Solid-phase synthesis of deoxynucleic guanidine (DNG) oligomersand melting point and circular dichroism analysis of binding fidelity of octameric thymidyl oligomers with DNA oligomers

Article abstract of DOI:10.1021/ja984212w

A practical solid-phase synthesis of deoxynucleic guanidine (DNG), a positively charged DNA backbone analogue, is reported. The nucleoside coupling step in the solid-phase synthesis of DNG involves the attack of a terminal 3′-amine upon an electronically

Full text of DOI:10.1021/ja984212w

3888
J. Am. Chem. Soc. 1999, 121, 3888-3896
Solid-Phase Synthesis of Deoxynucleic Guanidine (DNG) Oligomers
and Melting Point and Circular Dichroism Analysis of Binding
Fidelity of Octameric Thymidyl Oligomers with DNA Oligomers
Barry A. Linkletter, Istvan E. Szabo, and Thomas C. Bruice*
Contribution from the Department of Chemistry, UniVersity of California, Santa Barbara, California 93106
ReceiVed December 7, 1998
Abstract: A practical solid-phase synthesis of deoxynucleic guanidine (DNG), a positively charged DNA
backbone analogue, is reported. The nucleoside coupling step in the solid-phase synthesis of DNG involves
the attack of a terminal 3′-amine upon an electronically activated 5′-carbodiimide to create a protected
guanidinium internucleoside linkage. The activated carbodiimide is synthesized in situ by the mercury(II)
abstraction of sulfur from an unsymmetrically substituted thiourea in which one substituent is an electron-
withdrawing protecting group and the other is the 5′-nucleoside monomer. This produces, in addition to the
carbodiimide, a mercury sulfide precipitate which accumulates in the pores of the solid support, restricting
solvent and reagent access and reducing the coupling yields with each successive cycle. This obstacle is overcome
by a simple washing step involving a thiophenol solution which readily removes the mercury salt. The addition
of this step to the cycle enables DNG oligomers to be synthesized using standard macroporous SPS supports.
Coupling yields of 98% were estimated from the HPLC analysis of the product mixtures. An octameric thymidyl
oligomer (II) was synthesized and the fidelity of binding to octameric adenyl DNA oligomers containing
cytidyl mismatches was determined. Binding was studied by thermal denaturation (Tm), Job plots, and circular
dichroism spectrophotometry. The DNG oligomer (II) formed a 2:1 complex with octameric adenyl DNA
(III) with a melting temperature of 63 °C. Each cytidyl mismatch induced a penalty of 4 to 5 °C in the
observed melting temperatures. DNA sequences with four or more mismatches showed no base pairing in the
presence of II. No association was observed between II and octameric cytidyl DNA. These observations
demonstrate that DNG oligomers of moderate length are able to discriminate between complementary and
mismatched DNA oligomers.
Introduction
is to replace the phosphoribose backbone entirely, such as in
the cases of PNA,10,11 PHONA,12 or PNAA.13
Oligonucleoside analogues capable of arresting cellular
processes at the translational and transcriptional levels via
recognition and binding to complementary RNA or DNA are
known, respectively, as antisense and antigene agents.1-3
Important goals in designing antisense compounds include
increasing the binding affinity to the base sequence while
maintaining fidelity of recognition, resistance to degradation by
nucleases, and effective membrane permeability. A recurring
theme in many of these antisense compounds is the incorporation
of neutral internucleoside linkages which eliminate the mutual
repulsion between the negatively charged phosphate diester
backbones in duplex DNA. These unnatural linkages are likely
to be resistant to nucleases and may also be membrane
permeable. Oligonucleosides linked by amides,4 phosphonates,5
carbamates,6 methylenemethylimino (MMI),7 heterocycles,8 and
acetals9 are representative of this approach. Another approach
In recent years, approaches involving the incorporation of
positive charges in antisense oligomers have been developed.
Positive charges can be added to the bases14 or the sugar rings
to give zwitterionic DNA. The phosphate backbone can be
alkylated with alkylamines to produce positively charged
phosphate triester linkages between the nucleosides.15 Another
promising approach is to replace the internucleoside phosphate
linkage with a positively charged, achiral guanidinium group.16
The guanidinium linkage is resistant to nucleases17 and the
positive charges of the backbone may give rise to cell membrane
permeability through electrostatic attraction of the oligonucleo-
side to the negatively charged phosphate groups of the cell
surface. Recently, micelles featuring positively charged surfaces
(9) Wang, J.; Matteucci, M. D. Bioorg. Med. Chem. Lett. 1997, 7, 229.
(10) Uhlmann, E.; Peyman, A.; Breipohl, G.; Will, D. W. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 2796.
(1) Agrawal, S.; Zhao, Q. Y. Curr. Opin. Chem. Biol. 1998, 2, 519.
(2) Temsamani, J. G. P. Biotechnol. Appl. Biochem. 1997, 26, 665.
(3) De Mesmaeker, A.; Ha¨ner, R.; Martin, P.; Moser, H. E. Acc. Chem.
Res. 1995, 28, 366.
(4) De Mesmaeker, A.; Waldner, A.; Lebreton, J.; Hoffmann, P.; Fritsch,
V.; Wolf, R. M.; Freier, S. M. Angew. Chem., Int. Ed. Engl. 1994, 33, 226.
(5) Tseng, B. Y.; Ts’o, P. O. P. Antisense Res. DeV. 1995, 5, 251.
(6) Waldner, A.; De Mesmaeker, A.; Lebreton, J.; Fritsch, V.; Wolf, R.
M. Synlett 1994, 57.
(11) Nielsen, P. E. Pure Appl. Chem. 1998, 70, 105.
(12) Peyman, A.; Uhlmann, E.; Wagner, K.; Augustin, S.; Breipohl, G.;
Will, D. W.; Scha¨fer, A.; Wallmeier, H. Angew. Chem., Int. Ed. Engl. 1996,
35, 2636.
(13) Fujii, M.; Yoshida, K.; Hidaka, J. Bioorg. Med. Chem. Lett. 1997,
7, 637.
(14) Ueno, Y.; Mikawa, M.; Matsuda, A. Bioconj. Chem. 1998, 9, 33.
(15) Skibo, E. B.; Xing, C. Biochemistry 1998, 37, 15199.
(16) Dempcy, R. O.; Almarsson, O¨ .; Bruice, T. C. Proc. Natl. Acad.
Sci. U.S.A. 1994, 91, 7864.
(17) Barawkar, D. A.; Bruice, T. C. Proc. Natl. Acad. Sci. U.S.A. 1998,
95, 11047.
(7) Morvan, F.; Sangvhi, Y. S.; Perbost, M.; Vasseur, J.-J.; Bellon, L. J.
