Synthesis and stability of exocyclic triazine nucleosides

Article abstract of DOI:10.1039/b503757c

Synthesis and stability studies of exocyclic amino triazine nucleosides were performed. Stability of the nucleosides was found to be dependent on triazine ring electron density, solvent, and pH. The nucleosides were found to be more stable when the triazine ring was electron deficient, in high pH aqueous solutions and in aprotic solvents. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b503757c

View Article Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
Synthesis and stability of exocyclic triazine nucleosides
Michelle Hysell,a Jay S. Siegel*b and Yitzhak Tor*a
a Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla,
California 92093-0358. E-mail: ytor@ucsd.edu
b Institute for Organic Chemistry, University of Zu¨rich, Winterthu¨rerstrasse 190,
CH-8006 Zu¨rich, Switzerland. E-mail: jss@oci.unizh.ch
Received 14th March 2005, Accepted 18th May 2005
First published as an Advance Article on the web 19th July 2005
Synthesis and stability studies of exocyclic amino triazine nucleosides were performed. Stability of the nucleosides
was found to be dependent on triazine ring electron density, solvent, and pH. The nucleosides were found to be more
stable when the triazine ring was electron deficient, in high pH aqueous solutions and in aprotic solvents.
Murchison meteorite6,8 (Fig. 2(a)). Given their possible presence
Introduction
at an early stage of the evolution of biomolecules, is it plausible
Exocyclic amino nucleosides (EANs) represent a class of bi-
that these triazine heterocycles could have contributed to the
ologically relevant nucleobases for which little stability and
basis for an early coding alphabet?
physical data are available (Fig. 1).1,2 They result from oxidative
damage to natural DNA, in the form of formamidopyrimidines
that have been implicated in mutagenesis.2 Secondary fungal
metabolites, such as clitocine, fit the EAN motif and act as
adenosine mimics that display anti-leukemic and insecticidal
activities (Fig. 1).1–3 Synthetic EANs like 4-substituted 8-(D-
ribofuranosylamino) pyrimido[5,4-d]pyrimidines show antitu-
mor and antiviral properties.4 These selected examples would
indicate that electron deficient heterocycles aid in forming stable
EAN systems. s-Triazine represents such an electron deficient
heterocycle, the synthesis and structure of which will follow.5
Fig. 2 Triazine based nucleosides. (a) Trimerization of hydrogen
cyanide to give triazine derivatives. (b) Nucleosides based on a triazine
core.
Fig. 1 Exocyclic amino nucleosides.
Functionalized triazines, derivatized through the ring nitro-
Another question which arises in the study of isosteric non-
natural bases is to what extent they may have played a role in
molecular evolution. Hydrogen cyanide, a likely component in
a primordial soup, is known to form s-triazines.6,7 Substituted
triazines have been formed under conditions designed to simu-
late the prebiotic world. Melamine, ammelide, and cyanuric acid
or precursors that hydrolyze to such triazines were found in the
gens, mimic pyrimidine like Cꢀ and Tꢀ and as EANs they could
mimic purines like Aꢀ and Gꢀ (Fig. 2(b)). Indeed, aminotriazines
and their hydrolysis products could offer isosteres of both
modern base families. Thus, in contrast to present day DNA,
which requires pyrimidine and purine heterocyclic cores for a
complete code, a triazine code would require only one. Whether
this is an advantage, remains a philosophical issue.9
2 9 4 6
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 9 4 6 – 2 9 5 2
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Article Online
Possible complications with EANs in aqueous media include
epimerization, ring expansion, and hydrolysis.10 The electron
deficient triazines could conceivably decrease the incidence of
isomerization.11 Synthesis and study of the physical properties
of exocyclic amino triazine nucleotides (EATNs) should provide
insight into this potential prebiotic DNA coding system.12,13 In
aqueous solutions, studies on the isomerization and hydrolysis
of EATNs show the dependence of such reactions on the solvent,
pH, and the electron density of the heterocyclic base. From these
studies, a general picture for what makes a stable EATN system
thus develops.
Results and discussion
Synthesis of the desired EATNs began by selective TIPS
protection of the primary hydroxyl in 2-deoxy-D-ribose to give
aldehyde 1 in 56% yield (Scheme 1). Condensation of 1 with
methylamine and subsequent nucleophilic aromatic substitution
with cyanuric chloride afforded triazine nucleoside 2 in 58% yield
as a mixture of anomers (1 : 1, a: b). The anomers were separated
by chromatography and the anomer 2-b was identified by an
observed NOE between H-1ꢀ and H-4ꢀ. From 2-b, a diverse set
of compounds can be prepared through nucleophilic aromatic
substitutions.
Fig. 3 Crystal structure of 8. Carbon, oxygen and nitrogen are shown
in white, red and blue, respectively. Hydrogens were omitted for clarity.
Torsional angles C9–N1–C5–N4 and C4–N1–C5–N5 were found to be
1◦ and 12◦ respectively.
Dichlorotriazine 2-b was treated with a saturated solution of
NH3 in tetrahydrofuran to give 3 as a mixture of equilibrating
rotamers in 83% yield.14,15 TBAF deprotection of 3 gave the
desired electron poor derivative 4 in 79% yield. Further nucle-
ophilic aromatic substitution on 3 using NaOMe in refluxing
MeOH afforded 5 in 93% yield;16 deprotection of 5 with TBAF
gave 6a in 66% yield. The electron rich adenosine mimic, 8, was
synthesized directly from 2 by double SNAr with methylamine
and subsequent TBAF deprotection (overall yield for 2 steps
86%). A crystal structure of 8 reaffirms the assignment of b to
the anomeric position in the starting material 2 (Fig. 3).
A comparison of the crystal structures of 2ꢀ-deoxyadenosine
and 8 showed that the two compounds are structurally similar.
Each contains the base in the anti conformation. The hetero-
cyclic moieties overlap well, maintaining the desired distances
between the furanose ring and the hydrogen bonding faces
allowing for Watson–Crick base pairing. The glycosidic nitrogen
in 8 is sp2 hybridized as indicated by bond angles (Fig. 3). It
follows that the lone pair is fully conjugated with the triazine
ring system. Additional support is seen with a short C◦5–N1 bond
◦
˚
distance of 1.37 A and torsional angles of 1 and 12 (Fig. 3).
