A second-generation copper(II)-mediated metallo-DNA-base pair

Article abstract of DOI:10.1016/j.bioorg.2003.09.001

Metal-dependent pairing of nucleobases represents an alternative DNA base pairing scheme. Our first-generation copper(II)-mediated pyridine-2,6- dicarboxylate (Dipic) and pyridine (Py) metallo-base pair has a stability comparable to the natural base pairs dA:dT and dC:dG but does not have the selectivity of the Watson Crick base pairs. In order to increase the selectivity of base pair formation, a second-generation metallo-base pair was generated consisting of a pyridine-2,6-dicarboxamide (Dipam) and a pyridine (Py) nucleobase. This new metallo-base pair is more stable than the natural base pairs dA:dT and dC:dG and highly selective against mispairing. In addition, incorporation of multiple metallo-base pairs into DNA results in the formation of stable duplexes demonstrating that hydrogen bonding base pairs can efficiently be replaced by metal-dependent base pairs at multiple sites in DNA.

Full text of DOI:10.1016/j.bioorg.2003.09.001

BIOORGANIC
CHEMISTRY
Bioorganic Chemistry 32 (2004) 13–25
www.elsevier.com/locate/bioorg
A second-generation copper(II)-mediated
metallo-DNA-base pair
Nicole Zimmermann, Eric Meggers, and Peter G. Schultz*
Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road,
La Jolla, CA 92037, USA
Received 1 July 2003
Abstract
Metal-dependent pairing of nucleobases represents an alternative DNA base pairing
scheme. Our first-generation copper(II)-mediated pyridine-2,6-dicarboxylate (Dipic) and pyr-
idine (Py) metallo-base pair has a stability comparable to the natural base pairs dA:dT and
dC:dG but does not have the selectivity of the Watson Crick base pairs. In order to increase
the selectivity of base pair formation, a second-generation metallo-base pair was generated
consisting of a pyridine-2,6-dicarboxamide (Dipam) and a pyridine (Py) nucleobase. This
new metallo-base pair is more stable than the natural base pairs dA:dT and dC:dG and highly
selective against mispairing. In addition, incorporation of multiple metallo-base pairs into
DNA results in the formation of stable duplexes demonstrating that hydrogen bonding base
pairs can efficiently be replaced by metal-dependent base pairs at multiple sites in DNA.
Ó 2003 Elsevier Inc. All rights reserved.
Keywords: Nucleobases; Alternative binding schemes; Metallo-DNA-base pairs; Copper(II); Tridentate
pyridine ligands
1. Introduction
Watson–Crick hydrogen bonding of adenine with thymine (dA:dT base pair) and
guanine with cytosine (dG:dC base pair) is essential for the stability and specificity of
duplex formation. Additional, stable and selective base pairs would increase the
* Corresponding author. Fax: +1-858-784-9440.
E-mail address: schultz@scripps.edu (P.G. Schultz).
0045-2068/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.bioorg.2003.09.001

14
N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
Fig. 1. Copper(II)-mediated metallo-base pairing between the tridentate nucleobases Dipic, Dipam or
Me-Dipam and a monodentate pyridine (Py) nucleobase.
capacity of DNA for information storage and might allow the generation of DNAs
with novel functions [1,2]. Efforts to develop such a third unnatural base pair have
focused on two different strategies thus far: the design of bases with altered hydrogen
bonding patterns [3] or hydrophobic packing interactions [4,5]. We and others re-
cently reported another strategy in which hydrogen bonding interactions are re-
placed by metal-dependent pairing of two nucleobases [6,7]. Our first generation
metallo-base pair involved a pyridine-2,6-dicarboxylate nucleobase (Dipic) as a pla-
nar tridentate ligand, and a pyridine nucleobase (Py) as the complementary single
donor ligand (Fig. 1) [6]. When these bases are introduced into the middle of a
15-nucleotide duplex, the duplex displays almost comparable melting temperatures
(Tm) in the presence of equimolar Cu2þ to a duplex with a dA:dT pair. However, this
metallo-base pair has only modest pairing selectivity. Relatively stable mispairs are
formed when paired with natural bases, especially with the purine nucleobases dA
and dG. In order to store genetic information, an unnatural base pair should not
only be highly stable but also highly selective against mispairing with the natural nu-
cleobases. Herein we report the design and synthesis of a new copper(II)-mediated
metallo-base pair consisting of a tridentate pyridine-2,6-dicarboxamide ligand
(Dipam) and a monodentate pyridine ligand (Py) (Fig. 1). This second-generation
metallo-base pair shows high specificity against mispairing with the natural nucleo-
bases while maintaining a high pairing stability.
2. Experimental
2.1. General
All reactions were carried out in oven-dried glassware under inert atmosphere
with commercial dehydrated solvents (Sigma–Aldrich) unless otherwise stated. Alde-
hydo-D-ribose 6 was prepared according to known literature procedures [8]. The syn-
thesis of the Py nucleobase followed the same route as shown in Scheme 1 and is a
modification of the first reported synthesis for this nucleobase (for analytical data see

