
for a further 26 h. After cooling the reaction mixture to 20°C,
the precipitate was filtered and the filtrate was concentrated in
vacuo to yield a dark reddish-brown residue. This residue was
redissolved in a minimum amount of hot MeOH (ca. 40 mL),
which, after cooling to room temperature, was triturated with
Et2O to yield the product as a brown solid (1.25 g, 20%), mp
165–167°C. X-ray quality crystals were obtained by slow
evaporation of a MeOH:Et2O (1:20) solution at room tempera-
ture (rt). IR: 3600–2600 (br, νOH), 3355 (m, νNH), 1650 (s,
νCTO), 1580 (m, νCTC), 1530 (s, amide II), 1448 (m, νCTC),
1355, 1315, 1295, 1245, 1200, 1160, 810, 695, 660. 1H NMR:
N1,N5,N10-Tris(2-hydroxynicotinyl)spermidine
(H3NICSPERM)
To a suspension of HImNic (2.639 g, 13.96 mmol) in 1,4-di-
oxane (200 mL) was added spermidine (0.658 g, 4.53 mmol)
in one portion. The resulting mixture was refluxed for 19 h.
After cooling the reaction mixture to 20°C, the brown-orange
precipitate was filtered, washed with Et2O, and dried in vacuo
to yield 1.611 g of the crude product; 205 mg of this was col-
umn chromatographed on lipophilic Sephadex LH-20™ (3 ×
120 cm column; eluant: 50% MeOH in EtOAc) to afford pure
© 1998 NRC Canada

418
Can. J. Chem. Vol. 76, 1998
0.88 (s, 3H1), 3.20–3.41 (partially obscured d, 6H2, J = 6 Hz),
7.43 (dd, 3Hγ, J = 8, 0.8 Hz), 7.55 (dd, 3Hβ, J = 8, 4 Hz), 8.19
(dd, 3Hα, J = 4, 0.8 Hz), 9.62 (t, 3H, C(O)NH, J = 6 Hz), 12.46
(br s, 3H, OH). MS (FAB), m/z: 481 (M+H)+. Anal. calcd.
(found) for C23H24N6O6: C 57.50 (57.39), H 5.03 (4.90), N
17.49 (17.55).
2
1
CH
2
1
2
2
N
N
C
3
C CH
1
3
1
HN
HN
O
HN
O
HN
N
N
N
N
O
O
δ
δ
OH
OH
H3PICTAME
3
3
3
3
PICTREN
PICTAME
H3PICTREN
refluxed for a further 19 h. After cooling the reaction mixture
to 20°C, and removal of the solvent under reduced pressure,
the residue so obtained was column chromatographed on silica
gel (eluant: 5% MeOH:45% Et2O:50% EtOAc; Rf = 0.58) to
afford PICTAME (1.30 g, 70%) as a white solid, mp
207–208°C. IR: 1670 (s, νCTO), 1590 (m, νCTC), 1535 (s, am-
ide II), 1465 (m, νCTC), 1435, 1335, 1000, 760, 690. 1H NMR:
0.53 (s, 3H1), 3.22 (d, 6H2, J = 6 Hz), 7.62 (ddd, 3Hβ, J = 8, 4,
1.2 Hz), 8.01 (dt, 3Hγ, J = 8, 8, 1.6 Hz), 8.06-8.16 (m, 3Hδ),
8.66-8.74 (m, 3Hα), 9.37 (t, 3H, C(O)NH, J = 6 Hz). MS
(FAB), m/z: 433 (M+H)+. Anal. calcd. (found) for
C23H24N6O3: C 63.88 (63.63), H 5.59 (5.53), N 19.43 (19.20).
Tris{[(3-hydroxypicolinyl)carbonyl]-2-aminoethyl}amine
(H3PICTREN)
To a mixture of 3-hydroxypicolinic acid (2.52 g, 18.12 mmol)
in THF (150 mL) was added CDI (3.50 g, 21.6 mmol) in one
portion. The mixture was refluxed for 5 h whereupon TREN
(0.883 g, 6.04 mmol) was added and the solution was refluxed
for a further 18 h. After cooling the reaction mixture to 20°C,
the precipitate was filtered and the filtrate was concentrated in
vacuo to yield a dark reddish-brown residue. This residue was
redissolved in a minimum amount of hot MeOH (ca. 20 mL),
which, after cooling to rt, was triturated with Et2O to yield the
product as
a light-brown solid (2.15 g, 70%), mp
156–158°C. IR: 3600–2700 (br, νOH), 3350 (m, νNH), 1640 (s,
νCTO), 1590 (m, νCTC), 1540 (s, amide II), 1450 (m, νCTC),
1300, 1240, 1200, 1180, 1150, 860, 820, 685, 665. 1H NMR:
2.78 (t, 6H1, J = 6 Hz), 3.36–3.47 (partially obscured m, 6H2),
7.33 (dd, 3Hγ, J = 8, 0.8 Hz), 7.42 (dd, 3Hβ, J = 8, 4 Hz), 7.97
(dd, 3Hα, J = 4, 0.8 Hz), 8.96 (t, 3H, C(O)NH, J = 6 Hz), 12.52
(br s, 3H, OH). MS (FAB), m/z: 510 (M+H)+. Anal. calcd.
(found) for C24H27N7O6: C 56.58 (56.46), H 5.34 (5.41), N
19.24 (19.21).
''
1
3
N
''
''
N
O
OH
H
N
2
8
2
4
6
5
1
N
9
3
HN
7
'
N
N
H
O
'
OH O
N
N
O
'
OH
3
OH
H3PICSPERM
H3PICTRPN
Tris{[(picolinyl)carbonyl]-2-aminoethyl}amine (PICTREN)
To a mixture of picolinic acid (3.0 g, 24.4 mmol) in THF (200
mL) was added CDI (4.35 g, 26.8 mmol) in one portion. The
resultant mixture was refluxed for 1 h whereupon TREN
(1.17 g, 8.02 mmol) was added and the mixture was refluxed
for a further 19 h. After cooling the reaction mixture to 20°C
and removal of the solvent under reduced pressure, the dark-
brown residue thus obtained was column chromatographed on
lipophilic Sephadex LH-20™ (3 × 120 cm column; eluant:
MeOH) to afford PICTREN (3.43 g, 93%) as a deep red-brown
viscous liquid. IR (in CHCl3): 3390 (m, νNH), 1665 (s, νCTO).