Am. Chem. Soc. 1996, 118, 255.
(8) von Matt, P.; Altmann, K.-H. Bioorg. Med. Chem. Lett. 1997, 7, 1553.
10.1021/ja984212w CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/09/1999

Solid-Phase Synthesis of Deoxynucleic Guanidine
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3889
have been reported to carry impermeable compounds such as
DNA oligomers through the lipid bilayer of cells.18
Scheme 1. Synthesis of Monomers for Solid-Phase
Synthesis of DNG Oligomersa
Short pentameric thymidyl oligomers of DNG (I) with four
positively charged internucleoside guanidinium groups (Chart
1), have been synthesized by solution phase methods19,20 and
have been found to bind strongly to polyadenylic DNA [poly-
(dA)] in a 2:1 (DNG:DNA) complex.21-23 No interaction of
polyguanylic [poly(dG)], polycytidylic [poly(dC)], or poly-
thymidylic [poly(dT)] acids with I was observed.21-23 A detailed
study of the binding of I with short strand DNA oligomers
containing adenine and guanine tracts established that these short
DNG oligomers bind with fidelity in a triple helical fashion.22
The pentameric DNG I was the longest oligomer that could
be synthesized by solution-phase synthesis. To be an effective
antisense or antigene agent, DNG must be available in longer
sequences that can recognize unique sites in the human genome.
A solid-phase synthetic method was developed to make longer
oligomers.24 The single most important feature of solid-phase
synthesis is that purification after each synthetic cycle is
accomplished by washing the solid support to which the growing
oligomer is covalently attached. This eliminates losses of the
product while enabling the use of greater excesses of the
reagents, increasing coupling yields. Using this method, the
octameric thymidyl DNG oligomer II (Chart 1) was synthe-
sized.24 This paper presents, for the first time, the detailed solid-
phase synthesis of an octameric DNG oligomer and investigates
the binding of this new, much more positively charged oligomer
to complementary and mismatched short DNA oligomers.
Results and Discussion
a Key: (a) 20% TFA/DCM; (b) benzoyl isothiocyanate, DCM; (c)
Na2CO3, 2:1 MeOH/H2O; (d) H2S, pyridine; (e) 9-fluorenylmethyl-
chloroformate, DCM, TEA; (f) peracetic acid; (g) trichloroethoxycar-
bonyl isothiocyanate, TEA, DCM; (h) H2, 10% Pd/C, EtOH.
Solid-Phase Synthesis of DNG. Monomers for the various
solid-phase synthesis approaches were synthesized as outlined
in Scheme 1. The syntheses of 1, 2, and 3 have been described
previously16 and the details for the synthesis of 4, 5, 6, and 7
are described in the Experimental Section.
Scheme 2. Formation of an Electronically Activated
Carbodiimide from an Acyl-Protected Thiourea Enables
Synthesis of a Protected Guanidine Group
Initially, the synthetic approach was based on the origional
solution-phase synthesis.16 The oligomer was extended by the
addition of 4 to a Rink peptide amide resin25 loaded with a 3′-
amino thymidyl nucleside. The coupling reaction involved the
displacement of the sulfonyl group to produce a guanidinium
linkage.26 After coupling, the 3′-azide was reduced with H2S
in pyridine16 and the cycle was repeated. However, the coupling
yields dropped substantially as the oligomer was extended and
the maximum length of oligomer possible by this solid-phase
approach did not surpass that of the intial solution-phase method.
It was decided to protect the guanidinium linkages in a
manner that would enhance the solubility of the growing
oligomer in organic solvents and prevent the basic guanidinium
group from interfering with the coupling chemistry. A new
synthetic approach was adopted which results in protected
guanidinium linkages being formed in the coupling step. The
protecting group can then be removed after the synthesis is
completed. This reaction was adapted from a method developed
by Kim and Qian for the synthesis of terminal guanidines27
which involves converting a N,N′-bis-Boc-protected thiourea to
a carbodiimide by addition of HgCl2 followed by substitution
of an alkylamine. The two t-Boc groups on the carbodiimide
intermediate electronically activate it to the nucleophilic attack
of the amine without the need for protonation or any catalyst.
The resulting diprotected guanidinium group is then deprotected
by acid to produce a terminal guanidinium.
We have found that a single electron-withdrawing group is
sufficient to activate the carbodiimide to attack by an amino
group. A thiourea with an electron-withdrawing protecting group
on one nitrogen and a nucleoside on the other can be used as a
precursor to an activated carbodiimide suitable for producing a
protected guanidinium internucleoside linkage when reacted with
an 3′-amino nucleoside derivative as shown in Scheme 2.
The monomer, 5, was easily obtained by the reduction and
Fmoc protection of 2, an intermediate in the synthesis of 4.
(18) Huang, C. Y.; Uno, T.; Murphy, J. E.; Lee, S.; Hamer, J. D.;
Escobedo, J. A.; Cohen, F. E.; Radhakrishnan, R.; Dwarki, V.; Zuckermann,
R. N. Chem. Biol. 1998, 5, 345.
(19) Dempcy, R. O.; Browne, K. A.; Bruice, T. C. J. Am. Chem. Soc.
1995, 117, 6140.
(20) Dempcy, R. O.; Luo, J.; Bruice, T. C. Proc. Natl. Acad. Sci. U.S.A.
1996, 93, 4326.
(21) Blasko´, A.; Dempcy, R. O.; Minyat, E. E.; Bruice, T. C. J. Am.
Chem. Soc. 1996, 118, 7892.
(22) Blasko´, A.; Minyat, E. E.; Dempcy, R. O.; Bruice, T. C. Biochem-
istry 1997, 36, 1821.
(23) Brown, K. A.; Dempcy, R. O.; Bruice, T. C. Proc. Natl. Acad. Sci.
U.S.A. 1995, 92, 7051.
(24) Linkletter, B.; Bruice, T. C. Bioorg. Med. Chem. Lett. 1998, 8, 1285.
(25) Rink, H.; Sieber, P. Peptides 1988, 139.
(26) Maryanoff, C. A.; Stanzione, R. C.; Plampin, J. N. Phosphorus Sulfur
1986, 27, 221.
(27) Kim, K. S.; Qian, L. Tetrahedron Lett. 1993, 34, 7677.

3890 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Linkletter et al.
Chart 1
Figure 1. Cation exchange HPLC of the crude cleavage mixture for
the synthesis of 13 (A), 14 (B), and 15 (C): (A) cleaved after the first
coupling/loading step and subsequent deblockingspeak at 4.9 min is
13; (B) cleaved after the second coupling reaction and subsequent
deblockingspeak at 8.3 min is 14; and (C) cleaved after coupling the
third 3′-azido nucleoside and subsequent deblockingspeak at 15.2 min
is 15.