N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
15
Scheme 1. Synthetic route for phosphoramidite 1a. aConditions: (a) Br2, oleum (27–30% SO3), reflux
(57%). (b) KMnO4, H2O, 90 °C (60%). (c) SOCl2, reflux, then MeOH (78%). (d) NaBH4, EtOH, reflux
(77%). (e) TBDMS triflate, 2,6-lutidine, CH2Cl2, rt (98%). (f) n-BuLi, ether, )78 °C, then aldehydo-D-ri-
bose 6, THF (35% 7(1R)), (36% 7(1S)). (g) MeSO2Cl, NEt3, pyridine, 0 °C, then rt (60%). (h) TBAF,
THF, rt (98%). (i) TEMPO, NaOCl, NaClO2, CH3CN, phosphate buffer (pH 6.7), 35 °C (84%). (j) tri-
methylsilyl diazomethane, benzene, MeOH, rt (59%). (k) TFA, CH2Cl2, rt (98%). (l) DMTr-Cl, NEt3,
py, rt (53%). (m) (i-Pr)2NP(Cl)CH2CH2CN, (i-Pr)2NEt, CH2Cl2, rt (88%).
[8]). Oligonucleotides were synthesized using an Expedite nucleic acid synthesizer
(PerSeptive Biosystems). DNA synthesis reagents were purchased from Glen
Research, Sterling, VA. All other reagents were purchased from Sigma–Aldrich.
2.2. Synthesis
2.2.1. 3-Bromo-2,6-dimethyl-pyridine (2)
2,6-Lutidine (50 mL, 0.43 mol) was slowly added to oleum (27–30% SO3, 250 mL)
at 0 °C, followed by the addition of bromine (9.0 mL, 0.18 mol). The reaction mixture
was stirred at 150 °C for 18 h. After cooling to RT the reaction mixture was poured
into ice-water (1 L), neutralized by careful addition of 5 M NaOH (2 L) and extracted
with ethyl acetate (5 ꢀ 600 mL). The combined organic extracts were dried over
MgSO4 and concentrated. Vacuum destillation [bp: 95 °C (15 mbar)] afforded 2

16
N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
(37.9 g, 57%) as a colorless oil. 1H-NMR (500 MHz, CDCl3) d 7.62 (1H, d,
J ¼ 8:1 Hz), 6.82 (1H, d, J ¼ 8:1 Hz), 2.59 (3H, s), 2.44 (3H, s). 13C-NMR (150 MHz,
CDCl3) d 156.5, 156.3, 139.8, 122.1, 118.2, 24.9, 23.9.
2.2.2. 3-Bromo-pyridine-2,6-dicarboxylic acid dimethyl ester (3)
To a solution of compound 2 (37.9 g, 0.20 mol) in water (500 mL) was added
KMnO4 (171.0 g, 1.08 mol) in ten portions. The first five portions were added over
5 h at 70 °C and the second five portions over 5 h at 90 °C. After complete addition
stirring was continued at 90 °C for 12 h. The reaction mixture was filtered hot and the
residue was washed with hot water (4 ꢀ 100 mL). The filtrate was concentrated (ca.
400 mL) and conc. HCl (65 mL) was added. Heating to 100 °C redissolved the precip-
itated white solid, and the filtrate was stored at 4 °C overnight. The resulting white
solid was filtered, washed with ice-cold water (60 mL) and air-dried to give 5-bromo-
pyridine-2,5-dicarboxylic acid (29.8 g, 60%); (1H-NMR (250 MHz, d6-DMSO) d 8.18
(1 H, d, J ¼ 8:4 Hz), 7.82 (1H, d, J ¼ 8:4 Hz)). The dicarboxylic acid (28.8 g,
0.12 mol) was dissolved in thionyl chloride (100 mL) and heated to reflux for 4 h. Af-
ter cooling to RT thionyl chloride was removed in vacuo and MeOH (200 mL) was
added. The resulting solution was heated to reflux for 1 h, concentrated and dried
1
under high vacuum to give compound 3 (25.0 g, 78%) as a white solid. H-NMR
(500 MHz, CDCl3) d 8.13 (1H, d, J ¼ 8:4 Hz), 8.06 (1H, d, J ¼ 8:3 Hz), 4.00 (3H,
s), 3.99 (3H, s). ESMS calcd for C9H8BrNO4 (MHþ): 274.0; found: 274/276.
2.2.3. (5-Bromo-6-hydroxymethyl-pyridin-2-yl)-methanol (4)
To a suspension of compound 3 (25.0 g, 91 mmol) in EtOH (225 mL) at 0 °C was
added NaBH4 (16.2, 0.43 mol) in portions over 0.5 h. The reaction mixture was stir-
red for 3 h at RT and then heated to reflux for 16 h. After cooling to RT the reaction
mixture was poured into saturated aqueous NaHCO3 (500 mL) and extracted with
CH2Cl2 (5 ꢀ 200 mL). The combined organic extracts were dried over MgSO4. Chro-
matography over silica (CH2Cl2:MeOH; 10:1) afforded compound 4 (15.3 g, 77%) as
1
a white solid. H-NMR (500 MHz, CDCl3) d 7.81 (1H, d, J ¼ 8:1 Hz), 7.18 (1H, d,
J ¼ 8:1 Hz), 4.73 (2H, s), 4.71 (2H, s). 13C-NMR (150 MHz, CDCl3) d 157.6, 155.6,
140.8, 120.6, 116.9, 64.5, 63.2. HRMS (MALDI) calcd for C7H8BrNO2 (MH+):
217.9817; found: 217.9816.
2.2.4. 3-Bromo-2,6-bis-(tert-butyl-dimethyl-silanyloxymethyl)-pyridine (5)
To a solution of compound 4 (15.3 g, 70.2 mmol) and 2,6-lutidine (42 mL,
0.36 mol) in CH2Cl2 (150 mL) at 0 °C was added TBDMS trifluoromethanesulfonate
(43 mL, 0.19 mol) over 0.5 h. Stirring was continued at 0 °C for 1 h, followed by 2 h at
RT. The reaction mixture was poured into saturated aqueous NaHCO3 (200 mL)
and extracted with CH2Cl2 (3 ꢀ 150 mL). The combined organic extracts were
washed with brine (150 mL), dried over MgSO4 and concentrated. Chromatography
over silica (hexanes:ethyl acetate; 20:1) afforded compound 5 (30.7 g, 98%) as a col-
1
orless oil. H-NMR (500 MHz, CDCl3) d 7.80 (1H, d, J ¼ 8:4 Hz), 7.29 (1H, d,
J ¼ 8:1 Hz), 4.85 (2H, s), 4.76 (2H, s), 0.94 (9H, s), 0.89 (9H, s), 0.09 (6H, s), 0.08