Although structural similarity between 2ꢀ-deoxyadenosine
and 8 is important, its accommodation within a DNA duplex
is ultimately the key test. To incorporate nucleoside 8 into
DNA using standard solid phase oligonucleotide synthesis, the
protected precursor 12 was required (Scheme 1). TIPS protection
of 8 using TIPSTf in dichloromethane gave compound 9, which
Scheme 1 Synthesis of triazine nucleosides.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 9 4 6 – 2 9 5 2
2 9 4 7

View Article Online
was then protected using isobutyryl chloride in pyridine at 60 ◦C
to give 10 in 70% yield. TBAF deprotection and subsequent
DMT protection of the primary hydroxyl gave compound 12 in
80% yield (2 steps).
Nucleoside incorporation into DNA using standard protocols
requires the protected nucleoside be stable to oxidative, basic,
and acidic conditions. Compound 11 was stable to the oxidative
I2–pyridine conditions. However when 11 was briefly treated
with dichloroacetic acid in dichloromethane, simulating the
DMT deprotection protocol, significant degradation took place.
Although this acid instability is incompatible with standard
oligonucleotide synthesis protocols, there are alternative acid-
free routes for the synthesis of oligonucleotides.
Stability to aqueous environments is essential for testing
duplex stability and designing EATNs that structurally mimic
purines. Compound 11 was dissolved in D2O–THF-d8 and found
to be stable. The stability of 11, however, differed greatly from
that of the electron rich deprotected form 8, which was found to
decompose rapidly in water within 15 min, presumably through
the isomerization and hydrolysis path seen in the decomposition
of formamidopyrimidines.10 6a was then examined, as the tri-
azine moiety is intermediate in electron density, which provides a
slower, easier to study degradation. The 1H NMR of 6a after 30 h
in D2O showed significant degradation (Fig. 4(d)). A tentative
mechanism for isomerization and hydrolysis of exocyclic amino
triazine 6a is shown in Fig. 5.
To confirm that isomerization and hydrolysis were in fact the
occurring degradation pathways, pyranose 6c was independently
synthesized (Scheme 2). Condensation of 2-deoxy-D-ribose with
ammonia and subsequent nucleophilic aromatic substitution
with cyanuric chloride afforded triazine nucleoside 14 which was
treated with a saturated solution of NH3 in acetonitrile to give
15 in 79% yield. Further nucleophilic aromatic substitution on
15 using NaOMe in refluxing MeOH afforded 6c in 45% yield.
All protons and carbons of 6a and 6c were assigned using
COSY, DEPT, and HMQC data. The relative 13C NMR chemical
shift of C-4ꢀ distinguishes the difference between a pyranose and
furanose. As expected, the crowded C-4ꢀ in the furanose was
shifted 16 ppm downfield from that in the pyranose ring. In
1
addition, the large coupling constants seen in the H NMR
of pyranose 6c are consistent with a six-membered ring and
facilitated the assignment of 6c as the b isomer.17 The compound
synthesized and confirmed to be 6c was found to be identical to
the isomerization product of 6a seen in the 1H NMR.
With the rearrangement established, the effect of pH on the
triazine nucleosides was investigated. As suggested in Fig. 5,
tetrahydrofuran protonation is necessary for isomerization.
Proton NMR was used to study the stability of 6a at pH 6,
7, 8, and 11. Epimerization of 6a to 6b occurred rapidly relative
to isomerization as can be seen from the earlier appearance
of 6b relative to 2-deoxy-D-ribose and isomerization products
6c and 6d (Fig. 4). The pyranose forms 6c and 6d are the
thermodynamically favored product as is the case with the
formamidopyrimidines. A complete kinetic description of the
decomposition is complex due to the existence of several
competing reaction pathways: epimerization, ring expansion
and hydrolysis (Fig. 4); however, the half-lives can be estimated
by observing the decomposition to the point at which 50%
of the starting material was left. Compound 6a was found to
have a half-live of 10 h, 5 d, and ꢁ10 d at pH 6, 7, and 8
respectively. As expected from the proposed mechanism, the rate
of isomerization was found to increase at lower pH. In contrast
to previously studied formamidopyrimidines, which have been
shown to isomerize under basic conditions, 6a was found to be
completely stable at pH = 11. Unfortunately, 8 degraded within
15 min at pH 11. Thus, although pH can be used to control
isomerization in some systems, it is ineffective with electron rich
derivatives.
1
Fig. 4 Isomerization of 6a in D2O, pH = 6 followed by H NMR of
the anomeric protons at (a) t = 0, (b) t = 1 h, (c) t = 9 h, (d) t = 30 h.
Fig. 5 Proposed mechanism of epimerization.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 9 4 6 – 2 9 5 2
2 9 4 8

View Article Online
Scheme 2 Synthesis of pyranose 6c.
To assess solvent effects on stability, 6a was studied in water,
methanol, DMSO, acetonitrile, and acetone. Significant iso-
merization occurred in water as discussed above. Isomerization
was much slower in methanol (5% in 5 days). In the aprotic
solvents DMSO, acetonitrile, and acetone, no isomerization was
seen; thus, protic solvents seem to facilitate isomerization of
6a. With the electron rich compound 8, however, isomerization
was also seen in aprotic polar solvents. Degradation of 7
is less pronounced suggesting that when the 5ꢀ hydroxyl is
present, an intramolecular proton transfer might catalyze the
decomposition process.
In contrast to 8, the isostructural diaminopurine is stable to
aqueous conditions. Comparison of the structures showed that
they are similar which implies that an electronic effect must
be at the heart of the degradation path. For the molecule to
degrade (Fig. 5), a Schiff base intermediate must form requiring
a nucleophilic nitrogen. Whereas the nitrogen atom in position
7 of adenosine is sp2 hybridized and part of a cyclic conjugated
pi system, the exocyclic amine in 8 has a lone pair that is
interacting to a greater or lesser extent with the heterocycle.
