Synthesis of Two Stable Nitrogen Analogues of S-Adenosyl-L-methionine

Article abstract of DOI:10.1021/jo9907742

Homochiral syntheses of two stable nitrogen analogues of S-adenosyl-L-methionine (AdoMet) are described. In the first analogue, AzaAdoMet, the sulfonium center of AdoMet, is replaced by an N-methyl moiety whose pK_a is 7.08. This provides a char

Full text of DOI:10.1021/jo9907742

J . Org. Chem. 1999, 64, 7467-7473
7467
Syn th esis of Tw o Sta ble Nitr ogen An a logu es of
S-Ad en osyl-L-m eth ion in e
Mark J . Thompson, Abdelaziz Mekhalfia, David P. Hornby, and G. Michael Blackburn*
Krebs Institute, Departments of Chemistry, and Molecular Biology, University of Sheffield,
Brook Hill, Sheffield S3 7HF, U.K.
Received May 11, 1999
Homochiral syntheses of two stable nitrogen analogues of S-adenosyl-L-methionine (AdoMet) are
described. In the first analogue, AzaAdoMet, the sulfonium center of AdoMet, is replaced by an
N-methyl moiety whose pKa is 7.08. This provides a charge-switchable analogue of AdoMet whose
ionic state is a function of the pH. A second analogue, MeAzaAdoMet, has a quaternary
dimethylammonium group in place of the methylsulfonium center of AdoMet: thus, its ionic state
is independent of pH.
In tr od u ction
S-Adenosyl-L-methionine (AdoMet, Figure 1) 1 is an
essential coenzyme involved in many biochemical pro-
cesses, most notably as a methyl group donor. Stable
isosteric analogues of AdoMet having variable charge
character at the 5′-locus and that cannot donate a methyl
group should show a range of activities as inhibitors of
AdoMet-dependent methyl-transfer enzymes and thereby
support mechanistic and crystallographic studies of such
proteins. Moreover, the stability of such analogues rela-
tive to AdoMet should contribute to an understanding of
the hydrolytic lability of AdoMet.1 In a previous paper,2
we described the first homochiral synthesis of the aza-
desthio analogue of 1, AzaAdoMet 2, and demonstrated
that in its protonated form, this analogue binds to the
E. coli methionine repressor protein (MetJ ) by 1 order of
magnitude more tightly than does AdoMet itself. The
unusually low pKa of AzaAdoMet, 7.08, permits the com-
pound to be employed as a charge-switchable analogue
of 1 through control of pH. However, the switch to
AzaAdoMet into its protonated, cationic form requires a
pH well below 6.5, which may not be appropriate for some
proteins. We now report full details of the homochiral
synthesis of both AzaAdoMet 2 and the previously
unreported quaternary dimethylammonium analogue
MeAzaAdoMet 3, which bears a positive charge at the
5′-position that is independent of pH.
F igu r e 1. Structure of S-adenosyl-L-methionine (AdoMet) and
its nitrogen analogues AzaAdoMet 2 and MeAzaAdoMet 3.
Resu lts a n d Discu ssion
Two previous syntheses3,4 of 2 used the routes shown
in Figure 2. The 5′-methylamino-5′-deoxyadenosine de-
rivative 4 was alkylated with a protected iodide3,4 5 or 6
followed by removal of all protecting groups. However,
because both iodides 5 and 6 were employed in racemic
form, the final analogue 2 was obtained as a mixture of
epimers. We sought to modify these syntheses by replac-
ing 5 and 6 with a chiral iodide, leading to homochiral
AzaAdoMet 2.
F igu r e 2. Previous preparations of AzaAdoMet as mixed
epimers.3,4
Amine 4a was prepared by a modification of existing
procedures (Figure 3). D-Adenosine was protected5 as the
2′,3′-O,O-(1-methylethylidene) derivative 7. Conversion6
(1) (a) Parks, L. W.; Schlenk, F. J . Biol. Chem. 1958, 230, 295. (b)
Schlenk, F. In Biochemistry of Adenosylmethionine; Salvatore, F., et
al., Eds.; Columbia University Press: New York, 1977; p 7.
(2) Thompson, M. J .; Mekhalfia, A.; J akeman, D. L.; Phillips, S. E.
V.; Phillips, K.; Porter, J .; Blackburn, G. M. J . Chem. Soc., Chem.
Commun. 1996, 791.
(3) Minnick, A. A.; Kenyon, G. L. J . Org. Chem. 1988, 53, 4952.
(4) Davis, M.; Dudman, N. P. B.; White, H. F. Aust. J . Chem. 1983,
36, 1623.
(5) Tomasz, J . In Synthetic Procedures in Nucleic Acid Chemistry
(Vol. 2); Tipson, R. S., Zorbach, W. W., Eds.; Wiley-Interscience: New
York, 1973; p 765.
(6) J ahn, W. Chem. Ber. 1965, 98, 1705.
10.1021/jo9907742 CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/09/1999

7468 J . Org. Chem., Vol. 64, No. 20, 1999
Thompson et al.
F igu r e 3. Preparation of 5′-methylamino-5′-deoxy-2′,3′-O,O-
(1-methylethylidene) adenosine 4a and coproduct 4b from
D-adenosine.
F igu r e 5. Structure of iodide 10 as determined by X-ray
crystallography.
variable largely because of its instability on silica. Direct
conversion of 9 into the iodide 10 was attempted using
iodobenzene diacetate (IBDA) and iodine.12 Although this
reaction proceeded well, separation of iodide 10 from the
iodobenzene formed in the reaction proved difficult and
low yields were obtained.
The key intermediates 4a and 10 were heated together
in acetonitrile solution in the presence of N,N-diisopro-
pylethylamine to give the fully protected analogue 11.
Attempts to remove the benzyloxycarbonyl group from
11 by catalytic hydrogenation with 10% Pd-C proved
unsuccessful; reactions were incomplete after several
days, and mixtures of products were detected by TLC.