N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
17
(6H, s). 13C-NMR (150 MHz, CDCl3) d 159.8, 156.1, 140.9, 120.3, 118.1, 66.6, 65.6,
26.0, 25.9, 18.5, 18.3, )5.1, )5.4. ESMS calcd for C19H36BrNO2Si2 (MHþ): 446.2;
found: 446/448.
2.2.5. 1S-[2,6-Bis-(tert-butyl-dimethyl-silanyloxymethyl)-pyridin-3-yl]-3S, 4R,5-tris-
(4-methoxy-benzyloxy)-pentan-1-ol (71S) and 1R-[2,6-bis-(tert-butyl-dimethyl-
silanyloxymethyl)-pyridin-3-yl]-3S,4R,5-tris-(4-methoxy-benzyloxy)-pentan-1-ol (71R)
To a solution of compound 5 (5.5 g, 12.3 mmol) in ether (35 mL) at )78 °C was
added dropwise a solution of n-BuLi in hexanes (1.6 M, 8.5 mL, 13.6 mmol) over
0.5 h. After stirring for 1 h at )78 °C a solution of aldehyde 6 [8] (6.7 g, 13.5 mmol)
in THF (35 mL) was added. After 0.5 h at )78 °C the cooling bath was removed,
and stirring was continued at RT for 2 h. The reaction mixture was poured into
ice-cold saturated aqueous NaHCO3 and extracted with CH2Cl2 (3 ꢀ 150 mL).
The combined organic extracts were washed with brine (150 mL), dried over MgSO4
and concentrated. Separation of diastereomers was achieved by chromatography
over silica (hexanes:ethyl acetate; 8:1, then 5:1). Compound 71R (3.67 g, 35%) eluted
first, followed by compound 71S (3.81 g, 36%). Both compounds were obtained as
1
colorless oils. 71R: H-NMR (500 MHz, CDCl3) d 7.69 (1H, d, J ¼ 8:1 Hz), 7.39
(1H, d, J ¼ 8:1 Hz), 7.24 (6H, m), 6.84 (6H, m), 5.22 (1H, m), 4.95 (1H, d,
J ¼ 11:7 Hz), 4.78 (2H, d, J ¼ 4:0 Hz), 4.73 (2H, d, J ¼ 11:7 Hz), 4.62 (3H, m),
4.49 (1H, d, J ¼ 11:4 Hz), 4.44 (2H, d, J ¼ 3:3 Hz), 4.00 (1H, m), 3.79 (9H, m),
3.66 (1H, d, J ¼ 3:7 Hz), 3.62 (2H, d, J ¼ 5:1 Hz), 2.00 (2H, m), 0.95 (9H, s),
0.86 (9H, s), 0.06 (8H, s), 0.04 (4H, s). HRMS (MALDI) calcd for C48H71NO9Si2
(MNaþ): 884.4565; found: 884.4597. 71S: 1H-NMR (500 MHz, CDCl3) d 7.80
(1H, d, J ¼ 7:7 Hz), 7.42 (1H, d, J ¼ 7:7 Hz), 7.23 (6H, m), 6.85 (6H, m), 5.21
(1H, m), 4.80 (4H, m), 4.62 (3H, m), 4.42 (2H, m), 4.30 (1H, m), 3.90 (1H, m),
3.81 (9H, s), 3.62 (3H, m), 2.05 (2H, m), 0.96 (9H, s), 0.87 (9H, s), 0.12 (8H, s),
0.06 (4H, s). HRMS (MALDI) calcd for C48H71NO9Si2 (MHþ): 862.4746; found:
862.4761.
2.2.6. 2,6-Bis-(tert-butyl-dimethyl-silanyloxymethyl)-3-[4S-(4-methoxy- benzyloxy)-
5R-(4-methoxy-benzyloxy-methyl)-tetrahydro-furan-2R-yl]- pyridine (8)
To a solution of compound 71S (2.96 g, 3.43 mmol) in pyridine (180 mL) at 0 °C
was added triethylamine (4.8 mL, 34.3 mmol). A solution of methanesulfonyl chlo-
ride (1.06 mL, 13.7 mmol) in pyridine (15 mL) was slowly added over 1 h. After 1 h
at 0 °C stirring was continued at RT for 3 h. The reaction mixture was quenched with
water (10 mL) and concentrated. The residue was taken up in CH2Cl2 (50 mL) and
washed with 5% aqueous NaHCO3 (50 mL), and the aqueous layer was extracted
with CH2Cl2 (2 ꢀ 50 mL). The combined organic extracts were dried over MgSO4
and concentrated. Chromatography over silica (hexanes:ethyl acetate; 4:1) afforded
compound 8 (1.49 g, 60%) as a yellow oil. 1H-NMR (500 MHz, CDCl3) d 7.96
(1H, d, J ¼ 8:1 Hz), 7.44 (1H, d, J ¼ 8:1 Hz), 7.28 (4H, m), 6.90 (4H, m), 5.53
(1H, dd, J ¼ 5:1; 10:5 Hz), 4.93 (1H, d, J ¼ 11:7 Hz), 4.83 (2H, s), 4.76 (1H, d,
J ¼ 11:7 Hz), 4.56 (2H, m), 4.53 (1H, d, J ¼ 11:4 Hz), 4.47 (1H, d, J ¼ 11:7 Hz),
4.26 (1H, m), 4.14 (1H, m), 3.84 (6H, s), 3.65 (2H, m), 2.52 (1H, ddd, J ¼ 1:3,