Chart 2
collected in the pores of the resin, the beads became black in
color and no longer shrunk in solvents, such as methanol, that
normally deswell polystyrene resin. To avoid the problem of
pore blockage, it is necessary to remove the precipitate from
the resin after each coupling cycle. When a 20% thiophenol
solution in DMF was filtered through the precipitate-blocked
resin, the dark color immediately disappeared and the resin
returned to its original yellow color and swelling behavior. The
resin was then washed with copious amounts of DMF and
agitated in 20% piperidine in DMF for 20 min to de-block the
Fmoc-protected 3′-amine. Then the cycle was repeated until the
desired number of nucleosides had been coupled.
The chlorotrityl linker used in this SPS method is cleavable
by 3% TFA in DCM or AcOH/TFE/DCM.29 The latter cleavage
protocol is mild enough to prevent depurination of adenine
nucleosides, making this methodology practical for any base
sequence of DNG. In the case of thymidyl oligomers, cleavage
from the resin can be accomplished in 5 min with 3% TFA in
DCM. The product was precipitated from cold ether and purified
by reverse phase preparative LC. The fractions containing the
product were evaporated to dryness to obtain the purified
protected oligomer 16. Analysis of the crude cleavage product
showed that the major component was approximately 75% of
the integration signal (Figure 2). The peaks preceding the major
product were shorter oligomers resulting from incomplete
coupling reactions. No peaks indicative of longer oligomers,
caused by double additions, were observed. From this informa-
tion, the average coupling yields were estimated to be 95 to
98%.
Rink amide resin was again employed and loaded with glycine
to give a terminal primary amine to react with the activated
carbodiimide. This method allows loading of the first nucleoside
using the same chemistry as all the coupling steps in the
synthesis. The HgCl2-mediated coupling of 6 with the resin-
bound 3′-amino group was very successful. A trimeric oligomer
with benzoyl-protected guanidinium linkages, 15, was synthe-
sized with coupling yields of 95% as measured by Fmoc
colorimetry methods and HPLC analysis of cleaved resin
samples. A trisubstituted guanidine-linked DNA analogue such
as 15 may be of interest as it has a slightly positive character
(as evidenced by its retention on the cation exchange HPLC
column, Figure 1) and is soluble in organic solvents.
The trichloroethoxycarbonyl group was found to be more
suitable than the benzoyl group as a guanidinium protecting
group for the synthesis of longer oligomers. It is stable to the
coupling conditions and is easily removed by zinc powder and
acetic acid.28 A commercially available 2% cross-linked poly-
styrene 2-(2-aminoethoxy)ethanol 2-chlorotrityl resin was used
as the solid support. The amino functionality allows the loading
step to use the same chemistry as the subsequent coupling steps
and the chlorotrityl linker is cleavable under mild acidic
conditions. The synthesis cycle proceeds in the 5′-3′ direction
as outlined in Scheme 3.
The deprotection of 16 was performed using cadmium powder
in acetic acid. This procedure produces superior results to the
usual zinc deprotection method used for trichloroethyl protecting
groups.30 The deprotection reaction proceeded in a quantitative
yield with no detected side products. The final product was
desalted by preparative cation exchange HPLC and, after
evaporation and sublimation of the ammonium acetate buffer,
In the solid-phase synthesis of II, reported in our preliminary
communication, the coupling chemistry suffered from the
presence of the mercury sulfide precipitate.24 As the precipitate
(29) Barlos, K.; Gatos, D.; Kallitsis, J.; Papaphotiu, G.; Sotivic, P.;
Wenging, Y.; Scha¨fer, W. Tetrahedron Lett. 1989, 30, 3943.
(30) Hancock, G.; Galpin, I. J.; Morgan, B. A. Tetrahedron Lett. 1982,
249.
(28) Just, G.; Grozinger, K. Synthesis 1976, 457.

Solid-Phase Synthesis of Deoxynucleic Guanidine
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3891
Scheme 3
(Tm) recorded for each DNA oligomer after complexation with
2 equiv of II. The Tm for the 2:1 association of octameric DNG
II associated with the complementary octameric adenyl DNA
III is 62.8 °C in the presence of 0.1 M KCl. The Tm for the
A:T duplex of III and octameric thymidyl DNA is less than 10
°C under the same conditions. The large increase in the Tm is
due to the opposite charge attraction between the positively
charged DNG and negatively charged DNA strands. The thermal
denaturation plots of 2 equiv of II in the presence of 1 equiv of
various DNA oligomers containing mismatches at the terminal
ends in 0.1 M KCl are presented in Figure 3. As mismatches
are incorporated in the DNA oligomer, the Tm and the
magnitude of the hyperchromic effect are reduced. When two
cytidine mismatches are present at each end of the oligomer
(as in DNA VI), complex formation is no longer significant as
measured by the hyperchromic effect.
Figure 2. Reverse phase analytical HPLC of the crude cleavage
mixture in the synthesis of 16 showing few side products or truncated
oligomers. Percent area of major peak is ca. 70% of the total integrated
region.
Internal mismatches have a much more pronounced effect.
The single internal mismatch substitution in DNA oligomer VIII
reduced the Tm from 62.8 to 48 °C and inclusion of two internal
mismatches (IX) eliminated base pairing association (Figure 4).
The large positive charge is not enough to overcome the
disruption in base pairing that two internal mismatches cause.
The observed Tm for the DNG:DNA association drops when
the salt concentration is increased from 0.1 M KCl to 0.3 M
KCl since increased ionic strength reduces the importance of
the opposite charge attraction.21,23 Figures 5 and 6 show the
thermal denaturation plots for the DNA oligomers possessing
terminal and internal mismatches with DNG II in the presence
of 0.3 M KCl. The Tm measurements are lower and the
was stored as a solution in water at 4 °C. Electrospray mass
spectroscopic analysis indicates the expected masses for the
triply charged [m/z 743.2, calcd for C92H134N41O26 (M + 3H)3+
743.4] and quadruply charged [m/z 557.7, calcd for C92H135N41O26
(M + 4H)4+ 557.8] forms of the oligomeric DNG II.
Effect of Mismatches on Thermal Melting Behavior.
Melting studies were performed to determine if II would bind
with fidelity to DNA oligomers. Table 1 shows the DNA
oligomers chosen and the thermal denaturation temperatures

3892 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Linkletter et al.
