Synthesis and NMR studies of new DOTP-like lanthanide(III) complexes containing a hydrophobic substituent on one phosphonate side arm

Article abstract of DOI:10.1021/ic010291s

Three derivatives of 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrakis(methylene phosphonic acid) (DOTP) containing a hydrophobic substituent on one side chain were prepared and their lanthanide complexes examined by NMR. The new ligands include 1-(1-octyl-methyl-phosphonic acid)-4,7,10-tris(methylene phosphonic acid)-1,4,7,10-tetraazacyclododecane (C₈-DOTP),1-(1 -undecyl-methyl-phosphonic acid)-4,7,10-tris(methylene phosphonic acid)-1,4,7,10-tetraazacyclododecane (C₁₁-DOTP), and 1-(1-4-nitro-phenyl-methyl-phosphonic acid)-4,7,10-tris(methylene phosphonic acid)-1,4,7,10-tetraazacyclododecane (NO₂-Ph-DOTP). ¹H NMR spectra of the ytterbium(III) complexes were assigned by using a combination of COSY spectroscopy and a fitting procedure that matches experimental NMR hyperfine shifts with those estimated from a MMX-derived structure. The analysis showed that a single isomer is present in solution and that the bulky hydrophobic substituent occupies the less sterically demanding H₆ equatorial position in the YbL^5- complexes. Although the YbL^5- complexes have lower symmetry due to the added substituent, the average ¹H hyperfine shifts are 5-10% larger in these complexes compared to YbDOTP^5-. This was magnified further in the hyperfine ²³Na NMR shifts of ion-paired sodium ions where the extracellular Na⁺ signal in perfused rat hearts displayed a 28% larger hyperfine shift in the presence of Tm(C₁₁-DOTP)^5- than with an equivalent amount of TmDOTP^5-.

Full text of DOI:10.1021/ic010291s

6572
Inorg. Chem. 2001, 40, 6572-6579
Synthesis and NMR Studies of New DOTP-like Lanthanide(III) Complexes Containing a
Hydrophobic Substituent on One Phosphonate Side Arm
Xiaodong Li,† Shanrong Zhang,† Piyu Zhao,† Zoltan Kovacs,† and A. Dean Sherry*,†,‡
Department of Chemistry, University of Texas at Dallas, Richardson, Texas 75083, and Department of
Radiology, Rogers Magnetic Resonance Center, University of Texas Southwestern Medical Center,
5801 Forest Park Road, Dallas, Texas 75290
ReceiVed March 16, 2001
Three derivatives of 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrakis(methylene phosphonic acid) (DOTP) contain-
ing a hydrophobic substituent on one side chain were prepared and their lanthanide complexes examined by
NMR. The new ligands include 1-(1-octyl-methyl-phosphonic acid)-4,7,10-tris(methylene phosphonic acid)-1,4,7,10-
tetraazacyclododecane (C8-DOTP), 1-(1-undecyl-methyl-phosphonic acid)-4,7,10-tris(methylene phosphonic acid)-
1,4,7,10-tetraazacyclododecane (C11-DOTP), and 1-(1-4-nitro-phenyl-methyl-phosphonic acid)-4,7,10-tris(methylene
phosphonic acid)-1,4,7,10-tetraazacyclododecane (NO2-Ph-DOTP). 1H NMR spectra of the ytterbium(III) complexes
were assigned by using a combination of COSY spectroscopy and a fitting procedure that matches experimental
NMR hyperfine shifts with those estimated from a MMX-derived structure. The analysis showed that a single
isomer is present in solution and that the bulky hydrophobic substituent occupies the less sterically demanding
H6 equatorial position in the YbL5- complexes. Although the YbL5- complexes have lower symmetry due to the
added substituent, the average 1H hyperfine shifts are 5-10% larger in these complexes compared to YbDOTP5-
.
This was magnified further in the hyperfine 23Na NMR shifts of ion-paired sodium ions where the extracellular
Na+ signal in perfused rat hearts displayed a 28% larger hyperfine shift in the presence of Tm(C11-DOTP)5- than
with an equivalent amount of TmDOTP5-
.
Introduction
Notably, TmDOTP5- forms strong ion pairs with Ca2+, Mg2+
,
and other metal cations,20 and this binding competition not only
reduces the efficiency of the SR but also tends to reduce mean
arterial blood pressure (∼20%) when infused into laboratory
rats.16 TmDOTP5- quickly distributes throughout all extra-
cellular space and is filtered by the kidneys with a time constant
of about 12 min.18 This relatively short renal clearance time
requires that the SR is infused continually throughout a 23Na
NMR study to maintain an adequate shift separation between
the intra- and extracellular Na+ resonances.
The thulium complex of DOTP (Chart 1), TmDOTP5-, has
been widely applied as a hyperfine frequency shift reagent (SR)
in resolving the NMR resonances of Na+ in extracellular and
intracellular compartments.1,2 The intrinsic negative charge on
the complex (HTmDOTP4- is the predominant anionic species
at pH 7.43) prevents it from crossing cell membranes, and thus
the complex forms ion-pair complexes with Na+ in all extra-
cellular space. Since it was first introduced in the mid-1980s,
TmDOTP5- has been widely used to separate the intra- and
extracellular 23Na resonances from isolated cells,4-6 perfused
organs,7-14 and intact animals15-19 by 23Na NMR. This SR is
currently considered the best available for in vivo use even
though some of its chemical properties could be improved upon.
(8) Malloy, R. C.; Buster, D. C.; Castro, M. M. C. A.; Geraldes, C. F. G.
C.; Jeffrey, F. M. H.; Sherry, A. D. Magn. Reson. Med. 1990, 15,
33-44.
(9) Butwell, N. B.; Ramasamy, R.; Sherry, A. D.; Malloy, C. R. InVest.
Radiol. 1991, 26, 1079-1082.
* Author to whom correspondence should be sent to either address.
Telephone: 972-883-2907 or 214-648-5877. Fax: 972-883-2925 or 214-
648-5881. E-mail: sherry@utdallas.edu or dean.sherry@utsouthwestern.edu.
† University of Texas at Dallas.
(10) Butwell, N. B.; Ramasamy, R.; Lazar, I.; Sherry, A. D.; Malloy, C.
R. Am. J. Physiol. 1993, 264, H1884-H1889.
(11) Pike, M. M.; Luo, C. S.; Clark, M. D.; Kirk, K.; Kitakaze, M.; Madden,
M. C.; Gragoe, E. J., Jr.; Pohost, G. M. Am. J. Physiol. 1993, 265,
H2017-H2026.
‡ University of Texas Southwestern Medical Center.
(1) Sherry, A. D.; Geraldes, C. F. G. C. Shift Reagents in NMR
Spectroscopy. In Lanthanide Probes in Life, Chemical and Earth
Sciences: Theory and Practice; Bunzli, J.-C. G., Choppin, G. R., Eds.;
Elsevier: Amsterdam, 1989; pp 93-126.
(2) Sherry, A. D. J. Alloys Compd. 1997, 249, 153-157.
(3) Sherry, A. D.; Ren. J.; Huskens, J.; Brucher, E.; Toth, E.; Geraldes,
C. F. C. G.; Castro, M. M. C. A.; Cacheris, W. P. Inorg. Chem. 1996,
35, 4604-4612.