18
N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
5.1, 13.2 Hz), 1.82 (1H, ddd, J ¼ 5:9, 10.6, 13.5 Hz), 0.99 (9H, s), 0.92 (9H, s), 0.14
(6H, s), 0.13 (3H, s), 0.08 (3H, s). HRMS (MALDI) calcd for C40H61NO7Si2 (MHþ):
724.4065; found: 724.4087.
2.2.7. {6-Hydroxymethyl-5-[4S-(4-methoxy-benzyloxy)-5R-(4-methoxy-benzyloxy-
methyl)-tetrahydro-furan-2R-yl]-pyridin-2-yl}-methanol (9)
To a solution of compound 8 (2.07 g, 2.86 mmol) in THF (20 mL) at 0 °C was
added a solution of tetrabutylammonium fluoride in THF (1.0 M, 11.5 mL,
11.5 mmol). After stirring for 4 h at RT, ethyl acetate (150 mL) and water (80 mL)
were added to the reaction mixture, the phases were separated and the aqueous
phase was extracted with ethyl acetate (2 ꢀ 80 mL). The combined organic extracts
were washed with brine (80 mL), dried over MgSO4 and concentrated. Chromatog-
raphy over silica (CH2Cl2:MeOH:NEt3; 100:8:0.5) afforded compound 9 (1.39 g,
1
98%) as a white solid. H-NMR (500 MHz, CDCl3) d 8.08 (1H, d, J ¼ 7:7 Hz),
7.36 (1H, d, J ¼ 8:1 Hz), 7.25 (4H, m), 6.90 (4H, m), 5.21 (1H, dd, J ¼ 5:1; 10:6 Hz),
4.91 (1H, m), 4.80 (1H, m), 4.51 (2H, d, J ¼ 3:3 Hz), 4.49 (2H, s), 4.28 (1H, m), 4.16
(1H, d, J ¼ 6:2 Hz), 3.83 (6H, s), 3.61 (2H, d, J ¼ 4:0 Hz), 3.14 (4H, m), 2.38 (1H,
dd, J ¼ 5:1; 12:8 Hz), 1.83 (1H, ddd, J ¼ 5:9, 10.6, 13.2 Hz). HRMS (MALDI) calcd
for C28H33NO7 (MHþ): 496.2335; found: 496.2339.
2.2.8. 3-[4S-(4-Methoxy-benzyloxy)-5R-(4-methoxy-benzyloxymethyl)-tetrahydro-
furan-2R-yl]-pyridine-2,6-dicarboxylic acid (10)
To a solution of compound 9 (1.78 g, 3.59 mmol) in acetonitrile (18 mL) and phos-
phate buffer (pH 6.7, 13.5 mL) was added TEMPO (80 mg, 0.51 mmol) at RT. After
warming to 35 °C a solution of sodium chlorite (1.65 g, 18.2 mmol) in water (3.6 mL)
and a solution of sodium hypochlorite (1.00 g, 13.4 mmol) in water (1.8 mL) were
added simultaneously over 1 h. After stirring at 35 °C overnight, water (25 mL)
was added to the reaction mixture which was then adjusted to pH 8 by addition
of 1 M NaOH. The resulting mixture was poured into an ice-cold saturated aqueous
sodium thiosulfate solution (50 mL) and stirring was continued for 0.5 h. The pH
was adjusted to pH 3 by slow addition of 1 M HCl (30 mL) and the aqueous phase
was extracted with CH2Cl2 (6 ꢀ 100 mL). The combined organic extracts were
washed with water (2 ꢀ 100 mL) and brine (2 ꢀ 100 mL), dried over MgSO4 and con-
centrated to afford compound 10 (1.58 g, 84%) as a white solid. 1H-NMR (500 MHz,
CDCl3) d 8.51 (1 H, d, m), 8.25 (1H, m), 7.22 (4H, m), 6.84 (4H, m), 6.07 (1H, m),
4.52 (3H, m), 4.39 (1H, d, J ¼ 11 Hz), 4.29 (1H, m), 4.10 (1H, m), 3.77 (3H, s), 3.75
(3H, s), 3.65 (2H, m), 2.87 (1H, m), 1.48 (1H, m). HRMS (MALDI) calcd for
C28H29NO9 (MNaþ): 546.1735; found: 546.1749.
2.2.9. 3-[4S-(4-Methoxy-benzyloxy)-5R-(4-methoxy-benzyloxymethyl)-tetrahydro-
furan-2R-yl]-pyridine-2,6-dicarboxylic acid dimethyl ester (11)
To a solution of compound 10 (2.04 g, 3.88 mmol) in benzene (40 mL) and MeOH
(10 mL) was added a 2.0 M solution of diazomethane in hexanes (5.5 mL, 11.0 mmol)
at RT. Stirring was continued for 0.5 h and the resulting yellow solution was concen-
trated. Chromatography over silica (CH2Cl2:MeOH; 25:1) afforded compound 11

