A nucleotide dimer synthesis without protecting groups using montmorillonite as catalyst

Article abstract of DOI:10.1080/15257770.2012.701787

A synthesis has been developed providing nucleotide dimers comprising natural or unnatural nucleoside residues. A ribonucleoside 5-phosphorimidazolide is added to a nucleoside adsorbed on montmorillonite at neutral pH with the absence of protecting groups. Approximately 30% of the imidazolide is converted into each 2-5 dimer and 3-5 dimer with the rest hydrolyzed to the 5-monophosphate. Experiments with many combinations have suggested the limits to which this method may be applied, including heterochiral and chimeric syntheses. This greener chemistry has enabled the synthesis of dimers from activated nucleotides themselves, activated nucleotides with nucleosides, and activated nucleotides with nucleotide 5-monophosphates.

Full text of DOI:10.1080/15257770.2012.701787

This article was downloaded by: [NUS National University of Singapore]
On: 16 June 2013, At: 20:19
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Nucleosides, Nucleotides and Nucleic
Acids
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/lncn20
A Nucleotide Dimer Synthesis Without
Protecting Groups Using Montmorillonite
as Catalyst
Prakash C. Joshi a , Michael F. Aldersley a , Dmitri V. Zagorevskii b
James P. Ferris a
a Department of Chemistry & Chemical Biology and The New York
&
Center for Astrobiology, Rensselaer Polytechnic Institute, Troy, New
York, 12180, USA
b Center for Biotechnology and Interdisciplinary Studies, Rensselaer
Polytechnic Institute, Troy, New York, 12180, USA
Published online: 01 Aug 2012.
To cite this article: Prakash C. Joshi , Michael F. Aldersley , Dmitri V. Zagorevskii & James P. Ferris
(2012): A Nucleotide Dimer Synthesis Without Protecting Groups Using Montmorillonite as Catalyst,
Nucleosides, Nucleotides and Nucleic Acids, 31:7, 536-566
To link to this article: http://dx.doi.org/10.1080/15257770.2012.701787
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.

Nucleosides, Nucleotides and Nucleic Acids, 31:536–566, 2012
Copyright ꢀC Taylor and Francis Group, LLC
ISSN: 1525-7770 print / 1532-2335 online
DOI: 10.1080/15257770.2012.701787
A NUCLEOTIDE DIMER SYNTHESIS WITHOUT PROTECTING
GROUPS USING MONTMORILLONITE AS CATALYST
Prakash C. Joshi,1 Michael F. Aldersley,1 Dmitri V. Zagorevskii,2
and James P. Ferris1
1Department of Chemistry & Chemical Biology and The New York Center for Astrobiology,
Rensselaer Polytechnic Institute, Troy, New York 12180, USA
2Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute,
Troy, New York 12180, USA
A synthesis has been developed providing nucleotide dimers comprising natural or unnatu-
2
ral nucleoside residues. A ribonucleoside 5ꢁ-phosphorimidazolide is added to a nucleoside adsorbed
on montmorillonite at neutral pH with the absence of protecting groups. Approximately 30% of
the imidazolide is converted into each 2ꢁ-5ꢁ dimer and 3ꢁ-5ꢁ dimer with the rest hydrolyzed to the
5ꢁ-monophosphate. Experiments with many combinations have suggested the limits to which this
method may be applied, including heterochiral and chimeric syntheses. This greener chemistry has
enabled the synthesis of dimers from activated nucleotides themselves, activated nucleotides with
nucleosides, and activated nucleotides with nucleotide 5ꢁ-monophosphates.
[Supplemental materials are available for this article. Go to the publisher’s online edition
of Nucleosides, Nucleotides & Nucleic Acids to view the free supplemental files.]
Keywords Montmorillonite; phosphorimidazolide; RNA dimers; green chemistry
INTRODUCTION
The montmorillonite catalyzed oligomerization of 5ꢁ-activated nu-
cleotides leads to oligomers up to 50 residues in length.[1,2] However, all
oligomers must necessarily begin as dimers, so we considered it important to
study in detail the formation of these products under prebiotic conditions.[3]
Received 9 March 2012; accepted 6 June 2012.
5-Fluoro-uridine was a gift from the Dayang Chemical Company. D-Xyloadenosine was a gift from
Dr Gilles Gosselin. Volclay SPV-200 was a gift from the American Colloid Co., Arlington Heights, IL.
This research was supported by NASA Astrobiology Institute grant NNA09DA80A. The purchase of the
Bruker Ultraflex III MALDI-TOF-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) used
in this work was made possible by the Gen∗NY∗ sis (Generating Employment through New York State
Science) program.
Address correspondence to Michael F. Aldersley, Department of Chemistry & Chemical Biology and
The New York Center for Astrobiology, Rensselaer Polytechnic Institute, Troy, NY 12180, USA. E-mail:
alderm@rpi.edu
536

Nucleotide Dimer Synthesis
537
TABLE 1 Pseudo first order rate constants, h−1, for the formation of dimers from D-AMP and D,
L-ImpA
Dimer, k1, h−1
Dimer Product
0.021ꢁ
0.014ꢁ
0.076ꢁ
0.054ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
D,D-A2 -5 pA
D,L-A2 -5 pA
D,D-A3 -5 pA
D,L-A3 -5 pA
Then a meaningful comparison could be drawn between our syntheses and
the formation of long oligomers that is part of our studies of the origins of
life.[4] In the synthesis of trimers from these dimers, we looked for alterna-
tive synthetic methods via a 5ꢁ-phosphate dimer with activated nucleotides
as well as 5ꢁ-hydroxy nucleotide dimers with the same reactant.
Yields of the materials were important and we sought to optimize both
conditions and reactants in this regard, because clearly there would be a
pressing need for dimer starting material for any given trimer nucleotide
synthesis. Moreover, the dimer and trimer materials synthesized were fre-
quently used as reference compounds in our detailed product analysis of
the montmorillonite catalyzed RNA syntheses.[5] Preliminary experiments
with 5ꢁ-phosphates and activated nucleotides yielded generic pMpN dimers,
where M and N are nucleoside residues, both 2ꢁ-5ꢁ and 3ꢁ-5ꢁ (Figure 1),
whereas the corresponding nucleoside provided MpN dimers, both 2ꢁ-5ꢁ and
3ꢁ-5ꢁ. In the absence of Na+-montmorillonite, no dimers were detected. How-
ever, the rate of reaction of the 5ꢁ-phosphate with the activated nucleotide
was very slow compared with the corresponding nucleoside (Tables 1–3)
and the slow reaction was accompanied by extensive hydrolysis (>95%) of
the activated nucleotide.
These studies were extended to considerations of homochirality by using
an enantiomeric activated nucleotide, D, L-ImpA (Figure 2). An enzymatic
hydrolysis of the 5ꢁ-phosphate group of the pMpN by alkaline phosphatase
(APH) followed at pH 8 to provide dimers MpN and comparisons were made
in yield, rate, and selectivity with the corresponding reactions of nucleosides
and activated nucleotides (Tables 1–3).
A further notable feature of this method is the absence of protecting
groups, thus simplifying both product isolation and product identification.
Montmorillonite (Figure 3) has been used as a catalyst[6] and as a support
in other synthetic processes.[7] Nucleophilic substitution reactions using an
embedded crown[8] and the diverse uses of montmorillonite K-10 are notable
examples.
TABLE 2 Pseudo first order rate constants, h−1, for the formation of dimers and the second order rate
constants for their subsequent loss to trimers from D-adenosine and D, L-ImpA
Dimer, k1, h−1
0.85
4.1ꢁ5 ꢁ
0.49
0.62
4.9ꢁ2 ꢁ
0.59
Trimer, k2, M−1h−1
Dimer Product
3.59 ꢁ
3.62 ꢁ
ꢁ
ꢁ
D,D-A2 -5 pA
D,L-A2 -5 pA
D,D-A3 -5 pA
D,L-A3 -5 pA