Table 1. Melting Points of Mixtures of DNA Oligomers and
DNG IIa
Figure 5. Thermal denaturation plots for DNA:DNG association (1.25
µM DNA oligomer, 2.5 µM DNG 2, 10 mM KHPO4 buffer, pH 7.4,
0.3 M KCl). Rate of heating was 0.2 °C/min DNA oligomer: III, b;
IV, 9; V, ]; VI, O; VII, 0; and XV, 4.
a Samples consist of 1.25 µM DNA oligomer, 2.5 µM II, and 10
mM KHPO4 buffer at pH 7.4 and the indicated ionic strength controlled
by KCl. Melting points were determined by first derivative analysis of
samples that showed melting behavior.
Figure 6. Thermal denaturation plots for DNA:DNG association (1.25
µM DNA oligomer, 2.5 µM DNG 2, 10 mM KHPO4 buffer, pH 7.4,
0.3 M KCl). Rate of heating was 0.2 °C/min DNA oligomer: III, b;
VIII, 9; IX, ]; X, O; XI, ×; XII, +; XIII, O; XIV, 0; and XV, [.
Figure 3. Thermal denaturation plots for DNA:DNG association (1.25
µM DNA oligomer, 2.5 µM DNG II, 10 mM KHPO4 buffer, pH 7.4,
0.1 M KCl). Rate of heating was 0.2 °C/min DNA oligomer.
Absorbance measurements taken every 1 °C (for clarity only one in
five data points are represented by a symbol) in plots 1 through 4.
DNA oligomer: III, b; IV, 9; V, ]; VI, O; VII, 0; and XV, 4.
Figure 7. DNA:DNG thermal melting point (Tm) in relation to the
number of mismatches at the terminals (O, 0.1 M KCl; 0, 0.3 M KCl)
and internal mismatches (×, 0.1 M KCl; ], 0.3 M KCl). [DNA
oligomer] ) 1.25 µM, [II] ) 2.5 µM, [KHPO4] ) 10 mM, pH 7.4,
rate of heating 0.2 °C/min. Data from Table 1.
determined for all oligomers which showed a significant
hyperchromicity in the presence of II. In all the nonbinding
thermal plots a very weak hyperchromicity is seen in the range
of 35 to 45 °C. This is probably due to a nonspecific ionic
association of the backbones and not base pairing effects.
Effect of Mismatches and Ionic Strength on Continuous
Variation Plots. Job31 plot analyses were performed to measure
the stoicheometry of association. The absorbance at 260 nm
Figure 4. Thermal denaturation plots for DNA:DNG association (1.25
µM DNA oligomer, 2.5 µM DNG 2, 10 mM KHPO4 buffer, pH 7.4,
0.1 M KCl). Rate of heating was 0.2 °C/min DNA oligomer: III, b;
VIII, 9; IX, ]; X, O; XI, ×; XII, +; XIII, O; XIV, 0; and XV, [.
hyperchromicities are smaller than in the case of 0.1 M KCl.
Figure 7 shows the relationship between the number of
mismatches (terminally or internally placed) and the Tm

Solid-Phase Synthesis of Deoxynucleic Guanidine
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3893
Figure 9. CD spectra of HO-T(pT)7-OH (O) compared to DNG II
(b). Solution conditions: 2.0 × 10-6 M oligomer, 0.1 M KCL, 10
mM phosphate buffer, pH 7.3 at 20 °C.
that mismatches due to inclusion of pyrimidine bases do not
support the formation of a triplex under the experimental
conditions.
Circular Dichroism Spectroscopy of Complementary and
Mismatched DNA Oligomers with II. The CD spectrum of
an oligomer in solution can give valuable information about its
conformation as a single strand or in association with other DNA
oligomers. The bases of DNA are planar and have no intrinsic
CD signal of their own. Any CD signals observed in the
absorbance band of the bases (230-300 nm) are caused by the
spatial organization of the bases in a chiral structure, such as a
helix, under the influence of the chiral sugar backbone. Base
stacking interactions magnify this effect and give rise to the
strong CD signals observed for DNA and RNA oligomers and
polymers.32 The CD spectra of the octameric thymidyl DNA
oligomer HO-T(pT)7-OH (XVI) and DNG II are compared in
Figure 9. The DNA thymidyl oligomer shows a CD profile
consistent with polymeric thymidyl DNA CD spectra.33 The
DNG, however, is very different. The positive CD band at 280
nm is not present in the CD spectrum of II, and the negative
band at 255 nm is greatly attenuated. This seems to indicate
that the nucleobases interact with each other more weakly in
the oligothymidyl DNG II than in the case of the oligothymidyl
DNA XVI. This may be due to the steric effect of the
guanidinium linkage. The angle between the bonds of the
tetrahedral phosphate diesters in the backbone of DNA is ∼109°
whereas the angle between the bonds of a guanidium group is
∼120°. This wider angle may lead to a greater distance between
the nucleoside bases in single-stranded DNG and, consequently,
the bases would not stack as effectively as in single-stranded
helical DNA. If this is the case, the positive charge attraction
of the guanidinium backbone to the DNA target seems to
overcome the penalty for the lack of preorganization and any
strain induced in the sugar ring by the geometry of the
guanidinium.
Figure 8. Job plots for DNG II with DNAs III, V, VI, IX, and XV
plotted by percent hyperchromicity (top) and absorbance (bottom).
[oligomer] ) 2.0 uM, [KHPO4] ) 10 mM, [KCl] ) 0.1 M, pH 7.4,
observed at 260 nm.
was measured for samples containing a constant concentration
of 2 µM oligonucleside varying between 0 and 100% DNG II
with a given DNA oligomer making up the remainder. The
inflection point in the plot indicates the stoicheometry of the
DNG:DNA complexes. Job plots of oligomers III, V, VI, IX,
and XV complexed with DNG II in 10 mM phosphate buffer
and 0.1 M KCl salt for ionic strength are shown in Figure 8.
The Job plot for oligomer III with DNG II shows a clear
inflection point at 64% DNG indicating that III binds in a 1:2
ratio with DNG II. Oligomer V, with two terminal mismatches,
has an inflection point at 50% DNG II. Oligomer IX displays
a weaker hyperchromic effect with DNG II and the Job plot in
this case does not show a clear inflection point. Oligomer VI
has an even weaker hyperchromic effect and also displays no
clear break point in the Job plot. The maximum hyperchromicity,
as determined by the percent difference between the observed
optical density (OD) and that calculated for the sum of the two
separate oligomers at the given concentrations, is greatest for
the complementary oligomer III, with a hyperchromicity of 33%
at 66% DNG II. As mismatches are included, the hyperchro-
micity is weakened and the point of maximum deflection from
the calculated OD shifts toward 50% DNG II for oligomers V,
IX, and VI, with maximum hyperchromicities of 24%, 18%,
and 10%, respectively. The noncomplementary oligomer XV
shows little hyperchromic effect in the presence of DNG II.