(12) Van Emous, J. G.; Schreur, J. H. M.; Ruigrok, T. J. C.; Van Echteld,
C. J. A. J. Mol. Cell. Cardiol. 1998, 30, 337-348.
(13) Colet, J.-M.; Makos, J. D.; Malloy, C. R.; Sherry, A. D. Magn. Reson.
Med. 1998, 39, 155-159.
(14) Simor, T.; Lorand, T.; Szollosy, A.; Gaszner, B.; Digerness, S. B.;
Elgavish, G. A. NMR Biomed. 1999, 12, 267-274.
(15) Bansal, N.; Germann, M. J.; Lazar, I.; Malloy, C. R.; Sherry, A. D. J.
Magn. Reson. Imaging 1992, 2, 385-391.
(4) Sherry, A. D.; Malloy, C. R.; Jeffrey, F. M. H.; Cacheris, W. P.;
Geraldes, C. F. G. C. J. Magn. Reson. 1988, 76, 528-533.
(5) Ramasamy, R.; Mota de Freitas, D.; Jones, W.; Wezeman, F.; Labotka,
R.; Geraldes, C. F. G. C. Inorg. Chem. 1990, 29, 3979-3985.
(6) Wittenkeller, L.; Mota De Freitas, D.; Geraldes, C. F. G. C.; Tome,
A. J. R. Inorg. Chem. 1992, 31, 1135-1144.
(7) Buster, D. C.; Castro, M. M. C. A.; Geraldes, C. F. G. C.; Malloy, C.
R.; Sherry, A. D.; Siemers, T. C. Magn. Reson. Med. 1990, 15, 25-
32.
(16) Bansal, N.; Germann, M. J.; Seshan, V.; Shires, G. T., III; Malloy, C.
R.; Sherry, A. D. Biochemistry 1993, 32, 5638-5643.
(17) Xia, Z.-F.; Horton, J. W.; Zhao, P.-Y.; Bansal, N.; Babcock, E. E.;
Sherry, A. D.; Malloy, C. R. J. Appl. Physiol. 1994, 76, 1507-1511.
(18) Seshan, V.; Germann, M. J.; Preisig, P.; Malloy, C. R.; Sherry, A.
D.; Bansal, N. Magn. Reson. Med. 1995, 34, 25-31.
(19) Seshan, V.; Sherry, A. D.; Bansal, N. Magn. Reson. Med. 1997, 38,
821-827.
(20) Ren. J.; Sherry, A. D. Inorg. Chim. Acta 1996, 331-341.
10.1021/ic010291s CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/10/2001

New DOTP-like Lanthanide(III) Complexes
Inorganic Chemistry, Vol. 40, No. 26, 2001 6573
Chart 1
Preparation of Lanthanide Complexes. Lanthanide complexes used
in relaxometric and Na+ shift experiments were prepared by mixing
stoichiometic amounts of lanthanide chloride stock solutions with
5-10% excess ligand. Complexes prepared for high-resolution NMR
studies (1H, 13C, 31P, 1H-1H COSY) were prepared by reacting a ligand
with 10-15% excess lanthanide chloride. The ligands were suspended
in water and were solubilized by adding either tetramethylammonium
hydroxide (for the Tm3+ complex) or sodium hydroxide (for the other
lanthanide complexes) to pH ∼ 8-9. Lanthanide chloride solutions
were added in one portion, and the mixture was heated to ∼65 °C.
The pH of the mixture was maintained at ∼8-9 during complexation.
After the reaction was completed, as judged by stabilization of the
solution pH, the pH of the complex solution was adjusted to the desired
value. When excess lanthanide chloride was used, the mixture was
centrifuged to remove unreacted Ln(OH)3 precipitate before adjusting
the pH. The complexes were then lyophilized to a powdery solid and
redissolved into water or buffers as necessary.
There are three steps one might take to improve the properties
of TmDOTP5-: (1) reduce its affinity for Ca2+ and Mg2+; (2)
increase the magnitude of the 23Na hyperfine shift so that less
SR is required; and (3) reduce its in vivo clearance rate. The
first is difficult to solve because any SR with a reduced affinity
for divalent ions would likely also have reduced affinity for
sodium.21 In this work, we have attempted to address the second
and third approaches using a single ligand modification. It has
been reported that the hyperfine shifts in DOTA-like complexes
increase with increasing ligand rigidity.22,23 This suggested that
introduction of a substituent somewhere in the ligand backbone
might make the complex more rigid and hence magnify the
hyperfine shifts. Second, a hydrophobic substituent could help
to prolong the in vivo lifetime of a SR through noncovalent or
covalent interaction with plasma proteins.
In this work, we report the synthesis and characterization of
new DOTP-like ligands having an extended functional group
on a methylenic carbon of a phosphonate pendant arm. The new
ligands (Chart 1) included 1-(1-octyl-methyl-phosphonic acid)-
4,7,10-tris(methylene phosphonic acid)-1,4,7,10-tetraazacyclo-
dodecane (C8-DOTP), 1-(1-undecyl-methyl-phosphonic acid)-
4,7,10-tris(methylene phosphonic acid)-1,4,7,10-tetraazacyclo-
dodecane (C11-DOTP), and 1-(1-4-nitro-phenyl-methyl-phos-
phonic acid)-4,7,10-tris(methylene phosphonic acid)-1,4,7,10-
tetraazacyclododecane (NO2-Ph-DOTP). The latter derivative
is particularly interesting since the NO2 substituent on the phenyl
group can be reduced to NH2 and this provides a site for further
modification of the ligand. Paramagnetic lanthanide complexes
of these ligands were prepared and investigated by NMR
spectroscopy, and some preliminary 23Na NMR shifts were
measured in a perfused heart preparation.
NMR Experiments. NMR spectra of all synthetic intermediates,
the final products, and the lanthanide complexes were recorded on a
1
Varian INOVA 500 spectrometer. H, 13C{1H}, 23Na, and 31P spectra
were obtained using a 5 mm tunable broad-band probe at 499.8, 125.7,
1
132.2, and 202.3 MHz, respectively, at 25 °C. H-1H COSY spectra
were acquired using an ID/PFG probe. 1H and 13C NMR chemical shifts
were referenced to either TMS or solvent resonances (organic solvents)
and to tert-butyl alcohol (1H at 1.24 ppm, 13C at 30.29 ppm) in aqueous
solution. 31P NMR chemical shifts were referenced to external 85%
aqueous H3PO4 solution (δ ) 0 ppm) while 23Na NMR chemical shifts
were referenced to external 140 mM NaCl solution in D2O (δ ) 0
ppm).