538
P. C. Joshi et al.
FIGURE 1 Top: Synthesis of D, D-pApA dimers from D-AMP and D-ImpA. Bottom: Synthesis of D, D-ApA
dimers from D-adenosine and D-ImpA. The D, D-pApA dimers can be dephosphorylated to the D, D-ApA
dimers with APH at pH 8. (Color figure available online).
TABLE 3 The relative rates for individual dimer formation
From D-adenosine
From D-AMP
Relative Rate
1.38
0.27
41
0.8
0.18
36
1
1
8
0.95
0.71
11
Adenosine/AMP
Dimer Product
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
D,D-A2 -5 pA
D,L-A2 -5 pA
D,D-A3 -5 pA
D,L-A3 -5 pA

Nucleotide Dimer Synthesis
539
FIGURE 2 Synthesis of D, L and D, D-ApA dimers from D-AMP/D-adenosine and D, L-ImpA. (Color
figure available online).
FIGURE 3 Structure of montmorillonite.[4] (Color figure available online).

540
P. C. Joshi et al.
We propose this method of dimer synthesis as a greener alternative to
other more standard methods, e.g., those involving highly protected phos-
phoramidates.[9,10]
RESULTS
Dimers from D-AMP or D-Adenosine and an Activated Nucleotide
Syntheses using D-AMP with D, L-ImpA (23/24) (Figure 1 and Figure 4)
followed by removal of the 5ꢁ-phosphate revealed that the yield of any given
dimer was greater for the faster reactions with D-adenosine (and fastest for
formation of homochiral 2ꢁ-5ꢁ-dimer) compared with D-AMP (Figures 4 and
5, Tables 1 and 2). Relative rates are shown in Table 3. Thus, the synthesis
for dimers of the type MpN is more efficient than that for dimers of the type
pMpN.
Consequently, the mono-nucleotides (NMP) formed by hydrolysis of the
activated nucleotide, ImpN, do not subsequently react to any great extent to
form dimers pNpN with the same activated nucleotide. Such dimers are only
detected in our HPLC analyses of the reaction mixtures in which a low yield
of dimers is being formed. In these cases, where the activated nucleotide
does not react well with the nucleoside, the usual oligomerization products
FIGURE 4 The comparative reactions of D, L-ImpA with D-AMP. The initial product, pApA, was de-
phosphorylated to form isomers of ApA. (Color figure available online).

Nucleotide Dimer Synthesis
541
FIGURE 5 The comparative reactions of D, L-ImpA with D-adenosine to form ApA isomers. (Color
figure available online).
are also formed.[1,2] It is clear from Table 3 that considerable regioselectivity
and stereoselectivity occur.
Nucleoside Dimers Containing A, U, G, and/or C
Conditions have been optimized for each type of standard reaction (Fig-
ure 6a, X = Y = OH). Approximately 30% of the imidazolide is converted
into the 2ꢁ-5ꢁ dimer (Figure 7), 30% into the 3ꢁ-5ꢁ dimer (Figure 8) with
the rest hydrolyzed to the corresponding 5ꢁ-monophosphate (Figure 9).
Unusual mass spectrometry data for products from the oligomerization of
FIGURE 6 Dimer formation from: (a) a ribonucleoside (X = OH, Y = OH) or (b) ribonucleoside 5ꢁ-
monophosphate (X = OH, Y = OPO3H2) or (c) chimeric dimer formation (X = H, Y = OH) from a
deoxynucleoside and ImpN (N = A, U, G, C).

542
P. C. Joshi et al.
FIGURE 7 Yields of 2ꢁ-5ꢁ-products from our standard dimer syntheses. Data in Table 5.
ImpG (27) alone, especially showing elevated intensities of even mass nu-
cleotides (Figure 10a) and very high intensity peaks for tetramer and octamer
(Figure 10b), and the comparatively very low dimer yields with ImpG
(Table 5) prompted a more detailed investigation of this activated nucleotide
(see Section “The ImpG Anomaly”).
Other Ribonucleosides, Deoxyribonucleosides, and Chimeric
Dimers
Experiments with D-deoxynucleosides, D-ImpdA (30), and other non-
standard components have suggested some limits of this method, although
FIGURE 8 Yields of 3ꢁ-5ꢁ-products from our standard dimer syntheses. Data in Table 5.

Nucleotide Dimer Synthesis
543
FIGURE 9 Yields of 5ꢁ-NMP from our standard dimer syntheses. Data in Table 5.
the reaction is essentially the same for the formation of chimeric[11] prod-
ucts (Figure 6c). For example, D-inosine (10) readily forms dimers but D-aza-
nucleosides (8, 9) and D-xanthosine (11) do not. Neither D-deoxy-adenosine
(12) nor D-deoxy-guanosine (13) reacted with D-ImpdA (30). D-2ꢁ-methoxy
adenosine (22) did not react with any ImpN. Heterochiral syntheses from
enantiomeric nucleosides have been successfully carried out with methods
developed for the optimized syntheses of all dimers from D- and L-adenosine
and D- and L-uridine. Noteworthy are the differing yields of any given ho-
mochiral dimer involving A or U compared with its heterochiral counterpart.
Xylo- and arabino-nucleosides[12] were tested under identical conditions to
their ribo-counterparts but failed to react. Similarly, a selection of deoxynu-
cleosides was used in attempts to synthesize chimeric products.[11] Table 4
provides the list of materials used throughout this investigation. Table 5 pro-
vides an indication of the types of dimer synthesis explored, together with
the yields.
DISCUSSION
Scope and Versatility of the Synthesis
One of the starting materials for this dimer synthesis is a nucleoside
(Table 4 and Figure 11) and the other is an activated nucleotide, a 5ꢁ-
phosphorimidazolide (Table 4 and Figure 12). The nucleosides that have
been used in combination with activated nucleotides are shown in Table 5.
In all cases where a reaction occurred, two major products were formed from
the ribonucleoside and these were readily detected, identified by HPLC, and
isolated by preparative reverse phase HPLC (e.g., Figure 13). The products

544
P. C. Joshi et al.
FIGURE 10 (a) MALDI-MS of oligomers, (pG)n from ImpG in the presence of a lithium buffer and
Li-montmorillonite at pH = 8.00. Negative ion linear mode. (b) MALDI-MS of selected cyclic oligomers,
c-(pG)n from ImpG in the presence of lithium buffer and Li+-montmorillonite at pH = 8.00. Negative
ion reflector mode. (Color figure available online).