In 8, the presence of the two electron donating methyl amino
substituents on the triazine makes the lone pair relatively free
and available to form a Schiff base intermediate.
To maintain structural integrity, it is important to know the
optimal electron density in the aromatic ring needed to prevent
isomerization. Compounds 8, 6a and 4, varying in electron
density, were investigated. Both 8 and 6a isomerize in water,
although electron rich 8 isomerizes more than a 100 times
faster than 6a at neutral pH. Only slight isomerization was seen
with the most electron poor nucleoside 4, which was found to
epimerize less than 10% in 7 d. One can imagine that the EATN
from diamino triazine would have ample stability at neutral pH
to survive for days in solution.
therefore likely to present a range of structure types with varying
degrees of stability, which may or may not parallel clitocine, the
compound they seek to mimic. Nevertheless, they may have
interesting biological activity, even if a direct structure activity
relationship between the triazine library and clitocine is not
indicated. Our results establish a clear caveat for the assumption
of EAN structure and stability based on a few electron deficient
model compounds. Indeed, all EANs synthesized should be
carefully analyzed to confirm their structural identity until a
better empirical base is in place.
The stability of these EATNs does not preclude the possibility
for a prebiotic code containing EANs that mimic purines. The
isomerization of the exocylic amino nucleosides could provide
for a library of pyranoses and furanoses. Rapid hydrolysis re-
mains an issue. To circumvent the isomerization and hydrolysis,
the exocyclic amino nucleoside need only contain a relatively less
nucleophilic exocyclic amine. Reducing the number of donor
groups on the ring is one way to achieve this effect. In that
regard, the more direct adenosine analog, the EATN from
diamino triazine, should have ample stability at neutral pH
to survive days in solution. The nucleophilicity also could be
tuned through the use of an electron withdrawing group directly
attached to the glycoside nitrogen. Alternatively, an electron
deficient heterocycle, as was seen with 4, could be used to control
the isomerization event. From a completely different perspective,
the triazine could have been part of a coding system containing
an alternative backbone where none of the problems inherent
to the glycoside bond would exist. In any event, these molecules
provide a fruitful area of discussion, rich in research results.
Experimental
1H- and 13C NMR spectra were recorded on a Varian Mercury
400 MHz spectrometer. High resolution mass spectra were
obtained by the Scripps Research Institute using either MALDI
or ESI. All reactions were performed in oven-dried round
bottom flasks under argon. Anhydrous pyridine and methanol
were purchased from Fisher Scientific. Anhydrous tetrahydro-
furan and dichloroethane were passed through a column of
activated alumina. Commercial chemicals were used as supplied
from Aldrich or Acros chemicals. Thin layer chromatography
was performed using alumina glass-backed plates coated with
0.25 mm F254 silica gel. TLC plates were visualized using
either UV light or charring after staining with CAM. For NMR
experiments, buffered D2O solutions were made using 0.1 mM
phosphate buffers. The pH = 11 D2O solution was made using
K2CO3. All 1H NMR experiments were performed using a 400
Varian MHz NMR with a delay time of 5 s. The following abbre-
viations were used: DIPEA, diisopropylamine; DMTrCl, 4,4ꢀ-
dimethoxytritylchloride; DMAP, 4-(dimethylamino)pyridine;
Conclusions
In conclusion, we have found that isomerization and hydrolysis
are important degradation pathways for exocyclic amino triazine
nucleosides, which has significant implications on the design and
synthesis of future EATNs. Clearly, the EATNs, which contain
electron rich aromatic rings can be expected to isomerize and
hydrolyze over a wide range of pHs and in a variety of solvents
resulting in conversion to the pyranose or 2-deoxy-D-ribose. In
contrast, systems with intermediate electron density isomerize to
the pyranose form only in protic solvents and can be maintained
in aqueous conditions at high pH. Only the electron poor
systems remain structurally intact at physiological pHs. This has
implications on the homogeneity of libraries based on EATNs.
For example, the 1152 triazine library synthesized at Ribafarm
Inc., spans a wide range of electron richness in the rings and is
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 9 4 6 – 2 9 5 2
2 9 4 9

View Article Online
pyr, pyridine; TBAF, tetrabutylammonium fluoride; TIPSTf,
triisopropylsilyltriflate.
column chromatography (66% ethyl acetate–hexanes) to afford
3 as a mixture of rotamers (1.50 g, 83%). Rf (50% hexanes in
ethyl acetate) = 0.3; IR film (tmax/cm−1) 3327, 3222, 2943, 2873,
1641, 1571, 1501, 1492; dH(400 MHz; CDCl3; CHCl3) 1.04–1.09
(42H, m) 1.82–1.86 (2H, m), 2.16–2.23 (2H, m), 2.96 (3H, s), 3.02
(3H, s), 3.71–3.91 (6H, m), 4.48 (2H, s), 5.91 (4H, br s), 6.65–6.72
(2H, m); dC(100 MHz; CDCl3; CHCl3) 11.8, 17.9, 25.5, 27.9, 36.5,
36.8, 64.2, 67.8, 72.8, 77.2, 84.4, 85.2, 165.8, 166.4, 166.8, 169.1,
169.6; HRMS (MALDI-FTMS) m/z calcd for C18H35ClN5O3Si
[M + H]+ 432.2192, found 432.2201.