When DNG II is associated with complementary DNA, the
CD spectrum does not match that of the theoretical spectrum
calculated from the weighted sums of the CD spectra for the
individual oligomers. The CD spectrum of the triple helical
complex of DNG II with DNA III and the spectrum calculated
for the weighted combination of uncomplexed II and III are
shown in Figure 10. The differences in the spectra indicate that
It has been reported that the pentameric DNG I binds to
homopurine strands in a 2:1 DNG:DNA ratio even in the
presence of mismatches, although the mismatches significantly
reduce the strength of association.21,22 We see from the above
(32) Johnson, W. C., Jr. In Circular Dichroism and the Conformational
Analysis of Biomolecules; Fasman, G. D., Ed.; Plenum Press: New York,
1996; p 433.
(33) Gray, D. M.; Ratcliff, R. L.; Vaughan, M. R. Methods Enzymol.
1992, 211, 389.
(31) Job, P. Ann. Chim. 1928, 9, 113.

3894 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Linkletter et al.
oligomeric guanidyl-linked thymidines is not limited to nucleo-
sides and could, in principle, be useful for many oligomeric
guanidine compounds. It is now possible to synthesize DNG
oligomers of many different lengths and sequences in useful
quantities.
Binding studies with short complementary and mismatched
DNA oligomers indicate that the highly positively charged DNG
oligomer II is able to discriminate between complementary and
noncomplementary base pairs. Cytidyl mismatches in octameric
adenyl DNA oligomers lower the Tm melting points by
approximately 4 to 5 °C for each terminal mismatch and weaken
the hyperchromic effect. Four terminal mismatches result in the
loss of the hyperchomicity due to base pairing, leaving only
weak nonspecific ionic association. Internal mismatches have
a more profound effect with the majority of the hyperchromic
effect removed after two internal mismatches. A single substitu-
tion near the center of the octameric adenyl oligomer VIII
results in a reduction in the Tm of 15 °C in 0.1 M KCl. The
CD spectrum of single-stranded II indicates that the thymidyl
bases are not stacked in an organized helix, as in the case of
the analogous octameric thymidine. CD spectra of II, in the
presence of complementary and mismatched oligomers, com-
pared to the calculated CD signal for the nonassociated
oligomers indicate that the complex of II with the complemen-
tary oligomer III has the greatest change in CD spectra and the
difference is reduced with increasing mismatches until no
significant difference is seen for the association with non-
complementary oligomer XV.
Figure 10. CD spectra of a 2:1 mixture of DNG II and DNA III
compared to the calculated CD spectra for the combined unassociated
oligomers.
Experimental Section
Materials. HPLC grade solvents and sodium carbonate were
obtained from Fisher Scientific. Anhydrous solvents, trichloroethoxy-
chloroformate, triethylamine, 9-fluorenylmethylchloroformate, mer-
cury(II) chloride, and potassium thiocyanate were purchased from
Aldrich and used without further purification. Novabeads Rink amide
MBHA resin (0.30 mequiv/g substitution) and 2% cross-linked
polystyrene 2-(2-aminoethoxy)ethanol 2-chlorotrityl resin (0.31 meq/g
substitution) was purchased from Novabiochem. All DNA oligomers
were purchased prepurified from the Biological Resource Center at
UCSF. Compounds 1, 2, and 3 were synthesized as described
previously.16
Figure 11. Difference CD spectra resulting from the subtraction of
the calculated CD spectrum from that measured for 2:1 mixtures of
DNG II and DNA oligomers III, V, VI,IX, and XV.
structural changes have taken place in the two oligomers because
of their association. Difference spectra obtained from subtracting
the measured CD for samples containing DNG II in the presence
of DNA III, V, VI, IX, and XV from the calculated CD for the
corresponding oligomer combinations are shown in Figure 11.
The difference spectra show the largest changes for the
complementary sequence III with DNG II. As mismatches are
included, the difference spectra drop in amplitude, indicating
that the strength of association is weakening. For the non-
complementary oligomer XV, the difference indicates no
association.
Computer modeling studies of an octameric thymidyl DNG
oligomer with an octameric adenyl DNA oligomer for both
double and triple helical binding modes have been reported in
a previous study.34 These studies suggested that the duplex and
Watson-Crick portion of the triplex maintain an overall B-like
conformation. This is consistent with the CD spectra of
complementary DNA III with DNG II (Figure 10) which
indicates a normal B-DNA based triple helix with a strong
negative band at 247 nm and a weaker positive band above
260 nm.33
General. Hydrogenations were carried out with a Parr hydrogenator
equipped with a 500 mL hydrogenation vessel. Analytical reverse phase
HPLCs were performed on a Hewlett-Packard 1050 system equipped
with a quaternary solvent delivery system and UV detector set at 260
nm and a 2.1 × 250 mm C18 ODS-hypersil reverse phase column
purchased from Altech. A gradient from 100% eluent A (100 mM
triethylammonium acetate, pH 7.0) to 50% eluent A, 50% eluent B
(acetonitrile) over 15 min with a flow rate of 1.3 mL/min was used.
Analytical cation exchange HPLCs were performed with the same
system with a 2.1 × 250 mm SCX ion exchange column purchased
from Altech. A gradient from 100% eluent A (100 mM TRIS buffer,
pH 7.0) to 100% eluent B (100 mM TRIS, 1.0 M guanidinium
hydrochloride, pH 7.0) over 20 min with a flow rate of 1.3 mL/min
was used. Preparative reverse phase HPLC were performed on the same
system with a 10 × 250 mm C8 ODS-hypersil reverse phase column
purchased from Altech. A gradient from 66% eluent A (100 mM
triethylammonium acetate, pH 7.0), 33% eluent B (acetonitrile) to 20%
eluent A, 50% eluent B over 15 min with a flow rate of 3.0 mL/min
was used. Preparative cation exchange HPLCs were performed with
the same system with a 10 × 250 mm SCX ion exchange column
purchased from Altech. A gradient from 66% eluent A (water), 34%
eluent B (3 M ammonium acetate) to 100% eluent B over 15 min with
Conclusions
An effective and convenient solid-phase synthesis for oligo-
meric DNG has been developed. It features an efficient coupling
reaction and a mild cleavage and deprotection protocol that will
be compatible with all DNA bases. This method for synthesizing
1
a flow rate of 3.0 mL/min was used. H and 13C NMR spectra were
obtained on a Varian Unity 400 spectrometer at 400 and 100 MHz,
respectively. UV spectra were obtained on a Cary 100 Bio UV/vis
spectrophotometer equipped with a temperature programmable cell
block. IR spectra were in KBr pellets with a Perkin-Elmer 1300
(34) Luo, J.; Bruice, T. C. J. Chem. Soc. 1998, 120, 1115.

