Facile and efficient access to 2,6,9-tri-substituted purines through sequential N9, N2 Mitsunobu reactions

Article abstract of DOI:10.1016/j.tetlet.2009.04.137

A facile, efficient and mild synthesis of 2,6,9-tri-substituted purines is presented, starting from commercially available 2-amino-6-chloropurine, which employs sequential N9 then N2 Mitsunobu reactions as key steps. Importantly, our synthetic approach to

Full text of DOI:10.1016/j.tetlet.2009.04.137

Tetrahedron Letters 50 (2009) 4258–4261
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Facile and efficient access to 2,6,9-tri-substituted purines through
sequential N9, N2 Mitsunobu reactions
*
*
Steven Fletcher , Vijay M. Shahani, Patrick T. Gunning
Department of Chemistry, University of Toronto, Mississauga, ON, L5L 1C6, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
A facile, efficient and mild synthesis of 2,6,9-tri-substituted purines is presented, starting from commer-
cially available 2-amino-6-chloropurine, which employs sequential N9 then N2 Mitsunobu reactions as
key steps. Importantly, our synthetic approach to N2-functionalization of the purine nucleus obviates
the harsh conditions required by the traditional nucleophilic aromatic substitution of a 2-halo group with
primary amines. Benzylic, allylic, propargylic and aliphatic alcohols all coupled in very good to excellent
yields in both Mitsunobu reactions. Significantly, excellent chemoselectivity and N9-regioselectivity
were observed for the first coupling, and reactions were complete within 15 min at room temperature.
Our novel methodology may be readily adapted to furnish N9-mono- or N2,N9-di-functionalized guanine
analogues, and the utility of our protocol is further demonstrated by the efficient synthesis of the CDK
inhibitor bohemine.
Received 8 April 2009
Revised 29 April 2009
Accepted 30 April 2009
Available online 6 May 2009
Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
Protein kinases, which play critical roles in regulating the cell
cycle, utilize ATP as a co-substrate, and thus considerable research
into the design of kinase inhibitors has focused on the develop-
ment of purine-based ATP antagonists.1 Notably, 2,6,9-tri-substi-
tuted purines, including olomoucine, bohemine and (R)-
roscovitine (Fig. 1), have proven potent and selective inhibitors
of cyclin-dependent kinases (CDKs).2 In fact, Cyclacel Pharmaceuti-
cals has recently announced that (R)-roscovitine has progressed to
Phase IIb clinical trials against non-small cell lung cancer, and to
Phase II against nasopharyngeal cancer.3
The synthesis of 2,6,9-tri-substituted purines typically begins
with 2,6-dihalopurines. Whilst displacement of the 6-halo (usually
chloro) group with amine nucleophiles is relatively facile, nucleo-
philic aromatic substitution of the 2-halo group often demands a
large excess of the amine, high temperatures and/or extended reac-
tion times.4 Excepting recent research in microwave-assisted syn-
thesis,5 milder alternatives to the synthesis of 2,6,9-tri-substituted
purines would be of significant benefit. As part of our research on
peptide nucleic acids (PNAs), we have developed a synthetic route
for the preparation of novel N2,N9-di-functionalized guanines start-
ing from the popular guanine precursor 2-amino-6-chloropurine
(1), which we have found may be readily adapted to the prepara-
tion of other purines. In this Letter, we wish to report a novel meth-
odology for the facile and efficient synthesis of 2,6,9-tri-
substituted purines, including N9-mono and N2,N9-di-functional-
ized guanines, which exploits mild, sequential Mitsunobu reac-
tions as key steps.
Direct N9-functionalization of 1 under Mitsunobu conditions6
has been reported previously, and, although good N9-regioselectiv-
ity was observed, products were generally obtained in poor to
moderate yields,7a–c with higher yields being obtained when sev-
eral equivalents of reagents were employed.7d The poor solubility
of 1 in THF, the preferred solvent for Mitsunobu reactions, and
the competing nucleophilicity of the exocyclic N2 amino group
are likely primary reasons for the limited yields observed.