N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
19
(1.26 g, 59%) as a colorless oil. 1H-NMR (500 MHz, CDCl3) d 8.33 (1H, d,
J ¼ 8:4 Hz), 8.16 (1H, d, J ¼ 8:4 Hz), 7.23 (4H, m), 6.86 (4H, m), 5.69 (1H, dd,
J ¼ 5:9; 9:9 Hz), 4.51 (3H, m), 4.41 (1H, d, J ¼ 11:4 Hz), 4.24 (1H, m), 4.09 (1H,
m), 3.99 (3H, s), 3.97 (3H, s), 3.80 (3H, s), 3.79 (3H, s), 3.62 (2H, m), 2.72 (1H,
ddd, J ¼ 2:2, 6.2, 13.2 Hz), 1.74 (1H, ddd, J ¼ 6:2, 9.9, 13.2 Hz). 13C-NMR
(150 MHz, CDCl3) d 165.4, 165.0, 161.1, 159.2, 146.1, 145.6, 144.0, 136.7, 130.0,
129.4, 129.3, 127.5, 113.8, 84.1, 80.3, 76.7,73.1, 70.8, 70.4, 55.3, 53.1, 40.7. HRMS
(MALDI) calcd for C30H33NO9 (MHþ): 552.2228; found: 552.2225.
2.2.10. 3-(4S-Hydroxy-5R-hydroxymethyl-tetrahydro-furan-2R-yl)-pyridine-2,6-di-
carboxylic acid dimethyl ester (12)
To a solution of compound 11 (1.26 g, 2.28 mmol) in CH2Cl2 (25 mL) was added a
20% solution of TFA in CH2Cl2 (25 mL) at RT. After stirring for 0.5 h at RT MeOH
was added (100 mL) and the resulting mixture was concentrated and coevaporated
with toluene (3ꢀ). Chromatography over silica (CH2Cl2:MeOH:NEt3; 15:1:0.1) af-
1
forded compound 12 (700 mg, 98%) as a white solid. H-NMR (500 MHz, CDCl3)
d 8.23 (2H, s), 5.72 (1H, dd, J ¼ 5:9; 9:7 Hz), 4.42 (1H, m), 4.04 (1H, m), 3.99
(3H, s), 3.97 (3H, s), 3.88 (1H, m), 3.82 (1H, m), 2.57 (1H, ddd, J ¼ 2:4, 5.9,
13.4 Hz), 1.90 (1H, ddd, J ¼ 6:4, 9.7, 13.2 Hz). 13C-NMR (150 MHz, CDCl3) d
165.9, 164.9, 146.3, 142.8, 136.2, 128.9, 127.4, 87.2, 76.1, 73.3, 63.2, 53.2, 44.1.
HRMS (MALDI) calcd for C14H17NO7 (MNaþ): 334.0897; found: 334.0907.
2.2.11. 3-{5R-[Bis-(4-methoxy-phenyl)-phenyl-methoxymethyl]-4S-hydroxy-tetra-
hydro-furan-2R-yl}-pyridine-2,6-dicarboxylic acid dimethyl ester (13)
Compound 12 (700 mg, 2.25 mmol), azeotroped from pyridine (2 ꢀ 2 mL), was
dissolved in pyridine (4 mL). Triethylamine (1.43 mL, 10.25 mmol), followed by
DMTr-Cl (835 mg, 2.46 mmol) was added, and the reaction mixture was stirred over-
night at RT. After addition of MeOH (5 mL), the solvent was removed. Chromatog-
raphy over silica (50–66% ethyl acetate in hexanes, 1% NEt3) afforded compound 13
(736 mg, 53%) as a white foam. 1H-NMR (500 MHz, CDCl3) d 8.36 (1H, d,
J ¼ 8:1 Hz), 8.15 (1H, d, J ¼ 8:3 Hz), 7.43 (2H, m), 7.32 (4H, m), 7.26 (2H, m),
7.20 (1H, m), 6.81 (4H, m), 5.77 (1H, dd, J ¼ 6:2, 9.2 Hz), 4.39 (1H, m), 4.09 (1H,
m), 3.99 (3H, s), 3.96 (3H, s), 3.77 (6H, s), 3.38 (2H, m), 2.60 (1H, ddd, J ¼ 2:9,
6.2, 13.2 Hz), 1.90 (1H, ddd, J ¼ 6:2, 9.2, 13.2 Hz). HRMS (MALDI) calcd for
C35H35NO9 (MHþ): 614.6617; found: 614.6623.
2.2.12. 3-{5R-[Bis-(4-methoxy-phenyl)-phenyl-methoxymethyl]-4S-[(2-cyano-eth-
oxy)-diisopropylamino-phosphanyloxy]-tetrahydro-furan-2R-yl}-pyridine-2-6-dicarb-
oxylic acid dimethyl ester (1)
Compound 13 (405 mg, 0.66 mmol), azeotroped from pyridine (2 ꢀ 2 mL), was
dissolved in CH2Cl2 (10 mL). N,N-Diisopropylethylamine (0.60 mL, 3.44 mmol)
was added, followed by the dropwise addition of 2-cyanoethyl diisopropylamino-
chloro phosphoramidite (0.31 mL, 1.39 mmol). After 3 h at RT, the reaction mixture
was partitioned between CH2Cl2 (50 mL) and saturated aqueous NaHCO3 (50 mL).
The layers were separated, and the aqueous layer was extracted with CH2Cl2