1H NMR: 2.76 (t, 6H1, J = 6 Hz), 3.34–3.50 (partially ob-
scured m, 6H2), 7.49-7.56 (m, 3Hβ), 7.90–8.00 (m, 6Hγ,δ),
8.43-8.47 (m, 3Hα), 8.70 (t, 3H, C(O)NH, J = 6 Hz). MS
(FAB), m/z: 462 (M+H)+. Anal. calcd. (found) for
C24H27N7O3 0.5H2O: C 61.26 (60.99), H 6.00 (5.94), N 20.84
(20.92).
N1,N5,N10-Tris(3-hydroxypicolinyl)spermidine
(H3PICSPERM)
To a mixture of 3-hydroxypicolinic acid (2.657 g, 19.1 mmol)
in THF (275 mL), CDI (3.103 g, 19.1 mmol) was added and
the solution refluxed for 3 h. Spermidine (0.87 g, 5.97 mmol)
was added to the mixture, which was refluxed for an additional
15 h. After cooling the reaction mixture to 20°C, the solution
was filtered and the filtrate was concentrated in vacuo to a
yellowish oil. The slow-moving contaminants (Rf 0.05) in the
mixture were removed by column chromatography on silica
gel (eluant: 5% MeOH in EtOAc). Combining the appropriate
fractions and evaporating the solvent yielded a yellow oil. This
material was then triturated with MeOH (ca. 50 mL) to afford
H3PICSPERM (1.31 g, 42%) as a white solid, mp
99–100°C. TLC (eluant: 5% MeOH in EtOAc) Rf = 0.48. IR:
3600–2600 (br, νOH), 3360 (m, νNH), 1650 (s, νCTO), 1610 (m,
νCTC), 1535 (s, amide II), 1445 (m, νCTC), 1610, 1390, 1190,
990, 810, 795. 1H NMR (δ at 400 MHz in CDCl3): 1.62–2.13
(series of overlapping m, 6H3,7,8), 3.43–3.60 (m, 5H4,6′,9), 3.64,
4.04, 4.12 (three t each, 3H2,6", J = 6 Hz), 7.20–7.37 (m,
6Hβ,β′,β′′,γ,γ′,γ′′), 7.98–8.24 (m, 3Hα,α′,α′′), 8.33, 8.56 (two br s,
2H, C(O)NH, W1/2 = 14 Hz), 12.2 (br s, 2H, OH), 12.3 (br s,
1H, OH). 13C NMR (δ at 50 MHz in CDCl3): 169.4, 169.1,
168.9, 158.8, 158.7, 157.8, 139.4, 138.8, 134.3, 131.4, 128.6,
1,1,1-Tris{[(picolinyl)carbonyl]-2-aminomethyl}ethane
(PICTAME)
To a mixture of picolinic acid (1.576 g, 12.80 mmol) in THF
(150 mL) was added CDI (2.283 g, 14.08 mmol) in one por-
tion. The resultant mixture was refluxed for 2 h whereupon
TAME (0.5 g, 4.30 mmol) was added and the mixture was
© 1998 NRC Canada

419
Hoveyda et al.
127.4, 126.1, 102.9, 49.7, 47.8, 38.8, 36.3, 28.8, 27.6, 26.7,
24.6. MS (FAB), m/z: 509 (M+H)+. Anal. calcd. (found) for
C25H28N6O6: C 59.05 (59.06), H 5.55 (5.63), N 16.53 (16.60).
using the program BEST (26). The program sets up simultane-
ous mass-balance equations for all the components present at
each addition of base and calculates the pH at each data point
according to the current set of stability constants and total
concentrations of each component. The stability constants are
iteratively varied to minimize the sum of the square of the
Tris{[(3-hydroxypicolinyl)carbonyl]-2-aminopropyl}amine
(H3PICTRPN)
difference between observed (pHobs) and calculated (pHcalc
pH values. An indication of the fit is given by σfit, where σfit is
)
To a mixture of 3-hydroxypicolinic acid (1.145 g, 8.23 mmol)
in THF (120 mL) was added CDI (1.335 g, 8.23 mmol) in one
portion. The mixture was refluxed for 2 h whereupon TRPN
(0.5 g, 2.66 mmol) was added and the resultant mixture was
refluxed for a further 19 h. After cooling the reddish-brown
reaction mixture to 20°C, the solvent was removed under re-
duced pressure and the viscous dark-brown liquid was redis-
solved in a minimum amount of hot MeOH (ca.40 mL) and
triturated with Et2O to remove any unreacted starting acid
(0.45 g, 39%) that precipitated. The filtrate was concentrated
under reduced pressure and the brownish oil thus obtained was
column chromatographed twice on lipophilic Sephadex LH-
20™ (3 × 120 cm column; eluant: MeOH) and on silica gel
(eluant: 10% MeOH in EtOAc, Rf = 0.13) to afford
H3PICTRPN (0.13 g, 9%) as a glassy (sticky) brown solid. IR:
3600–2600 (br, νOH), 3350 (m, νNH), 1655 (s, νCTO), 1595 (m,
νCTC), 1540 (s, amide II), 1445 (m, νCTC), 1350, 1289, 1248,
σfit = Σ (pHcalc – pHobs)2/(N–1)
and N is the total number of data points (typically ~100). In all
the titrations considered, σfit < 0.01. The final calculated de-
protonation constants are the average of the deprotonation con-
stants determined from individual titrations; an estimate of the
error in these values is given by one standard deviation (1σ)
for these results.
X-ray crystallographic analysis of H3PICTAME
Crystallographic data for H3PICTAME appear in Table 1. The
final unit-cell parameters were obtained by least squares on the
setting angles for 25 reflections with 2θ = 55.1°–82.4°. The
intensities of three standard reflections, measured every 200
reflections throughout the data collection, showed only small
random fluctuations. The data were processed,2 corrected for
Lorentz and polarization effects, and for absorption (empirical,
based on azimuthal scans).