We reasoned that protection of the amine of 1 with a hydropho-
bic group such as tert-butoxycarbonyl (Boc) should enhance its sol-
ubility in organic solvents. In addition, the resultant NHBoc
carbamate should function as an activated amine suitable for par-
ticipation in the Mitsunobu reaction, wherein the inherent bulk of
the Boc group should, at the same time, exert chemoselective con-
trol, allowing the Mitsunobu reaction to proceed first at the endocy-
clic N9 position and then, in a subsequent step, at the exocyclic N2
position. To this end, treatment of 1 with 1 equiv of Boc2O in the
presence of catalytic DMAP in DMSO led to N9-Boc-2-amino-6-chlo-
ropurine (2) quantitatively within 30 min (Scheme 1). Although
here not significant, the N9 versus N7 regiochemistry was assumed,
based on the reported N9 regiochemistry of the tris-Boc protected
analogue of 2.8 Generation of the anilide anion of 2 with an excess
of NaH in THF invoked the desired shift of the Boc group from N9
to N2, furnishing N2-Boc-2-amino-6-chloropurine (3) in near-
quantitative yield, and thereby providing a higher yielding and
more economical route to the previously reported compound 3.8
Purine 3 exhibited much improved solubility in organic sol-
vents, especially in THF, and, employing n-butanol as a represen-
tative alcohol, the desired Mitsunobu reaction proceeded in
both excellent yield, chemoselectivity and N9-regioselectivity9
(4a: R1 = n-Bu, 88%), with the diisopropylazodicarboxylate
* Corresponding authors. Tel.: +1 905 569 4588; fax: +1 905 569 4929 (P.T.G.).
E-mail addresses: steven.fletcher@utoronto.ca (S. Fletcher), patrick.gunning@
utoronto.ca (P.T. Gunning).
0040-4039/$ - see front matter Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.04.137

S. Fletcher et al. / Tetrahedron Letters 50 (2009) 4258–4261
4259
HN
HN
N
HN
N
OH
N
N
N
N
N
N
N
N
OH
N
N
N
H
OH
N
H
N
H
R-roscovitine
(seliciclib or CYC202)
olomoucine
bohemine
Figure 1. 2,6,9-Tri-substituted purine CDK inhibitors.
Table 1
(DIAD)/triphenylphosphine (PPh3) redox combination (Scheme 2).
The excellent N9-regioselectivity is consistent with the findings
of others working on related purines.10
Reaction substrate scope in the N9 Mitsunobu coupling of purine 3 with R1OH
alcoholsa
Cl
Cl
N
R1OH, PPh3
DIAD
Reports on the Mitsunobu-mediated alkylation of the purine N2
amino group are scarce. Of those reports, activation was achieved
by converting the amino group to its acetamide11,12 or to its more
reactive (more acidic) trifluoroacetamide.13 Yields were poor to
moderate with aliphatic alcohols;11,13 highest yields were limited
to allylic and benzylic alcohols.12 The use of the Boc group to pro-
tect/activate the aromatic amino group of functionalized heterocy-
cles towards the Mitsunobu reaction has also been reported in
moderate yield,14 giving us confidence that the likewise alkylation
of the NHBoc unit in purine 4 may be successful. Indeed, 4 was
smoothly alkylated with n-butanol under modified Mitsunobu
conditions (2.5 equiv of each of n-butanol, PPh3 and DIAD at
35 °C for 30 min) to give 5aa (R1 = R2 = n-Bu) in excellent yield
(93%). Nucleophilic aromatic substitution of 5aa with benzylamine,
followed by deprotection of the Boc group with TFA furnished the
2,6,9-tri-substituted purine 6aa (R1 = R2 = n-Bu, R3 = Bn) in a note-
worthy overall yield of 72% (six steps).