20
N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
(2 ꢀ 50 mL). The combined organic phases were dried over MgSO4 and concen-
trated. Chromatography over silica (33–50% ethyl acetate in hexanes, 1% NEt3) af-
forded compound 2 (472 mg, 88%) as a white foam.
1H-NMR (500 MHz, CDCl3) d 8.40 (1H, m), 8.14 (1H, m), 7.44 (2H, m), 7.26 (7H,
m), 6.80 (4H, m), 5.75 (1H, m), 4.48 (1H, m), 4.22 (1H, m), 4.00 (3H, s), 3.97 (3H, s),
3.85 (1H, m), 3.77 (6H, s), 3.60 (3H, m), 3.35 (2H, m), 2.66 (1H, m), 2.42 (1H, m),
1.90 (1H, m), 1.28 (1H, m), 1.18 (8H, m), 1.08 (4H, M). HRMS (MALDI) calcd for
C44H52N3O10P (MHþ): 814.3469; found: 814.3463.
2.3. Oligonucleotide synthesis
All oligonucleotides were prepared on a 1.0 lM scale. A standard protocol for
2-cyanoethyl phosphoramidites was used, except that the coupling of the phospho-
ramidites 1 and dPy was extended to 15 min. Phosphoramidites for phenoxacetyl
protected dA (Pac-dA), 4-isopropyl-phenoxacetyl protected dG (iPr-Pac-dG), and
acetyl protected dC (Ac-dC) were used (ultramild synthesis reagents from Glen Re-
search) and the oligonucleotides were deprotected for 6 h with 50 mM potassium car-
bonate in methanol at RT. The tritylated oligonucleotides were then purified by
reversed phase HPLC (5–60% acetonitrile in 0.05 M aqueous TEAA, TEAA ¼ tetra-
ethyl ammonium acetate), detritylated with 80% acetic acid, and again purified by
reversed phase HPLC (5–20% acetonitrile in 0.05 M aqueous TEAA). Identities of
all oligonucleotides was confirmed by MALDI-TOF MS.
2.4. Thermal denaturation studies
All UV melting experiments were carried out in 5 mM sodium phosphate, 50 mM
sodium perchlorate, pH 7.0, with 15 lM copper(II) acetate and 1 lM DNA duplex
concentration. For the preparation of all samples and buffers analytical grade water
free of metal ion contaminations (Fluka) was used. All samples were treated with
Chelex-resin at RT overnight to remove Cu2þ contaminations prior to measure-
ments. Melting temperature experiments were carried out on a Cary 300-Bio UV/
Vis spectrophotometer. Melting curves were recorded at 260 nm for a consecutive
heating (15–65 °C)-cooling-heating protocol with a linear gradient of 0.5 °C/min.
3. Results and discussion
3.1. Ligand design and synthesis
Based on the high pairing stability of our first-generation Dipic–Py metallo-base
pair, we envisioned that small variations in the structure of the ligands might result
in further improvements in pairing selectivity. We decided to further optimize the
ligand structure of the tridentate Dipic ligand without changing the structure of the
monodentate Py ligand. Phosphoramidite 1 serves as a precursor for the synthesis
of the Dipic ligand. After incorporation of this phosphoramidite into oligonucleo-

N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
21
Fig. 2. Structure of phosphoramidite 1 and, after incorporation into oligonucleotides, conversion to Dipic
by treatment with NaOH or nucleobases Dipam and Me-Dipam by treatment with NH4OH or MeNH2.
tides using a DNA synthesizer, the two dimethylester groups were saponified to
give the two metal-coordinating carboxylate groups of the Dipic ligand (Fig. 2).
By using a convertible nucleoside approach we replaced the ester groups of this
same precursor with two different amide groups: the two carboxylate groups of
the O, N, O-Dipic ligand (X ¼ O) were replaced with amide substituents to afford
the two new N,N,N-ligand systems Dipam (X ¼ NH) and Me-Dipam (X ¼ NCH3)
(Fig. 2). In the case of pyridine-containing ligands with amide side chains it has
been reported that complexation of Cu2þ probably occurs through the deproto-
nated amide N-atom under neutral conditions [9]. Since the deprotonated amide
group and the carboxylate group are isoelectronic, we expected very similar elec-
tronic properties of both new ligand systems. On the other hand the steric proper-
ties differ since Dipam and Me-Dipam contain additional substituents (H or Me,
respectively) on the coordinating N atom which might result in different pairing
stability and/or pairing selectivity.
The synthetic route for phosphoramidite 1 is shown in Scheme 1. Compound 5
was synthesized in five steps starting from 2,6-lutidine. Lithiation of 5 followed
by nucleophilic addition to aldehydo-D-ribose 6 [8] gave a diastereomeric mixture
of alcohols 7(1R) and 7(1S) in equal amounts. After separation of the diastereomers
by chromatography, 7(1S) was cyclized to give the b-nucleoside 8. Cleavage of the
TBDMS-ethers followed by oxidation and esterification afforded compound 11. The
PMB protecting groups were removed, and tritylation followed by phosphitylation
provided phosphoramidite 1 which was then incorporated into the middle of a
15-nucleotide strand using an automated DNA synthesizer. Treatment of the oligo-
nucleotide with concentrated ammonium hydroxide or methylamine at 56 °C for
12 h afforded the nucleobases Dipam and Me-Dipam in 85% and 84% yield, respec-
tively [10]. The identities of the oligonucleotides were confirmed by MALDI-TOF
spectroscopy.