The structure was solved by direct methods. Non-hydrogen
atoms were refined with anisotropic thermal parameters, the
O-bound hydrogen atoms were refined with isotropic thermal
parameters, and the remaining hydrogen atoms were fixed in
calculated positions with N—H = 0.91 Å, C—H = 0.98 Å, and
BH = 1.2 Bbonded atom. A correction for secondary extinction was
applied (Zachariasen type, isotropic), the final value of the
extinction coefficient being 5.7(3) × 10–7. Neutral atom scat-
tering factors for all atoms and anomalous dispersion correc-
tions for the non-hydrogen atoms were taken from the
International tables for X-ray crystallography (27). Final
atomic coordinates and equivalent isotropic thermal parame-
ters, bond lengths, bond angles, and hydrogen bond geometry
appear in Tables 2–5, respectively. Hydrogen atom parame-
ters, anisotropic thermal parameters, torsion angles, intermo-
lecular contacts, and least-squares planes are available as
supplementary material.3
1
1198, 1150, 807, 695. H NMR: 1.75 (t, 6H1, J = 6 Hz),
2.35–2.70 (partially obscured m, 6H2), 3.35 (dt, 6H3, J = 8,
7 Hz), 7.38 (d, 3Hγ, J = 8 Hz), 7.48 (dd, 3Hβ, J = 8, 3 Hz), 8.10
(d, 3Hα, J = 3 Hz), 9.22 (br s, 3H, C(O)NH, W1/2 = 20 Hz),
12.62 (br s, 3H, OH). MS (FAB), m/z: 552 (M+H)+. Anal.
calcd. (found) for C27H33N7O6 0.25H2O: C 58.32 (58.43), H
6.07 (5.90), N 17.63 (17.55).
Potentiometric equilibrium measurements
All potentiometric measurements on H3NICTREN and
H3NICTAME were made with an automatic titrator system
consisting of a Fisher model 950 digital pH meter equipped
with Orion–Ross glass and calomel reference electrodes, a
Metrohm Dosimat model 665 autoburet, and water-jacketed
titration vessels maintained at 25.0 ± 0.1°C with a Julabo UC
circulating bath (25). Both autoburet and pH meter were interfaced
to an IBM-compatible PC so that the titrations were automated.
During the titrations, the solutions were kept under an argon
atmosphere that was first passed through 2 M NaOH solution.
The electrodes were calibrated before each titration, and
often afterwards, by titrating a known amount of aqueous HCl
with a known concentration of NaOH. A plot of mV (calculated)
vs. pH gave a working slope and intercept so that the pH could
be read directly as –log[H+]. The value of pKw used was 13.76.
All solutions were prepared at an ionic strength of 0.16 M
NaCl. Carbonate-free NaOH (~0.16 M) was prepared with
deionized, distilled water that was boiled and purged with ar-
gon while cooling to ensure removal of CO2. This solution was
standardized potentiometrically with potassium hydrogen
phthalate. Standard HCl solutions were prepared similarly and
were standardized with the NaOH prepared above. Stock li-
gand solutions (~ 0.5 mM) were prepared using deionized, dis-
tilled water and were used within 3–4 days. The compounds
were titrated in the pH range 2–11.5; a minimum of five titra-
tions was performed for each compound.
Results and discussion
Synthetic procedures
HOSu/DCC method (Method A) (28)
Depending on the nature of the starting materials, N,N′-dicy-
clohexylcarbodiimide (DCC) by itself can act in situ as a
2
teXan: Crystal Structure Analysis Package. Molecular Structure
Corp., The Woodlands, Texas. 1985 and 1992.
Supplementary material mentioned in the text can be purchased
3
from: The Depository of Unpublished Data, Document Delivery,
CISTI, National Research Council Canada, Ottawa, Canada K1A
0S2. Tables of hydrogen atom coordinates and bond distances
have also been deposited with the Cambridge Crystallographic
Data Centre, and can be obtained on request from The Director,
Cambridge Crystallographic Centre, 12 Union Road, Cambridge,
CB1 1EZ. U.K.
Potentiometric calculations
The deprotonation constants of the ligands were calculated
© 1998 NRC Canada

420
Can. J. Chem. Vol. 76, 1998
Table 1. Crystallographic data for H3PICTAME.a
Table 2. Final atomic coordinates (fractional) and Beq (Å2)
for H3PICTAME.a
Formula
fw
C23H24N6O6
480.48
Atom
x
y
z
Beq
0.22904(12) 5.56(5)
0.1759(2) 7.88(8)
Crystal system
Space group
a, Å
b, Å
c, Å
Monoclinic
P21/a
O(1) 0.5379(2)
O(2) 0.5327(3)
O(3) 0.3223(2)
O(4) 0.1540(2)
O(5) 0.4835(2)
O(6) 0.6443(4)
N(1) 0.4708(2)
N(2) 0.3842(3)
N(3) 0.3942(2)
N(4) 0.2565(3)
N(5) 0.4902(2)
N(6) 0.6955(3)
C(1) 0.4177(3)
C(2) 0.2945(3)
C(3) 0.5197(3)
C(4) 0.4863(3)
C(5) 0.4439(3)
C(6) 0.4710(3)
0.19534(12)
0.0389(2)
0.57156(12)
0.69186(15)
0.51490(15)
0.5781(2)
0.22750(13)
0.0661(2)
0.54131(13)
0.67422(15)
0.37287(15)
0.3500(2)