Solid-Phase Synthesis of Deoxynucleic Guanidine
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3895
spectrophotometer. TLC was carried out on silica gel (Kieselger 60
MHz, DMSO-d6) δ 12.2, 35.2, 43.5, 60.7, 80.2, 83.1, 110.4, 135.2,
F
254) glass backed commercial plates and visualized by UV light.
Melting Studies. Thermal denaturation (Tm) plots were obtained
150.5, 163.8, 166.5; IR (KBr pellet) 3197, 3074, 2108 (azide), 1687
(CdO), 1479, 1269, 1055 (-SO3H); MS (FAB) m/z 374 (M + H)+
HRMS (FAB) m/z 374.09001, calcd for C11H15N7O6S 374.08828.
;
by observing the absorbance at 260 nm of a solution of the oligomers
in 1 cm path-length quartz cuvettes as the temperature was raised 0.2
°C/min from 20 to 95 °C. All samples had been previously annealed
by cooling from 90 to 10 °C at 0.2 °C/min and stored at 10 °C for 2
days. Samples consisted of 2.5 µM DNG II and 1.25 µM DNA oligomer
with 10 mM potassium phosphate buffer at pH 7.4 and either 0.1 or
0.3 M KCl for ionic strength control. For Tm determinations hyper-
chromicity was used. Data were recorded every 1 °C. Samples were
covered with mineral oil to prevent evaporation.
Job Plots. Job31 or continuous variation plots were obtained by
mixing six samples at various ratios of DNG oligomer to DNA oligomer
while maintaining the total concentration of nucleoside base at 2 µM.
All solutions contained 10 mM potassium phosphate buffer at pH 7.4
and 0.1 M KCl. Solutions were mixed and left to equilibrate at room
temperature for 24 h. DNA oligomer concentrations were determined
spectrophotometrically using extinction coefficients provided by the
manufacturer. The value of ꢀ268 ) 8700 M-1 cm-1 was used for the
extinction coefficient (per nucleoside base) of DNG II.21
Circular Dichroism Spectra. CD spectra were obtained on an OLIS
RSM circular dichroism spectrophotometer. Scans were run from 350
to 210 nm taking a measurement every 1 nm. The integration time for
each data point was 1.4 s. Ten scans were made of each sample and
then averaged and smoothed using a 15 point exponential fitting
algorithm. Samples were held in a 1 cm path-length cuvette and the
temperature was maintained at 20 °C.
Methods: 5′-Benzoylthiourea-3-(9-fluorenylmethoxycarbamoyl)-
3′,5′-deoxythymidine (5). Hydrogen sulfide gas was bubbled through
a solution of 2 (2.05 g, 4.8 mmol) in 80 mL of 1:1 pyridine/water for
5 min until saturated, then the mixture was left to stir for 3 h. Argon
gas was then bubbled through the mixture to remove the hydrogen
sulfide. A fine yellow precipitate formed which was removed by
filtration through a pad of Celite. The filtrate was evaporated to dryness
and the residue was dissolved in 100 mL of dioxane. TEA (1.0 mL)
was added and then 9-fluorenylmethoxychloroformate (1.15 g, 4.4
mmol). After 30 min TLC analysis indicated the reaction was complete.
The mixture was evaporated to dryness and 100 mL of water was added,
and the resulting slurry was extracted with ethyl acetate (3 × 150 mL).
The organic extracts were dried with sodium sulfate, evaporated, and
then purified by silica gel flash chromatography (3 × 20 cm silica
column with 75% to 100% gradient of ethyl acetate in hexanes). TLC
5′-Trichloroethoxycarbonylthiourea-3′-azido-3′,5′-deoxythymi-
dine (6). First, trichloroethoxycarbonylisothiocyanate (Troc-NCS) was
synthesized using an adapted literature procedure.35 A solution of
potassium thiocyanate (10 g, 0.10 mole), potassium acetate (25 mg,
25 mmol), and pyridine (0.05 mL, 0.06 mmol) in 10 mL of water was
made to which was added trichloroethylchloroformate (10 mL, 47
mmol) slowly over a period of 10 min while the temperature was
maintained at 7 °C by a water bath. The resulting heterogeneous mixture
was stirred vigorously for 9 h at 7 °C, and the progress of the reaction
was monitored by gas chromatography. When all starting material was
consumed, the reaction mixture was extracted once with 100 mL of
DCM and the organic layer was collected, dried with sodium sulfate,
and evaporated to a red oil. The product proved to be too unstable to
purify or store effectively so the crude material was made as needed
and used immediately without further purification. To a solution of 1
(1.1 g, 3.0 mmol) in 100 mL of dichloromethane was added 30 mL of
trifluoroacetic acid. The solution was stirred for 20 min at room
temperature and then evaporated to dryness using a rotary evaporator
and kept under high vacuum overnight. The residue was a pale yellow
foam. This was slurried in 50 mL of dichloromethane and cooled in
an ice bath and then 1.3 mL of triethylamine (TEA) was added. Upon
addition of the TEA the mixture became clear. To this mixture was
added 0.70 g of the crude Troc-NCS and the solution was left to stir
in the ice bath for 10 min. Then the solution was washed with 100 mL
of water and the water layer was extracted with dichloromethane (2 ×
50 mL). The combined organic layers were dried over sodium sulfate
and concentrated to a foam using a rotary evaporator. The residue was
purified by silica gel flash column chromatography (5 × 20 cm column
with 75% to 100% gradient of ethyl acetate in hexanes). Fractions
containing the product, 6, were collected and evaporated under vacuum
to give 0.98 g of yellow foam (65%). TLC (3:1, EtOAc:hexanes) Rf )
1
0.72; H NMR (400 MHz DMSO-d6) δ (ppm) 1.78 (3 H, s, thymine-
CH3), 2.33 (m, 1 H; 2′-H), 2.46 (m, 1 H; 2′-H), 3.75 (dt, 1 H; 5′-H, J
) 4, 14 Hz), 4.00 (m, 1 H, 4′-H), 4.08 (m, 1 H; 5′-H), 4.38 (q, 1 H,
3′-H, J ) 7 Hz), 4.91 (dd, 2 H; Troc-CH2, J ) 12.4, 16.8 Hz), 6.10 (t,
J ) 7 Hz, 1 H; 1′-H), 7.47 (d, 1 H, 6-H, J ) 1 Hz), 9.85 (t, J ) 6 Hz,
1 H, 5′-NH), 11.37 (s, 1 H; thymine-NH), 11.70 (s, 1 H; Troc-NH);
13C NMR (100 MHz, DMSO-d6) δ 12.1, 35.2, 45.7, 60.5, 73.8, 80.1,
83.5, 94.9, 110.0, 135.8, 150.4, 151.9, 163.8, 180.0; IR (KBr pellet)
3044, 2953, 2830, 2499, 2116 (azide), 1700 (CdO), 1524 (CdS), 1470,
1369, 1276, 1218, 1077, 1036, 959, 738; MS (FAB) m/z 500 (M +
H)+; HRMS (FAB) m/z 500.006931, calcd for C14H16H7O5SCl3
500.00775.