22
N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
3.2. Base pairing stability and selectivity
Our first generation Dipic–Py metallo-base pair shows high pairing stability but
lacks selectivity. When introduced into the middle of the 15-nucleotide duplex 14,
the duplex displays almost comparable melting temperatures (Tm) in the presence
of Cu2þ to a duplex with a dA:dT pair at the same site (Tm ¼ 36:5 and 39.1 °C for
Dipic:Py and dA:dT, respectively, Table 1) [11]. However, the range in Tm of the Di-
pic base when paired with natural bases is 26.2–34.1 °C. Relatively stable mispairs
are formed with dA and dG; Dipic:dA forms the most stable mispair and is destabi-
lized by only 2.4 °C relative to Dipic:Py. For comparison, mispairs between the nat-
ural nucleobases are 9.6–15.8 °C less stable in this sequence context. Nucleobase
Dipam (X ¼ NH) shows both a very high pairing stability and selectivity. In the pres-
ence of fifteen equivalents Cu2þ ions, the Dipam:Py base pair is even more stable
than a dG:dC and a dA:dT base pair (Tm ¼ 43:0, 40.1 and 39.1 °C for Dipam:Py,
dG:dC and dA:dT, respectively, Table 1) [12]. Like the Dipic nucleobase, the purine
nucleobases dA and dG still form the most stable mispairs with Dipam but the base
pairing selectivity has been significantly enhanced: the most stable mispair (Di-
pam:dG) is destabilized by 8.5 °C relative to Dipam:Py, and the Dipam:dA mispair
is destabilized by 13.0 °C. Although the most stable mispair is with dA for Dipic
and with dG for Dipam, the absolute values of the melting temperatures remain al-
most the same (34.1 and 34.5 °C for Dipic:dA and Dipam:dG, respectively). Since the
pairing stability of Dipam:Py exceeds the stability of the natural dA:dT and dC:dG
base pairs, pairing selectivity is gained at the same time. Thus the Dipam:Py base pair
exhibits pairing stability and selectivity comparable to the natural base pairs. The
rationale for these improved properties of the Dipam:Py metallo-base pair over
Table 1
Sequence of the 15-mer DNA duplex 14 and denaturation temperatures (Tm) for various combinations of
X and Ya
Y
X
Dipam
Dipic
A
T
C
G
Me-Dipam
b
Dipam
Dipic
Py
n.d.c
n.d.c
43.0 (28.0)d
30.0
b
36.5
34.1
23.6
b
29.5
39.1
27.6
27.3
29.5
23.5
24.0
25.7
27.7
30.5
26.5 (26.6)d
A
b
T
27.5
25.5
39.1
23.3
28.1
26.1
b
C
26.2
28.6
40.1
b
G
34.5
39.5
a Melting temperature experiments were carried out on a Cary 300-Bio UV/Vis spectrophotometer. The
heating rates were 0.5 °C minꢁ1. Conditions: 5 mM sodium phosphate, 50 mM sodium perchlorate, pH 7,
with 15 M copper(II) acetate and 1 lM DNA duplex (50-CACATTAXTGTTGTA-30 and 30-
GTGTAATYACAACAT-50 (14)
b No sigmoidal melting curve was observed in the temperature range between 15 and 65 °C.
c Not determined.
d Melting temperature in the absence of copper(II) acetate.

N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
23
the Dipic:Py remains unclear at this point. Efforts are underway to obtain a structure
of a DNA duplex containing the Dipam:Py metallo-base pair to gain further insight
into the increased pairing selectivity and stability of this metallo-base pair. X-ray
analysis would also clarify the exact coordination mode of the amide groups to
the metal center since coordination through the N-atoms is more likely but coordi-
nation through the O-atoms cannot be excluded at this point. UV-melting experi-
ments revealed that nucleobase Me-Dipam (X ¼ NCH3) does not form a stable
metallo-base pair with the Py nucleobase, as there is no change in Tm in the presence
of Cu2þ (Table 1). The two methyl substituents of the amide groups might be steri-
cally too demanding to properly fit into a DNA duplex. This might increase the C10–
C10 distance to an extent that metal-coordination of the Py base to the Cu(II) ion
becomes less effective which might result in the loss of pairing stability. In addition,
the pK value of the Me-Dipam ligand is likely to be shifted to a higher value in
comparison to the Dipam ligand due to the introduction of the methyl groups. This
change in pK value makes the deprotonation of the amide group more difficult which
might result in destabilization of the corresponding metal complex.
3.3. DNA containing multiple metallo-base pairs
Metallo-base pairing in combination with automated DNA synthesis offers a con-
venient way to construct DNA duplexes containing metal ions in defined locations
[13]. The construction of oligomers consisting of multiple metallo-base pairs could
lead to materials with interesting electronic properties based on the controlled and
periodic spacing of metal ions along the helix axis. The high pairing stabilities of
both the Dipam:Py and the Dipic:Py base pair prompted us to study the UV-melting
behaviour of duplexes containing multiple metallo-base pairs. Fifteen-nucleotide du-
plexes with two and four adjacent metallo-base pairs (16a/b–17a/b) were synthesized
and their thermal stability in the presence of 15 equivalents of Cu2þ ions per metallo-
base pair was determined (Fig. 3) [14,15]. In duplexes 16a/b one dA:dT base pair
adjacent to the metallo-base pair of duplexes 15a/b was replaced by a second
metallo-base pair. This substitution resulted in an increased duplex stability of
0.4/2.0 °C (43.0/43.4 °C for 15a/16a and 36.5/38.5 °C for 15b/16b, respectively) in
comparison to duplexes 15a/b containing one metallo-base pair. Further substitution
Fig. 3. Fifteen-nucleotide duplexes containing adjacent metallo-base pairs (P ¼ Py, D ¼ Dipam or Dipic)
and their stabilities determined by melting temperature experiments.