5-O-Triisopropylsilyl-2-deoxy-D-ribose 1
To a solution of 2-deoxy-D-ribose (25.0 g, 0.186 mol) in
anhydrous pyridine (300 mL) was added triisopropylsilyl chlo-
ride (43.9 mL, 0.205 mol). The reaction was stirred at room
temperature for 4 days. The solution was concentrated under
reduced pressure and partitioned between ethyl acetate (600 mL)
and water (200 mL). The organic layer was separated then
washed with 0.05 M HCl(aq), NaHCO3(aq), and brine. The solution
was dried over Na2SO4, filtered, and concentrated. The crude
product was purified by flash column chromatography (33% to
50% hexanes in ethyl acetate) to afford the desired oil (30.5 g
of a,b, 56%). IR film (tmax/cm−1) 3389, 2943, 2866, 1461; major
isomer dH(400 MHz; CDCl3; CHCl3) 1.35–1.75 (21H, m), 1.87–
2.30 (2H, m), 2.97 (1H, s, NMe), 3.55–3.82 (2H, m, H-5ꢀ), 3.90–
3.99 (1H, m, H-4ꢀ), 4.40 (1H, br s, H-3ꢀ), 5.55 (1H, s, H-1ꢀ);
dC(100 MHz; CDCl3; CHCl3) 11.8, 17.9, 41.4, 64.0, 73.4, 87.9,
99.4; HRMS (MALDI-FTMS) m/z calcd for C14H30NaO4Si
[M + Na]+ 313.1805, found 313.1807; Rf (50% hexanes in ethyl
acetate) = 0.5.
6-Chloro-N2-methyl-N2-(2ꢀ-deoxy-b-D-erythro-pentofuranosyl)-
[1,3,5]triazine-2,4-diamine 4
To a solution of 3 (512 mg, 1.18 mmol) in tetrahydrofuran
(1 mL) was added 1.0 M TBAF in tetrahydrofuran (1.19 mL,
1.19 mmol). After 15 min the solution was concentrated and
the crude product was purified by flash chromatography (25%
acetonitrile in ethyl acetate) to afford a white solid (256 mg,
79%). dH(400 MHz; D2O) 2.32 (1H, dt, J 14.0 and 7.0), 2.94
(3H, s), 3.75 (1H, dd, J 12.0 and 4.0), 3.85 (1H, dt, J 8.4 and
4.6), 4.36 (1H, dt, J 6.3 and 3.5), 6.57 (1H, br s); dC(100 MHz;
D2O) 28.3, 35.6, 61.8, 71.1, 84.9, 85.1, 165.3, 166.4, 166.5, 167.6,
167.7; (MALDI-FTMS) m/z calcd for C9H15ClN5O4 [M + H]+
276.0863, found 276.0863.
4,6-Dichloro-N2-methyl-N2-(5ꢀ-O-triisopropylsilyl-2ꢀ-deoxy-a,b -
D-erythro-pentofuranosyl)-[1,3,5]triazin-2-yl-amine, 2a and 2b
4-Chloro-6-methoxy-N2-methyl-N2-(5ꢀ-O-triisopropylsilyl-2ꢀ-
deoxy-b-D-erythro-pentofuranosyl)-[1,3,5]triazin-2-yl-amine 5
5ꢀ-O-Triisopropylsilyl-deoxy-D-ribose (25.8 g, 89 mmol) was dis-
solved in 2.0 M methylamine in methanol (800 mL) and stirred at
room temperature for 16 h. The solution was concentrated under
reduced pressure and residual methanol was removed by adding
3 × 150 mL of tetrahydrofuran and concentrating. The resulting
oil was dissolved in tetrahydrofuran (800 mL) and cooled to
−78 ◦C. To the solution was added DIPEA (17.1 mL, 98 mmol)
and cyanuric chloride (12.63 g, 68.5 mmol). The reaction was
stirred at −20 ◦C for 14 h. The tetrahydrofuran was removed
under reduced pressure and the residue was partitioned between
ethyl acetate and brine. The organic layer was separated and
dried with Na2SO4, concentrated, and purified via flash column
chromatography (20% hexanes in ethyl acetate) to afford an oil
(23.3 g, 1 : 1 a : b, 58%). Ratio of anomers was determined by 1H
NMR. Separation of the epimers is accomplished using column
chromatography (20% hexanes in ethyl acetate) resulting in pure
fractions containing pure aor b and some mixed fractions of the
two.
2b (Rf = 75% hexanes in ethyl acetate = 0.5). IR
film (tmax/cm−1) 3414, 2943, 2867, 1612, 1565, 1484, 1417;
dH(400 MHz; CDCl3; CHCl3) 1.04–1.08 (21H, m), 2.14 (1H, ddd,
J 13.2, 6.6 and 3.8, H2ꢀ), 2.20–2.27 (1H, m, H2ꢀ), 3.10 (3H, s),
3.82 (1H, dd, J 10.4 and 4.8, H5ꢀ), 3.87–3.90 (1H, m, H5ꢀ), 3.95
(1H, dd, J 10.4 and 2.8, H4ꢀ), 4.50–4.53 (1H, m, H3ꢀ), 6.70 (1H,
t, J 6.8, H1ꢀ); dC(100 MHz; CDCl3; CHCl3) 11.7, 17.8, 28.4,
37.2, 84.7, 85.7, 164.9, 169.4, 170.3; HRMS (MALDI-FTMS)
m/z calcd for C18H32Cl2NaN4O3Si [M + Na]+ 473.1513, found
473.1525. 2a (Rf = 75% hexanes in ethyl acetate = 0.4). IR film
(tmax/cm−1) 3423, 2951, 2864, 1571, 1475, 1414; dH(300 MHz;
CDCl3; CHCl3) 1.04–1.08 (21H, m), 1.90 (1H, dt, J 13.5, 5.7),
2.62 (1H, dt, J 13.8, 6.9), 3.15 (3H, s), 3.71 (1H, dd, J 10.5 and
4.5), 3.79 (1H, dd, J 10.8 and 3.3), 4.08–4.12 (1H, m), 4.43–
4.51 (1H, m), 6.54 (1H, dd, J 7.2 and 6.8); dC(100 MHz; CDCl3;
CHCl3) 11.8, 17.9, 29.1, 38.7, 64.9, 72.3, 86.6, 87.0, 164.1, 169.1,
169.9; HRMS (MALDI-FTMS) m/z calcd for C18H33Cl2N4O3Si
[M + H]+ 451.1693, found 451.1704.