N
N
N
N
THF
rt, 15 min
N
H
N
N
NHBoc
NHBoc
R1
3
4a - 4f
Entry
1
R1OH
Product
Yieldb (%)
4a
88
89
90
OH
2
3
4b
4c
TBDMSO
OH
OH
4
5
4d
4e
86
85
OH
OH
Given the success of the Mitsunobu-mediated, N9-functionali-
zation of purine 3 with n-butanol, we were interested in examining
the scope of this reaction. The results of our investigations are
shown in Table 1. Primary and secondary aliphatic, allylic, propar-
gylic and benzylic alcohols all gave very good to excellent yields of
OH
6
4f
85
Reaction conditions: purine 3a (1 equiv), R1OH (1.1 equiv) and PPh3 (1.1 equiv)
were dissolved in anhydrous THF (0.07 M) at rt. After 2 min, DIAD (1.1 equiv) was
added dropwise.
a
Cl
Cl
N
7N
N
6
b
5
4
Isolated yield after silica gel flash column chromatography.
N1
N
a
8
N
N
Boc
2
N
NH2
b
NH2
9 H
3
N9 alkylated products, with only trace to minor amounts (<5–10%)
of the more polar (TLC, silica gel) N7 alkylated products observed.15
N9-Regioselectivity was confirmed by comparing the 1H and 13C
NMR spectra of the two regioisomers, as described in the litera-
ture.9 Importantly, reaction conditions were especially mild,
requiring just 1.1 equiv of each of the alcohol, PPh3 and DIAD,
and proceeding to completion within 15 min at room temperature,
suggesting this Mitsunobu reaction should also be compatible with
more complex purine substrates. Moreover, it is noteworthy that,
whilst yields are approximately the same, our reaction conditions
1
2
Cl
N
N
N
N
H
NHBoc
3
Scheme 1. Reagents and conditions: (a) Boc2O, cat. DMAP, DMSO, 0 °C, 30 min,
99%; (b) NaH, THF, rt, 2 h, 96%.
R3
Cl
Cl
Cl
HN
.TFA
N
N
N
a
b
c, d
N
N
N
N
N
R2
Boc
R2
N
N
H
N
N
N
NHBoc
N
NHBoc
N
N
N
N
R1
R1
R1
H
3
4
5
6
Scheme 2. Reagents and conditions: (a) (1) R1OH, PPh3, THF, rt, 2 min; (2) DIAD, rt, 15 min; (b) (1) R2OH, PPh3, THF, rt, 2 min; (2) DIAD, 35 °C, 30 min; or (1) R2OH, PBu3, THF,
rt, 2 min; (2) ADDP, rt, 4–8 h; (c) R3NH2, DIPEA, 75 °C, 6 h; (d) TFA/CH2Cl2, 1:1, rt, 1 h.

4260
S. Fletcher et al. / Tetrahedron Letters 50 (2009) 4258–4261
are significantly milder than those reported in a recent, N9 Mitsun-
obu methodology paper on a related purine derivative, likely due in
part to the enhanced solubility of our purine substrate.10
5aa–af with benzylamine as a representative primary amine pro-
ceeded in excellent yields, and subsequent Boc deprotections with
50% TFA in CH2Cl2 were quantitative (Table 3, yields are for two
steps), furnishing the 2,6,9-tri-substituted purines 6aa–af as their
TFA salts.18 6-Chloropurine derivatives such as 5 are versatile
intermediates; the chlorine atom can be readily displaced by other
nucleophiles, including alcohols,17,19 thiols17,20 and water.17,7b For
example, quantitative hydrolytic dechlorination7b of 4a or 5aa
with TFA/H2O afforded the N9-functionalized guanine 7a or the
N2,N9-di-functionalized guanine 7aa, respectively (Scheme 3),
demonstrating how our methodology may be readily adapted to
the synthesis of novel guanine analogues.
Finally, we wish to report the utility of our methodology in the
synthesis of the 2,6,9-tri-substituted purine CDK inhibitor bohe-
mine (8cb), which was thus furnished in excellent overall yield
(64%) from commercially available purine 1 (Scheme 4).