Nucleotide Dimer Synthesis
545
TABLE 4 Nucleosides and activated nucleotides
Nucleoside
Activated nucleotide
1
2
3
4
5
6
7
8
9
D-Adenosine
L-Adenosine
D-Uridine
L-Uridine
D-Guanosine
D-Cytidine
D-5-Fluorouridine
D-5-Aza-cytidine
D-6-Aza-uridine
D-Inosine
23
24
25
26
27
28
29
30
D-ImpA
L-ImpA
D-ImpU
L-ImpU
D-ImpG
D-ImpC
D-ImpF-U
D-ImpdA
10
11
12
13
14
15
16
17
18
19
20
21
22
D-Xanthosine
D-2ꢁ-Deoxy-adenosine
D-2ꢁ-Deoxy-guanosine
D-Arabino-cytidine
D-Iododeoxyuridine
D-Ribavirin
D-Trifluorothymidine
D-Arabino-adenosine
D-Xylo-adenosine
D-Doxifluridine
D-Floxuridine
D-2ꢁ-OCH3 Adenosine
ꢁ
ꢁ
ꢁ
ꢁ
are the M2 -5 pN and M3 -5 pN isomers where M is derived from the nucleoside
and N is derived from the activated nucleotide. For example, ꢁthe synthesis
ꢁ
of (48) from D-adenosine (1) and L-ImpU (26) gives D, L-A2 -5 pU and D,
ꢁ
ꢁ
L-A3 -5 pU.
The dimers always elute in the order 2ꢁ-5ꢁ and then 3ꢁ-5ꢁ (see the Experi-
mental section and Figure 13). Deoxyribonucleosides gave only one chimeric
product when a reaction occurred with an activated ribonucleotide. Simi-
larly, a ribonucleoside gave two chimeric products with the activated deoxyri-
bonucleotide, D-ImpdA (30): with D-inosine (10), D-ImpdA (30) gave the
isomeric 2ꢁ-5ꢁ and 3ꢁ-5ꢁproducts (74) in 22% and 26% yields, respectively.
For other examples, see Table 5, Part C.
The Role of Montmorillonite in the Catalytic Process
The mechanism for nucleotide dimer synthesis invokes general
acid/base catalysis and the presence of protons to make the imidazole a bet-
ter leaving group (Figure 14) in the same way as the synthesis of oligomers
directly from an activated nucleotide, D-ImpN.[13] The relative absence of
protons in the sodium-saturated clay, compared with pH 7 montmorillonite,
results in the absence of catalytic activity in all our work with the clay cat-
alysts. In addition, the clay layers constrain the adsorbed materials to two-
dimensional movements.

546
P. C. Joshi et al.
TABLE 5 Summary of the identities of nucleotide dimers prepared (∗optimized)
Nucleotide
Nucleoside
ImpN
% 2ꢁ-5ꢁ
%3ꢁ-5ꢁ
Part A
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
D,D-ApA∗
D,D-ApU∗
D,D-ApC
D,D-ApG
D,D-UpA∗
D,D-UpU∗
D,D-UpC
D,D-UpG
D,D-CpA
D,D-CpU
D,D-CpC
D,D-CpG
D,D-GpA
D,D-GpU
D,D-GpC
D,D-GpG
1
1
1
1
3
3
3
3
6
6
6
6
5
5
5
5
D-Adenosine
D-Adenosine
D-Adenosine
D-Adenosine
D-Uridine
D-Uridine
D-Uridine
D-Uridine
D-Cytidine
D-Cytidine
D-Cytidine
D-Cytidine
D-Guanosine
D-Guanosine
D-Guanosine
D-Guanosine
23
25
28
27
23
25
28
27
23
25
28
27
23
25
28
27
D-ImpA
D-ImpU
D-ImpC
D-ImpG
D-ImpA
D-ImpU
D-ImpC
D-ImpG
D-ImpA
D-ImpU
D-ImpC
D-ImpG
D-ImpA
D-lmpU
D-ImpC
D-ImpG
32
30
51
4.8
34
46
53
8.3
40
46
46
6.1
19
38
36
5
32
16
22
6.1
27
21
22
15
29
15
22
8
13
18
18
4.8
Part B
Part C
47
48
49
50
D,L-ApA∗
D,L-ApU∗
D,L-UpA∗
D,L-UpU∗
1
1
3
3
D-Adenosine
D-Adenosine
D-Uridine
24
26
24
26
L-ImpA
L-ImpU
L-ImpA
L-ImpU
19
52
22
25
46
14
24
35
D-Uridine
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
D,D-ApF-U
D,D-UpF-U
D,D-CpF-U
D,D-GpF-U
D,D-F-UpA
D,D-F-UpU
D,D-F-UpC
D,D-F-UpG
D,D-F-UpF-U
D,D-dApA
D,D-dApU
D,D-dApC
D,D-dApG
D,D-dApF-U
D,D-dGpA
D,D-dGpU
D,D-dGpC
D,D-dGpG
D,D-dGpF-U
D,D-IpA
D,D-IpU
D,D-IpC
D,D-IpG
D,D-IpdA
D,D-ApdA
D,D-UpdA
D,D-F-UpdA
D,D-CpdA
D,D-GpdA
D,D-RibpA
D,D-RibpU
1
3
6
5
7
7
7
7
7
12
12
12
12
12
12
12
12
12
12
10
10
10
10
10
1
D-Adenosine
D-Uridine
D-Cytidine
D-Guanosine
29
29
29
29
23
25
28
27
29
23
25
28
27
29
23
25
28
27
29
23
25
28
27
30
30
30
30
30
30
23
25
D-ImpF-U
D-ImpF-U
D-ImpF-U
D-ImpF-U
D-ImpA
D-ImpU
D-ImpC
D-ImpG
D-ImpF-U
D-ImpA
D-ImpU
D-ImpC
D-ImpG
D-ImpF-U
D-ImpA
D-ImpU
D-ImpC
D-ImpG
D-ImpF-U
D-ImpA
D-ImpU
D-ImpC
D-ImpG
D-ImpdA
D-ImpdA
D-ImpdA
D-ImpdA
D-ImpdA
D-ImpdA
D-ImpA
32
47
43
37
29
28
44
8.1
32
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
35
46
45
7.6
22
23
17
22
19
19
34
21
24
13
21
45
32
5
13
35
34
27
44
12
30
32
17
20
9.1
26
27
21
27
22
14
9
D-5-Fluorouridine
D-5-Fluorouridine
D-5-Fluorouridine
D-5-Fluorouridine
D-5-Fluorouridine
D-Deoxyadenosine
D-Deoxyadenosine
D-Deoxyadenosine
D-Deoxyadenosine
D-Deoxyadenosine
D-Deoxyguanosine
D-Deoxyguanosine
D-Deoxyguanosine
D-Deoxyguanosine
D-Deoxyguanosine
D-Inosine
D-Inosine
D-Inosine
D-Inosine
D-Inosine
D-Adenosine
D-Uridine
D-5-Fluorouridine
D-Cytidine
D-Guanosine
3
7
6
5
16
16
26
23
27
30
4
7.3
D-Ribavirin (Rib)
D-Ribavirin (Rib)
D-ImpU
4.6
(Continued on next page)