Heart Perfusions. Male Sprague-Dawley rats weighing 350-400
g were anesthetized in an air/diethyl ether chamber. Hearts were rapidly
excised and placed in ice-cold phosphate-free Krebs-Henseleit (KH)
perfusate. After cannulation of the aorta, perfusion was started at a
constant pressure of 100 cm of water. Hearts were initially perfused
with KH buffer containing 147 mM Na+, 124.6 mM Cl-, 29 mM
HCO3-, 5 mM K+, 1.5 mM Ca2+, 1.2 mM Mg2+, 1.2 mM SO42-, and
10 mM glucose. For Ca2+-free perfusions an identical buffer without
Ca2+ was used. The perfusion apparatus had two buffer chambers to
allow easy switching from one perfusate to another without interruption
of flow. Buffer was recirculated, bubbled continuously with 95% O2-
5% CO2 and maintained at 37 °C. The perfusate was filtered on each
pass through a 45 µm filter. After an initial stabilization period with
KH buffer (15-20 min), the perfusate was switched to Ca2+-free KH
buffer containing 1.8 mM TmDOTP5- and a 23Na spectrum was
recorded. The perfusate was then switched to a second Ca2+-free KH
buffer containing 1.8 mM Tm(C11-DOTP)5-, and a second spectrum
was recorded. Triple quantum filtered (TQF) 23Na NMR spectra were
acquired using the parameters described previously by Seshan.19
Ligand Synthesis. Diethyl (()-(1-Hydroxynonyl)-phosphonate
(1a). Diethyl phosphite (15.338 g, 11.11 mmol) was added to nonyl
aldehyde (15.046 g, 105.78 mmol), followed by the addition of
triethylamine (1.07 g, 10.6 mmol). The mixture was heated at 65 °C
for 26 h. The crude product was subjected to high vacuum at 65 °C
overnight to afford product 1a as a pale yellow oil, which solidified
into an off-white waxlike solid (27.2 g, 91.7%). The product was used
Experimental Section
General. 1,4,7,10-Tetraazacyclododecane (cyclen) and DOTP were
obtained from Macrocyclics (Dallas, TX). BSA (fatty acid free and
regular) was purchased from Sigma and was used without further
purification. Lanthanide chloride stock solutions (∼0.2 M) were
standardized with EDTA (0.005 M) using xylenol orange as the
endpoint indicator in NaAC-HOAC buffer (pH ) 5.2). All other
chemicals and solvents were purchased from commercial sources and
were used without further purification. Silica gel (200-400 mesh, 60
Å) for column chromatography was purchased from Aldrich. TLC was
conducted on Whatman precoated silica gel on polyester plates. All
hydrogenation reactions were carried out in a Parr hydrogenation
apparatus. pH measurements were made at room temperature with a
Fisher Scientific Accumet 925 pH/ion meter and a calibrated Orion
8103 Ross combination pH electrode. Molecular models and molecular
mechanics (MM+) calculations were conducted using HyperChem 5.0
(Hypercube Inc., Gainesville, FL). Elemental analysis was performed
by Galbraith Laboratories (Knoxville, TN).
1
(21) Sink, R. M.; Buster, D. C.; Sherry, A. D. Inorg. Chem. 1990, 29,
3645-3649.
in the next step without further purification. H (CDCl3): δ 4.16 (m,
4H, OCH2), 3.85 (m, 1H, HOCHP(O)), 1.34 (tt, 6H, OCH2CH3), 1.72
(m), 1.62 (b), 1.30 (b), 0.88 (t) (17H, (CH2)7CH3). 13C (CDCl3): δ
67.62 (d, 1C, HOCHP(O)), 62.43 (dd, 2C, OCH2), 16.38 (dd, 2C, OCH2-
CH3), 31.75, 31.25, 29.35, 29.22, 29.15, 25.64, 22.54, 13.97 (8C,
(CH2)7CH3).
(22) Jacques, V.; Gilsoul, D.; Comblin, V.; Desreux, J. F. J. Alloys Compd.
1997, 249, 173-177.
(23) Comblin, V.; Gilsoul, D.; Hermann, M.; Humblet, V.; Jacques, V.;
Mesbahi, M.; Sauvage, C.; Desreux, J. F. Coord. Chem. ReV. 1999,
185-186, 451-470.

6574 Inorganic Chemistry, Vol. 40, No. 26, 2001
Li et al.
Diethyl (()-(1-Hydroxydodecyl)-phosphonate (1b). The title
compound was prepared following the method described above by
mixing dodecyl aldehyde (12.71 g, 68.96 mmol), diethyl phosphite (10.0
g, 72.41 mmol), and triethylamine (0.733 g, 7.24 mmol). The product
was obtained as an off-white solid (17.3 g, 84.6%) and was used in
the next step without further purification. 1H (CDCl3): δ 4.15 (m, 4H,
OCH2), 3.84 (m, 1H, HOCHP(O)), 1.35 (t, 6H, OCH2CH3), 1.71 (m),
1.62 (b), 1.28 (b), 0.88 (t) (23H, (CH2)10CH3). 13C (CDCl3): δ 67.72
(d, 1C, HOCHP(O)), 62.47 (dd, 2C, OCH2), 16.43 (dd, 2C, OCH2CH3),
31.85, 31.29, 29.58, 29.56, 29.54, 29.44, 29.30, 29.12, 25.68 (d), 22.62,
14.04 (11C, (CH2)10CH3).
Ar), 5.99 (d, 1H, TfOCHP(O)), 4.33-4.02 (m, 4H, OCH2), 1.31 (q,
6H, CH3). 13C (CDCl3): δ 148.86, 137.7, 128.66, 124.10 (6C, Ar),
118.26 (q, 1C, CF3), 80.76 (d, 1C, TfOCHP(O)), 64.80 (dd, 2C, OCH2),
16.36 (dd, 2C, CH3).
1-(1-Octyl-methyl-phosphonate)-1,4,7,10-tetraazacyclododec-
ane (3a). Cyclen (23.545 g, 136.65 mmol) was dissolved in CHCl3
(70 mL). The solution was cooled in an ice bath, and compound 2a
(11.271 g, 27.33 mmol) in CHCl3 (50 mL) was added dropwise. After
the addition was completed, the mixture was stirred at room temperature
for 2 days. The solvent was removed under reduced pressure, and the
residue was extracted with diethyl ether (4 × 100 mL). The ether
extracts were combined and washed with H2O (3 × 10 mL). Ether
was separated, dried (Na2SO4), and filtered, and the removal of the
solvent yielded a red-brown oil. The crude product was purified by
silica gel flash column chromatography (80 g, the oil was first soaked
onto 10 g of silica gel) eluting successively with (C2H5)2O/CH3OH
(60:40) and (C2H5)2O/CH3OH/NH4OH (42:50:8) to afford a yellow oil
(5.496 g, 48.3%). The product was dissolved in 2 N NaOH (30 mL)
and extracted with CHCl3 (3 × 50 mL). The CHCl3 extracts were
combined, dried (Na2SO4), filtered, and concentrated to give a yellow
oil (4.817 g, 40.55%). 1H (CDCl3): δ 4.11 (m, 4H, OCH2), 2.97-2.49
(m, b, 16H, NCH2CH2N), 1.32 (tt, 6H, OCH2CH3), 1.67 (m, b), 1.29
(b), 0.87 (t) (17H, (CH2)7CH3). 13C (CDCl3): δ 60.99 (dd, 2C, OCH2),
58.33 (d, 1C, CHP(O)), 49.30 (d), 47.25 (b), 45.84 (b), 45.53 (8C,
NCH2CH2N), 16.43 (dd, 2C, OCH2CH3), 31.71, 29.34, 29.19, 29.08,
27.92 (d), 27.39 (d), 22.52, 13.93 (8C, (CH2)7CH3).