5′-(N′-Trichloroethoxycarbonylthiourea)-3′-(9-fluorenylmethoxy-
carboxamide)-3′,5′-deoxythymidine (7). To a solution of 6 (0.25 g,
0.50 mmol) in 50 mL of 95% ethanol was added 10 mg of 10% Pd/C
and the mixture was hydrogenated at 50 PSI for 2 h then filtered through
Celite. The filtrate was evaporated to dryness using a rotary evaporator
and kept under high vacuum overnight. The solid residue was dissolved
in a mixture of 20 mL of dioxane and 10 mL of 10% sodium carbonate
in water. This mixture was cooled in an ice bath and 9-fluorenyl-
methylchloroformate (0.23 g, 0.89 mmol) was added slowly. The
mixture was left to stir in the ice bath for 1 h until TLC analysis
indicated that the reaction was complete. A crystalline precipitate had
appeared during the reaction and 10 mL of water were added until it
completely dissolved. Then the mixture was extracted with ethyl acetate
(3 × 100 mL) and the organic layer was dried with sodium sulfate and
evaporated to a white foam which was purified by silica gel flash
column chromatography (5 × 20 cm column with 0 to 5% gradient of
methanol in ethyl acetate). Fractions containing the product, 7, were
collected and evaporated to give 0.21 g of white foam (60% yield).
TLC (3:1, EtOAc:hexanes) Rf ) 0.35; 1H NMR (400 MHz DMSO-d6)
δ (ppm) 1.79 (3 H, s, thymine-CH3), 2.16 (m, 1 H; 2′-H), 2.30 (m, 1
1
(EtOAc) Rf ) 0.62; H NMR (400 MHz DMSO-d6) δ (ppm) 1.79 (3
H, s, thymine-CH3), 2.20 (m, 1 H; 2′-H), 2.33 (m, 1 H; 2′-H), 3.90 (m,
1 H; 5′-H), 4.00 (m, 2 H; 5′-H, 4′-H), 4.16 (t, J ) 8 Hz, 1 H; 3′-H),
4.22 (t, J ) 7 Hz, 1 H; Fmoc-CH), 4.35 (m, 2 H; Fmoc-CH2), 6.18 (t,
J ) 7 Hz, 1 H; 1′-H), 7.32 (t, J ) 7 Hz, 2 H; Fmoc-2′′-H), 7.40 (t, J
) 7 Hz, 2 H; Fmoc-3′′-H), 7.48 (t, J ) 8 Hz, 2 H, benzoyl-m-H), 7.58
(s, 1 H; 6-H), 7.61 (t, J ) 8 Hz, 1 H, benzoyl-p-H), 7.68 (m, 2 H;
Fmoc-1′′-H), 7.77 (d, J ) 8 Hz, 1 H; Fmoc-NH), 7.88 (d, J ) 7 Hz,
4 H; Fmoc-4′′-H, benzoyl-o-H), 11.06 (t, J ) 5 Hz, 1 H; 5′-NH), 11.36
(s, 1 H; thymine-NH), 11.46 (s, 1 H; Troc-NH); 13C NMR (100 MHz,
DMSO-d6) δ 12.1, 36.0, 46.7, 51.7, 65.5, 73.8, 80.7, 83.3, 94.9, 109.9,
120.1, 125.1, 127.1, 127.6, 136.2, 140.8, 143.8, 150.4, 152.0, 155.8,
163.7, 179.6; IR (KBr pellet) 3034, 2956, 2857, 1704 (CdO), 1658
(CdO), 1516 (CdS), 1449, 1252, 1167, 1077, 740; MS (FAB) m/z
626 (M + H)+; HRMS (FAB) m/z 626.20820, calcd for C33H31N5O6S
626.20733.
5′-(Aminoiminomethanesulfonic acid)-3′-azido-3′,5′-deoxythymi-
dine (4). To a 37% (w/w) solution of peracetic acid in acetic acid in
an ice bath was added, by spatula, 250 mg of 3. All the material
dissolved and over the following 10 min a white precipitate appeared.
The mixture was poured into 50 mL of cold ether and the precipitate
was collected by filtration and dried under vacuum. While solid (227
mg) was collected (79% yield). 1H NMR (400 MHz DMSO-d6) δ (ppm)
1.79 (3 H, s, thymine-CH3), 2.29 (m, 1 H; 2′-H), 2.34 (m, 1 H; 2′-H),
3.58 (m, 1 H; 5′-H), 3.62 (m, 1 H; 5′-H), 3.93 (dd, 1 H, 4′-H, J ) 6,
10 Hz,), 4.38 (m, 1 H, 3′-H), 6.08 (t, J ) 7 Hz, 1 H; 1′-H), 7.36 (d, 1
H, 6-H, J ) 1 Hz), 9.31 (s, 1 H, SO3H), 9.41 (s, 1H, imine-NH), 9.78
(t, J ) 6 Hz, 1 H, 5′-NH), 11.37 (s, 1 H; thymine-NH); 13C NMR (100
(35) Wang, S. S.; Magliocco, L. G. U.S. Patent, American Cyanimid
Company, 5194673, 1993.
(36) Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal.
Biochem. 1970, 34, 595.

3896 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Linkletter et al.
H; 2′-H), 3.83 (m, 1 H; 5′-H), 3.95 (m, 2 H; 5′-H, 4′-H), 4.10 (p, J )
7 Hz, 1 H; 4′-H), 4.22 (t, J ) 7 Hz, 1 H; Fmoc-CH), 4.35 (m, 2 H;
Fmoc-CH2), 4.91 (dd, 2 H; Troc-CH2, J ) 12, 16.8 Hz), 6.13 (t, J )
6.8 Hz, 1 H; 1′-H), 7.32 (t, J ) 7 Hz, 2 H; Fmoc-2′′-H), 7.40 (t, J )
7 Hz, 2 H; Fmoc-3′′-H), 7.52 (s, 1 H; 6-H), 7.69 (m, 2 H; Fmoc-1′′-
H), 7.75 (d, J ) 8 Hz, 1 H; Fmoc-NH), 7.88 (d, J ) 7 Hz, 2 H; Fmoc-
4′′-H), 9.83 (t, J ) 5 Hz, 1 H; 5′-NH), 11.35 (s, 1 H; thymine-NH),
11.66 (s, 1 H; Troc-NH); 13C NMR (100 MHz, DMSO-d6) δ 12.1, 36.0,
46.7, 51.7, 65.5, 73.8, 80.7, 83.3, 94.9, 109.9, 120.1, 125.1, 127.1, 127.6,
136.2, 140.8, 143.8, 150.4, 152.0, 155.8, 163.7, 179.6; IR (KBr pellet)
3302, 3179, 3047, 2957, 1703 (C ) O), 1683 (CdO), 1528 (CdS),
1451, 1257, 1218, 1080, 1036, 738; MS (FAB) m/z:: 696 (M + H)+
HRMS m/z 696.0844, calcd for C29H29N5O7SCl3 696.0853.