Use of freshly prepared Pd black13 as the catalyst drove
the reaction to completion, but these forcing conditions
resulted in fragmentation of the molecule. Catalytic
transfer hydrogenation using cyclohexene,14 1,4-cyclo-
hexadiene,15 or formic acid16 as hydrogen donor also
afforded mixtures of products. However, treatment of 11
with BF3-EtSH complex17 was found to effect complete
removal of all protecting groups. The final product
AzaAdoMet 2 was purified on Dowex 50WX4-400 ion-
exchange resin.3
F igu r e 4. Synthesis of the chiral protected iodide 10 from
L-glutamic acid and its use in the preparation of AzaAdoMet
2.
into 5′-O-tosylate 8 was followed by reaction with neat
liquid methylamine7 to give amine 4a in 46% overall
yield. In the reaction of the 5′-O-tosylate 8 with methyl-
amine, a 3% yield of the 5′-dimethylamino derivative 4b
was also isolated. This material was identified positively
as 4b by comparison with a sample prepared by the
reaction of 5′-O-tosylate 8 with dimethylamine. The 5′-
O-mesylate derivative was also explored but gave con-
siderably lower yields of 4a .
The pKa of 2 was determined by NMR titration by
1
monitoring the NMe H signal (singlet at δ 2.3 at high
pH changing to a doublet at δ 2.9 at low pH). The data
so obtained (Figure 6) reveal a major shift response to
the titration of the tertiary MeN group (pH 5.5-8.5) and
a smaller response to a second ionization (pH 8.5-10.5),
which we assign to the R-amino function. These data were
analyzed using18 Kaleidagraph to give pKa values of 7.08
( 0.07 (>NMe) and 9.66 ( 0.31 (-NH2). Thus, analogue
2 may be employed either in its protonated (net cationic)
form or in its unprotonated (net zwitterionic) form by
controlling pH in the range 5.5-8.5. The NMR solutions
of 2 (H2O with 10% D2O, ionic strength 10 mM, 25 °C)
used in these titrations showed no change in the signals
(S)-Glutamic acid was protected8 as the oxazolidinone
9 (Figure 4), which was treated with oxalyl chloride to
afford the corresponding acid chloride and subsequently
converted into the alkyl iodide 10 via a Barton-type
decarboxylative iodination9 using 2-iodo-1,1,1-trifluoro-
ethane as iodine donor.10 While the transformation of (S)-
glutamic acid into 10 uses procedures of reliable stereo-
chemical integrity, the (S)-absolute configuration of 10
was confirmed by anomalous dispersion X-ray crystal-
lography (Figure 5).11 The yield of 10 was found to be
(12) Concepcio´n, J . I.; Francisco, C. G.; Friere, R.; Herna´ndez, R.;
Salazar, J . A.; Sua´rez, E. J . Org. Chem. 1986, 51, 402.
(13) J ung, M. E. J . Org. Chem. 1992, 57, 3528.
(7) Murayama, A.; J astorff, B.; Cramer, F.; Hettler, H. J . Org. Chem.
1971, 36, 3029.
(8) Itoh, M. Chem. Pharm. Bull. 1969, 17, 1679.
(9) Barton, D. H. R.; Crich, D.; Motherwell, W. B. Tetrahedron Lett.
1983, 24, 4979.
(14) J ackson, A. E.; J ohnstone, R. A. W. Synthesis 1976, 685.
(15) Felix, A. M.; Heimer, E. P.; Lambros, T. J .; Tzougraki, C.;
Meienhofer, J . J . Org. Chem. 1978, 43, 4194.
(16) ElAmin, B.; Anantharamaiah, G. M.; Royer, G. P.; Means, G.
E. J . Org. Chem. 1979, 44, 3442.
(10) Eaton, P. E.; Tsanaktsidis, J . Tetrahedron Lett. 1989, 30, 6967.
(11) Coordinates for this structure are provided in the Supporting
Information.
(17) Bose, D. S.; Thurston, D. E. Tetrahedron Lett. 1990, 31, 6903.
(18) KaleidaGraph 3.0.5, Abelbeck Software, Reading, PA, 1986-
1994.

Synthesis of Nitrogen Analogues of AdoMet
J . Org. Chem., Vol. 64, No. 20, 1999 7469
F igu r e 7. Syn conformations of (a ) AdoMet 1 and (b)
AzaAdoMet 2 excised from X-ray structures of binary com-
plexes with MetJ repressor protein.2
F igu r e 6. Plot of N-Me 1H chemical shift for 2 vs pH at 25
°C ionic strength 10 mM in H2O/D2O (9:1 v/v). Curve calculated
1
2
for overlapping pKa 7.0756 ( 0.076 (step 0.5228 ppm) and pKa
9.662 ( 0.259 (step 0.0566 ppm).
either for the nucleoside or for the diaminobutyrate
moieties over the course of these titrations or for several
months afterward. It thus appears that AzaAdoMet
exhibits none of the hydrolytic instability of AdoMet itself
either with respect to glycosidic hydrolysis or to homo-
serine lactone formation,19 both of which must therefore
be attributable to the activity of the sulfonium center in
AdoMet.
This charge-switchable nature of AzaAdoMet 2 has
been utilized in an investigation of the MetJ -DNA-
AdoMet complex,20 where Phillips proposed2 that the
positive charge at the 5′-position of the corepressor
(AdoMet) is the major feature responsible for the en-
hanced stability of the ternary repressor/corepressor/
operator complex relative to the binary repressor/operator
complex.21,22 At pH 7.4, in its uncharged form, the binding
of AzaAdoMet to the MetJ protein was found to be very
weak. However, at pH 5.5 when the protonated form
predominates, the binding of AzaAdoMet to MetJ was
found to be 10 times stronger than for AdoMet itself at
the same pH.22 This result amply demonstrates the
charge-switchable nature of AzaAdoMet.
F igu r e 8. Bromoaminobutyrate 12 was prepared and inves-
tigated as an alternative to iodide 10 for synthesis of 2. This
was used to prepare protected AzaAdoMet 15.
AdoMet and AzaAdoMet 2 lies approximately 6 kcal
mol-1 above the favored C2′-endo conformation, which
shows that the MetJ protein binds the corepressor in a
high-energy conformation.