Diethyl (()-(1-Hydroxy-1-(4-nitro-phenyl)-methyl)-phosphonate
(1c). The title compound was prepared following a method similar to
that described above by mixing 4-nitrobenzyl aldehyde (20.017 g,
132.458 mmol), diethyl phosphite (19.207 g, 139.08 mmol), and
triethylamine (1.407 g, 13.9 mmol) at 70 °C for 38 h. The crude product
was dissolved in CH2Cl2 (250 mL). Diethyl ether (200 mL) was added,
and the mixture was filtered to remove unreacted aldehyde. The removal
of the solvent gave the compound 1c as a brown solid (35.42 g,
92.50%). The product was used in the next step without further
1
purification. H (CDCl3): δ 8.21, 8.18, 7.69 (d), 7.67 (d) (4H, Ar),
5.19 (d, HOCHP(O)), 4.58 (b, OH), 4.12 (m, 4H, OCH2), 1.29 (t, 6H,
CH3). 13C (CDCl3): δ 147.33 (d), 144.65 (d), 127.65 (d), 123.14 (6C,
Ar), 69.86 (d, 1C, (HO)CHP(O)), 63.47 (dd, 2C, OCH2), 16.27 (t, 2C,
CH3).
Triflate of Diethyl (()-(1-Hydroxynonyl)-phosphonate (2a).
Compound 1a (27.2 g, 99.56 mmol) was dissolved in CH2Cl2 (160
mL). Pyridine (10.9 g, 139.4 mmol) was added, and the mixture was
cooled in liquid nitrogen to near freezing. Triflic anhydride (32.3 g,
114.5 mmol) was added dropwise with stirring. The mixture was
warmed to 0 °C in an ice bath for 2 h. The solvent was removed, and
the residue was extracted with diethyl ether (4 × 50 mL). The ether
extracts were filtered, washed successively with cold H2O (2 × 30 mL)
and 0.5 N HCl (2 × 30 mL), dried (Na2SO4), and filtered. The solvent
was removed under reduced pressure to afford a red-brown oil (34.5
g, 84.1%). The product was used in the next step without further
purification. 1H (CDCl3): δ 5.0 (m, 1H, CHOTf), 4.23 (m, 4H, OCH2),
1.38 (t, 6H, OCH2CH3), 2.02 (b, m), 1.28 (b), 0.88 (t) (17H, (CH2)7CH3).
13C (CDCl3): δ 118.25 (q, 1C, CF3), 82.02 (d, 1C, CHOTf), 63.54
(dd, 2C, OCH2), 16.16 (dd, 2C, OCH2CH3), 31.62, 30.15, 28.97, 28.94,
28.82, 24.90 (d), 24.47, 13.88 (8C, (CH2)7CH3).
Triflate of Diethyl (()-(1-Hydroxydodecyl)-phosphonate (2b).
Compound 1b (10.0 g, 31.02 mol) was dissolved in CH2Cl2 (100 mL).
Pyridine (3.9 g, 50.0 mmol) was added, and the mixture was cooled in
liquid nitrogen to near freezing. Triflic anhydride (10.06 g, 35.67 mmol)
was added dropwise with stirring. The reaction was warmed to 0 °C in
an ice bath for 2.5 h. The solvent was removed under reduced pressure,
and the residue was extracted with pentane (4 × 50 mL). The pentane
extracts were filtered, washed successively with cold H2O (2 × 20 mL)
and 0.5 N HCl (1 × 20 mL), dried (Na2SO4), and filtered. Pentane
was removed under reduced pressure to yield a red-orange oil (12.56
g, 89.1%). The product was used in the next step without further
purification. 1H (CDCl3): δ 4.99 (m, 1H, CHOTf), 4.23 (m, 4H, OCH2),
1.38 (t, 6H, OCH2CH3), 2.04 (m), 1.56 (b), 1.46 (b), 1.29 (b), 1.26 (b),
0.88 (t) (23H, (CH2)10CH3). 13C (CDCl3): δ 118.36 (q, 1C, CF3), 82.11
(d, 1C, CHOTf), 63.70 (dd, 2C, OCH2), 16.26 (dd, 2C, OCH2CH3),
31.84, 30.22, 29.50, 29.48, 29.35, 29.25, 29.09, 28.90, 24.98 (d), 22.61,
14.02 (11C, (CH2)10CH3).
1-(1-Undecyl-methyl-phosphonate)-1,4,7,10-tetraazacyclododec-
ane (3b). Cyclen (9.5 g, 55.0 mmol) was dissolved in CHCl3 (75 mL),
and the solution was cooled in an ice bath. Compound 2b (12.5 g,
27.5 mmol) in CHCl3 (75 mL) was added dropwise. After the addition,
the mixture was stirred at room temperature for 3 days. The solvent
was removed under reduced pressure, and the residue was extracted
with diethyl ether (4 × 100 mL). The ether extracts were combined
and washed with H2O (5 × 4 mL). Ether was separated, and the aqueous
extracts were combined and extracted with diethyl ether (5 × 100 mL).
The ether extracts from the above two processes were combined, dried
(Na2SO4), and filtered, and the removal of ether gave a red-brown oil
(11.964 g, 91%). Purification by silica gel flash column chromatography
(120 g, the crude product was first soaked onto 20 g of silica gel) eluting
successively with (C2H5)2O/CH3OH (70:30), (C2H5)2O/CH3OH/NH4OH
(55:40:5) yielded a yellow oil. The product was dissolved in 3 N NaOH
(20 mL) and extracted with CHCl3 (3 × 50 mL). The CHCl3 was
separated, combined, dried (Na2SO4), filtered, and concentrated to afford
1
a yellow oil (8.01 g, 60.93%). H (CDCl3): δ 4.11 (m, 4H, OCH2),
2.97-2.50 (m, b, 16H, NCH2CH2N), 1.32 (t, 6H, OCH2CH3), 1.68 (m),
1.26 (b), 0.88 (t) (23H, (CH2)10CH3). 13C (CDCl3): δ 61.0 (dd, 2C,
OCH2), 58.29 (d, 1C, CHP(O)), 16.45 (dd, 2C, OCH2CH3), 49.34, 47.31
(b), 45.88, 45.60 (8C, NCH2CH2N), 31.74, 29.52, 29.46, 29.42, 29.19,
27.98 (d), 27.40 (d), 22.51, 13.95 (11C, (CH2)10CH3).