Nucleotide Dimer Synthesis
547
TABLE 5 Summary of the identities of nucleotide dimers prepared (Continued)
Nucleotide
Nucleoside
ImpN
% 2ꢁ-5ꢁ
%3ꢁ-5ꢁ
82 D,D-RibpC
83 D,D-RibpG
84 D,D-RibpdA 16 D-Ribavirin (Rib)
85 D,D-DoxpA
86 D,D-DoxpU 20 D-Doxifluridine (Dox) 25 D-ImpU
16 D-Ribavirin (Rib)
16 D-Ribavirin (Rib)
25 D-ImpC
27 D-ImpG
30 D-ImpdA
31
Trace
12
4.2
12
22
Trace
4.7
4.6
5.8
20 D-Doxifluridine (Dox) 23 D-ImpA
87 D,D-DoxpC
88 D,D-DoxpG
89 D,D-DoxpdA 20 D-Doxifluridine (Dox) 30 D-ImpdA
20 D-Doxifluridine (Dox) 28 D-ImpC
20 D-Doxifluridine (Dox) 27 D-ImpG
19
Trace
7.4
13
Trace
5
Part D
90
91
92
93
94
95
96
97
98
8
8
8
8
8
9
9
9
9
9
D-5-Aza-cytidine
D-5-Aza-cytidine
D-5-Aza-cytidine
D-5-Aza-cytidine
D-5-Aza-cytidine
D-6-Aza-uridine
D-6-Aza-uridine
D-6-Aza-uridine
D-6-Aza-uridine
D-6-Aza-uridine
23 D-ImpA
25 D-ImpU
27 D-ImpG
28 D-ImpC
no products no products
no products no products
no products no products
no products no products
29 D-ImpF-U no products no products
23 D-ImpA
25 D-ImpU
27 D-ImpG
28 D-ImpC
no products no products
no products no products
no products no products
no products no products
99
29 D-ImpF-U no products no products
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
11 D-Xanthosine
11 D-Xanthosine
11 D-Xanthosine
11 D-Xanthosine
23 D-ImpA
25 D-lmpU
27 D-ImpG
28 D-ImpC
29 D-ImpF-U no products no products
30 D-ImpdA no products no products
30 D-ImpdA no products no products
23 D-ImpA
25 D-ImpU
27 D-ImpG
28 D-ImpC
no products no products
no products no products
no products no products
no products no products
11 D-Xanthosine
12 D-Deoxyadenosine
13 D-Deoxyguanosine
14 D-Arabinocytidine
14 D-Arabinocytidine
14 D-Arabinocytidine
14 D-Arabinocytidine
14 D-Arabinocytidine
18 D-Arabinoadenosine
18 D-Arabinoadenosine
18 D-Arabinoadenosine
18 D-Arabinoadenosine
18 D-Arabinoadenosine
19 D-Xyloadenosine
19 D-Xyloadenosine
19 D-Xyloadenosine
19 D-Xyloadenosine
19 D-Xyloadenosine
15 D-Iododeoxyuridine
15 D-Iododeoxyuridine
15 D-Iododeoxyuridine
15 D-Iododeoxyuridine
15 D-Iododeoxyuridine
no products no products
no products no products
no products no products
no products no products
30 D-ImpdA no products no products
23 D-ImpA
25 D-ImpU
27 D-lmpG
28 D-ImpC
no products no products
no products no products
no products no products
no products no products
30 D-ImpdA no products no products
23 D-ImpA
25 D-ImpU
27 D-ImpG
28 D-ImpC
no products no products
no products no products
no products no products
no products no products
30 D-ImpdA no products no products
23 D-ImpA
25 D-lmpU
27 D-ImpG
28 D-ImpC
no products no products
no products no products
no products no products
no products no products
30 D-ImpdA no products no products
17 D-Trifluorothymidine 23 D-ImpA
17 D-Trifluorothymidine 25 D-ImpU
17 D-Trifluorothymidine 27 D-ImpG
17 D-Trifluorothymidine 28 D-ImpC
no products no products
no products no products
no products no products
no products no products
17 D-Trifluorothymidine 30 D-ImpdA no products no products
22 2ꢁ-OCH3 D-Adenosine 23 D-ImpA
22 2ꢁ-OCH3 D-Adenosine 25 D-ImpU
no products no products
no products no products
22 2ꢁ-OCH3 D-Adenosine 30 D-ImpdA no products no products

548
P. C. Joshi et al.
FIGURE 11 Nucleosides used in the synthesis of dimers: to be used in conjunction with Table 4.
Limitations of the Synthetic Method
Using HPLC and/or HPLC/MS, there were no products detected (Table
5, Part D) from the following reactions:
• The polar nucleosides D-5-aza-cytidine (8), D-6-aza-uridine (9), and D-
xanthosine (11) with ImpN (N = A, U, C, G, F-U).
• Deoxyadenosine (12) and deoxyguanosine (13) with ImpdA (30).
• The nonstandard deoxynucleosides, iododeoxyuridine (15), and trifluo-
rothymidine (17) with ImpN (N = A, U, C, G, dA).
• ImpA, ImpU or ImpdA with 2ꢁ-OCH3-adenosine (22).
• Arabinonucleosides (14, 18) with ImpN (N = A, U, C, G, dA).
• Xylo-adenosine (19) with ImpN (N = A, U, C, G, dA).
We attribute the reasons below for some of the failures of this synthetic
method:

Nucleotide Dimer Synthesis
549
FIGURE 12 Activated nucleotide (anions) used in the synthesis of dimers: to be used in conjunction
with Table 4.
1. With the most polar nucleosides, their inability to satisfactorily bind to the
pH 7 Na+ montmorillonite surfaces. We already know that the binding
of typical pyrimidines is poorer than that of the typical purine-based
adenosine.[4]
2. The 3ꢁ-hydroxy groups of deoxyribonucleosides do not have such a low
pKa value as the hydroxyl groups of ribonucleosides.[8] The latter can
form an intramolecular hydrogen bond between the 2ꢁ and 3ꢁ-OH groups.
Thus, the deoxyribonucleosides fail to react with ImpdA. Moreover, the
activated deoxyribonucleotide, D-ImpdA (30), is altogether less reactive
than its ribonucleotide counterpart with little oligomer formation from
itself on montmorillonite over 4 days; in dimer syntheses, extensive unre-
acted D-ImpdA was similarly found. The ability to form an intramolecular
hydrogen bond between the 2ꢁ and 3ꢁ-OH groups of ribonucleosides en-
ables mutual pKa reduction.[14,15] We believe that the lack of this ability

550
P. C. Joshi et al.
FIGURE 13 HPLC of the products from D-uridine and D-ImpC. (See also Table 6.)
and a possible additional steric factor are responsible for the complete
lack of reaction with the 2ꢁ-OCH3 ribonucleoside (22).
3. Nonstandard deoxynucleosides such as iododeoxyuridine (15) and tri-
fluorothymidine (17) are hydrophobic (e.g., iododeoxyuridine (15) is
FIGURE 14 Dimer formation illustrating the role of montmorillonite in providing protons and acting
as a general base. (Color figure available online).

Nucleotide Dimer Synthesis
551
only slightly soluble in aqueous solvents) and to this we attribute their
individual failure to react under standard conditions.
4. The poorer yields and slower reaction of D-AMP compared with D-
adenosine are readily understood because AMP will be anionic on the
clay surface at pH 7 whereas D-adenosine is a neutral species. Thus,
there is no repulsion between the reactants, D-adenosine and D-ImpN
(an anion), which promotes a faster reaction. Notwithstanding, dimers
from D-adenosine do react further to trimers. Here, only one negative
charge is dispersed over the nucleotide dimer molecule compared with
two negative charges over the D-AMP molecule. The combined rate of
formation of trimers can be estimated by numerical analysis, assuming a
simple consecutive reaction (Table 2). Thus, in this example, any given
dimer will react to form a pair of trimers with a given ImpN. In the case
being presented in Table 2, four dimers are formed from D-adenosine
and D, L-ImpA. Each forms four trimers because there is a mixture of D,
L-ImpA present, each capable of forming 2ꢁ-5ꢁ and 3ꢁ-5ꢁ linkages to any of
the four dimers.
On the other hand, amongst the nonstandard combinations, inosine
(10) reacted satisfactorily to form dimers in all cases tried; most interest-
ingly, ribavirin (16) formed products (137) up to tetramers under stan-
dard conditions (Table 5). D-5ꢁ-deoxy-5-fluorouridine (doxifluridine), (20)
formed dimers (Table 5, Part C). For example, 5ꢁ-deoxy-5-fluorouridine (20)
is shown reacting with D-ImpC (28) to form the dimers (87) (Figure 15).
Calibration of Mass Spectrometry Detection Limit
It can be safely concluded that the lack of detection of products was
because the products had not formed under the conditions employed: The
high sensitivity of the detection method would identify any products formed
FIGURE 15 Dimer formation from a nonstandard ribonucleoside, doxifluridine (20) and D-ImpC (28).
(Color figure available online).