0.3826(2)
0.3734(2)
0.3153(2)
0.1737(2)
0.0838(2)
0.0208(2)
10.257(2)
15.572(3)
15.208(2)
96.124(15)
2415.2(8)
4
0.45428(12) 5.52(5)
0.46730(15) 6.83(7)
0.14869(13) 6.53(6)
0.0469(2)
0.36175(14) 4.62(6)
0.3798(2) 6.38(8)
0.32155(14) 4.60(6)
0.24658(15) 5.64(7)
0.18000(14) 5.10(6)
β,°
V, Å3
9.72(9)
Z
ρcalc, g/cm3
1.321
F(000)
1008
Radiation (λ, Å)
µ, cm–1
Cu (1.54178)
8.21
0.0871(2)
0.3313(2)
0.3794(2)
0.3664(2)
0.2961(2)
0.3080(2)
0.2472(2)
0.2641(3)
0.3373(3)
0.3938(3)
0.3516(2)
0.3741(2)
0.3317(2)
0.3813(2)
0.3397(2)
0.2529(3)
0.2085(2)
0.2318(2)
0.1416(2)
0.0898(2)
0.0444(2)
–0.0053(3)
–0.0085(3)
0.0397(3)
8.09(10)
3.88(6)
5.10(7)
4.18(6)
4.52(7)
4.87(8)
5.76(9)
7.41(11)
8.10(12)
8.06(12)
4.30(6)
4.49(7)
4.49(7)
5.04(8)
6.52(10)
7.03(10)
6.81(10)
4.63(7)
5.60(9)
6.42(10)
8.12(13)
10.5(2)
11.9(2)
10.36(15)
Crystal size, mm
Transmission factors
Scan type
0.12 × 0.25 × 0.30
0.97–1.00
ω–2θ
Scan range, deg in ω
Scan speed, deg/min
Data collected
2θmax, deg
0.89 + 0.20 tan θ
32 (up to 8 rescans)
+h, +k, ±l
155
C(7) 0.4358(4) –0.0632(2)
C(8) 0.3773(4) –0.0809(2)
C(9) 0.3524(4) –0.0152(2)
Crystal decay, %
Total reflections
Total unique reflections
Rmerge
Negligible
5368
5099
C(10) 0.4804(3)
C(11) 0.3232(3)
C(12) 0.2444(3)
C(13) 0.1655(3)
C(14) 0.0950(3)
C(15) 0.1050(4)
C(16) 0.1874(4)
C(17) 0.3797(3)
C(18) 0.5335(3)
C(19) 0.6459(4)
C(20) 0.6950(4)
C(21) 0.7968(5)
C(22) 0.8473(5)
C(23) 0.7953(5)
0.4707(2)
0.5867(2)
0.6585(2)
0.7071(2)
0.7739(2)
0.7895(2)
0.7392(2)
0.3717(2)
0.4421(2)
0.4294(3)
0.4972(3)
0.4827(4)
0.4034(5)
0.3367(4)
0.022
Reflections with I ≥ 3σ(I)
No. of variables
R
2506
329
0.041
Rw
0.038
gof
2.27
Max ∆/σ (final cycle)
Residual density e/Å3
0.0003
–0.12 to 0.25
a Temperature 294 K, Rigaku AFC6S diffractometer,
graphite monochromator, takeoff angle 6.0°, aperture 6.0 ×
6.0 mm at a distance of 285 mm from the crystal, stationary
background counts at each end of the scan (scan/background
time ratio 2:1), σ2(F2) = [S2(C + 4B)]/Lp2 (S = scan rate, C =
scan count, B = normalized background count), function
minimized Σw(|Fo| − |Fc|)2 where w = 4Fo/σ2(Fo2),
aBeq = (8⁄3)π2ΣΣ(Uij ai aj (ai aj)).
R = Σ ||F | − |F || ⁄ Σ |F |, Rw = (Σw(|F | − |F | ⁄Σw|F | ) ⁄ , and
2
)
o
c
o
o
c
o
gof =[Σw(|Fo| − |Fc| ⁄(m − n)] ⁄ . Values given for R, Rw, and
gof are based on those reflections with I ≥ 3σ(I).
)
the isolation of the intermediate imidazolide. Due to the lack
of sufficient solubility in THF (bp 67°C), the N-acylimida-
zolide of 2-hydroxynicotinic acid was isolated, and the next
step performed in a more suitable, i.e., more polar, solvent,
dioxane. Conveniently enough, these crystalline intermedi-
ates, although highly reactive in solution towards nucleophiles
such as amine nitrogen, are stable in air at room temperature
for at least a week.
dehydrative coupling reagent, leading to the formation of the
amide bond. In the present work, however, it was important
that an activated ester be isolated, as the reaction between the
amines (e.g., TREN or TAME) and the pyridine carboxylic
acids used was fast enough to result in the quantitative precipi-
tation of the corresponding ammonium salt. The main draw-
back in using this method was the practical difficulty of
purifying the activated ester from the dicyclohexylurea that
formed as a side product. This problem was circumvented by
adopting the more efficient CDI method.
X-ray crystal structure of H3PICTAME
In X-ray crystallography, the main criterion for identification
of an H-bond is the distance between the hydrogen and the
acceptor atom (that is, the distance H A in B—H A where A
and B are frequently nitrogen or oxygen atoms). However, in
using X-ray diffraction data to extract bond distances involv-
ing a hydrogen atom, one must be aware of the fact that such
bond lengths are systematically shortened (relative to the
“true” internuclear separation) since these distances are,
strictly speaking, a measure of the distance between the
CDI method (Method B) (31)
The advantage of method B is that it obviates the intermediacy
of DCC (and hence the DCU side product). The side products
formed in method B (imidazole and CO2) are easier to remove.
This method, although it involves two steps, allows for a sin-
gle-step synthesis of the amide unless practical reasons call for
© 1998 NRC Canada

421
Hoveyda et al.
Table 4. Bond angles (deg) in H3PICTAME with estimated
Table 3. Bond lengths (Å) in H3PICTAME with estimated
standard deviations in parentheses.
standard deviations in parentheses.