1-(1-4-Nitro-phenyl-methyl phosphonate)-1,4,7,10-tetraazacyclo-
dodecane (3c). Cyclen (2.16 g, 12.5 mmol) was dissolved in CHCl3
(15 mL), to which compound 2c (4.8 g, 11.4 mmol) in CHCl3 (25 mL)
was added dropwise over 4 h. After 5 days the solvent was removed,
acetone (150 mL) was added, and the unreacted cyclen was filtered
off. While the solvent was being removed, the product started to
precipitate. The crude product was dissolved in H2O/acetone and soaked
onto ∼10 g of silica gel, and the solvent was removed by rotary
evaporation. The silica gel was loaded on top of another 65 g of silica
gel. The column was eluted successively with (C2H5)2O/CH3OH (60:
40) and (C2H5)2O/CH3OH/NH4OH (55:40:5). Fractions containing the
product were combined, and the removal of the solvent afforded product
Triflate of Diethyl (()-(1-Hydroxy-1-(4-nitro-phenyl)methyl)-
phosphonate (2c). Compound 1c (10.0 g, 34.6 mmol) was dissolved
in CH2Cl2 (100 mL). Pyridine (3.78 g, 48.4 mmol) was added, and the
mixture was cooled in liquid nitrogen to near freezing. Triflic anhydride
(11.2 g, 39.8 mmol) was added dropwise with stirring. The mixture
was warmed to 0 °C in an ice bath for 80 min. The solvent was removed
under reduced pressure, and the residue was extracted with diethyl ether
(3 × 50 mL). The ether extracts were filtered, washed with cold H2O
(2 × 20 mL) and cold 0.5 N HCl (2 × 20 mL), dried (Na2SO4), and
filtered. The solvent was removed, and the crude product was purified
by silica gel flash column chromatography (250 g) eluting successively
with CH2Cl2 and CH2Cl2/(C2H5)2O (4:1) to afford a light yellow solid
1
3c as a light yellow solid (4.6 g, 91%). H (CDCl3): δ 8.23, 8.22,
7.76, 7.74 (4H, Ar), 4.30 (d, 1H, CHP(O)), 4.25 (m), 4.0 (m), 3.87
(m) (4H, OCH2), 3.25 (b), 2.80 (b), 2.66 (b), 2.52 (b, m) (16H, NCH2-
CH2N), 1.38 (t, 3H, CH3), 1.10 (t, 3H, CH3). 13C (CDCl3): δ 147.45,
140.89 (d), 131.24 (d), 123.26 (6C, Ar), 62.34 (dd, 2C, OCH2), 60.79
(d, 1C, CHP(O)), 49.06 (b), 47.20, 45.98, 45.06 (8C, CH2CH2N), 16.3
(dd, 2C, CH3).
1-(1-Octyl-methyl-phosphonate)-4,7,10-tris(methylene phosphon-
ate)-1,4,7,10-tetraazacyclododecane (4a). To compound 3a (4.817 g,
11.083 mmol) were added paraformaldehyde (1.166 g, 36.573 mmol)
1
(7.07 g, 48.52%). H (CDCl3): δ 8.34, 8.29, 7.75 (d), 7.71 (d) (4H,

New DOTP-like Lanthanide(III) Complexes
Inorganic Chemistry, Vol. 40, No. 26, 2001 6575
and triethyl phosphite (6.353 g, 38.263 mmol). The mixture was stirred
at room temperature for 4 days and then subjected to high vacuum
first at room temperature and again at 65 °C overnight to remove excess
reactants and volatile impurities. The resultant brown oil was purified
by silica gel flash column chromatography (200 g, the crude product
was first soaked onto 15 g of silica gel) eluting successively with
(C2H5)2O/CH3OH (70:30) and (C2H5)2O/CH3OH/NH4OH (65:30:5) to
give a yellow oil (4.673 g, 47.65%). 1H (CDCl3): δ 4.18 (m, 4H,
OCH2), 4.11 (m, 12H, OCH2), 3.91 (d, 1H, CHP(O)), 2.97 (d, 2H,
CH2P(O)), 2.94 (d, 4H, CH2P(O)), 3.01-2.83 (m, b, 16H, NCH2CH2N),
1.33(t, 24H, OCH2CH3), 1.61 (m, b), 1.45 (b), 1.35 (t), 1.27 (b), 0.88
(t) (17H, (CH2)7CH3). 13C (CDCl3): δ 61,45 (m, 6C, OCH2), 60.87 (d,
2C, OCH2), 60.19 (d, 1C, CHP(O)), 50.17 (d, 1C, CH2P(O)), 50.06 (d,
2C, CH2P(O)), 54.55 (d), 53.16 (d), 52.73(b), 50.25(b) (8C, NCH2-
CH2N), 16.43 (m, 8C, OCH2CH3), 31.78, 29.51, 29.44 (d), 29.26, 27.27
(d), 22.57, 14.0 (8C, (CH2)7CH3).
1-(1-Undecyl-methyl-phosphonate)-4,7,10-tris(methylene phos-
phonate)-1,4,7,10-tetraazacyclododecane (4b). The title compound
was prepared following the method described above by mixing
compound 3b (8.0 g, 16.75 mmol), paraformaldehyde (1.682 g, 55.53
mmol), and triethyl phosphite (9.184 g, 55.27 mmol). Purification with
silica gel flash column chromatography (150 g) eluting with (C2H5)2O/
CH3OH (60:40) afforded compound 4b as a brown oil (11.795 g,
75.7%). 1H (CDCl3): δ 4.21-4.07 (m, 16H, OCH2), 3.91 (d, 1H,
CHP(O)), 2.97 (d, 2H, CH2P(O)), 2.94 (d, 4H, CH2P(O)), 3.02-2.82
(m, 16H, NCH2CH2N), 1.33 (t, 24H, OCH2CH3), 1.61 (m, b), 1.45 (b),
1.35 (t), 1.26 (b), 0.88 (t) (23H, (CH2)10CH3). 13C (CDCl3): δ 61.39
(m, 6C, OCH2), 60.80 (d, 2C, OCH2), 60.16 (d, 1C, CHP(O)), 50.16
(d, 1C, CH2P(O)), 50.04 (d, 2C, CH2P(O)), 54.53 (d), 53.15 (d), 52.72
(d), 50.22 (b) (8C, NCH2CH2N), 31.79, 29.61, 29.56, 29.53, 29.47,
29.23, 27.91 (d), 27.26 (d), 25.55, 13.99 (11C, (CH2)11CH3), 16.46 (m,
8C, OCH2CH3).
1-(1-4-Nitro-phenylmethyl phosphonate)-4,7,10-tris(methylene
phosphonate)-1,4,7,10-tetraazacyclododecane (4c). The title com-
pound was prepared following the method described above by mixing
compound 3c (3.0 g, 6.76 mmol), paraformaldehyde (0.679 g, 7.102
mmol), and triethyl phosphite (3.709 g, 7.441 mmol). Purification with
a silica gel flash column (70 g, C2H5)2O/CH3OH (4:1)) afforded the
product 4c as a brown oil (3.504 g, 58.0%). 1H (CDCl3): δ 8.21, 8.19,
7.86, 7.84 (4H, Ar), 4.36 (d, CHP(O)), 4.27-4.17 (m, 6H, OCH2),
4.15-4.05 (m, 18H, OCH2), 3.23 (b, m), 2.99-2.76 (m), 2.65 (m, b)
(16H, NCH2CH2N), 1.37-1.28 (m, 24H, CH3). 13C (CDCl3): δ 147.23,
142.30 (d), 131.31 (d), 122.97 (6C, Ar), 61.16 (d, 1C, CHP(O)), 62.26
(d, 2C, OCH2), 61.45 (m, 6C, OCH2), 50.24 (d, 2C, CH2P(O)), 50.2
(d, 1C, CH2P(O)), 53.72 (d), 53.54 (d), 53.39 (d), 49.64 (8C, NCH2-
CH2N), 16.82 (d, 8C, CH3).
stirring. After the addition, diethyl ether (100 mL) was added. The
product was filtered, washed with diethyl ether (4 × 20 mL), and dried
under vacuum to give compound 5b as a pale yellow solid (7.6 g,
84.8%). A sample for elemental analysis was prepared by reprecipitating
the product from base by addition of acid. 1H NMR (D2O, pH ) 7.4):
δ 3.45-2.78 (b, m, 16H, NCH2CH2N), 1.63 (b), 1.44 (b), 1.30 (b),
1.21 (b), 1.17 (b), 0.75 (t) (23H, (CH2)10CH3). 13C (D2O, pH ) 7.4):
δ 58.35 (d, b, 1C, CHP(O)), 52.41 (d, 1C, CH2P(O)), 52.03 (d, b, 2C,
CH2P(O)), 53.2 (b), 51.4 (b), 49.62 (b), 47.48 (b), 45.52 (b) (8C, NCH2-
CH2N), 31.85, 29.97, 29.48, 29.44, 29.34, 29.18, 28.64 (d), 25.47, 22.70,
14.1 (11C, (CH2)10CH3). Anal. Calcd for C23H54N4O12P4‚2.56H2O: C,
36.87; H, 7.90; N, 7.48. Found: C, 36.97; H, 8.14; N, 7.49.