A solution of 3 (1.20 g, 2.77 mmol) and NaOMe (643 mg,
11.1 mmol) in methanol (20 mL) was refluxed for 14 h. The
solution was concentrated under vacuum and the crude product
was purified via column chromatography (66% ethyl acetate in
hexanes) to afford a solid (911 mg, 77%). IR film (tmax/cm−1)
3414, 2943, 2867, 1612, 1565, 1484, 1417; dH(400 MHz; CDCl3;
CHCl3) 1.00–1.20 (21H, m), 2.06–2.08 (1H, m), 2.16–2.23 (1H,
m), 2.99 (3H, s), 3.69–3.77 (3H, m), 3.91–3.94 (1H, m), 4.46 (1H,
m), 5.38 (2H, br s), 6.77 (1H, br s, J 6.8); dC(100 MHz; CDCl3;
CHCl3) 11.9, 18.0, 27.7, 36.5, 54.0, 64.3, 73.1, 84.2, 167.0, 167,6,
171.0; HRMS (MALDI-FTMS) m/z calcd for C19H38N5O4Si
[M + H]+ 428.2687, found 428.2686.
4-Amino-6-methoxy-N2-methyl-N2-(2ꢀ-deoxy-b-D-erythro-
pentofuranosyl)-[1,3,5]triazin-2-yl-amine 6a
To a solution of 5 (911 mg, 2.13 mmol) in tetrahydrofuran
(2 mL) was added 1.0 M TBAF in tetrahydrofuran (2.10 mL,
2.10 mmol). After 15 min the solution was concentrated and
purified via flash chromatography (50% acetonitrile in ethyl
acetate) to afford a white solid (380 mg, 66%). UV kmax(H2O)/nm
214; dH(400 MHz; CDCl3; CHCl3) 2.09 (1H, ddd, J 14.0, 6.6 and
3.4, H2ꢀ), 2.35 (1H, dt, J 14.0 and 7.4, H2ꢀ), 2.97 (3H, s, NMe),
3.70 (1H, dd, J 12.2 and 5.4, H4ꢀ), 3.77 (1H, dd, J 12.2 and
4.2, H5ꢀ), 3.87–3.88 (1H, m, H4ꢀ), 3.90 (3H, s, OMe), 4.39 (1H,
dt, J 6.9 and 3.4, H3ꢀ), 6.70 (1H, t, J 6.8, H1ꢀ); dC(100 MHz;
CDCl3; CHCl3) 25.6 (NMe), 33.0 (C2ꢀ), 52.2 (OMe), 59.6 (C5ꢀ),
68.9 (C3ꢀ), 82.4 (C4ꢀ), 82.8 (C1ꢀ), 164.7, 165.6, 168.6; HRMS
(MALDI-FTMS) m/z calcd for C10H18N5O4 [M + H]+ 272.1353,
found 272.1354; Rf (20% methanol in chloroform) = 0.1.
N4,N6-dimethyl-N2-methyl-N2-(5ꢀ-O-triisopropylsilyl-2ꢀ-deoxy-
b-D-erythro-pentofuranosyl)-[1,3,5]triazine-2,4,6-triamine 7
To a solution of 2b (4.16 g, 9.21 mmol) in tetrahydrofuran
(18 mL) was added 2.0 M methylamine in tetrahydrofuran
(55 mL). The solution was heated to 60 ◦C in a pressure
tube for 4 h. The solution was concentrated under reduced
pressure and the crude product was purified using flash column
chromatography (33% hexanes in ethyl acetate) to give an oil
(3.48 g, 86%). IR (tmax/cm−1) 3336, 3231, 2943, 2864, 1650,
1580, 1545; dH(400 MHz; CDCl3; CHCl3) 1.04–1.10 (21H, m),
2.08 (1H, m), 2.35 (1H, papp, J 6.8 Hz), 2.91 (6H, br s), 2.98
6-Chloro-N2-methyl-N2-(5ꢀ-O-triisopropylsilyl-2ꢀ-deoxy-b-D-
erythro-pentofuranosyl)-[1,3,5]triazin-2,4-diamine 3
A solution of 2b (1.90 g, 4.21 mmol) and DIPEA (0.808 mL,
4.63 mmol) in tetrahydrofuran (40 mL) was saturated with
NH3(g) at 0 ◦C. The reaction was stirred at room temperature
overnight, concentrated under vacuum and purified by flash
2 9 5 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 9 4 6 – 2 9 5 2

View Article Online
(3H, s), 3.70–3.74 (1H, m), 3.77–3.80 (1H, m), 3.98 (1H, dd, J
9.6 and 3.2), 4.38 (1H, dt, J 6.4 and 4.8), 4.84 (2H, br s), 6.88
(1H, br s); dC(100 MHz; CDCl3; CHCl3) 12.0, 18.0, 27.4, 27.6,
36.4, 64.7, 73.6, 83.6, 84.0, 85.1, 165.6, 166.2; HRMS (MALDI-
FTMS) m/z calcd for C20H40N6O3Si [M + H]+ 441.3004, found
441.3016.
N4,N6-diisobutyryl-N4,N6-dimethyl-N2-methyl-N2-(2ꢀ-deoxy-
b-D-erythro-pentofuranosyl)-[1,3,5]triazine-2,4,6-triamine 11
To a solution of 10 (1.72 g, 2.33 mmol) in tetrahydrofuran
(5 mL) was added 1.0 M TBAF in tetrahydrofuran (4.69 mL,
4.69 mmol). The reaction was stirred at room temp for 15 min,
concentrated and the crude product was purified via flash
chromatography (100% ethyl acetate) to afford an oil (873 mg,
88%). IR film (tmax/cm−1) 3441, 2960, 2871, 1684, 1572, 1529;
dH(400 MHz; CDCl3; CHCl3) 1.14–1.26 (m, 14H), 1.67 (1H,
br s), 1.90 (1H, br s), 2.14–2.20 (1H, m), 2.31–2.42 (1H, m), 3.06
(3H, s), 3.39 (6H, s), 3.71–3.87 (3H, m), 4.46 (1H, m), 6.76 (1H,
t, J 7.0); HRMS (MALDI-FTMS) m/z calcd for C19H38N5O4
[M + H]+ 425.2507, found 425.2500.