In a study of the binding of AzaAdoMet to cobalt
precorrin-4-methyltransferase (CbiF), preliminary X-ray
analysis of the binary complex shows the nucleoside
adopts the more favored anti conformation.24 The ob-
served pucker of CbiF-bound AzaAdoMet is C2′-exo-C3′-
endo (3T2), very close in energy to the C3′-endo minimum.
C2′-endo and C3′-endo are equally populated in anti-purine
nucleosides, unlike the syn case where C2′-endo is pre-
ferred.
We next sought to improve our synthesis of 2 because
of difficulties encountered in the deprotection of 11, so
we explored alternative protecting group strategies. We
envisaged that iodide 10 might be replaced by a bromide
such as 12 (Figure 8), derived from homoserine lactone
(R-amino-γ-butyrolactone). The synthesis25 of 12 was
carried out with racemic lactone in order to evaluate the
viability of this route. Ring opening of DL-homoserine
lactone under strongly acidic conditions gave R-(2-bro-
moethyl)glycine25,26 13. Esterification25,26 to give 14 fol-
lowed by N-Boc protection25 gave bromide 12 in moderate
yield. This compound was then used in the alkylation of
amine 4a under the same conditions as those described
above. However, the maximum yield of alkylated product
An X-ray crystal structure2 of the resulting binary
complex provides independent confirmation of the abso-
lute configuration of 2. The details of this structure show
that for both the natural co-repressor, AdoMet (Figure
7a), and the isosteric analogue, AzaAdoMet 2 (Figure 7b),
the glycosylic bond of the bound nucleoside adopts the
higher energy syn-conformation (+synclinal) with torsion
angle ø +76° for AdoMet (+77° for AzaAdoMet). The
0
ribose ring pucker is T4, with largely O4′-endo character.
In the case of purine nucleosides adopting the syn
conformation, the observed sugar pucker of O4′-endo lies
at the energy maximum of the interconversion between
C2′-endo and C3′-endo, the C3′-endo conformation being
1-2 kcal mol-1 higher in energy; hence, C2′-endo rather
than C3′-endo pucker is more often observed in these
cases.23 The O4′-endo pucker seen here for protein-bound
(19) Hoffman, J . L. Biochemistry 1986, 25, 4444.
(20) The coordinates of both the MetJ -DNA-AdoMet and MetJ -
DNA-AzaAdoMet complexes have been deposited in the Brookhaven
Protein Database.
(21) Phillips, K.; Phillips, S. E. V. Structure 1994, 2, 309.
(22) Parsons, I. D.; Persson, B.; Mekhalfia, M. A.; Blackburn, G. M.;
Stockley, P. G. Nucleic Acids Res. 1995, 23, 211.
(23) Saenger, W. Principles of Nucleic Acid Structure; Verlag: New
York, 1983; p 69.
(24) Schubert, H. Unpublished results.
(25) Bajgrowicz, J . A.; El Hallaoui, A.; J acquier, R.; Pigie`re, C.;
Viallefont, P. Tetrahedron Lett. 1984, 25, 2231.
(26) Koch, T.; Buchardt, O. Synthesis 1993, 1065.

7470 J . Org. Chem., Vol. 64, No. 20, 1999
Thompson et al.
F igu r e 9. Use of alternative 2′,3′-O-ribose protecting groups
in the synthesis of 5′-amino derivatives of adenosine.
15 was found to be only 33%, suggesting that bromide
12 may be unstable at the reaction temperature (55 °C).
Furthermore, attempts to remove the N-Boc and 1-me-
thylethylidene protecting groups from 15 using TMSBr
or TMSI proved unsuccessful.
Alternative protecting groups were then considered for
the ribose moiety. 2′,3′-O,O-Benzylidene adenosine27 was
converted into the 5′-O-tosylate 16, which was treated
with liquid methylamine (Figure 9), but the desired 5′-
methylamino compound 17a was only isolated as minor
product. Unexpectedly, the major product (2:1 ratio) from
this reaction (51% yield) was found to be a 5′-dimethyl-
amino derivative, shown to be 17b by NMR analysis and
unambiguous synthesis. 2′,3′-Di-O-TBDMS adenosine 18
was synthesized by the method of Ogilvie et al.28 and then
treated with p-toluenesulfonyl chloride followed by liquid
methylamine. The only product isolated after chroma-
tography was again found to be 5′-dimethylamino com-
pound 19, shown as previously to be identical by 1H NMR
and MS to a sample prepared by using liquid dimethyl-
amine in place of liquid methylamine. This formation of
the 5′-dimethylamino compounds 17b and 19 from the
corresponding 5′-O-tosylates and pure methylamine29
appears unprecedented. While such a methyl transfer
might involve intermediate formation of an N1-methyl-
adenosine species via a Dimroth-like transamination30
followed by transmethylation from N1 to N5′, such a
process might well be expected to lead to some N6-
methylation product and none was observed. An alterna-
tive possibility is that toluenesulfonic acid produced in
the reaction catalyzes the formation of some dimethyl-
amine in situ (by disproportionation of methylamine),
which can then react with 5′-O-tosyladenosine.31
F igu r e 10. Synthesis of the MeAzaAdoMet 3 and side-
reaction generating pyrrolidone 22.
The synthesis of MeAzaAdoMet 3 used 2′,3′-di-O-
TBDMS protection for the adenosine ribose moiety. For
the amino acid, protecting groups removable by hydro-
genation were selected, the N-Z-group being preferred.
2′,3′-Di-O-TBDMS adenosine 18 was treated with p-
toluenesulfonyl chloride in dry pyridine (Figure 10), and
the resulting crude 5′-O-tosylate was stirred in liquid
dimethylamine to afford the 5′-(dimethylamino)-5′-deoxy
compound 19 in excellent overall yield. N-Z-(S)-glutamic
acid R-benzyl ester 20 was treated with oxalyl chloride
to generate the acid chloride and the crude product
reacted under the Barton decarboxylative iodination
conditions previously described for the synthesis of 10.
However, the product from this procedure was not iodide
21 but rather the γ-lactam product 22 resulting from
intramolecular cyclization of the acid chloride. Synthesis
of iodide 21 from the acid 20 was achieved directly in
moderate yield by use of IBDA-I2.12
(27) Ponpipom, M. M.; Hanessian, S. Can. J . Chem. 1972, 50, 246.