1-(1-4-Nitro-phenyl-methyl-phosphonic acid)-4,7,10-tris(methyl-
ene phosphonic acid)-1,4,7,10-tetraazacyclododecane (5c, NO2-Ph-
DOTP). The title compound was prepared following a method similar
to that described above by dissolving compound 4c (3.504 g, 3.92
mmol) in 30 wt % hydrogen bromide in acetic acid (35 mL). The crude
product was recrystallized in a mixture of H2O (6 mL), ethyl alcohol
(45 mL), and diethyl ether (30 mL). The precipitate was filtered, washed
with diethyl ether, and dried under vacuum to give 5c as a pale yellow
powder (2.30 g, 87.7%). The product was further purified by acidifying
1
the basic solution of the ligand. H (D2O, pH ) 7.4): δ 8.17, 8.15,
7.67, 7.66 (4H, Ar), 4.23-2.51 (b, m, 16H, NCH2CH2N), 4.08 (d, 1H,
CHP(O)). 13C (D2O, pH ) 7.4): δ 149.42, 145.28, 133.98 (d), 125.82
(6C, Ar), 64.07 (d, 1C, CHP(O)), 54.29 (d, 2C, CH2P(O)), 53.31 (d,
1C, CH2P(O)), 54.60, 52.75, 52.11, 51.0, 50.94, 48.51, 47.28 (8C,
NCH2CH2N). Anal. Calcd for C18H35N5O14P4‚6.2H2O: C, 27.65; H,
6.07; N, 8.96. Found: C, 27.64; H, 5.80; N, 9.01.
Results and Discussion
Ligand Synthesis. The synthetic route for preparation of the
ligands is shown in Scheme 1. The method of Kluge24 was used
to prepare the alcohol phosphonates, 1a-1c. The reaction
proceeded by a Pudovik mechanism in which triethylamine
catalyzed the reaction by deprotonating diethyl phosphite.25 All
attempts to purify the product by the standard distillation
techniques were accompanied by extensive decomposition. No
attempt was made to separate the enantiomers. Conversion of
1a-1c into the corresponding triflates 2a-2c was accomplished
using a literature method.26
Reactions of 2a, 2b with cyclen to make the monoalkylated
cyclens 3a, 3b were complicated by competitive elimination of
the triflates to form diethyl alkenephosphonates. The elimination
reaction could be suppressed by using excess cyclen but could
not be totally eliminated even with a 500% excess of cyclen.
The monoalkylated products were purified from cyclen on a
silica column.
3a-3c were converted to 4a-4c using paraformaldehyde and
triethyl phosphite in a Mannich-type reaction. Conversion of
phosphonates 4a-4c into phosphonic acids 5a-5c was achieved
in 30 wt % HBr/glacial acetic acid (hydrolysis in 6 N
hydrochloric acid cleaved the pendant arm to give DO3P). The
ligands were purified to >95% by the recrystallization of the
hydrolysis products in a mixture of ethanol and water. 31P NMR
spectra of recrystallized 5a-5c showed that a small amount of
DO3P (<5%) remained. These products could be further purified
by dissolving in aqueous base followed by reacidification, but
this resulted in substantially reduced yields (∼50%) for 5a, 5b.
The yield was higher for ligand 5c (66.5%).
1-(1-Octyl-methyl-phosphonic acid)-4,7,10-tris(methylene phos-
phonic acid)-1,4,7,10-tetraazacyclododecane (5a, C8-DOTP). A
portion of compound 4a (4.027 g, 4.551 mmol) was dissolved in a 30
wt % hydrogen bromide solution in acetic acid (40 mL). The solution
was stirred at room temperature for 40 h. The solvent and excess
hydrogen bromide were removed under reduced pressure. The crude
product was purified by recrystallization in ethanol (75 mL) and H2O
(15 mL) mixture. The precipitate was filtered, washed with diethyl ether,
and dried under vacuum to give the product 5a as an off-white solid
(2.71 g, 67.3%). A sample for elemental analysis was prepared by
1
reprecipitating the product from base by addition of acid. H NMR
(D2O, pH ) 7.4) δ 3.47-2.79 (b, m, 16H, NCH2CH2N), 1.63 (b), 1.43
(b), 1.28 (b), 1.16 (b), 0.74 (b) (17H, (CH2)7CH3). 13C (D2O, pH )
7.4) δ 58.37 (d, b, 1C, CHP(O)), 52.42 (d, 1C, CH2P(O)), 52.05 (d, b,
2C, CH2P(O)), 52.9 (b), 51.8 (b), 49.65 (b), 47.58 (b) (8C, NCH2CH2N),
31.83, 29.93, 29.28, 29.14, 28.58, 25.47, 22.68, 14.11 (8C, (CH2)7CH3).
Anal. Calcd for C20H48N4O12P4‚2.84H2O: C, 33.72; H, 7.54; N, 7.87.
Found: C, 34.0; H, 7.62; N, 7.84.
1-(1-Undecyl-methyl-phosphonic acid)-4,7,10-tris(methylene phos-
phonic acid)-1,4,7,10-tetraazacyclododecane (5b, C11-DOTP). The
title compound was prepared following a method similar to that
described above by dissolving compound 4b (11.795 g, 12.67 mmol)
in a 30 wt % hydrogen bromide solution in acetic acid (80 mL). The
crude product was dissolved in ethyl alcohol (150 mL). H2O (20 mL)
was added dropwise to the above solution to induce precipitation while
NO2-Ph-DOTP (5c) was reduced to NH2-Ph-DOTP in a Parr
hydrogenation apparatus using palladium/activated carbon as
catalyst in H2O. A 1H NMR spectrum of the product indicated
(24) Kluge, A. F.; Cloudsdale, I. S. J. Org. Chem. 1979, 44, 4847-4851.
(25) Engel, R. Synthesis of carbon-phosphorus bonds; CRC Press Inc.:
Boca Raton, FL, 1988; p 101.
(26) Xavier, C.; Underiner, T. L. J. Org. Chem. 1985, 50, 2165-2170.

6576 Inorganic Chemistry, Vol. 40, No. 26, 2001
Li et al.
Scheme 1
by the different H-P coupling constants. The phosphonate
pendant arm trans to the chiral center gave rise to a quartet (a
2
doublet of doublets) at 17.4 ppm (2JH(1)P ) 14.6 Hz, JH(2)P
)
7.3 Hz). The other two phosphonate pendant arms trans to each
other gave rise to the pseudotriplet at 7.2 ppm (2JHP(1) ≈ JHP(2)
2
) 9.8 Hz) at 80 °C.