N4,N6-dimethyl-N2-methyl-N2-(2ꢀ-deoxy-b-D-erythro-
pentofuranosyl)-[1,3,5]triazine-2,4,6-triamine 8
To a solution of 7 (200 mg, 0.34 mmol) in tetrahydrofuran
(0.5 mL) was added 1.0 M TBAF in tetrahydrofuran (0.369 mL,
0.369 mmol). The solution was stirred for 15 min at room
temp, concentrated, and purified crude via flash chromatography
(100% ethyl acetate) using basic alumina to afford 8 (71 mg,
76%). 1H NMR shows some epimerization. Crystals were grown
by slow evaporation of a solution in 5% methanol in chloroform.
The crystal was dissolved in DMSO-d6 right before the NMR
was taken. dH(400 MHz; DMSO-d6) 1.62–1.75 (1H, m), 1.84–
2.18 (1H, m), 2.70 (6H, br s), 2.75–2.85 (3H, br s), 3.40 (1H,
br s), 3.52 (1H, br s), 4.05–4.12 (1H, m), 4.71 (1H, t, J 7.4),
5.04 (1H, d, J 3.6), 5.74 (1H, s), 6.55 (1H, br s), 6.90 (2H,
br s); HRMS (MALDI-FTMS) m/z calcd for C11H21N6O3 [M +
H]+ 285.1670, found 285.1667. Rapid degradation in solution
prevented further characterization.
N4,N6-diisobutyryl-N4,N6-dimethyl-N2-methyl-N2-(5ꢀ-O-(4,4ꢀ-
dimethoxytriphenylmethyl)-2ꢀ-deoxy-b-D-erythro-
pentofuranosyl)-[1,3,5]triazine-2,4,6-triamine 12
To a 0 ◦C solution of 11 (25 mg, 0.059 mmol) in dichloromethane
(0.3 mL) was added DIPEA (15.6 uL, 0.089 mmol) and DMTrCl
(22 mg, 0.066 mmol). The solution was warmed to room temp
and stirred for 2 h. The solution was concentrated and purified
using flash column chromatography (0.5% Et3N in 50% ethyl
acetate–hexanes) to afford a white solid (39 mg, 91%). IR film
(tmax/cm−1) 3463, 3058, 2965, 2872, 2836, 1735, 1685, 1607,
1562; dH(400 MHz; CDCl3; CHCl3) 1.16–1.22 (14H, m), 2.04
(1H, ddd, J 13.2, 6.0, and 3.2), 2.19–2.22 (1H, m), 3.02 (3H, s),
3.36 (3H, s), 3.72 (6H, s), 3.90 (2H, m), 3.95 (1H, dd, J 7.6 and
4.4), 4.40 (1H, m, J 3.2), 6.73 (1H, dd, J 8.0 and 6.0), 6.74–6.78
(4H, m), 7.15 (1H, d, J = 7.2), 7.19–7.24 (2H, m), 7.27–7.31 (4H
m), 7.41 (2H, t, J 6.8); dC(100 MHz; CDCl3; CHCl3) 14.0, 20.1,
28.4, 33.1, 36.9, 55.0, 63.9, 72.3, 85.3, 85.6, 86.1, 126.5, 127.4,
127.7, 129.7, 135.4, 144.4, 158.1, 165.2, 166.2, 166.7, 180.94,
180.97.
Crystal data. C11H22N6O4, M = 302.35, orthorhombic, a =
3
˚
˚
12.9893(12), b = 22.218(2), c = 5.2165(5) A, U = 1505.5(2) A ,
T = 100 K, space group P212121, Z = 4, l = 0.103 mm−1, 12910
reflections measured, 3420 unique (Rint = 0.034) which were used
in all calculations. The final wR (F2) was 0.136 [I > 2r(I)].†
N4,N6-dimethyl-N2-methyl-N2-(3ꢀ,5ꢀ-di-O-triisopropylsilyl-2ꢀ-
deoxy-b-D-erythro-pentofuranosyl)-[1,3,5]triazine-2,4,6-
triamine 9
6-Chloro-4-amino-N2-methyl-N2-(5ꢀ-O-triisopropylsilyl-2ꢀ-
deoxy-a-D-erythro-pentofuranosyl)-[1,3,5]triazin-2,4-diamine
(13 = 3a)14
To a solution of 7 (2.53 g, 5.74 mmol) in dichloromethane
(57 mL) in an ice bath was added DIPEA (1.10 mL, 6.31 mmol)
and TIPSTf (1.37 mL, 5.10 mmol). After 15 min the solution
was concentrated and the crude product was purified using flash
chromatography (33% hexanes in ethyl acetate) to give an oil
(3.42 g, 100%). IR film (tmax/cm−1) 3467, 3284, 2943, 2866, 1553,
1494, 1467, 1408; dH(400 MHz; CDCl3; CHCl3) 1.03–1.09 (42H,
m), 1.88–2.06 (1H, ddd, J 12.8, 5.6 and 1.8), 2.09 (1H, ddd, J
14.8, 9.2 and 5.6,), 2.89 (3H, s), 2.90 (3H, s), 2.98 (3H, s), 3.73
(1H, dd, J = 10.6 and 5.0), 3.82 (1H, dd, J 10.6 and 3.0), 3.87
(1H, br s), 4.57 (1H, dt, J 3.6 and 1.6), 4.97 (2H, br s), 6.90
(1H, br s); dC(100 MHz; CDCl3; CHCl3) 12.1, 18.09, 18.13, 27.6,
37.2, 64.0, 73.0, 84.5, 86.0, 165.7; HRMS (MALDI-FTMS) m/z
calcd for C29H61N6O3Si2 [M + H]+ 597.4344, found 597.4345.