552
P. C. Joshi et al.
above 8 nanomoles per injection of the diluted reaction mixture. This value
was obtained from calibration experiments with standard solutions of pdCdC
enabling an estimate of the limit of detection of the HPLC/MS analytical
technique. “Trace” in Table 5 refers to detection by MS only, where the
product is formed below the detection limit of the HPLC system (<0.1%).
The ImpG Anomaly
We believe that the low yields of products obtained in reactions involving
D-ImpG (27) are caused by the formation of its tetraplex species (Figure 16).
The large tetraplexes are simply unable to insert themselves into the mont-
morillonite lattice to react with the adsorbed nucleoside. Matrix-assisted
laser desorption/ionization (MALDI) mass spectral evidence from the sep-
arate oligomerization reactions of D-ImpA (23) and D-ImpU (25) reveals
FIGURE 16 ImpG anion tetraplex with bound potassium ion. (Color figure available online).

Nucleotide Dimer Synthesis
553
long oligomers of A and U forming with a monotonic decreasing yield of
the oligomers.[2] Oligomerization of D-ImpG (27) alone on montmorillonite
demonstrates a very different outcome. Even numbered oligomers dominate
as shown by the MS evidence (Figure 10), so we again consider the oligomers
from D-ImpG (27) are based largely upon tetraplex or similar formation. D-
ImpG forms a thixotropic gel upon solution in water, typical of guanosine
derivatives[16,17] and particularly with potassium ions available.[18] Additional
evidence concerning unusual delivery of groups of four nucleotides to the
reacting site is provided by:
1. The cyclic nucleotide c-(pG)4 (m/z = 1379) (with scavenged potassium
ions, (m/z = 1417), even from a reaction nominally done in the presence
of only sodium) (Figure 10b) and is formed at an estimated 100 times
that expected. This unusual species has been modeled to confirm its
ability to engulf and retain a potassium ion (Figure 17 and Supplementary
Information).
2. The cyclic species c-(pG)8 (Figure 10b) is formed at a concentration
some 1000 times that expected. This species may be a dimer of the cyclic
tetramer, with a cage-like structure as shown (Figure 18 and Supplemen-
tary Information), again able to engulf and retain two (m/z = 2797) or
three (m/z = 2836) potassium ions (Figure 10b).
FIGURE 17 Model of c-(pG)4K: (a) The ring of sugar phosphates traps a potassium ion. (b) Four
hydrogen-bonded guanines trap a potassium ion. (Color figure available online).

554
P. C. Joshi et al.
FIGURE 18 Multiplex derived cyclic oligomers of G (a) Molecular model of a hydrogen bonded c-
(pG)42− dimer (b) Molecular model of [c-pG)4]2K−. (Color figure available online).
3. Extensions of our method to other dimer syntheses involving ImpG and
nonstandard nucleosides (Table 5, Part C) further confirmed the lower
yields with this reagent. Numerical analysis of the yields of D-GMP from
the use of D-ImpG (eight separate standard nucleotide syntheses) indi-
cates that approximately 85% hydrolysis occurs with a yield of only 15%
of the associated dimers. Each dimer is produced in approximately equal
amounts and it can be assumed that a similar amount of D-ImpG is hy-
drolyzed to D-GMP within the clay (as occurs with the other D-ImpN).
This suggests that the amount of D-ImpG hydrolyzed within the clay is
about 7.5% with the remaining 77.5% produced by the hydrolysis of mul-
tiplexes of the activated nucleotide, (D-ImpG)n. The ratio of these two
values, (77.5/7.5 =) 10.3, suggests that the multiplexes of D-ImpG could
consist of (D-ImpG)8 and (D-ImpG)12 in similar amounts although clearly
other compositional mixes could occur. This is an area of ongoing mass
spectrometry analysis to identify multiplex derived products similar to
c-(pG)4 and c-(pG)8 (Figures 17 and 18).

Nucleotide Dimer Synthesis
Oligomers from D-ImpdA (30)
555
In the presence of Na+-montmorillonite at pH 7, oligomers formed up to
the tetramer (trace); additional experiments over several weeks in which the
reactions were carried out at −2◦C failed to enhance the oligomer length,
irrespective of the cation present. Nevertheless, D-ImpdA (30) performs
satisfactorily in its reactions with nucleosides to form dimers with the provisos
already discussed.
Enzymatic Hydrolysis of Dimers (Table 7)
Alkaline phosphatase cleaved the 5ꢁ-terminal phosphate group of all
the pMpN dimers (synthesized from a 5ꢁ-nucleotide monophosphate and
D-ImpN) to produce MpN. However, RNaseT2 was selective in hydrolyzing
the 3ꢁ-5ꢁ linked dimers MpN to their respective 3ꢁ- nucleotide monophos-
phate (3ꢁ-Mp) and the nucleoside (N) in 1:1 proportion. While all the 3ꢁ-5ꢁ
linked D, D dimers were hydrolyzed by RNaseT2, the L, L-enantiomers were
completely resistant to enzymatic hydrolysis. Likewise, the L, D-enantiomers
also were not hydrolyzed by RNaseT2, but interestingly the 3ꢁ-5ꢁ linked D, L-
enantiomers hydrolyzed to their respective 3ꢁ- mononucleotide and a nucle-
oside. The nonstandard dimers 3ꢁ-5ꢁ linked D, D enantiomers, pMpN, were
also hydrolyzed by alkaline phosphatase to 3ꢁ-5ꢁ-MpN. Again, RNaseT2 diges-
tion of 3ꢁ-5ꢁ-MpN provides 3ꢁ-Mp and N. For example, the 3ꢁ-5ꢁ-doxifluridine
pC dimer (87) (see Figure 15) was hydrolyzed to doxifluridine-3ꢁ-phosphate
and cytidine (6) while the 2ꢁ-5ꢁ-isomer was unaffected. The products were
identified by HPLC and confirmed by HPLC/MS whenever necessary.
These selective enzymatic hydrolysis reactions facilitated identification of
the dimers and were responsible for determining their regioselectivity be-
cause in every reaction involving ribosyl nucleotides, formation of both 2ꢁ-5ꢁ
and 3ꢁ-5ꢁ linked dimers is observed. That the 2ꢁ-5ꢁ-linked dimers were eluted
first in a reverse phase HPLC column is an empirically derived result based
on every RNaseT2 enzymatic hydrolysis reaction studied in this work. The use
of D-deoxy-adenosine 5ꢁ-phosphorimidazolide, D-ImpdA (30) with ribosyl-
nucleosides, either standard or nonstandard, provided chimeric products
that also responded to RNase T2 enzymatic hydrolysis. In this way, the iden-
tity of the dimers was again directly confirmed, and once more 2ꢁ-5ꢁ-linked
dimers eluted from HPLC faster than the 3ꢁ-5ꢁ counterpart. For example,
the 3ꢁ-5ꢁ-doxifluridinepdA dimer (89) was hydrolyzed to doxifluridine-3ꢁ-
phosphate and D-deoxyadenosine (12) while the faster eluting 2ꢁ-5ꢁ isomer
was unaffected by enzymatic hydrolysis.
Arabinose- and Xylose-Based Nucleosides
Intramolecular hydrogen bonding between the 2ꢁ and 3ꢁ-OH groups
affects the pKa values of these hydroxyl groups in ribose.[8,14,15] Comparable
effects are present in neither the arabinose-based nucleosides (14, 18 in this