NMR Spectra. 1H NMR spectra of several ytterbium
complexes are compared in Figure 2. The spectra of Yb(C8-
DOTP)5- and Yb(C11-DOTP)5- had nearly identical chemical
shifts although the resonances of the later complex were
somewhat broader apparently due to slower tumbling of the
larger chelate. The spectra of Yb(NO2-Ph-DOTP)5- and Yb(NH2-
Ph-DOTP)5- were similar as well.
The 1H NMR spectrum of YbDOTP5- shows one set of H1-
H6 proton resonances due to C4 symmetry. The assignments
were reported previously.27 The single substitution on one
pendant arm of the DOTP-like ligands reported here lowers
symmetry of the complex, resulting in 4 sets of H1-H5
resonances (this is most evident for the most highly shifted H4
resonances) but only 3 H6 resonances. This indicates that the
single R substituent in these complexes occupies a normal H6
equatorial position. The H1, H4, H5, and H6 protons were easily
identified by comparison of their chemical shifts with the
reference compound YbDOTP5-. The H2 and H3 protons are
less highly shifted and cannot be identified by simple inspection.
Like the LnDOTP5- and LnDOTBP- complexes,27,28 the DOTP-
like complexes shown here exist as a single coordination isomer,
likely in a twisted square antiprismatic geometry.3
Further assignments of the proton resonances came from
COSY spectra (the spectrum of Yb(C8-DOTP)5- is illustrated
in Figure 3). The geminal (H3-H4 and H1-H2) and vicinal (H1-
H4, H2-H3) cross peaks established the correlations necessary
to assign each of the individual ethylene fragments within the
macrocyclic backbone. Only three H5-H6 correlation peaks
were observed, further confirming the absence of the fourth H6
due to the substitution by the hydrophobic side group. From
the cross peaks in the COSY spectrum, the connectivity
information between individual H1-H4 groups and between
individual H5-H6 groups was determined. Nonetheless, the H1-
H4 and H5-H6 groupings cannot be provided by COSY due to
the lack of the J couplings between the protons in these two
groups. A complete assignment of all protons was accomplished
by the observed hyperfine shifts with those calculated from a
MMX-derived model of these complexes.
that the nitro group was completely reduced but that a second
product was also present. 31P NMR spectra collected as a
function of time showed that NH2-Ph-DOTP gradually decom-
posed to DO3P in aqueous solution. A subsequent experiment
revealed that Ln3+ complexes of 5c could be reduced to the
corresponding Ln(NH2-Ph-DOTP)5- complex without further
degradation.
1H Hyperfine Shift Analysis. The observed chemical shift
(∆obs) of a ligand nucleus in a paramagnetic lanthanide complex
can be expressed by
31P NMR of the Ligands. Figure 1 shows temperature-
variable 31P NMR spectra of ligands 5b, 5c in D2O at pH )
7.4 without proton decoupling (spectra of 5a were nearly
identical to those of 5b). For 5a and 5b three phosphorus peaks
were observed with an area integration ratio of 1:1:2. The broad
peaks of ligand 5b indicate that the molecular tumbling rate is
slowed by the long aliphatic chain substituent. The peaks
sharpened at higher temperatures but remained too broad to
reveal H-P couplings. In comparison, the 31P NMR resonances
of ligand 5c were much sharper, likely due to faster rotational
averaging. Four phosphorus peaks were observed for 5c at 25
°C with a 1:1:1:1 area integration ratio. At higher temperature,
two of the four 31P peaks coalesced to give three phosphorus
peaks with a 1:1:2 area ratio. The spectrum at 80 °C allows the
observation of the full H-P coupling pattern. The doublet at
14.4 ppm (2JHP ) 25.3 Hz) could be directly assigned to the
pendant arm with the nitrophenyl substituent. The chiral center
at this pendant arm caused the nonequivalency of the two
methylene protons at the other three pendant arms as reflected
∆obs ) ∆p + ∆dia
(1)
where ∆p is the paramagnetic shift contribution and ∆dia the
diamagnetic shift contribution. The ∆dia term represents the
inductive and conformational effects that occur upon complex-
ation; this term is normally evaluated from the chemical shift
difference between free ligand and that of diamagnetic lantha-
num or lutetium complexes. In this study we did not attempt to
1
assign all H chemical shifts in spectra of the La3+ or Lu3+
complexes due to the complexity of those spectra, so a ∆dia
value of 3 ppm, as was reported previously,29 was used to
(27) Geraldes, C. F. G. C.; Sherry, A. D.; Kieffer, G. J. Magn. Reson.
1992, 97, 290-304.
(28) Aime, S.; Batsanov, S.; Botta, M.; Howard, J. A. K.; Parker, D.;
Senanayake, K.; Williams, G. Inorg. Chem. 1994, 33, 4696-4706.
(29) Zhang, S.; Kovacs, Z.; Burgess, S.; Aime, S.; Terreno, E.; Sherry, A.
D. Chem. Eur. J. 2001, 7, 288-296.

New DOTP-like Lanthanide(III) Complexes
Inorganic Chemistry, Vol. 40, No. 26, 2001 6577
Figure 1. Temperature-variable 31P NMR spectra of (a) C11-DOTP and (b) NO2-Ph-DOTP in D2O at pH 7.4. The spectra were obtained without
P{H} decoupling. C11-DOTP was obtained by acid purification.
1
Figure 2. High-resolution H NMR spectra of YbDOTP5- (pH ) 9.0) and the YbL5- complexes at pH ) 10.0 and 25 °C.
represent the average of 1H shifts in the diamagnetic complexes.
The paramagnetic shift contribution (∆p) can be further divided
in origin and can thereby be correlated to complex structure
by30
into contact (∆p-c) and pseudocontact components (∆p-pc
)
depending on the nature of the electron-nucleus interactions
between the lanthanide ion and the observed nucleus. It is
generally accepted that the paramagnetic shifts observed in the
1H spectrum of an ytterbium complex are largely pseudocontact
∆
p-pc ) D1[(3 cos2 θ - 1)/r3] + D2[(sin2 θ cos2 æ)/r3] (2)
where D1 and D2 are proportional to the magnetic susceptibility
differences, øzz - 1/3(øxx + øyy + øzz), and øxx - øyy, respectively.

6578 Inorganic Chemistry, Vol. 40, No. 26, 2001
Li et al.
Table 1. Summary of Axial (D1) and Nonaxial (D2) Susceptibility
Terms Obtained by Analysis of Proton Hyperfine NMR Shifts of
the Yb3+ Complexes
Yb3+ complex
D1
D2
DOTP
2931 ( 54
3101 ( 101
3187 ( 135
3021 ( 108
7.2 ( 54
C8-DOTP
C11-DOTP
NO2-Ph-DOTP
460 ( 104
403 ( 139
141 ( 112
the remaining atomic lengths and bond angles of the molecule
to vary. The Yb3+-N (2.63 Å) and Yb3+-O (2.26 Å) distances
were fixed to those reported in the crystal structure of
TmDOTP5- 32
.
Figure 4 illustrates the MMX-minimized structure of Yb(C8-
DOTP)5-. The H1-H6 protons were divided into four subgroups
as indicated by superscripts. The COSY spectrum reports which
H1-H4 protons are in the same subgroup, but does not identify
the subgroup number. Furthermore, the H5-H6 can be identified
on the basis of the magnitude of their ∆obs values (in comparison
to YbDOTP5-) but cannot be correlated to a particular H1-H4
Figure 3. 1H-1H COSY spectrum of Yb(C8-DOTP)5- at 0.223 M in
D2O at pH ) 10.0 and 25 °C. For clarity, only one H1-H4 connectivity
was drawn.