24
N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
of two more dA:dT base pairs with two metallo-base pairs gave duplexes 17a/b con-
taining four adjacent metallo-base pairs. In comparison to duplexes 16a/b with two
metallo-base pairs, duplex stabilities were further increased by 3.9/1.1 °C (43.4/
47.3 °C for 16a/17a and 38.5/39.6 °C for 16b/17b, respectively). These results demon-
strate that natural base pairs can be efficiently replaced with metallo-base pairs at
multiple sites in DNA duplexes. Efforts are underway to investigate the potentially
novel electronic properties of these DNA duplexes containing multiple metallo-base
pairs.
4. Conclusions
In summary we have generated an improved copper(II)-mediated metallo-base
pair with a high pairing stability and selectivity: the pyridine-2,6-dicarboxamide–
pyridine metallo-base pair (Dipam:Py) is more stable than the natural dC:dG
and dA:dT base pairs, and shows high selectivity against mispairing with the nat-
ural bases. Construction of stable DNA duplexes containing multiple metallo-base
pairs demonstrates that hydrogen bonding of natural base pairs can efficiently be
replaced with copper(II)-mediated metallo-base pairing. Efforts are underway to in-
vestigate the electronic properties of DNA duplexes containing multiple metallo-
base pairs.
Acknowledgments
Funding was provided by the National Institutes of Health (GM 64403), the
Alexander von Humboldt Foundation (Feodor-Lynen Fellowship to NZ.) and the
German Research Foundation (Fellowship to E.M.). This is manuscript number
15168-CH of The Scripps Research Institute.
References
[1] J.D. Bain, C. Switzer, A.R. Chamberlin, S.A. Benner, Nature 356 (1992) 537–539.
[2] G.F. Joyce, Proc. Natl. Acad. Sci. USA 95 (1998) 5845–5847.
[3] (a) S. Switzer, S.E. Moroney, S.A. Benner, J. Am. Chem. Soc. 111 (1989) 8322–8323;
(b) J.A. Piccirilli, T. Krauch, S.E. Moroney, S.A. Benner, Nature 343 (1990) 33–37;
(c) C.Y. Switzer, S.E. Moroney, S.A. Benner, Biochemistry 32 (1993) 10489–10496;
€
(d) J. Horlacher, M. Hottiger, V.N. Podust, U. Hubscher, S.A. Benner, Proc. Natl. Acad. Sci. USA
92 (1995) 6329–6333;
(e) M.J. Lutz, H.A. Held, M. Hottiger, U. Hubscher, S.A. Benner, Nucleic Acids Res. 24 (1996)
€
1308–1313.
[4] (a) E.T. Kool, Biopolymers 48 (1998) 3–17;
(b) K.M. Guckian, T.R. Krugh, E.T. Kool, Nat. Struct. Biol. 11 (1998) 954–959;
(c) E.T. Kool, J.C. Morales, K.M. Guckian, Angew. Chem., Int. Ed. 39 (2000) 990–1009.
[5] (a) D.L. McMinn, A.K. Ogawa, Y. Wu, J. Liu, P.G. Schultz, F.E. Romesberg, J. Am. Chem. Soc. 121
(1999) 11585–11586;

N. Zimmermann et al. / Bioorganic Chemistry 32 (2004) 13–25
25
(b) A.K. Ogawa, Y. Wu, D.L. McMinn, J. Liu, P.G. Schultz, F.E. Romesberg, J. Am. Chem. Soc.
122 (2000) 3274–3287;
(c) Y. Wu, A.K. Ogawa, M. Berger, D.L. McMinn, P.G. Schultz, F.E. Romesberg, J. Am. Chem.
Soc. 122 (2000) 7621–7632;
(d) M. Berger, A.K. Ogawa, D.L. McMinn, Y. Wu, P.G. Schultz, F.E. Romesberg, Angew. Chem.,
Int. Ed. 39 (2000) 2940–2942;
(e) M. Berger, S.D. Luzzi, A.A. Henry, F.E. Romesberg, J. Am. Chem. Soc. 124 (2002) 1222–1226.
[6] (a) E. Meggers, P.L. Holland, W.B. Tolman, F.E. Romesberg, P.G. Schultz, J. Am. Chem. Soc. 122
(2000) 10714–10715;
(b) S. Atwell, E. Meggers, G. Spraggon, P.G. Schultz, J. Am. Chem. Soc. 123 (2001) 12364–12367;
(c) N. Zimmermann, E. Meggers, P.G. Schultz, J. Am. Chem. Soc. 124 (2002) 13684–13685.
[7] (a) H. Weizman, Y. Tor, J. Am. Chem. Soc. 123 (2001) 3375–3376;
(b) K. Tanaka, Y. Yamada, M. Shionoya, J. Am. Chem. Soc. 124 (2002) 8802–8803;
(c) K. Tanaka, A. Tengeiji, T. Kato, M.S. Toyama, M. Shionoya, J. Am. Chem. Soc. 124 (2002)
12494–12498.
[8] M.A.W. Eaton, T.A. Millican, J. Chem. Soc., Perkin Trans. 1 (1988) 545–547.
[9] (a) H. Sigel, B.E. Fischer, B. Prijs, J. Am. Chem. Soc. 99 (1977) 4489–4496;
(b) H. Sigel, R.B. Martin, Chem. Rev. 82 (1982) 385–426.
[10] Yield is given after purification by reversed phase HPLC. Identities were confirmed by MALDI-TOF
MS. 15-mer oligonucleotides containing Dipic-Ester or Me-Dipam have the same mass but could be
distinguished by their different retention times using reversed phase HPLC.
[11] In [6a] slightly higher Tm-values were reported. Tm-values reported herein were confirmed by repeated
experiments and were consistent under the experimental conditions reported here.
[12] In comparison to the Dipic:Py metallo-base pair, the Dipam:Py base pair required the presence of at
least five equivalents of Cu2þ ions for the formation of a stable duplex. Interestingly, the Dipam:Py
base pair forms a slightly stable base pair even in the absence of Cu2þ ions (Tm ¼ 28:0 °C) which is not
the case for the Dipic:Py base pair. This stabilizing effect may be caused by weak hydrogen bonding
interactions. It might also explain the observed difference in required metal ion concentrations of
both metallo-base pairs due to different kinetics for the formation of the corresponding metal
complexes.
[13] Shionoya and coworkers recently reported a self-assembled metal array in artificial DNA consisting
of copper(II)-mediated base pairs between hydroxypyridone nucleobases: K. Tanaka, A. Tengeiji, T.
Kato, N. Toyama, M. Shionoya, Science 299 (2003) 1212–1213.
[14] Conditions: 5 mM sodium phosphate, 50 mM sodium perchlorate, pH 7, 1 lM DNA duplex, 30 lM
(15a/b)/60 lM (16a/b) copper(II) acetate..
[15] No sigmoidal melting curves were observed for duplexes 16a/b and 17a/b in the absence of Cu2þ ions.