556
P. C. Joshi et al.
study) nor xyloadenosine (19). Preliminary modeling of these nucleosides
and their 3ꢁ-anions reveals that in an aqueous environment, the oxygen
atoms in the 2ꢁ and 3ꢁ-OH groups in adenosine (−0.733, −0.696, respectively)
bear significantly smaller charges than their arabinose (−0.825, −0.727) and
xylose (−0.757, −0.723) counterparts, thus strengthening the case for the
greater stability and greater reactivity of the ribose species. The 3ꢁ-OH group
of 2ꢁ-deoxyadenosine (12 in this study) similarly bears a much larger charge
(−0.816) on the oxygen. This modeling is the subject of ongoing research.
Thus, it would seem that a delicate balance between the structure of the
pentose and the overall hydrophobicity of the nucleoside determines the
outcome of any given dimer synthesis. The chemistry here can be usefully
pursued as a probe of the interactions of the nucleosides and nucleotides;
our study of dimer synthesis clearly singles out ribose over its rivals. This is
important in contributing to our understanding of the catalysis of formation
of longer RNA oligomers by montmorillonite on the prebiotic Earth, as
required by the RNA world concept.[19]
CONCLUSION
A standard methodology has been developed for nucleotide dimer syn-
thesis, which is both simple and yet extremely versatile, providing the product
from readily accessible components.
Furthermore, the method has shown promise in targeting trimer syn-
thesis and the procedure lends itself to the development of combinatorial
libraries. The use of enzymatic hydrolysis has played a crucial role in this
work, facilitating product identity across the spectrum of products prepared.
The yields of the corresponding homochiral and heterochiral dimers from
A and U will require careful modeling of the reactants in their interactions
with both the clay and one another to locate the source of the similarities and
differences. The lack of reactivity of arabino- and xylo-nucleosides also poses
interesting structural, modeling, and origins of life issues. There are clearly
some limitations in the chemistry but we feel that many other potentially
useful but as yet untested nucleoside/activated nucleotide combinations
will provide further successful dimer or even trimer syntheses.
EXPERIMENTAL
Preliminary Considerations
The dimer syntheses were generally carried out in one of two ways. Either
a 10 mg portion of catalyst was used with a total volume of reactants of 200 µL
in an Eppendorf tube, or on a smaller scale, a 2 mg catalyst sample was used
with a total volume of reactants of 40 µL.

Nucleotide Dimer Synthesis
557
The second method was best performed in a centrifuge tube filter to
facilitate extraction of the dimers from the montmorillonite. By control-
ling the concentrations of the reagents and reaction time, trimer formation
could be minimized or even eliminated if an exclusive dimer synthesis was
intended. Conversely, conditions could be devised where both the dimers
and in lesser yield, the related trimers could be simultaneously synthesized.
In all syntheses, the identity of the product(s) formed, if any, was never
in doubt because of the nature of the synthetic method. Nevertheless, in
all cases, the product identity was confirmed by a combination of HPLC,
HPLC/MS, preparative HPLC, and enzymatic studies.
Materials and Methods
General
Standard ribonucleosides and 5ꢁ-ribonucleotides were obtained from
Sigma (St. Louis, MO, USA). L-AMP, L-adenosine, L-UMP and L-uridine
were purchased from ChemGenes Corporation (Wilmington, MA, USA).
Sodium pyrophosphate and perchloric acid were purchased from Aldrich
(St. Louis, MO, USA). HPLC grade NaClO4 and enzyme grade ammonium
acetate was purchased from Fisher (Pittsburgh, PA, USA). Other nucleosides
were from TCI America (Portland, OR, USA). The Na+ form of montmoril-
lonite was prepared by titration to pH 7 of the H+ form of montmorillonite.[4]
Homoionic Li+-montmorillonite and homoionic K+-montmorillonite were
similarly prepared. All montmorillonite catalyzed reactions, both synthesis
and control studies, were performed in a solution containing 0.2 M sodium
chloride and 0.075 M magnesium chloride.
Mass Spectrometric Analysis
Electrospray mass spectra were obtained using either (a) an Agilent 1100
LC/MSD ion trap system; samples were introduced into the ion source using
a syringe pump at a flow rate of ∼7 µL/min or using an auto-sampler, or (b)
Thermo LTQ Orbitrap mass spectrometer (Thermo Scientific, West Palm
Beach, FL, USA) coupled with an Agilent 1200 nanoflow HPLC system; the
column was an Agilent Technologies (Santa Clara, CA, USA) Zorbax Eclipse
SB-C18 5 µm (150 mm × 0.5 mm) at a flow rate of 20 µL/min. In all cases,
(M−H)− was identified in the ESI(−) mass spectrum and/or (M+H)+ was
identified in the ESI(+) mass spectrum. HPLC buffers were A) 0.2% formic
acid and B) 0.2% formic acid in 30% acetonitrile at an incremental rate of
1%/minute for 40 minutes of buffer B.
Matrix-assisted laser desorption/ionization (MALDI) mass spectra were
recorded on a MALDI-TOF-TOF mass spectrometer operating with a nitro-
gen laser (337 nm) in reflector mode. Samples were spotted onto a polished
stainless steel target plate with 2, 4, 6-trihydroxyacetophenone as a matrix
and dibasic ammonium citrate was added.[2]

558
P. C. Joshi et al.
High Performance Liquid Chromatography
High performance liquid chromatography (HPLC) analysis was per-
formed either using a UV detector operating at 260 nm or using a diode
array detector (DAD). The products as anions were separated on a Dionex
DNAPac-100 (4.0 × 250 mm) analytical anion exchange column (Dionex
Corporation, Sunnyvale, CA, USA), using a gradient of 0 ∼ 0.4 M NaClO4
with 2 mM Tris at pH 8 (1%/minute). The analytical separation of individ-
ual dimers and trimers was generally performed on a reverse phase Alltima
C-18, 5 µ (4.6 mm × 250 mm) column (Grace, Deerfield, IL, USA) either:
1. using a gradient of 0.02 M NaH2PO4 with 0.2% trifluoroacetic acid (pH
2.5) and 30% CH3CN in water with 0.2% trifluoroacetic acid (pH 2.5)
(1%/minute) or
2. using a gradient of 0.2% formic acid and 30% CH3CN in water with 0.2%
formic acid (1%/minute). Alternatively,
3. isocratically using 0.2% formic acid (85%) and 30% CH3CN in water with
0.2% formic acid (15%).
The analysis of the isomers of U was performed on a reverse phase
Jupiter C-18, 4µ (4.6 mm × 250 mm) column (Phenomenex; Torrance, CA,
USA) and a µBondapak C-18 (3.9 mm × 300 mm) column (Waters; Milford,
MA, USA) by using a gradient of 0.005 M NaH2PO4 with 5% CH3OH, pH
3.5 and 0.01 M NaH2PO4 with 40% CH3OH, pH 4. Large scale separations
and purification of D- and L- mononucleotides and selected dimers were
carried out on a preparative Alltima C-18, 5µ column (10 mm × 300 mm)
isocratically at pH 2.3 by using 0.075% trifluoroacetic acid (98%) and 30%
CH3CN in water with 0.075% trifluoroacetic acid (2%).
MpN could be identified by their relatively long retention times com-
pared with dimers pNpN from direct oligomerization reactions and by com-
parison of the characteristic ultraviolet spectra (e.g., cytidine (6) λmax 280 nm
and ribavirin (16) λmax 220 nm) obtained with the DAD.
Phosphorylation of D-5-Fluorouridine (7) and L-Nucleosides[20]
D-5-fluorouridine (7) and the L-nucleosides of adenine and uracil
were phosphorylated at the 5ꢁ-position. A mixture of D-5-fluorouridine (7)
(30 mmol) or L-nucleoside (30 mmol) and water (15 mmol) in TEP (1.0 mL)
was treated with POCl3 (60 mmol) at 0◦C for 5 h (L-adenosine (2)), 25 h
(L-uridine (4)), or 72 h (D-5-fluorouridine (7)). Progress of the reaction
was monitored by HPLC by withdrawing 2 µL samples at intervals, diluting
to 1.0 mL with water, and analyzing a 10 µL aliquot on an Alltima reverse
phase column. The reaction was terminated by adding the reaction mixture
dropwise into l.5 mL ice cold water and continuing stirring for an additional