(28) Ogilvie, K. K.; Beaucage, S. L.; Schifman, A. L.; Theriault, N.
Y.; Sadana, K. L. Can. J . Chem. 1978, 56, 2768.
(29) The possibility of a Dimroth-type rearrangement to give N6-
methyl-5′-(dimethylamino)-5′-deoxy-2′,3′-O,O-(1-methylethylidine)ad-
enosine was considered, but no substitution of the adenine ring could
be identified. 1H NMR (400 MHz, DMSO-d6) of 4a and 4b showed the
two methyl groups to be equivalent (s, 6H in both cases). Mass
spectroscopy data (CI+) showed peaks at 321 (MH+), 186 (MH - Ade+),
and 136 (AdeH+), indicating that no substitution of the adenine ring
had occurred. The purity of the methylamine used in the reaction was
determined by GC and its tR (5.31 min) compared with that of pure
dimethylamine (7.80 min). We found that the methylamine used in
the reactions described was greater than 99% pure and contained no
dimethylamine.
The tertiary amine 19 was reacted with iodide 21 to
form the quaternary dimethylammonium compound,
fully protected analogue 23. This reaction was initially
attempted at reflux, but at this temperature the major
process was that of benzyl transfer onto the 5′-nitrogen.
At lower temperature, the reaction was slower but more
selective and gave the quaternary dimethylammonium
salt 23 as the major product, purified by column chro-
matography.
(30) For a recent review, see: Fujii, T.; Itaya, T. Heterocycles 1998,
48, 359.
(31) A reviewer has suggested that dimethylamine might arise
through methyl transfer from N,N-ditosylmethylamine. Experiments
to explore these alternative possibilities are in progress.

Synthesis of Nitrogen Analogues of AdoMet
J . Org. Chem., Vol. 64, No. 20, 1999 7471
ase HhaI and is competitive with AdoMet 1.34 The
affinities of the analogues AzaAdoMet 2 and MeAzaA-
doMet 3 for a range of proteins that utilize AdoMet and
crystal structures of their complexes are under investiga-
tion by a number of research groups, and results of such
studies will be reported in due course.
Exp er im en ta l Section
DCM was refluxed over P2O5 or CaH2 and distilled prior to
use. Pyridine was refluxed over CaH2 and distilled. Anhydrous
grade benzene, carbon tetrachloride, DMF, and acetonitrile
and HPLC grade methanol and acetone were purchased from
Aldrich Chemical Co. and used as supplied. J values are given
in Hz. [R]D values are given in deg cm2 g-1. Melting points are
uncorrected.
5′-p-Tolu en esu lfon yl-2′,3′-O,O-(1-m eth yleth ylid en e)a d -
en osin e6 8. 2′,3′-O,O-(1-Methylethylidene)adenosine 12 (10.05
g, 32.7 mmol)5 was dissolved in dry pyridine (60 mL) with
warming, and then the solution was cooled to -20 °C.
p-Toluenesulfonyl chloride (7.06 g, 37.0 mmol, 1.1 equiv) was
added and then the solution stored at -20 °C for 3 d with
occasional shaking. This mixture was poured into chloroform
(200 mL) and extracted with cold 3 M sulfuric acid (360 mL).
The chloroform solution was concentrated to 50 mL by rotary
evaporation at room temperature and then the product pre-
cipitated by addition of light petroleum (500 mL). The solid
was collected by filtration, dissolved in methanol (60 mL), and
placed in the freezer for 2 h. The product was again filtered,
washed with cold methanol and then light petroleum, and
transferred to a 1 L round-bottom flask. Chloroform (15 mL)
was added followed by light petroleum (300 mL) and the
suspension allowed to stand in the freezer for 3 h. The solid
was collected by filtration, washed with light petroleum and
then ether, and dried thoroughly under high vacuum giving 8
as an off-white to pale yellow powder (13.26 g, 88%): 1H NMR
(250 MHz, DMSO-d6) δ 8.51 (s, 1H), 8.39 (s, 1H), 7.62 (d, 2H,
J ) 8.6), 7.30 (d, 2H, J ) 8.2), 6.25 (d, 1H, J ) 2.1), 5.35 (dd,
1H, J ) 2.1, 6.1), 4.97 (dd, 1H, J ) 3.4, 6.1), 4.45-4.37 (m,
1H), 4.37-4.16 (m, 2H), 2.38 (s, 3H), 1.52 (s, 3H), 1.30 (s, 3H);
m/z (FAB) 462 (MH+).
F igu r e 11. Improved synthesis of AzaAdoMeMet 3.
Treatment of 23 with TBAF removed both TBDMS
groups and the benzyl ester, but this crude product could
not be separated from TBAF itself. Use of ammonium
fluoride in methanol at 50 °C resulted in a mixture of
both TBDMS deprotection and transesterification of the
benzyl ester. Catalytic hydrogenation using 10% Pd-C
catalyst again proved slow, taking 3 d for complete
disappearance of starting material while giving a mixture
of products, and indeed, electrospray MS of the crude
mixture showed that the required product 23 was not
present. Catalytic transfer hydrogenation of 23 was found
to result in transesterification of the benzyl ester leaving
the Z group intact.
Finally, the fully protected species 23 was treated
BF3-EtSH complex, which effected complete removal of
the Z group and both TBDMS groups together with
partial removal of the benzyl ester (Figure 10). This crude
mixture of analogue 3 and its benzyl ester was treated
under catalytic transfer hydrogenation conditions in
aqueous methanol to avoid transesterification, and the
final product was purified by ion exchange. Catalytic
(Pd-BaSO4 in MeOH-H2O, 14%) and transfer (MeOH-
H2O, 27%) hydrogenation conditions were not as effective
as treatment with tetra-n-butylammonium hydroxide32
(MeOH-H2O, 42%), although it could not be established
whether debenzylation was complete.
Because of these difficulties, iodide 21 was replaced
by 10 to give a different quaternary dimethylammonium,
fully protected product 24 (Figure 11) which was depro-
tected by BF3-EtSH treatment alone. Although purifica-
tion of 3 proved problematic, it was isolated in a final
yield of 44%. The X-ray crystal structure of a ternary
MetJ -DNA-MeAzaAdoMet complex has recently been
solved33 to 2.0 Å, which confirms the absolute configu-
ration of analogue 3 and establishes its binding confor-
mation as close to that of AdoMet itself.