In conclusion, we have developed a novel methodology for the
facile and efficient synthesis of 2,6,9-tri-substituted purines,
including N9-mono- and N2,N9-di-functionalized guanines, which
exploits mild, sequential Mitsunobu reactions as key steps in the
syntheses. Benzylic, allylic, propargylic and aliphatic alcohols all
As shown in Table 2, we then examined the ability of the N2
NHBoc group in 4a to undergo the Mitsunobu reaction with the
same set of alcohols featured in Table 1. In this case, two different
reaction conditions were required, dependent on the nature of the
alcohol, in order for all purine 4a to be consumed. After some
experimentation, it was found that the less reactive primary and
secondary aliphatic alcohols n-butanol, 3-(tert-butyldimethylsilyl-
oxy)-propanol and iso-propanol (2.5 equiv) all effected smooth
conversion of 4a (1 equiv) to the products 5aa, 5ab and 5ac,
respectively, with the DIAD/PPh3 redox system (2.5 equiv of each)
at 35 °C. Reactions proceeded in excellent yields and were com-
plete within 30 min (Table 2, entries 1–3). In contrast, treatment
of purine 4a with the more reactive allylic, progargylic and ben-
zylic alcohols using the previously described protocol led to around
only 70% of the starting material being consumed. In addition to
the steric hindrance about the NHBoc group, which was effectively
countered, in the case of the aliphatic alcohols, by employing an
excess of the Mitsunobu reagents to compensate for their con-
sumption in side reactions, we also considered that the pKa of
the NHBoc group might be on the cusp of the ‘allowed pKa’ for
the Mitsunobu reaction. Thus, we investigated the alternative
Table 3
C6 Nucleophilic aromatic substitution followed by Boc-deprotection of purines 5aa–
1,10-azodicarbonyldipiperidine
(ADDP)/tri-n-butylphosphine
af to furnish 2,6,9-tri-substituted purines 6aa–afa
(TBP) combination, which has been shown to successfully couple
primary alcohols to less acidic HA nucleophiles (ADDP forms a
more basic betaine than DIAD).16 Gratifyingly, after optimization,
we found that allylic, propargylic and benzylic alcohols all coupled
smoothly with purine 4a in the presence of 2.5 equiv of the alcohol,
ADDP and PBu3, proceeding to completion over 4–8 h at room tem-
perature and in very good yields (Table 2, entries 4–6).
Cl
N
NHBn
N
1) BnNH2, DIPEA,
DMSO, 75 ºC, 6 h
.TFA
R2
N
N
N
N
N
R2
Boc
2) TFA/CH2Cl2,
1:1, rt, 1 h
N
N
N
H
3
3
5aa - 5af
6aa - 6af
Nucleophilic aromatic substitution of 6-chloropurines with pri-
Entry
1
Substrate
R2
Product
Yieldb (%)
mary amines is well documented.17 Indeed, reaction of compounds
5aa
5ab
5ac
5ad
5ae
6aa
6abc
6ac
6ad
6ae
92
85
93
90
91
Table 2
Reaction substrate scope in the N2 Mitsunobu coupling of purine 4a with R2OH
alcoholsa
2
3
4
5
TBDMSO
A) R2OH, PPh3
Cl
N
Cl
N
DIAD, THF, 35 ºC
N
N
N
N
or
N
N
R2
Boc
B) R2OH, PBu3
ADDP, THF, rt
NHBoc
N
3
3
4a
R2OH
5aa - 5af
Entry
1
Method
A
Time
Product
Yieldb (%)
93
6
5af
6af
93
30 min
30 min
5aa
OH
a
5abc
93
92
86
88
87
Reaction conditions: (1) a solution of purine 5aa (1 equiv), DIPEA (3 equiv) and
BnNH2 (1.5 equiv) in anhydrous DMSO (0.15 M) was heated at 75 °C for 6 h; (2) Boc-
protected analogue of purine 6aa was deprotected in a 1:1 mixture of TFA/CH2Cl2
(0.05 M) at rt for 1 h.