1
subgroup (with the exception of H5 , which can be assigned
from the COSY spectrum). Therefore, the coordinates of all
protons and ytterbium (from MMX) and the ∆p,exp ()∆obs - 3
ppm) values were compared using Forsberg’s method, allowing
reassignment of all H1-H6 subgroups (while satisfying all the
proton connectivities established by the COSY spectrum) until
a best agreement between ∆p,exp and ∆p,cal for all protons was
found. The agreement between calculated and observed hyper-
fine shifts was good for all three complexes (R2 < 1%, see
Supporting Information). The numerical values of D1 and D2
found by this analysis are summarized in Table 1. The nonaxial
magnetic susceptibility terms, D2, were larger for the asymmetric
YbL5- complexes, as expected, while the symmetric D1 terms
for the new complexes were slightly higher (3-9%) than the
value found for YbDOTP5-. Interestingly, the D1 values
θ, æ, and r are the polar coordinates of the ligand nucleus with
Ln(III) at the origin and with the principal magnetic axis of the
system as the z axis.
1
An analysis of H hyperfine shifts was carried out using a
method described by Forsberg.31 First, a molecular model of
the chelate with Yb3+ bound to four macrocyclic nitrogens and
four phosphonic oxygens was built using HyperChem 5.0. It
was assumed that Yb3+ does not have an inner-sphere water
molecule as reported for the LnDOTP5- complexes.3 The
structure of the molecule was then optimized by molecular
mechanics (MMX), holding the Yb3+-N and Yb3+-O bond
distances fixed during energy minimization while allowing all
Figure 4. (a) MMX-minimized molecular model of Yb(C8-DOTP)5-. H1-H6 protons were divided into four subgroups, and the superscript denotes
the subgroup number. (b) Tentative assignment of the protons of the Yb(C8-DOTP)5- spectrum, a result of the combination of 2D COSY spectrum
and the proton shift analysis.

New DOTP-like Lanthanide(III) Complexes
Inorganic Chemistry, Vol. 40, No. 26, 2001 6579
are located within the same dipolar shift cone but ∼180° apart20).
To test this, sodium hyperfine shifts were measured as a function
of TmL5- concentration in aqueous buffer and compared to
those measured for an equivalent amount of TmDOTP5- (Figure
5). All 23Na hyperfine shifts increased linearly with SR
concentration, with the slopes falling in the order Tm(C11-
DOTP)5- (3.6) > Tm(C8-DOTP)5- (3.5) > Tm(NO2-Ph-
DOTP)5- (3.2) > TmDOTP5- (3.1). This demonstrates that the
Na+ shifting ability correlates with the magnitude of the
hyperfine shifts observed for the H4 protons and implies that
the number of Na+ binding sites and binding constants is similar
in the new TmL5- complexes as compared to TmDOTP5-. Since
Ca2+ reduces the efficiency of TmDOTP5- as a SR by direct
competition with Na+ for sites on the SR,20 we also examined
the effects of adding Ca2+ to the new SRs. Nearly identical
decreases in 23Na shift were observed when each of the new
SRs and TmDOTP5- were titrated with Ca2+ (data not shown).
This suggests that, like Na+, Ca2+ has a similar binding affinity
for the new TmDOTP-like systems. Tm(C8-DOTP)5- and
Tm(C11-DOTP)5- also precipitated from solution upon addition
of Ca2+ at a much lower concentration than required for
precipitation of TmDOTP5-, consistent with a lower water
solubility for the Ca:TmL5- ternary complexes.
Triple quantum filtered (TQF) 23Na spectra from a single rat
heart perfused with two different SRs are also shown in Figure
5. 23Na NMR signals were collected first in the presence of
TmDOTP5- and then, after a brief washout period, collected
again in the presence of Tm(C11DOTP)5-. Interestingly, the
larger 23Na hyperfine shift provided by Tm(C11-DOTP)5- in
aqueous solution appears to be magnified in the tissue experi-
ment (compare 3.6 versus 3.1, a 16% difference in H2O with
3.6 versus 2.8, a 29% difference in the tissue preparation). The
origin of these differences is unknown at present but could
reflect an accumulation of the hydrophobic reagent in hydro-
phobic tissue regions thereby providing a higher “effective”
concentration of this SR in the extracellular space. Similar results
were obtained for hearts perfused first with TmDOTP5-
followed by Tm(C8-DOTP)5- (data not shown). Although the
origins of the enhanced hyperfine shifts in perfused tissue versus
aqueous solution are not known and will require further
experimentation, the nearly 30% increase in Na+ hyperfine shift
observed for these systems should prove advantageous in future
in vivo 23Na NMR experiments.
Figure 5. (a) 23Na shift as a function of (Me4N)4HTmL in aqueous
buffer: (0) (Me4N)4HTm(C8-DOTP), (4) (Me4N)4HTm(C11-DOTP),
(O) (Me4N)4HTm(NO2-Ph-DOTP); (]) (Me4N)4HTmDOTP. The lines
correspond to the linear regression of the experimental data. Buffer:
118 mM NaCl, 25 mM NaHCO3, 5 mM KCl, 10 mM glucose, pH )
9.0. (b) Triple quantum filtered (TQF) 23Na spectra collected from an
in vitro rat heart perfusion experiment first with TmDOTP5- and then
with Tm(C11-DOTP)5- as the shift reagent.
correlated well with the hyperfine shifts (average of four values)
of the H4 protons of these complexes, suggesting that the
nonaxial D2 term contributes little to the observed hyperfine
1H shifts.
23Na NMR. As the hyperfine shifts of the H4 protons were
consistently larger in the new YbL5- complexes compared to
YbDOTP5- (the average downfield increases were 2.9, 5.2, and
0.1 ppm for Yb(C8-DOTP)5-, Yb(C11-DOTP)5-, and Yb(NO2-
Ph-DOTP)5-, respectively), we expected that the hyperfine shifts
for ion-paired Na+ ions might also be larger because both nuclei
are positioned near the highest-fold symmetry axis of the
complex (the H4 protons and the presumptive Na+ binding sites
Acknowledgment. This research was supported in part by
grants from the Robert A. Welch Foundation (AT-584) and the
National Institutes of Health (HL-34557 and RR-02584).
Supporting Information Available: Tables S1 and S2 summarizing
1
observed and calculated H hyperfine shifts for Yb3+ complexes of
DOTP, C8-DOTP, C11-DOTP, and NO2-Ph-DOTP. This material is
available free of charge via the Internet at http://pubs.acs.org.
IC010291S
(30) Peters, J. A.; Huskens, J.; Raber, D. J. Prog. Nucl. Magn. Reson.
Spectrosc. 1996, 28, 283-350.
(31) Forsberg, J. H.; Delaney, R. M.; Zhao, Q.; Harakas, G.; Chandran, R.
Inorg. Chem. 1995, 34, 3705-3715.
(32) Paulus, E. F.; Juretschke, P.; Long, J. 3. Jahrestagung der Deutschen
Gesellschaft fu¨r Kristallografie; Oldenbourg: Darmstadt, 1995.