Angle
(deg)
Angle
(deg)
Bond
Length
Bond
Length
Bonds
Bonds
O(1)—C(4)
O(3)—C(11)
O(5)—C(18)
N(1)—C(3)
N(2)—C(5)
N(3)—C(10)
N(4)—C(12)
N(5)—C(17)
N(6)—C(19)
C(1)—C(2)
1.245(3)
1.243(3)
1.254(3)
1.456(3)
1.336(3)
1.454(3)
1.336(3)
1.447(3)
1.339(4)
1.532(3)
1.533(3)
1.482(4)
1.388(4)
1.377(5)
1.388(4)
1.357(4)
1.478(4)
1.370(6)
1.409(7)
O(2)—C(6)
O(4)—C(13)
O(6)—C(20)
N(1)—C(4)
N(2)—C(9)
N(3)—C(11)
N(4)—C(16)
N(5)—C(18)
N(6)—C(23)
C(1)—C(3)
C(1)—C(17)
C(5)—C(6)
C(7)—C(8)
C(11)—C(12)
C(13)—C(14)
C(15)—C(16)
C(19)—C(20)
C(21)—C(22)
1.343(4)
1.347(3)
1.365(5)
1.326(3)
1.331(4)
1.339(3)
1.332(4)
1.325(4)
1.330(5)
1.535(3)
1.531(3)
1.396(4)
1.349(5)
1.486(4)
1.381(4)
1.381(4)
1.385(5)
1.342(7)
C(3)-N(1)-C(4)
C(10)-N(3)-C(11)
C(17)-N(5)-C(18)
C(2)-C(1)-C(3)
C(2)-C(1)-C(17)
C(3)-C(1)-C(17)
N(1)-C(3)-C(1)
O(1)-C(4)-C(5)
N(2)-C(5)-C(4)
C(4)-C(5)-C(6)
O(2)-C(6)-C(7)
C(6)-C(7)-C(8)
N(2)-C(9)-C(8)
O(3)-C(11)-N(3)
124.2(2) C(5)-N(2)-C(9)
124.2(2) C(12)-N(4)-C(16) 117.4(3)
124.7(3) C(19)-N(6)-C(23) 118.5(4)
109.7(2) C(2)-C(1)-C(10)
109.0(2) C(3)-C(1)-C(10)
111.1(2) C(10)-C(1)-C(17) 110.8(2)
113.9(2) O(1)-C(4)-N(1)
120.8(3) N(1)-C(4)-C(5)
117.4(3) N(2)-C(5)-C(6)
120.1(3) O(2)-C(6)-C(5)
120.1(3) C(5)-C(6)-C(7)
119.5(3) C(7)-C(8)-C(9)
122.8(3) N(3)-C(10)-C(1)
117.9(3)
109.7(2)
106.5(2)
122.7(3)
116.4(3)
122.5(3)
122.1(3)
117.8(3)
119.5(3)
112.7(2)
C(1)—C(10)
C(4)—C(5)
C(6)—C(7)
C(8)—C(9)
123.2(3) O(3)-C(11)-C(12) 120.5(3)
C(12)—C(13)
C(14)—C(15)
C(18)—C(19)
C(20)—C(21)
C(22)—C(23)
N(3)—C(11)-C(12) 116.3(3) N(4)-C(12)-C(11) 117.1(3)
N(4)-C(12)-C(13)
O(4)-C(13)-C(12)
123.1(3) C(11)-C(12)-C(13) 119.7(3)
123.2(3) O(4)-C(13)-C(14) 118.8(3)
C(12)-C(13)-C(14) 117.9(3) C(13)-C(14)-C(15) 119.4(3)
C(14)-C(15)-C(16) 119.2(3) N(4)-C(16)-C(15) 122.9(3)
N(5)-C(17)-C(1)
O(5)-C(18)-C(19)
N(6)-C(19)-C(18)
C(18)-C(19)-C(20) 120.7(4) O(6)-C(20)-C(19) 121.9(4)
O(6)-C(20)-C(21) 118.8(5) C(19)-C(20)-C(21) 119.3(5)
C(20)-C(21)-C(22) 119.4(6) C(21)-C(22)-C(23) 119.3(5)
N(6)-C(23)-C(22) 121.5(5)
113.9(2) O(5)-C(18)-N(5)
121.1(3) N(5)-C(18)-C(19) 116.4(3)
117.4(3) N(6)-C(19)-C(20) 121.9(4)
122.6(3)
centroids of electron density of the two atoms concerned. In
the case of hydrogen, this does not coincide with its nuclear
position, but is displaced significantly in the direction of the
covalent bond (i.e., B—H). For example, O—H and N—H
covalent bond lengths, which are observed to be 0.97 and 1.03
Å by neutron diffraction, can be determined to be as short as
0.6 and 0.7 Å, respectively, by X-ray diffraction (33). Hence,
in order to identify an H-bond based on X-ray diffraction data,
the H A bond length should be at least 0.09 Å (34) shorter than
the sum of the van der Waals radii (35) of H and A (36).
Angular orientations are also of importance, viz. the extent of
deviation from linearity of the A—H B bond. Statistical sur-
veys of H-bond geometry in small molecules have revealed a
mean value close to 180° (37), although exceptions to this
finding are plentiful and are often due to the geometric con-
straints imposed elsewhere in the structure.
In the solid state structure, an H-bond will form whenever
it can because its formation will raise the stability of a molecu-
lar crystal by a few kilocalories per mole (38). The H3PIC-
TAME solid state structure nicely demonstrates this
observation, with eight intra- and one intermolecular H-bonds
per molecule. Following the nomenclature of H-bonding ge-
ometry, two types of such interactions are present in H3PIC-
TAME: two-center (chelate) and three-center (bifurcated)
H-bonds (Fig. 2).
The atom labelling and numbering scheme used is shown
in the ORTEP diagram of the X-ray crystal structure of H3PIC-
TAME (Fig. 3a). In each arm of the molecule, the arylamide
moiety is held strictly planar (Fig. 3c) due to the extended π
network of the aromatic amide linkage. The H-bond distances
and some of the relevant bond angles are listed in Table 5, a
perusal of which clearly indicates that all the H-bonds identi-
fied are well within the H A distance criterion (H O = 2.5
Å; H N = 2.6 Å) and, as expected, the three-center H-bonds
(H(5) H-bonded to N(3), N(4), and O(5) and H(6) H-bonded
to N(5), N(6), and O(1), respectively) display a greater devia-
tion from linearity.
Table 5. Hydrogen bonds (Å) and C—H O interaction
geometries (deg) in H3PICTAME with estimated standard
deviations in parentheses.
Interaction
D—H H A
D A D—H A
O(2)—H(1) O(1)b 0.96(3) 1.68(4) 2.565(3) 152(3)
O(4)—H(2)O(3)b
O(6)—H(3)O(5)b
N(1)—H(4)O(1)a,b
1.01(4) 1.65(4) 2.569(3) 148(3)
1.29(5) 1.28(5) 2.575(3) 178(4)
0.91
2.18
2.29
2.18
2.29
2.15
2.27
2.51
2.49
2.51
3.069(3) 164
2.689(3) 106
2.904(3) 136
2.690(3) 107
2.891(3) 138
2.682(3) 107
2.824(3)
2.842(3) 101
2.827(3) 98
N(1)—H(4) N(2)b 0.91
N(3)—H(5) O(5)c 0.91
N(3)—H(5) N(4)c 0.91
N(5)—H(6) O(1)c 0.91
N(5)—H(6) N(6)c 0.91
C(3)—H(11) O(1) 0.98
C(10)—H(16) O(3) 0.98
C(17)—H(20) O(5) 0.98
99
a Symmetry operation: ½ + x, –1/2 + y, 1 – z.
b Two centred H-bond.
c Three centred H-bond.