Nucleotide Dimer Synthesis
559
1 hour. The crude yields of L-AMP, L-UMP, and D-F-UMP at the comple-
tion of the reactions were 90 5%, 75 5% and 85 5%, respectively.
TEP was removed from the reaction mixture by extraction with ether (3 ×
5 mL). A portion of mononucleotide (below 10%) retained in the ether
extract was recovered quantitatively by extraction with water (2 × l mL). The
L-mononucleotides were purified on a preparative Alltima column (purity >
99.5%). Similarly, the phosphorylation of 5-fluorouridine yielded 5-FUMP
(purity > 99%).
Preparation of Activated Nucleotides (23–30)[21]
A mixture of 5ꢁ-mononucleotide (free acid, 0.16 mmol) and imidazole
(2.2 mmol) was dissolved in DMF (2 mL) and DMSO (2 mL) in a 50 mL
flask and the solvent was evaporated to near dryness at a reduced pressure.
The evaporation was repeated twice with DMF (2 × 2 mL) to remove resid-
ual water. The residue was dissolved in DMF (2 mL) and DMSO (2 mL)
and stirred with 2, 2ꢁ-dithiodipyridine (0.64 mmol), triphenyl phosphine
(0.46 mmol), and TEA (0.65 mmol) for 3 hours. The resulting product
was recovered from the clear yellow reaction mixture as a precipitate by
adding the reaction mixture dropwise to a solution of anhydrous sodium
perchlorate (0.35 g) in a mixture of ether (30 mL), acetone (20 mL), and
TEA (2 mL) with stirring. The stirring was continued for 2 hours (to ensure
complete formation of Na+ salt) and a white, flocculent solid precipitated
(15 minutes). The supernatant was decanted and the remaining reaction
mixture was centrifuged. The resulting white pellet was washed thrice with
a mixture of ether (50 mL) and acetone (50 mL) and then with ether
(2 × 50 mL) and dried overnight in a vacuum desiccator (yields > 85%).
Their purities (>99.5%) were determined by reverse phase HPLC.
Synthesis of Dimers
Only representative examples of synthesis are provided. All syntheses
were carried out via similar methodology.[13] Optimized outcomes are in-
dicated in Table 5. Note that products, where formed, always had a known
identity because of the synthetic route.
Example 1. Synthesis of Homochiral and Heterochiral Linear Dimers of A
(31, 47)
ꢁ
ꢁ
ꢁ
ꢁ
Method 1. D, D-A2 -5 pA and D, D-A3 -5 pA were prepared by the addition
of 15 mM D-adenosine (1) (200 µL) to 20 mg Na+-montmorillonite and
incubating the mixture at 24◦C for l hour. 10 mM D-ImpA (23) (200 µL)
was added to the mixture and the reactioꢁn was allowed to stand for 48 hours
ꢁ
ꢁ
ꢁ
at 24◦C. The D, D-A2 -5 pA and D, D-A3 -5 pA were purified on a preparative
reverse phase HPLC column, eluting in that order (Figure 19). The products

560
P. C. Joshi et al.
TABLE 6 Examples of trimer formation accompanying dimer synthesis
Yields%
2ꢁ5ꢁ, 2ꢁ5ꢁ
2ꢁ5ꢁ, 3ꢁ5ꢁ
3ꢁ5ꢁ, 2ꢁ5ꢁ
3ꢁ5ꢁ, 3ꢁ5ꢁ
135
136
137
D,D,D-UpCpC
D-Uridine
D-Uridine
D-Ribavirin
D-ImpC
D-ImpA
D-ImpA
0.7
0.9
7.7
0.8
1.6
6.8
0.6
1.7
4.5
2.0
0.8
5.5
D,D,D-UpApA
D,D,D-RibavirinpApA∗
∗This reaction also produced 4.9% of the tetramers (D,D,D,D-RibavirinpApApA).
ꢁ
ꢁ
were identified by hydrolysis with RNase T2 and APH. D, D-A2 -5ꢁpA was
ꢁ
unaffected by the ꢁenzymatic digestion (Table 7, line 1). D, D-A3 -5 pA was
ꢁ
hydrolyzed to D-A3 p and D-adenosine. D-A3 p was hydrolyzed to D-adenosine
by APH (Table 7, line 2). The products at each stage were identified by HPLC
by their retention times on a reverse phase Alltima C-l8 column.
ꢁ
ꢁ
ꢁ
ꢁ
Similarly, D, L-A2 -5 pA and D, L-A3 -5 pA were prepared by the addition
of 15 mM D-adenosine (1) (200 µL) to 20 mg Na+-montmorillonite and
incubating the mixture for l hour.[22] 10 mM L-ImpA (24) (200 µL) was
added to the mixture and the reaction wasꢁ allowed to stand for 48 hours
ꢁ
ꢁ
ꢁ
at 24◦C to yield D, L-A2 -5 pA and D, L-A3 -5 pA, which were purified on a
preparative reverse phase HPLC column, eluting in that order (Figure 19).
FIGURE 19 HPLC of homo- and hetero-chiral dimers of A.

Nucleotide Dimer Synthesis
561
TABLE 7 Enzymatic hydrolysis of MpN dimers
RNaseT2
APH
ꢁ
D, D & L, L-M2 pN
ꢁ
1.
2.
→
→
No hydrolysis
→
→
No hydrolysis
ꢁ
D, D & L, L-M3 pN
D-M3 p + ꢁD-N +
D-M + D-N +
ꢁ
L, L-M3 pN
50%
L, L-M3 pN
50%
No hydrolysis
ꢁ
D, L & L, D-M2 pN
ꢁ
3.
4.
→
→
No hydrolysis
→
→
ꢁ
D, L & L, D-pM3 pN
D-M3 p + L-N +
D-M + L-N +
ꢁ
ꢁ
L, D-M3 pN
L, D-M3 pN
50%
50%
Key: M, N = nucleosides; RNaseT2 = hydrolysis by ribonuclease T2; APH = hydrolysis by alkaline
phosphatase.
ꢁ
ꢁ
The structures were confirmed by enzymatic hydrolysis: D, L-A2 -5ꢁpA was
ꢁ
unaffected by the ꢁenzymatic digestion (Table 7, line 3). D, L-A3 -5 pA was
ꢁ
hydrolyzed to D-A3 p and L-adenosine. D-A3 p was hydrolyzed to D-adenosine
by APH (Table 7, line 4). The products at each stage were identified by HPLC
by their retention times on a reverse phase Alltima C-18 column. Further
confirmation of the identity of the dimers was by:
1. co-injection HPLC (Figure 19) and
2. electrospray HPLC/MS, where the dimers were separated on a reverse
phase Alltima C-l8 column using a gradient of 0–40% of 0.2% formic
acid and 0.2% formic acid in 30% CH3CN (1%/minute). Calculated for
neutral ApA (C20H25N10O10P) 596. Found: m/z 597, [M+1]+ in ESI(+)
mass spectrum and m/z 595, [M−1]− in ESI(−) mass spectrum.
Example 2. Typical RNA Dimer Synthesis: UpC (37) (See Figure 13)
Method 2. On a smaller scale, using Na+-montmorillonite (2 mg) and
carrying out the experiment in a prerinsed (40 µL water) centrifuge tube
filter with a total volume of reactants of 40 µL, a wide range of nucleosides
were treated with an activated nucleotide. For the synthesis of D, D-UpC
(37), the clay mineral was incubated with 20 µL of 20 mM D-uridine (3) for
1 hour. An equal volume of the activated nucleotide D-ImpC (28) (10 mM)
was added. After 3 days the reaction mixture was analyzed as described
ꢁ
ꢁ
ꢁ
ꢁ
in Method 1. The yields of D, D-U2 -5 pC and D, D-U3 -5 pC were 53% and
22% (Table 5), respectively, based upon consumed D-ImpC. Calculated for
neutral UpC (37) (C18H24N5O13P) 549. Found: m/z 550, [M+1]+ in ESI(+)
mass spectrum and m/z 548, [M−1]− in ESI(−) mass spectrum. In either
method, the ratio of concentrations of nucleoside: activated nucleotide of
approximately 2: l ensured a very low yield of higher oligomers. The reaction
still worked well with concentrations of nucleoside as low as 1 mM or 2 mM,
and activated nucleotide half that, with almost no change in the relative
percentages of the 2ꢁ-5ꢁ and 3ꢁ-5ꢁ isomers.