(4S)-N-Ben zyloxyca r bon yl-4-(2-iod oeth yl)oxa zolid in -
5-on e 10. The oxazolidin-5-one 9 (3.73 g, 12.7 mmol) was
dissolved in dry DCM (125 mL) under nitrogen. Distilled oxalyl
chloride (1.69 g, 1.16 mL, 13.3 mmol, 1.05 equiv) was added
followed by DMF (2 drops) and the mixture stirred for 1 h.
TLC (MeOH/DCM 1:19) showed complete disappearance of
starting material so the solvent was evaporated under vacuum.
The residue was twice redissolved in DCM and evaporated and
then dried under high vacuum to give a pale yellow solid (3.74
g, 94%). This crude acid chloride was used in the next step
without further purification. Dry N-hydroxypyridine-2-thione
sodium salt (1.50 g, 10.1 mmol, 1.1 equiv), DMAP (65 mg, 0.53
mmol), and 2,2,2-trifluoro-1-iodoethane (8.6 g, 4.05 mL, 41
mmol, 5.0 equiv) were added to dry DCM (55 mL) under argon.
The mixture was brought to reflux (heated with an aluminum
ring), and then a solution of the crude acid chloride (2.56 g,
8.2 mmol) in dry DCM (29 mL) was added over 10 min. As
addition was started the mixture was irradiated with two
tungsten filament lamps (100 and 150 W), and after addition
was complete, reflux was continued under irradiation for a
further 1 h. The reaction mixture was then washed with water
(50 mL), concentrated HCl (25 mL), and water (50 mL) and
evaporated under vacuum. The product was purified by flash
column chromatography on silica (DCM) and crystallized from
dry ether to give 10 as a white solid (1.33 g, 42%): 1H NMR
(250 MHz, CDCl3) δ 7.34-7.28 (br s, 5H), 5.54-5.46 (s, 1H),
5.20 (d, 1H, J ) 4.0), 5.14 (m, 2H), 4.31 (t, 1H, J ) 6.0), 3.15
(t, 2H, J ) 7.3), 2.5-2.3 (br m, 2H); 13C NMR (60 MHz, CDCl3)
δ 171.2, 153.0, 137.5, 135.2, 130.3, 128.8, 128.5, 127.5, 78.0,
68.2, 55.3, 34.8, -2.3; m/z (FAB) 375 (M+).
In preliminary experiments, MeAzaAdoMet 3 has been
confirmed as a nanomolar inhibitor of the DNA methyl-
(32) Abdel-Magid, A.; Cohen, J . M.; Maryanoff, C. A.; Shah, R. D.;
Villani, F. J .; Zhang, F. D. Tetrahedron Lett. 1998, 39, 3391.
(33) Porter, J . Uunpublished results.
(34) Hornby, D. P. Personal communication.

7472 J . Org. Chem., Vol. 64, No. 20, 1999
Thompson et al.
(4S)-N-Ben zyloxyca r bon yl-4-[2-[N-2′,3′-O,O-(1-m eth yl-
eth ylid en e)a d en osyl-N-m eth yl]a m in oeth yl]oxa zolid in -5-
on e 11. A solution of iodide 10 (1.32 g, 3.52 mmol, 1.1 equiv)
in anhydrous acetonitrile (25 mL) was added to a stirred
solution of amine 4 (1.03 g, 3.22 mmol) and N,N-diisopropyl-
ethylamine (0.45 g, 0.61 mL, 3.5 mmol, 1.1 equiv) in the same
solvent (40 mL) under argon. The solution was heated at 55
°C for 2 d and then the solvent removed under vacuum and
the residue twice purified by flash column chromatography
on silica using (Et3N/MeOH/EtOAc, 1:1:98) and then (propan-
2-ol/DCM, 7.5:92.5) as eluent to give a pale yellow oil. This
product was twice dissolved in DCM and evaporated to yield
the title compound as a white foam (0.93 g, 51%): 1H NMR
(250 MHz, CDCl3) δ 8.29 (s, 1H), 7.87 (s, 1H), 7.40-7.20 (br s,
5H), 5.96 (d, 1H, J ) 2.1), 5.84 (s, 2H), 5.50-5.35 (br m, 2H),
5.20 (d, 1H, J ) 4.0), 5.13 (m, 2H), 4.83 (dd, 1H, J ) 4.0, 6.4),
4.35-4.19 (br m, 2H), 2.62-1.78 (br m, 9H), 1.54 (s, 3H), 1.33
(s, 3H); 13C NMR (60 MHz, CDCl3) δ 172.1, 155.7, 149.2, 128.7,
128.4, 128.3, 128.0, 127.7, 120.2, 114.5, 90.8, 84.9, 83.9, 83.3,
81.7, 67.5, 60.1, 59.3, 57.6, 57.0, 54.5, 52.9, 42.2, 41.0, 27.1,
(MeOH/EtOAc, 1:19) giving an oil. Ether was twice added and
evaporated to give the title compound as a white foam (125
mg, 33%): 1H NMR (250 MHz, CDCl3) δ 8.28 (s, 1H), 7.90 (s,
1H), 6.01 (d, 1H, J ) 1.8), 5.90 (s, 2H), 5.85-5.74 (br m, 1H),
5.44 (t, 1H, J ) 6.7), 4.4-4.2 (br m, 1H), 3.65 (d, 3H, J ) 3.7),
2.62-2.21 (br m, 4H), 2.16 (s, 3H), 1.94-1.66 (br m, 2H), 1.55
(s, 3H), 1.39-1.24 (br m, 12H); HRMS calcd for C24H37N7O7
536.2827, found 536.2833.