Fig. 2. The two types of H-bond present in the H3PICTAME X-ray
crystal structure.
A
B
H
B
H
A
A'
two-center (chelate)
three-center (bifurcated)
© 1998 NRC Canada

422
Can. J. Chem. Vol. 76, 1998
Fig. 3. (a) ORTEP representation of the X-ray crystal structure of H3PICTAME; (b) torsion angles between amide nitrogens and apical C(2);
(c) wire frame diagram showing only the three arylamide moieties with arrows representing the dihedral angles between the least-squares
planes. In the wire framework representation, the arrangement and orientation of atoms are identical to that in (a). All values shown are based
on the X-ray diffraction data; the planes include refined hydrogen atoms as well as all other non-hydrogen atoms of the arylamide moiety, but
the backbone carbons are omitted for clarity.
© 1998 NRC Canada

423
Hoveyda et al.
Table 6. pKa valuesa of H3NICTAME, H3NICTREN, and related
Fig. 4. The amide resonance and H-bonding (see text).
ligands (0.16 M NaCl, 25°C).
pKa H3NICTAME H3NICTREN H2nice H2pice H6TRNSf H6TAMSf
H
H
-
O
O
O
10.68 (2)b
10.58 (4)b
9.71 (3)b
11.2 (2)b 11.99b 11.06b 11.2 (1)c 11.19 (4)c
O
C
1
2
3
4
5
6
10.7 (2)b
10.0 (1)b
6.25 (4)c
4.98d 5.03d 10.6 (1)c 9.81 (4)c
C
+
N
< 0
< 0
9.59 (3)c 8.91 (2)b
8.07 (3)b 7.95 (3)b
7.29 (3)b 6.56 (2)b
6.17 (3)b 2.92 (2)c
N
H
H
a The numbers in parentheses refer to 1σ.
b Phenol deprotonation.
c Ammonium deprotonation.
d Pyridinium deprotonation.
e Reference 44.
i
ii
General characterization
The preorganization of a multidentate ligand is governed by
two factors, each capable of diminishing the entropic penalty
attendant upon metal ion complexation: (i) the conformation
of each metal binding moiety in the molecule; (ii) the confor-
mation of the “arms” (containing the binding moiety) with
respect to each other. The first factor determines, for instance,
the ease with which a chelate ring will form, while the second
factor contributes to the complete encapsulation of the metal
ion. Based on the solid state structure of H3PICTAME, it is
evident that although the structure is preorganized in the first
sense, the three arms are not favorably disposed to define the
cavity necessary for the coordination of a group 13 metal ion
in a hexacoordinate octahedral polyhedron. In fact, in solution,
f Reference 45.
The molecule in the solid state lacks a C3 axis of symmetry;
the torsion angles about the apical C(2), the intervening
methylene carbon, and the amidic nitrogen succinctly depict
this fact (Fig. 3b). The apical carbon (C2) is in a staggered
conformation with respect to N5; whereas the analogous tor-
sion angles in the other two arms are roughly 60°, correspond-
ing to a gauche conformation with one in a positive
(clockwise) and the other in a negative (counterclockwise)
sense. Since nearly all H-bond donors and acceptors are “bur-
ied” intramolecularly, preorganizing the molecule for chela-
tion, only one intermolecular H-bond forms in each molecule
as shown in Fig. 3. It is reasonable to conjecture, in the absence
of an X-ray structure, that the structural isomer, H3NICTAME,
should show a greater propensity toward aggregation (as op-
posed to intramolecular H-bond formation) based merely on
geometric considerations.
1
as evinced by H NMR spectroscopy, one set of signals is
observed for the tripodal arylamides prepared, indicating that
each arm rotates freely. The exceptions to this finding are, of
course, H3PICSPERM and H3NICSPERM, which are asym-
metric by virtue of their spermidine backbones. Therefore, for
the purposes of solution coordination chemistry, the latter facet
of preorganization is of no relevance in the case of the aryl-
amides.
A noteworthy detail in the solid state H-bonding network
present in H3PICTAME is the variation in the H-bond dis-
tances (and angles) of the two-center O—H O (Table 5).
Thus, whereas O(6)—H(3) O(5) constitutes a symmetrical
H-bonding site, the other two analogous H-bonds are asym-
metric and, accordingly, the corresponding H O distances are
significantly longer. The most interesting aspect about the ge-
ometry of the symmetrical site is that, as a result of its perfect
linearity (within experimental error), an unseemly angle of
87(2)° is imposed upon the C(20)-O(6)-H(3) bond — in con-
trast to 105(2)° for the other two analogous angles (Fig. 3). In
spite of this anomalous bond angle, however, the symmetrical
disposition of the H3, with respect to the two adjacent oxygens
(O6 and O5), is corroborated by the significant lengthening of
the respective C(18)—O(5) bond to 1.254 Å, (C(4)—O(1) =
1.245 Å, C(11)—O(3) = 1.243 Å) and a decrease in the
C(18)—N(5) bond to 1.324 Å (C(4)—N(1) = 1.327 Å,
C(11)—N(3) = 1.333 Å). This observation would be expected
if the amide resonance structure ii (Fig. 4) were more predomi-
nant; the more the type ii resonance contributes to the overall
resonance hybrid structure, the better the H-bonding acceptor
ability of that amidic oxygen. The structural data summarized
above clearly reflect this fact: the concomitant increase in CTO
and decrease in C—N bond lengths is paralleled by a shorten-
ing of the H-bond distance. Thus, not too surprisingly, the
longest CTO bond length, C(18)—O(5), corresponds to the
shortest H-bond (1.28(5) Å, O(5) H(3)).
In considering the local conformation of each individual
metal binding moiety, it is possible to predict which conformer
is preferred (Fig. 5) by comparing the strengths of the hydro-
gen bonds available to the two forms. The strength of an H-
bond (A—H B) is known to depend on the relative acidity of
A and the basicity of B; the strength of the H-bond should
increase as the acid becomes more acidic and the base more
basic (39). Since the phenolic hydrogens (pKa ca. 10) (40) are
more acidic than the amide (pKa ca. 15) (41), the strongest
H-bonds ought to occur in conformations 1a and 2a (Fig. 5).
The 1H NMR and IR spectroscopic results, although not
unequivocally conclusive, appear to be in accord with the
qualitative prediction that conformers 1a and 2a are prevalent
in solution.