562
P. C. Joshi et al.
TABLE 8 Optimization of the yield of trimers of A (See Figure 20.)
Reactants
ImpA
(mM)
Products (%)
ꢁ
A3 pA
Clay
(mg)
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
(mM)
A3 pA3 pA
A3 pA2 pA
AMP
A3 pA
ImpA∗
1
2
3
4
5
6
7
8
1
1
0.5
1
0.5
5
1
1
0.5
1
0.5
5
5
10
1
1
0.5
0
24
26
17
11
9.4
0
22
23
13
9.2
8.2
0
28
26
24
22
21
16
8.8
5.3
16
25
10
0.9
38
54
56
80
58
87
7.8
4.5
5.9
3.6
26
0.1
1
0.1
1
0.5
0
3.1
0
4.5
0
7.9
∗Unreacted ImpA.
In the absence of Na+-montmorillonite, no dimers were detected when
the reactions were carried out under the standard conditions described
above.
Synthesis of Linear Trimers
The method has further been modified to target trimer synthesis; for
example, preliminary experiments were optimized whereby a ꢁdimer, e.g.,
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
A3 -5 pA was converted into trimers, A3 -5 pA2 -5 pA and A3 -5 pA3 -5 pA.
Linear Trimers of ApApA
Yields of trimers can be optimized from a nucleoside or from a starting
dimer and an appropriate activated nucleotide. Sample data are provided for
ꢁ
ꢁ
the optimization experiments for D, D-A3 -5 pA reacting with D-ImpA (Table
8, Figure 20) to form linear trimers of A.
Trimers from (i) D-Uridine and D-ImpC (ii) D-Uridine and D-ImpA
Careful inspection of the HPLC analysis of the reaction mixtures from
most of the dimer syntheses of this paper reveal small amounts of trimers;
e.g., see Figure 13 in the synthesis of D, D-UpC (37) from D-uridine (3)
and D-ImpC (28), UpA (35) (from D-uridine (3) and D-ImpA (23)), there
are traces of all four trimers in nonoptimized experiments. Sample data
(Table 6) are provided for D, D, D-UpCpC (135) and UpApA(136) trimers
(representative structures given in Figure 21). Calculated for neutral UpCpC
(135) (C27H36N8O20P2) 854. Found: m/z 855, [M + 1]+ in ESI(+) mass
spectrum and m/z 853, [M−1]− in ESI(−) mass spectrum. Calculated for
UpApA (136) (C29H36N12O18P2) 902. Found: m/z 903, [M+1]+ in ESI(+)
mass spectrum and m/z 901, [M − 1]− in ESI(−) mass spectrum.

Nucleotide Dimer Synthesis
563
FIGURE 20 Optimization of the yield of trimers of A. See Table 8.
Use of Enzymes in Characterizing the Dimers
Identification of dimers containing both D- and L-nucleotides was, where
appropriate, done by enzymatic hydrolysis (Table 7). This technique was
especially valuable in the cases of homochiral and heterochiral dimers of
the same pair of nucleoside residues (Table 7) and confirming the identity
of nonstandard dimers, including chimeric dimers such as 2ꢁ-5ꢁ and 3ꢁ-5ꢁ-
doxifluridinepdA (89).
Alkaline phosphatase hydrolysis was performed on either the entire re-
action mixture (after appropriate dilution) or individually isolated dimers
and trimers. An alkaline phosphatase (APH) solution (0.5 units, total vol-
ume 500 µL) was adjusted to pH 8 (Tris buffer) and incubated for 1 hour
at 37◦C to remove 5ꢁ-terminal phosphate groups from the products from
FIGURE 21 Selected trimers (135) and (136) referred to in the Experimental section. (Color figure
available online).

564
P. C. Joshi et al.
the reactions of an NMP and ImpN. APH was also used to cleave the 3ꢁ-
phosphate residues as shown in Table 7, lines 2 and 4. RNase T2 hydrolysis
of RNA oligomers was performed in 100 µL at 37◦C and pH 4 using 2 units
of the enzyme for 1–4 hours. The samples were analyzed by HPLC prior to
the enzymatic digestion to ensure a suitable concentration of the material
undergoing hydrolysis.
Rates of Formation of Selected Dimers
The syntheses of dimers from D-AMP and from D-adenosine (1) were
compared. 10 mM D/L ImpA (100 µL) was added to Na+-montmorillonite
(10 mg), which had been incubated with D-AMP (15 mM, 100 µL for 1 hour,
centrifuged and a fresh portion of D-AMP added). Portions of the reaction
mixture were removed hourly, centrifuged, and the supernatant treated
with APH to remove terminal 5ꢁ-phosphates. Reverse phase HPLC analysis
enabled simultaneous measurement of the four products formed (sample
HPLC traces Figure 19), whereby rates of formation were determined (Ta-
bles 1–3). The use of the enantiomeric mixture of activated nucleotides
enabled the determination of any homochiral/heterochiral tendencies by a
simple comparison of the relative rates of formation, An analogous experi-
ment with adenosine clearly demonstrated the increased rate of reaction of
the nucleoside (a neutral species) compared with the nucleotide (an anion)
under identical conditions. Similarly, the curves obtained from the D- adeno-
sine reactions not only enabled the rates of formation of all four dimers to be
determined, but also enabled the total rates of conversion of a given dimer
to its four related trimers. This was achieved by standard numerical analysis
of the consecutive reactions to provide the pseudo first and second order
rate constants (Tables 1–3).
Oligomerization Reactions of ImpG (27)
D-ImpG (27) (40 µL, l5 mM in 1 M MCl) was incubated with montmo-
rillonite (2 mg) (from the H+ form that had been titrated to pH 7 with an
alkali metal hydroxide (MOH, M = Li, Na, K)). The D-ImpG (27) formed a
thixotropic gel, typical of guanosine gels in general.[14] The reactions were
conveniently carried out in centrifuge filter tubes. Typically, after 3 days, the
supernatant was removed by centrifugation and the clay extracted four times
(2 × 1 hour, overnight, 1 hour) with 40 µL sodium chloride (0.1 M) in 30%
aqueous acetonitrile. The combined extract was used for HPLC analyses and
following zip-tipping (PerfectPure C-18, Eppendorf) for mass spectroscopic
analysis.