2
3
4
5
A
A
B
B
B
TBDMSO
OH
OH
30 min
8 h
5ac
5ad
5ae
5af
b
Isolated yield (over two steps) of TFA salt.
Partial cleavage of the TBDMS group was also observed after 1 h in TFA/CH2Cl2;
c
OH
OH
OH
complete removal was accomplished by extending the reaction time to 2 h. Product
6ab was furnished as the free base after work-up and purification by silica gel flash
column chromatography.
8 h
6
4 h
Cl
N
O
N
.TFA
R
N
N
N
N
a
N
NH
Reaction conditions: method (A) purine 4a (1 equiv), R2OH (2.5 equiv) and PPh3
(2.5 equiv) were dissolved in anhydrous THF (0.07 M) at rt. After 2 min, DIAD
(2.5 equiv) was added dropwise, then reaction was heated at 35 °C for 30 min.
Method (B) purine 4a (1 equiv), R2OH (2.5 equiv) and PBu3 (2.5 equiv) were
dissolved in anhydrous THF (0.07 M). After 2 min, ADDP (2.5 equiv) was added in
one portion. Reaction was stirred at rt for time indicated.
a
R
N
N
H
Boc
3
3
4a (R = H)
5aa (R = (CH2)3CH3)
7a (R = H)
7aa (R = (CH2)3CH3)
b
Isolated yield after silica gel flash column chromatography.
Product 5ab contaminated with ca. 10% DIAD-H2 by-product.
c
Scheme 3. Reagents and conditions: (a) TFA/H2O, 3:1, rt, 48 h, 100%.

S. Fletcher et al. / Tetrahedron Letters 50 (2009) 4258–4261
4261
Cl
N
Cl
N
8cb) associated with this article can be found, in the online version,
at doi:10.1016/j.tetlet.2009.04.137.
N
N
N
c, d
N
N
N
H
NHR
NBoc
References and notes
1 (R = H)
3 (R = Boc)
1. Legraverend, M.; Grierson, D. S. Bioorg. Med. Chem. 2006, 14, 3987–4006. and
references 132–152 therein.
5cb
a, b
OTBDMS
2. Haesslein, J. L.; Jullian, N. Curr. Top. Med. Chem. 2002, 2, 1037–1050.
3. www.cyclacel.com/cyc/investors/news/pressreleases/2006/2006-06-29
www.cyclacel.com/cyc/investors/news/pressreleases/2009/2009-03-05.
4. Bork, J. T.; Lee, J. W.; Chang, Y.-T. QSAR Comb. Sci. 2004, 23, 245–260.
5. (a) Austin, R. E.; Okonya, J. F.; Bond, D. R. S.; Al-Obeidi, F. Tetrahedron Lett. 2002,
43, 6171–6196; (b) Huang, H.; Liu, H.; Chen, K.; Jiang, H. J. Comb. Chem. 2007, 9,
197–199.
e, f
HN
N
N
N
N
6. Mitsunobu, O. Synthesis 1981, 1, 1–28.
7. (a) Choo, H.; Beadle, J. R.; Chong, Y.; Trahan, J.; Hostetler, K. Y. Bioorg. Med.
Chem. 2007, 15, 1771–1779; (b) Tang, Y.; Muthyala, R.; Vince, R. Bioorg. Med.
Chem. 2006, 14, 5866–5875; (c) Hikishima, S.; Isobe, M.; Koyanagi, S.; Soeda, S.;
Shimeno, H.; Shibuya, S.; Yokomatsu, T. Bioorg. Med. Chem. 2006, 14, 1660–
1670; (d) Kim, H. S.; Ohno, M.; Xu, B.; Kim, H. O.; Choi, Y.; Ji, X. D.; Maddileti, S.;
Marquez, V. E.; Harden, T. K.; Jacobson, K. A. J. Med. Chem. 2003, 46, 4974–4987.