A solution of 2a (0.870 g, 1.93 mmol) and DIPEA (0.37 mL,
2.12 mmol) in tetrahydrofuran (19 mL) was saturated with
NH3(g) at 0 ◦C. The reaction was stirred at room temperature
overnight, concentrated under vacuum and purified by flash
column chromatography (66% ethyl acetate–hexanes) to afford
13 as a mixture of rotamers (0.819 g, 98%). Rf (50% hexanes
in ethyl acetate) = 0.3; IR film (tmax/cm−1) 3333, 3225, 2943,
2865, 1641, 1566, 1507, 1467; dH(400 MHz; CDCl3; CHCl3)
1.03–1.08 (21H, m), 1.97–2.05 (1H, m), 2.52–2.61 (1H, m), 3.10
(3H, s), 3.70 (1H, dd, J 13.6 and 8.4), 3.88 (1H, dd, J 13.6
and 4.0), 4.03–4.08 (1H, m), 4.42–4.48 (1H, m), 5.25–5.40 (2H,
br s), 6.49 (1H, t, J 7); dC(100 MHz; CDCl3; CHCl3) 11.9, 17.8,
18.0, 29.9 (br), 37.8, 65.3, 73.7, 85.5, 86.0 (br), 165.0, 166.8 (br);
HRMS (MALDI-FTMS) m/z calcd for C18H35ClN5O3Si [M +
H]+ 432.2192, found 432.2197.
N4,N6-diisobutyryl-N4,N6-dimethyl-N2-methyl-N2-(3ꢀ,5ꢀ-di-O-
triisopropylsilyl-2ꢀ-deoxy-b-D-erythro-pentofuranosyl)-
[1,3,5]triazine-2,4,6-triamine 10
To a solution of 9 (2.24 g, 3.75 mmol) in 1,2-dichloroethane
(33 mL) was added DMAP (8.21 g, 67.2 mmol) and isobutyryl
chloride (1.49 mL, 14.1 mmol). The solution was heated at 60 ◦C
for 3 h then concentrated. The crude product was purified using
flash chromatography (15% ethyl acetate in hexanes) to afford an
oil (1.73 g, 63%). dH(400 MHz; CDCl3; CHCl3) 1.03–1.08 (44H,
m,), 1.20 (1H, t, J 4.8), 1.94 (1H, dd, J 12.8 and 4.8), 2.14–2.20
(1H, m), 3.06 (3H, s), 3.37 (3H, s), 3.39 (3H, s), 3.70–3.90 (3H,
m), 4.60 (1H, d, J 5.6), 6.74 (1H, dd, J 9.6, 5.6); dC(100 MHz;
CDCl3; CHCl3) 12.0, 18.0, 20.3, 28.1, 33.1, 34.8, 37.7, 63.9,
72.9, 84.7, 86.8, 165.5, 166.2, 166.7, 180.9; HRMS (MALDI-
FTMS) m/z calcd for C37H73N6O5Si2 [M + H]+ 737.5175, found
737.5145.
4,6-Dichloro-N2-methyl-N2-(2ꢀ-deoxy-a,b-D-erythro-
pentopyranosyl)-[1,3,5]triazin-2-yl-amine 14
2-Deoxy-D-ribose (2.02 g, 15.1 mmol) was dissolved in 2.0 M
methylamine in methanol (80 mL) and stirred at room temper-
ature for 16 h. The solution was concentrated under reduced
pressure. The resulting oil was dissolved in acetonitrile (80 mL)
and cooled to −20 ◦C. To the solution was added DIPEA
(2.9 mL, 16.6 mmol) and cyanuric chloride (3.06 g, 16.6 mmol).
The reaction was stirred at −20 ◦C for 24 hours. The acetonitrile
was removed and the crude was purified via flash column
chromatography (20% CH3CN in EtOAc) to afford an oil (1.47 g,
33%). IR (thin film) (tmax/cm−1) 3404, 2968, 2935, 2879, 1556,
1488, 1414; dH(400 MHz; DMSO-d6) 6.02 (1H, d, J 11.2), 3.98
(1H, br s), 3.61–3.54 (3H, m), 2.98 (3H, s), 2.49 (2H, br s), 1.98
(1H, dt, J 13.6 and 2.4), 1.68 (1H, d, J 10.4); dC(100 MHz;
† CCDC reference number 266147. See http://dx.doi.org/10.1039/
b503757c for crystallographic data in CIF or other electronic format.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 9 4 6 – 2 9 5 2
2 9 5 1

View Article Online
CD3CN) 29.4, 34.3, 66.2, 66.8, 67.6, 78.9, 165.4, 169.9, 170.5;
HRMS (MALDI-FTMS) m/z calcd for C9H12Cl2N4O3 [M +
Na]+ 317.0179 found 317.0182.
Chem., 1988, 31(4), 786; A. D. Baxter, C. R. Penn, R. Storer, N. G.
Weir and J. M. Woods, Nucleosides Nucleotides, 1991, 10(1–3), 393; P.
Franchetti, L. Cappellacci, G. Cristalli, M. Grifantini and S. Vittori,
Nucleosides Nucleotides, 1991, 10(1–3), 543; I. Kubo, M. Kim, W. F.
Wood and H. Naoki, Tetrahedron Lett., 1986, 27, 4277; C.-H. Lee,
J. F. Daanen, M. Jiang, H. Yu, K. L. Kohlhaas, K. Alexander, M. F.
Jarvis, E. L. Kowaluk and S. S. Bhagwat, Tetrahedron Lett., 2001, 11,
2419.
4 A. K. Ghose, Y. S. Sanghvi, S. B. Larson, G. R. Revankar and R. K.
Robins, J. Am. Chem. Soc., 1990, 112, 3622; A. K. Ghose, V. N.
Viswanadhan, Y. S. Sanghvi, L. D. Nord, R. C. Willis, G. R. Revankar
and R. K. Robins, Proc. Natl. Acad. Sci. USA, 1989, 86, 8242.
5 E. M. Smolin and L. RapoportS-Triazine and Derivatives, Inter-
science Publishers, New York, 1959.
6 R. Hayatsu, Science, 1964, 146, 1291; R. Hayatsu, Geochim. Cos-
mochim. Acta, 1975, 39, 471.
7 S. L. Miller and L. E. Orgel, The Origins of Life on the Earth, Prentice-
Hall, Englewood Cliffs, N. J., 1st edn., 1974.