The 1H NMR chemical shifts of the phenolic hydrogens of
the arylamides show extensive downfield shifts (ca. 12 ppm in
DMSO-d6) indicative of H-bonding. The same observation
was found to hold true in the less polar solvent CDCl3, in the
case of the picolinic amides; the nicotinic derivatives were not
soluble enough in chloroform to allow for 1H NMR spectros-
copy (vide infra). On the other hand, the IR spectrum of
H3PICTAME in CCl4 shows a broad band ranging from ca.
3200 to 2700 cm–1 characteristic of an intramolecular chelated
phenolic OH stretching frequency (42). In the solid state IR
spectrum of H3PICTAME (KBr pellet), however, the phenolic
stretching frequency ranges over a greater span from ca. 3600
© 1998 NRC Canada

424
Can. J. Chem. Vol. 76, 1998
Fig. 5. The two conformers of 3-hydroxypicolinyl (1a, 1b) and 2-hydroxynicotinyl amide derivatives (2a, 2b). (The dipole moments are
depicted merely in a qualitative sense; they are not drawn to scale).
H
H
N
O
N
O
O
O
N
N
O
O
N
H
N
H
H
H
N
O
H
N
O
H
1a
2a
1b
2b
µnet
net
net
net
<
>
Observed water solubility trend: 1 < 2
appear in Table 6, and correspond to the successive deprotona-
tion equilibria of the three phenols (pKa1, pKa2, pKa3) and, in
the case of H3NICTREN, the deprotonation of the protonated
apical nitrogen. The pKa values of the pyridine ring nitrogens
are too low to be measured in the pH range examined, while
those of the amides fall outside the upper pH limit of the meas-
urements. The water solubility of H3PICTAME and H3PIC-
TREN was too low for potentiometric measurements.
It is appropriate to compare the pKa’s of H3NICTAME and
H3NICTREN with those of the parent acids and some related
aminephenol ligands (Table 6). The phenolic pKa’s of H3NIC-
TAME, H3NICTREN, H2nic, and H2pic are particularly high.
This would be anticipated for H3NICTAME on the basis of its
crystal structure, where all three phenol groups are involved
in hydrogen bonding with the adjacent amide carbonyl oxy-
gen. A similar explanation would account for the phenolic
pKa’s of H3NICTREN, and the high pKa1 of H2nic and H2pic
is consistent with the presence of a six-membered H-bond ring
involving the phenolic group and the deprotonated carboxylate
oxygen of these ligands. These arguments are reinforced with
the observation that for the aminephenols H6TRNS and
H6TAMS, the phenolic pKa’s are much lower as there is no
possibility of hydrogen bonding for the protonated phenol. In
fact, the deprotonated phenolate oxygen may H-bond with an
adjacent ammonium group, which would tend to favour depro-
tonation of the phenol and a low pKa.
to 2700 cm–1, in agreement with the solid state structure pro-
vided by X-ray crystallography wherein both intra- and inter-
molecular H-bonds are found to exist (the intermolecularly
H-bonded OH stretching frequencies are found in the
3550–3450 cm–1 range) (42).
It is worthwhile to reflect on the significant solubility dif-
ference between the 2-hydroxynicotinyl and the 3-hydroxypi-
colinyl amide derivatives. The water solubility (at neutral pH)
of the 3-hydroxypicolinyl amide derivatives is typically on the
order of tens of µM as opposed to the 2-hydroxynictonyl coun-
terparts that are soluble to the extent of ca. 0.5 mM, which is
in accord with the polarity differences expected based on a
qualitative consideration of dipole moments. In calculating the
dipole moments of the molecule (µnet) it is convenient to em-
ploy the moments of groups of atoms (amide, pyridine, OH)
as an approximation (43) as shown in Fig. 5, where the mag-
nitude of the three group moments are taken to be the same. It
is clear that only type a conformers (Fig. 5) corroborate the
observed water solubility trend.
The fact that 2-hydroxynicotinic amides adopt the type 2a
conformation is unexpected from the point of view of the elec-
troneutrality principle (ref. 35, pp. 172 and 273), which would
favor a conformation with a smaller electric dipole moment,
i.e., the 2b conformer. The reason for this conformational pref-
erence lies in H-bonding (cf. Fig. 5): the H-bonding in 2a is
more fortified, as discussed previously. Thus, the O—H O
hydrogen bond should still be stronger than N—H O in water,
even though this effect is mitigated in the presence of such a
polar solvent; the H-bond energy decreases in more polar en-
vironments, since the H-bond has a substantial electrostatic
component. Based on the evidence discussed here, the 2-hy-
droxynicotinic and the 3-hydroxypicolinic arylamides display
an intramolecular H-bonding network that is sustained in
water. Therefore, it can be concluded that these arylamide li-
gands are significantly partially preorganized for metal ion
chelation.
Concluding remarks
A series of potentially polydentate arylamide ligands have
been synthesized from the coupling of several tripodal amines
with nicotinic and picolinic acids and their derivatives. Two
synthetic procedures were utilized in this study but the HO-
SuNic method was not favoured due to the difficulty of sepa-
rating the DCU side product from the activated ester. Hence,
the CDI method was used in the synthesis of the compounds in
this work.
The X-ray crystal structure of H3PICTAME revealed
extensive hydrogen bonding, with eight intra- and one
Ligand deprotonation constants
The deprotonation constants4 were obtained on the basis of a
three- or four-proton model for H3NICTAME and H3NIC-
TREN, respectively, in the pH 2.2–11.5 region at 25°C and in
0.16 M NaCl. The deprotonation constants thus obtained
4
The acid dissociation constant, Ka, is defined as:
(m–1)+
HnLm+ U H
L
+ H+
n–1
L
(m–1)+
Kan = ([H
] [H+]) / [HnLm+
]
n–1
pKa = –log Ka
© 1998 NRC Canada

425
Hoveyda et al.
19. T.G. Lutz, D.J. Clevette, H.R. Hoveyda, V. Karunaratne,
A. Nordin, S. Sjöberg, M.Winter, and C. Orvig. Can. J. Chem.
72, 1362 (1994).