Solid-Phase Syntheses. (a) Synthesis of Benzoyl-Protected Guanidyl
Trimer (15). Rink amide resin (15 mg) was swelled in DMF for 2 h.
Then the resin was loaded with N-R-Fmoc-glycine by addition of 1.0
mL of a 0.6 M solution of the peptide in DMF and 1.0 mL of a 1.6 M
solution of DCC in DMF. The vial was agitated for 1 h and the resin
was filtered and washed. 5 (16 mg, 0.025 mmol) was dissolved in 1
mL of DMF and poured over the beads. Then 0.5 mL of a 50 mM
HgCl2 solution in DMF and 0.5 mL of a 100 mM TEA solution in
DMF were added rapidly and simultaneously to the monomer/resin
mixture in the reaction tube via two syringes. A fine white precipitate
formed immediately and the solution turned pale yellow. The tube was
capped tightly and agitated for 2 h at room temperature. After this
interval a black precipitate had replaced the white. The supernatant in
the reaction tube was removed by filtration. Most of the fine black
precipitate in this mixture passed through the filter. Repeated washing
with DMF (3 × 2 mL) was able to remove all visible precipitate, but
the resin beads were darkened from the precipitate contained within.
A solution of 10% piperidine in DMF was poured over the beads and
the reaction vial was agitated for 20 min. The supernatant was removed
by filtration and the beads were washed with DMF (3 × 2 mL). At
this point one nucleoside unit had been added to the resin and the 3′-
terminal amine deprotected (cleavage would give 10, see Scheme 3).
The coupling/deprotection cycle was repeated to give 11 and repeated
again to give 16. Resin samples were cleaved after each piperidine
deprotection step of the cycle to analyze the products. HPLC indicated
high coupling yields and FAB MS confirmed the presence of the
expected molecular ions for the products. Figure 1 shows the HPLC
chromatogram for each step of the trimer synthesis. m/z for 10, 444;
m/z for 11, 813; m/z for 16, 1182.
beads were briefly treated with a 20% thiophenol solution in DMF (2
mL for 10 s) and the beads instantly returned to their original yellow
color as the black precipitate was removed. The supernatant was
removed by filtration of the beads and the resin was washed with DMF
repeatedly (5 × 2 mL). The coupling/washing/deprotection cycle was
repeated seven more times to give the octameric thymidyl oligomer
with protected guanidinium linkages.
(c) Cleavage and Deprotection of Troc-Protected Guanidyl
Octamer. The resin was washed with DCM (2 × 2 mL), acetic acid
(2 × 2 mL), and methanol (2 × 2 mL) and dried under vacuum in the
presence of KOH desiccant. Then the beads were swelled in 2 mL of
DCM for 1 h. Then 0.5 mL of TFA was added and the beads
immediately turned dark red in color indicating the formation of the
chlorotrityl cation. The reaction tube was capped and agitated for 10
min then filtered. The filtrate was added dropwise to 30 mL of cold
ether. A white precipitate immediately formed and was collected by
centrifugation of the ether mixture. The supernatant was poured off
and the pellet was dried by a stream of nitrogen. Reverse-phase HPLC
showed this crude product to be approximately 75% pure indicating
an average coupling yield of 97%. This crude mixture was purified
preparative HPLC (10 × 250 mm Macrosphere 300A 7µ C8 silica
column, 0.1 M TEAA, pH 7.0 with a gradient of 0 to 80% acetonitrile
over 20 min, 3.0 mL/min) and the fractions containing the product were
evaporated using a centrifugal evaporator. The residue was then
dissolved in 2 mL of glacial acetic acid, 100 mg of cadmium powder
was added, and the mixture was vigorously shaken for 1 h. The mixture
was centrifuged and the supernatant was decanted and lyophilized.
Analytical cation exchange HPLC (4.6 × 100 mm Macrosphere 300A
7µ SCX column, buffer A: 50 mM Tris, pH 7.0; buffer B: 50 mM
Tris, 1 M guanidinium hydrochloride, gradient 0 to 50% B over 15
min, 1.3 mL/min) showed complete deprotection with no side products.
The deprotected product mixture was purified by cation exchange HPLC
on a 10 × 250 mm SCX column (10 × 250 mm Macrosphere 300A
7µ SCX column, 0 to 3 M ammonium acetate gradient over 15 min,
3.0 mL/min). The fractions containing the product were lyophilized,
redissolved in water, and lyophilized again to evaporate the ammonium
acetate buffer. The residue was dissolved in 2 mL of water and optical
density indicated that 1.6 mg of purified II was present.
Acknowledgment. This research was supported by grants
from the National Institute of Health and Genelabs Technologies,
Inc.
(b) Synthesis of Troc-Protected Guanidyl Octamer (16). Fifteen
milligrams of the commercially available 2% cross-linked polystyrene
2-(2-aminoethoxy)ethanol 2-chlorotrityl resin were placed in a screw-
capped reaction tube fitted with a coarse glass frit filter and a stopcock
and swelled in dry DMF for 2 h. Then the resin was filtered and enough
DMF was added to just barely cover the beads. 7 (22 mg) was placed
in a small vial and dissolved in 1 mL of dry DMF. This was transferred
via syringe to the reaction tube. Then 0.5 mL of a 50 mM HgCl2
solution in DMF and 0.5 mL of a 100 mM TEA solution in DMF were
added as described above. After 2 h the resin was filtered and washed
and the beads were gray from the mercury salt within. Then the resin
Supporting Information Available: Details of the initial
solid-phase synthesis via oxidized thiourea 4 is described in
detail and the cation exchange HPLC of purified DNG II and
CD spectra of a 2:1 solution of DNG II and DNA oligomers
V, VI, IX, and XV and their comparison to the CD spectrum
calculated from the CD spectra of the constituent oligomers
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA984212W