5′-Meth yla m in o-5′-d eoxy-2′,3′-O,O-ben zylid en ea d en o-
sin e 17a a n d 5′-Dim eth yla m in o-5′-d eoxy-2′,3′-O,O-ben -
zylid en ea d en osin e 17b. 2′,3′-O,O-Benzylidene adenosine
(0.31 g, 0.87 mmol) was dissolved in dry pyridine (4 mL) with
warming. This solution was cooled to -20 °C (freezer), and
then p-toluenesulfonyl chloride (183 mg, 0.96 mmol, 1.1 equiv)
was added. The mixture was kept at -20 °C for 3 d with
occasional shaking and then diluted with DCM (20 mL) and
washed with cold 3 M sulfuric acid (30 mL). The 5′-O-tosylate
16 (161 mg, 36%) was precipitated by addition of light
petroleum, filtered off, recrystallized from methanol, and used
directly in the next step without further purification. 5′-O-
Tosylate 16 (161 mg, 0.32 mmol) was transferred to a bomb,
and liquid methylamine (8 mL) was condensed onto it. The
bomb was sealed and the mixture stirred at room temperature
for 3 d. The excess amine was allowed to evaporate, and then
the residue was taken up in DCM (30 mL) and washed with
aqueous NaOH (0.75 M, 10 mL). The aqueous washings were
back-extracted with DCM (30 mL) and the combined organic
solutions evaporated. The crude mixture was then purified by
flash column chromatography (MeOH/DCM, 1:4). The major
product 17b eluted first (59 mg, 51%): 1H NMR (250 MHz,
CDCl3) δ 8.31 (s, 1H), 7.93 (s, 1H), 7.55-7.47 (m, 2H), 7.43-
7.32 (m, 3H), 6.16 (d, 1H, J ) 3), 5.98 (s, 1H), 5.65-5.53 (m,
3H), 5.02 (dd, 1H, J ) 2, 6), 4.57-4.48 (m, 1H), 2.70-2.48 (m,
2H), 2.23 (s, 6H); m/z (FAB) 383 (MH+). The anticipated
product 17a eluted later (31 mg, 27%): 1H NMR (250 MHz,
CDCl3) δ 8.28(s, 1H), 7.75 (s, 1H), 7.56-7.45 (m, 2H), 7.43-
7.30 (m, 3H), 6.08 (d, 1H, J ) 3), 5.99 (s, 1H), 5.90 (s, 2H),
5.62 (dd, 1H, J ) 3, 6.5), 5.11 (dd, 1H, J ) 2.1, 6.4), 4.51 (dt,
1H, J ) 2.1, 5), 2.95-2.82 (m, 2H), 2.40 (s, 3H); HRMS calcd
for C18H20N6O3 369.1675, found 369.1695.
5′-Dim eth yla m in o-5′-d eoxy-2′,3′-bis(ter t-bu tyld im eth -
ylsilyl)a d en osin e 19. 2′,3′-Bis(tert-butyldimethylsilyl)adenos-
ine 12 (230 mg, 0.46 mmol) was dissolved in dry pyridine (5
mL), and the solution was cooled to -20 °C. p-Toluenesulfonyl
chloride (133 mg, 0.70 mmol, 1.5 equiv) was added and the
mixture kept at -20 °C for 3 d with occasional shaking. The
solution was then diluted with DCM (40 mL) and washed with
cold 3 M sulfuric acid (30 mL). The DCM layer was transferred
to a bomb and the solvent evaporated under a stream of
nitrogen. Dimethylamine (5 mL) was then condensed onto the
crude 5′-O-tosylate and the bomb sealed. The mixture was
allowed to stir for 3 d at room temperature then the excess
amine allowed to evaporate. The residue was taken up in DCM
(30 mL) and washed with aqueous NaOH (0.75 M, 10 mL).
The aqueous layer was back-extracted with further DCM (10
mL), and then the combined DCM extracts were evaporated.
This crude product was purified by flash column chromatog-
raphy (MeOH/DCM, 1:9) to give the title compound as an off-
white solid (174 mg, 72% from 12): 1H NMR (250 MHz, CDCl3)
δ 8.21 (s, 1H,), 7.79 (s, 1H), 5.73 (d, 1H, J ) 5.2), 5.54 (s, 2H),
4.86 (t, 1H, J ) 4.4), 4.14-4.05 (m, 2H), 2.81-2.68 (m, 1H),
2.52-2.40 (m, 1H), 2.20 (s, 6H), 0.82 (s, 9H), 0.69 (s, 9H), 0.00
(s, 3H), -0.01 (s, 3H), -0.16 (s, 3H), -0.36 (s, 3H); 13C NMR
(100 MHz, CDCl3) δ 155.4, 152.8, 149.6, 140.6, 120.8, 90.0,
82.7, 74.3, 73.9, 61.7, 46.0, 25.9, 25.7, 18.1, 17.9, -4.3, -4.7,
-5.0; HRMS calcd for C24H47N6O3Si2 523.3248, found 523.3255.
Ben zyl N-Ben zyloxyca r b on yl-L-r-(2-iod oet h yl)glyci-
n a te 21. N-Benzyloxycarbonyl-L-glutamic acid R-benzyl ester
14 (0.37 g, 1.0 mmol) was dissolved in anhydrous carbon
tetrachloride (40 mL) under argon, and then iodobenzene
diacetate (0.18 g, 0.56 mmol, 0.55 equiv) and iodine (0.13 g,
0.51 mmol, 1.0 equiv) were added. The mixture was heated
(using an aluminum ring) at reflux for 90 min under irradia-
tion from two tungsten filament lamps (100 and 150 W).
Further portions of iodobenzene diacetate (0.18 g, 0.56 mmol,
25.3, 14.1; HRMS calcd for
568.2511.