8 R. D. Minard, P. G. Hatcher and R. C. Gourley, Origins Life Evol.
Biosphere, 1998, 28(4–6), 461; H. Cottin, M. C. Gazeau and F. Raulin,
Planet. Space Sci., 1999, 47(8–9), 1141.
N4-amino-6-chloro-N2-methyl-N2-(2ꢀ-deoxy-b-D-erythro-
pentopyranosyl)-[1,3,5]triazin-2-yl-amine 15
A solution of 14 (0.900 g, 3.05 mmol) and DIPEA (0.604 mL,
3.35 mmol) in CH3CN (30 mL) was saturated with NH3(g) at
◦
0 C. The reaction was stirred at room temperature overnight,
concentrated under vacuum and purified by flash column
chromatography (100% CH3CN) to afford 15 (0.662 mg, 79%).
dH(400 MHz; CD3OD) 1.01 (1H, dd, J 3.6, 2.4), 1.77 (1H, dd,
J 2.8, 2.4), 2.98 (3H, s), 3.65–3.70 (2H, m), 3.77–3.79 (1H, m),
4.14 (1H, d, J 6.0), 6.25 (1H, br s); dC(100 MHz; CD3OD) 28.7,
35.2, 35.5, 66.5, 68.0, 68.5, 79.0, 166.6, 167.8, 168.2, 170.0, 170.3;
HRMS (ESI-TOF) m/z [M + H]+ calcd for C9H14ClN5O3 for
276.0863, found 276.0863.
9 J. S. Siegel and Y. Tor, Org. Biomol. Chem., 2005, 3, 1591.
10 S. Raoul, M. Bardet and J. Cadet, Chem. Res. Toxicol., 1995, 8(7),
924; M. Berger and J. Cadet, Z. Naturforsch., B: Chem. Sci., 1985,
40(11), 1519.
N4-amino-6-methoxy-N2-methyl-N2-(2ꢀ-deoxy-b-D-erythro-
pentopyranosyl)-[1,3,5]triazin-2-yl-amine 6c
A solution of 15 (0.662 g, 2.40 mmol) and NaOMe (1.30 g,
24.3 mmol) in methanol (20 mL) was refluxed for 24 hours. The
solution was concentrated under vacuum and the crude product
was purified via column chromatography (100% CH3CN) to
afford a solid (292 mg, 45%). NMR taken in D2O, pH = 11. UV
kmax (H2O)/nm 215; dH(400 MHz; D2O) 1.92 (1H, d, J 14.0), 2.32
(1H, td, J 12.8, 2.4, H-2ꢀ), 3.00 (3H, s, NMe), 3.78–3.86 (3H, m,
H-4ꢀ, H-5ꢀ), 3.91 (3H, s, OMe), 4.29 (br s, 1H, H-3ꢀ), 6.15 (1H,
d, J 11.6, H-1ꢀ); dC(400 MHz; D2O) 28.6, 33.9, 54.6, 64.8, 66.4,
67.2, 78.6, 166.7, 167.7, 171.8; HRMS (MALDI-FTMS) m/z
calcd for C10H17N5O4[M + H]+ 272.1353, found 272.1359.
11 K. Haraguchi and M. M. Greenberg, J. Am. Chem. Soc., 2001,
123, 8636; K. Haraguchi, M. O. Delaney, C. J. Wiederholt, A.
Sambandam, Z. Hantosi and M. M. Greenberg, J. Am. Chem. Soc.,
2002, 124, 3263.
12 In addition to our studies reported here, triazines Cꢀ and Tꢀ have
previously been synthesized: A. Piskala, M. Masojidkova and D.
Saman, Collect. Czech. Chem. Commun., 1996, 61; A. Piskala, N. B.
Hanna, M. Masojidkova, M. Otmar, P. Fiedler and K. Ubik, Collect.
Czech. Chem. Commun., 2003, 68(4), 711; A. Piskala, N. B. Hanna,
M. Masojidkova, P. Fiedler and I. Votruba, Collect. Czech. Chem.
Commun., 2004, 69(4), 905.
13 Recently, Ribafarm, Inc. published the synthesis of a library of
s-triazine derivatives designed to mimic the structure of clitocine,
butlittle characterization of the individual components was reported.
See: C. V. Varaprasad, Q. Habib, D. Y. Li, J. F. Huang, J. W. Abt, F.
Rong, Z. Hong and H. Y. An, Tetrahedron, 2003, 59, 2297.
14 To ensure that the complex spectra were due to rotamers and not
epimers, the alpha anomer of 3 (13) was synthesized and shown to
have different spectra.
15 Triazine rotamers have been reported, see: H. E. Birkett, R. K. Harris,
P. Hodgkinson, K. Carr, M. H. Charlton, J. C. Cherryman, A. M.
Chippendale and R. P. Glover, Magn. Reson. Chem., 2000, 38(7),
504; A. R. Katritzky, D. C. Oniciu, I. Ghiviriga and R. A. Barcock,
J. Chem. Soc., Perkin Trans. 2, 1995, 4, 785.
Acknowledgements
This work was supported by the National Institutes of Health
(grant number GM 069773 to YT), the National Science
Foundation (CHE 0213323 to JSS and YT) and the Swiss
National Science Foundation (to JSS).
References
1 T. Kamikawa, S. Fujie, Y. Yamagiwa, M. Kim and H. Kawaguchi,
J. Chem. Soc., Chem. Commun., 1988, 3, 195.
2 M. A. Graziewicz, T. H. Zastawny, R. Olinski and B. Tudek, Mutat.
Res.-DNA Repair., 1999, 434(1), 41.
3 R. J. Moss, C. R. Petrie, R. B. Meyer, L. D. Nord, R. C. Willis,
R. A. Smith, S. B. Larson, G. D. Kini and R. K. Robins, J. Med.
16 No rotamers were observed due to free rotation about the glycoside–
nitrogen triazine bond.
17 Both 6a and 6c are identical by MALDI-FTMS supporting that they
are isomers. kmax of 6a and 6c are also identical indicating that the
triazine moiety remains intact.
2 9 5 2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 9 4 6 – 2 9 5 2