20. D. Kahne and W.C. Still. J. Am. Chem. Soc. 110, 7529 (1988).
21. E.B. Fleischer, A.E. Gebala, A. Levey, and P.A. Tasker. J. Org.
Chem. 36, 3042 (1971).
intermolecular H-bonds per molecule. In solution each arm of
the tripodal arylamides rotates freely. Nevertheless, the NMR,
IR, and solubility results indicate that the intramolecular H-
bonding network of these arylamides is sustained in solution
and, thus, these ligands are partially preorganized for the com-
plexation of metal ions.
22. W.C. Still, M. Kahn, and A. Mitra. J. Org. Chem. 43, 2923
(1978).
Acknowledgments
23. F.E. Hahn and M. Tamm. Angew. Chem. Int. Ed. Engl. 30, 203
(1991).
Acknowledgement is made to the Natural Sciences and Engi-
neering Research Council of Canada (NSERC), to the
U.S. National Institutes of Health (CA48964), and to the
B.C. Health Research Foundation for operating funds as well
as to the University of British Columbia for fellowships to
H.R.H. (University Graduate Fellowship) and A.K.W.S. (Kil-
lam Trust).
24. T. Miyazawa, T. Shimanouchi, and S.-I. Mizuchima. J. Chem.
Phys. 24, 408 (1956).
25. P. Caravan, T. Hedlund, S. Liu, S. Sjoberg, and C. Orvig. J. Am.
Chem. Soc. 117, 11230 (1995).
26. A.E. Martell and R.J. Motekaitis. Determination and use of sta-
bility constants. VCH, New York. 1988. p. 4.
27. (a) International tables for X-ray crystallography. Vol.
IV. Kynoch Press, Birmingham, U.K. (present distributor Klu-
wer Academic Publishers: Boston, Mass., U.S.A.). 1974.
pp. 99–102; (b) International tables for crystallography. Vol.
C. Kluwer Academic Publishers, Boston, Mass., U.S.A. 1992.
pp. 200–206.
References
1. R.D. Shannon. Acta Crystallogr. Sect. A: Cryst. Phys. Diffr.
Theor. Gen. Crystallogr. A32, 751 (1976).
2. F.A. Cotton and G. Wilkinson. Advanced inorganic chemistry.
28. G.W. Anderson, J.,E. Zimmerman and F.M. Callahan. J. Am.
Chem. Soc. 86, 1839 (1964).
4th ed. Wiley, Toronto. 1980. p. 682.
29. J.C. Sheehan and G.P. Hess. J. Am. Chem. Soc. 77, 1067 (1955).
30. A. Williams and I.T. Ibrahim. Chem. Rev. 81, 589 (1981).
31. R. Paul and G.W. Anderson. J. Am. Chem. Soc. 82, 4596 (1960).
32. H.A. Staab and W. Rohr. In Newer methods of preparative or-
ganic chemistry. Edited by W. Foerst. Academic, New York.
1968. pp. 72–74.
3. H. Kido and K. Saito. J. Am. Chem. Soc. 110, 3187 (1980).
4. Struct. Bonding. 58, (1984).
5. K.N. Raymond. Coord. Chem. Rev. 105, 135 (1990).
6. S.M. Moerlien, M.J. Welch, K.N. Raymond, and F.L. Weitl. J.
Nucl. Med. 22, 720 (1982).
7. M.A. Green and M.J. Welch. J. Nucl. Med. Biol. 16, 435 (1989).
8. T.B. Karpishin, T.M. Dewey, and K.N. Raymond. J. Am. Chem.
Soc. 115, 1842 (1993).
33. G.A. Jeffrey and W. Saenger. Hydrogen bonding in biological
structures. Springer, Berlin. 1991. p. 108.
34. M.R. Churchill. Inorg. Chem. 12, 1213 (1973).
35. L. Pauling. The nature of the chemical bond. 3rd ed. Cornell
University Press, Ithaca, N.Y. 1960. p. 449.
36. W.C. Hamilton and J.A. Ibers. Hydrogen bonding in solids. Ben-
jamin, New York. 1968. p. 14.
9. L.D. Loomis and K.N. Raymond. Inorg. Chem. 30, 906 (1991).
10. K. Spartalian, W.R. Oosterhuis, and J.B. Neilands. J. Chem.
Phys. 62, 3538 (1975).
11. W.R. Harris, C.J. Carrano, S.R. Cooper, S.R. Sofen, A. Avdeef,
J.V. McArdle, and K.N. Raymond. J. Am. Chem. Soc. 101, 6097
(1979).
37. E.N. Baker and R.E. Hubbard. Prog. Biophys. Mol. Biol. 44, 98
(1984), and references cited therein.
12. K.N. Raymond and C.J. Carrano. Acc. Chem. Res. 12, 183
38. J. Donohue. J. Phys. Chem. 56, 502 (1952).
39. T.M. Garrett, M.E. Cass, and K.N. Raymond. J. Coord. Chem.
25, 241 (1992).
(1979).
13. W.R. Harris and K.N. Raymond. J. Am. Chem. Soc. 101, 6534
(1979).
40. H.O. House. Modern synthetic reactions. 2nd ed. Benjamin,
Menlo Park, N.J. 1972. p. 494.
14. V.L. Pecoraro, W.R. Harris, G.B. Wong, C.J. Carrano, and
K.N. Raymond. J. Am. Chem. Soc. 105, 4623 (1983).
15. M.E. Cass, M. Garrett, and K.N. Raymond. J. Am. Chem. Soc.
111, 1677 (1989).
16. M.M. Finnegan, S.J. Rettig, and C. Orvig. J. Am. Chem. Soc.
108, 5033 (1986).
17. M. Finnegan, T.G. Lutz, W.O. Nelson, A. Smith, and C. Orvig.
Inorg. Chem. 26, 2171 (1987).
18. T.G. Lutz, D.J. Clevette, S.J. Rettig, and C. Orvig. Inorg. Chem.
28, 715 (1989).
41. H. Sigel and R.B. Martin. Chem. Rev. 82, 385 (1982).
42. L.J. Bellamy. Infrared spectra of complex molecules. Vol. 1. 3rd
ed. Chapman and Hall, New York. 1980. p. 108.
43. O. Exner. Dipole moments in organic chemistry. Georg Thieme,
Stuttgart. 1975. pp. 30–34.
44. A.K. W. Stephens and C. Orvig. Inorg. Chim. Acta, 273, 47
(1998).
45. P. Caravan and C. Orvig. Inorg. Chem. 36, 236 (1997).
© 1998 NRC Canada