C27H33N7O7 568.2520, found
5′-[N-[(3S)-3-Am in o-3-car boxypr opyl]-N-m eth ylam in o]-
5′-d eoxya d en osin e (Aza Ad oMet) 2. A solution of 11 (1.17
g, 2.06 mmol) in dry DCM (16 mL) was added via syringe to a
stirred mixture of boron trifluoride etherate (3.0 g, 2.6 mL,
21 mmol, 10 equiv) and ethanethiol (4.0 g, 4.8 mL, 62 mmol,
30 equiv) under nitrogen at 0 °C (ice bath). After 30 min, the
ice bath was removed and the reaction was continued at room
temperature for a further 23 h. Solvent was removed under
vacuum (using a KMnO4 trap) and the residue coevaporated
with methanol (3 × 10 mL) and then taken up in water (10
mL). This suspension was filtered and then lyophilized. The
sticky gum formed was taken up in water, filtered again, and
then loaded to a column of Dowex 50WX4-400 ion-exchange
resin (100 g). The column was washed with water (250 mL)
and then eluted with a gradient of NH4HCO3 (0-1.2 M in 3.0
L). Fractions from the major UV-active peak were collected
and evaporated under vacuum and then coevaporated with
water. The residue was redissolved in a small volume of water
and lyophilized to give pure analogue 2 as a white powder (535
mg, 68%): 1H NMR (400 MHz, D2O) δ 8.10 (s, 1H), 7.98 (s,
1H), 5.88 (d, 1H, J ) 5), 4.61 (t, 1H, J ) 5), 4.25-4.19 (m,
1H), 4.11 (t, 1H, J ) 5), 3.60 (dd, 1H), 2.88-2.76 (m, 2H), 2.73-
2.61 (m, 2H), 2.27 (s, 3H), 2.01-1.91 (m, 1H), 1.88-1.78 (m,
1H); 13C NMR (100 MHz, D2O) δ 174.0, 154.7, 152.0, 147.9,
139.4, 118.1, 87.5, 79.9, 72.6, 71.3, 58.5, 53.8, 53.5, 40.2, 25.7;
HRMS calcd for C15H24N7O5 382.1839, found 382.1933.
p Ka Deter m in a tion for Aza Ad oMet 2. AzaAdoMet 2 (20
mg) was dissolved in 10% D2O in H2O (500 mm3) at pH 3.0
ionic strength 10 mM (KCl) and acetone (2 mm3) added as the
NMR internal reference (δ ) 2.214). Proton spectra were
recorded at 500.13 MHz on a Bruker AMX 500 spectrometer
at 298 K. An aliquot of KOH (0.01 M) was added, the solution
homogenized, and the pH determined at 298 K in the NMR
tube using a Russell RL150 pH meter model fitted with a
combination glass electrode (error ( 0.03 pH unit). This
process was repeated stepwise to give 26 data points in the
range 3 < pH < 11.4. The chemical shift for the N-methyl
signal of 2 was plotted as a function of pH (adjusted to
compensate for the D2O-H2O mixture)35 (Figure 6) and
analyzed for two overlapping pKa values using18 Kaleidagraph
to give values of 7.08 ( 0.07 and 9.66 ( 0.31 (Figure 6).
Meth yl [4-[N-2′,3′-O,O-(1-m eth yleth ylid en e)a d en osyl-
N -m e t h yl]a m in o]-2-(t er t -b u t yloxyca r b on yla m in o)b u -
ta n oa te 15. Amine 4 (0.23 g, 0.72 mmol) was dissolved in dry
acetonitrile (15 mL) under nitrogen and then N,N-diisopro-
pylethylamine (0.15 g, 0.20 mL, 1.16 mmol, 1.6 equiv) added
followed by a solution of bromide 12 (0.34 g, 1.15 mmol, 1.6
equiv) in dry acetonitrile (10 mL). The solution was heated at
55 °C for 3 d, and then the solvent was removed under vacuum
and the residue purified by flash column chromatography
(35) Glasoe, P. K.; Long, F. A. J . Phys. Chem. 1960, 64, 188.

Synthesis of Nitrogen Analogues of AdoMet
J . Org. Chem., Vol. 64, No. 20, 1999 7473
0.55 equiv) and iodine (0.13 g, 0.51 mmol, 1.0 equiv) were
added, and then the mixture was refluxed under irradiation
for a further 90 min. The solution was allowed to cool to room
temperature, washed with aqueous sodium thiosulfate until
colorless, and evaporated under vacuum. The residue was
purified by flash column chromatography (EtOAc/light petro-
leum, 1:4) to give the title compound as a white solid (181 mg,
40%): 1H NMR (250 MHz, CDCl3) δ 7.37-7.15 (br m, 10H),
5.30 (s, 1H, J ) 12), 5.12 (s, 2H), 5.05 (s, 2H), 4.45-4.36 (m,
1H), 3.13-2.98 (m, 2H), 2.47-2.29 (m, 1H), 2.23-2.05 (m, 1H);
13C NMR (100 MHz, CDCl3) δ 171.0, 155.8, 135.9, 134.9,
128.69, 128.63, 128.56, 128.33, 128.28, 128.15, 67.6, 67.2, 54.7,
36.8, -1.1; HRMS (CI) calcd for C19H21INO4 454.0515, found
454.0496.
to give the quaternary dimethylammonium salt 24 as a white
powder (254 mg, 59%): 1H NMR (250 MHz, CDCl3) δ 8.50 (s,
1H), 8.17 (s, 1H), 7.48-7.24 (m, 6H), 6.79-6.53 (br m, 1H),
5.93 (d, 1H, J ) 6.4), 5.29-5.06 (m, 3H), 4.90 (d, 1H, J ) 11.3),
4.77 (t, 1H, J ) 12.8), 4.64-4.48 (m, 2H), 4.48-4.40 (br m,
1H), 4.17-3.98 (m, 2H), 3.33 (s, 3H), 3.15 (s, 3H), 2.7-2.3 (br
m, 4H), 1.01 (s, 9H), 0.81 (s, 9H), 0.28 (s, 3H), 0.22 (s, 3H),
0.00 (s, 3H), -0.38 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 170.7,
155.6, 155.3, 153.0, 149.0, 141.2, 136.1, 128.7, 128.6, 128.4,
128.1, 128.0, 120.9, 90.6, 80.0, 74.9, 72.7, 72.1, 67.8, 65.2, 64.5,
57.0, 52.8, 51.5, 51.4, 25.8, 25.6, 23.6, 18.0, 17.7; HRMS calcd
for C37H60N7O7Si2 771.4171, found 771.4117.
N-(5′-Aden osyl)-N-[(3S)-3-am in o-3-car boxypr opyl]-N,N-
d im eth yla m m on iu m iod id e (MeAza Ad oMet) 3, fr om 24.