8. Dey, S.; Garner, P. J. Org. Chem. 2000, 65, 7697–7699.
N
OH
H
bohemine (8cb)
Scheme 4. Reagents and conditions: (a) Boc2O, cat. DMAP, DMSO, 0 °C, 30 min,
99%; (b) NaH, THF, rt, 2 h, 96%; (c) (1) i-PrOH, PPh3, THF, rt, 2 min; (2) DIAD, rt,
15 min, 90%; (d) (1) TBDMSOCH2CH2CH2OH, PPh3, THF, rt, 2 min; (2) DIAD, 35 °C,
30 min, 93%; (e) BnNH2, DIPEA, DMSO, 75 °C, 6 h, 91%; (f) TFA/CH2Cl2, 1:1, rt, 2 h,
89%.
9. N9-Regioselectivity was confirmed by comparing the 1H and 13C NMR spectra
(in DMSO-d6) of the N9 and N7 isomers, as according to Kjellberg, J.; Johansson,
N. G. Tetrahedron 1986, 42, 6541–6544.
10. Lu, W.; Sengupta, S.; Petersen, J. L.; Akhmedov, N. G.; Shi, X. J. Org. Chem. 2007,
72, 5012–5015.
11. Maruyama, T.; Yorikane, A.; Kozai, S. Nucleosides Nucleotides Nucleic Acids 2001,
20, 935–936.
coupled in very good to excellent yields in both reactions, the first
of which was swift (reactions were complete within 15 min),
chemoselective and N9-regioselective. Significantly, our synthetic
approach to N2-functionalization of the purine nucleus obviates
the harsh conditions required by the traditional nucleophilic aro-
matic substitution of a 2-halo group with primary amines.
12. Véliz, E. A.; Beal, E. A. Tetrahedron Lett. 2006, 47, 3153–3156.
13. Chang, Y.-T.; Gray, N. S.; Rsania, G. R.; Sutherlin, D. P.; Kwon, S.; Norman, T. C.;
Sarohia, R.; Leost, M.; Meijer, L.; Schultz, P. G. Chem. Biol. 1999, 6, 361–375.
14. Perron-Sierra, F.; Dizier, D. S.; Bertrand, M.; Genton, A.; Tucker, G. C.; Casara, P.
Bioorg. Med. Chem. Lett. 2002, 12, 3291–3296.
15. For example, N9-4f: Rf = 0.46 (Hex/EtOAc, 1:9); N7-4f: Rf = 0.17 (Hex/EtOAc,
1:9). Our findings that N7-alkylated purines are more polar than their
regioisomeric N9 counterparts is consistent with reports in the literature,
such as: Geen, G. R.; Grinter, T. J.; Kincey, P. M.; Jarvest, R. L. Tetrahedron 1990,
46, 6903–6914.
Acknowledgments
16. Tsunoda, T.; Yamamiya, Y.; Itô, S. Tetrahedron Lett. 1993, 34, 1639–1642.
17. For a review, see: Legraverend, M. Tetrahedron 2008, 64, 8585–8603.
18. Purines 6aa, 6ac–af as their TFA salts were of >95% purity, as judged by 1H
NMR. The purineꢀTFA salt can be neutralized and extracted into an organic
solvent (CH2Cl2 or EtOAc) by partitioning against a saturated aqueous solution
of NaHCO3.
The authors gratefully acknowledge financial support for this
work from the Canadian Foundation of Innovation and the Univer-
sity of Toronto (Connaught Foundation).
19. For example: Lembicz, N. K.; Grant, S.; Clegg, W.; Griffin, R. J.; Heath, S. L.;
Golding, B. T. J. Chem. Soc., Perkin Trans. 1 1997, 185–186.
Supplementary data
20. For example: Sekiya, K.; Takashima, H.; Ueda, N.; Kamiya, N.; Yuasa, S.;
Fujimura, Y.; Ubasawa, M. J. Med. Chem. 2002, 45, 3138–3142.
Supplementary data (synthetic procedures and characterization
data (1H NMR, 13C NMR and MS) for purines 2, 3, 4a, 5aa, 5af and