A solution of 24 (318 mg, 0.35 mmol) in dry DCM (13 mL) was
added dropwise to a stirred mixture of boron trifluoride
etherate (0.45 mL, 0. 51 g, 3.6 mmol, 10 equiv) and ethanethiol
(0.80 mL, 0.68 g, 10.9 mmol, 30 equiv) under argon at 0 °C.
The ice bath was removed 10 min after addition was complete
and then the mixture stirred at room temperature overnight.
Solvent was evaporated under vacuum (using a KMnO4 trap)
and the residue coevaporated with methanol (3 × 10 mL). The
crude residue was dissolved in a small volume of water and
loaded onto a column of Dowex 50WX4-400 (20 g, NH4+ form).
The column was washed with water (100 mL) and then eluted
with 1.0M NH4OH (1.5 l), and the UV-active fractions were
collected and evaporated to give 46 mg of product. Further
product remained on the column so it was eluted with 1.0 M
NaOH (100 mL), which was immediately neutralized with 1.0
M HCl and evaporated to dryness. The first 46 mg portion of
product was loaded onto a second Dowex 50WX4-400 (100 g,
NH4+ form) column and eluted with 0-1.5 M NH4OH (1.0 L).
The major peak (0.7-1.0 M) was evaporated to give a pale
yellow solid, redissolved in water, and decolorized by passing
through Dowex 1X4-400 (10 g, Cl- form). The UV-active
fractions were evaporated, and the solid was redissolved in a
small volume of water and lyophilized to give 3 (34 mg). The
portion of product from the NaOH wash was dissolved in water
N1,5-Bis(ben zyloxyca r bon yl)-2-p yr r olid on e 22. N-Ben-
zyloxycarbonyl-L-glutamic acid R-benzyl ester (2.0 g, 5.39
mmol) was dissolved in dry DCM (50 mL) under nitrogen.
Distilled oxalyl chloride (0.75 g, 0.52 mL, 13.3 mmol, 5.91
mmol, 1.1 equiv) was added followed by DMF (2 drops) and
the mixture stirred at room temperature for 2 h. The solvent
was evaporated, and then the yellowish oil was dried under
high vacuum give a pale yellow solid used in the next step
without further purification. DMAP (165 mg, 1.35 mmol) and
2,2,2-trifluoro-1-iodoethane (5.7 g, 2.66 mL, 27 mmol, 5.0
equiv) were added to a suspension of dry N-hydroxypyridine-
2-thione sodium salt (0.89 g, 5.9 mmol, 1.1 equiv) in dry DCM
(30 mL) under nitrogen. The mixture was brought to reflux
(heated using an aluminum ring), and then a solution of the
above crude acid chloride in dry DCM (12 mL) was added
dropwise over 10 min under irradiation with two tungsten
filament lamps (100 and 150 W). Reflux was continued under
irradiation for a further 90 min after addition was complete.
The reaction mixture was then washed with water (50 mL), 1
M HCl (30 mL), and water (30 mL) and then evaporated under
vacuum. The product was purified by flash column chroma-
tography (EtOAc/light petroleum, 1:1) to give an off-white,
crystalline solid (1.44 g, 76%) identified as the γ-lactam
product 22 of intramolecular cyclization of the acid chloride:
1H NMR (250 MHz, CDCl3) δ 7.37-7.24 (br s, 5H), 5.21 (s,
2H), 5.12 (s, 2H), 4.71 (dd, 1H, J ) 2.7, 9.2), 2.71-2.25 (m,
3H), 2.11-1.99 (m, 1H); m/z (EI) 353 (M+); Rf 0.18 (EtOAc/
light petroleum, 1:2); [R]D -0.11° (c 6.6 × 10-3 in DCM) (lit.36
[R]D -12.6° (c 0.9 in CHCl3)).
N-[2′,3′-Bis(ter t-bu tyld im eth ylsilyl)a d en osyl]-N-[3-(S)-
3-N-ben zyloxycar bon ylam in o-3-(ben zyloxycar bon yl)pr o-
p yl]-N,N-d im eth yla m m on iu m Iod id e 23. Iodide 15 (0.86
g, 1.90 mmol, 1.5 equiv) in dry acetonitrile (10 mL) was added
to a suspension of the tertiary amine 13 (0.66 g, 1.26 mmol)
in the same solvent (5 mL) under nitrogen. This mixture was
heated at 45-50 °C for 4 d and then allowed to cool to room
temperature. Solvent was evaporated under vacuum and the
residue purified by flash column chromatography (MeOH/
DCM, 1:9) to give the title compound as an off-white foam (0.72
g, 58%): 1H NMR (250 MHz, CDCl3) δ 8.17 (s, 1H), 7.83 (s,
1H), 7.19-7.02 (m, 10H), 6.22-6.10 (1H), 6.03-5.75 (2H), 5.67
(d, 1H, J ) 6.7), 5.06-4.72 (6H), 4.52 (t, 1H, J ) 13.1), 4.26-
4.09 (4H), 3.75-3.41 (m, 2H), 3.07 (s, 3H), 2.99 (s, 3H), 2.33-
2.08 (2H), 2.06-1.80 (2H), 0.77 (s, 9H), 0.55 (s, 9H), 0.04 (s,
3H), 0.00 (s, 3H), -0.24 (s, 3H), -0.63 (s, 3H3); 13C NMR (100
MHz, CDCl3) δ 170.2, 156.2, 155.7, 153.0, 149.0, 141.2, 136.1,
135.0, 128.6, 128.53, 128.47, 128.12, 128.07, 120.7, 90.0, 80.2,
74.7, 72.8, 67.9, 67.0, 65.0, 63.1, 52.1, 51.9, 51.5, 25.8, 25.6,
25.2, 17.9, 17.7, -1.3, -2.3, -2.8; HRMS calcd for C43H66N7O7-
Si2 848.4562, found 848.4616.
+
and loaded onto the Dowex 50WX4-400 column (100 g, NH4
form), desalted by washing with 1.0 M NH4Cl (300 mL) and
then water (200 mL), and then eluted with the same gradient
of NH4OH used above and further purified on Dowex 1X4-400
in the same manner to give 3 (28 mg): a combined yield of 3
of (62 mg, 44%); 1H NMR (400 MHz, D2O) δ 8.10 (s, 2H), 5.92
(d, 1H, J ) 3), 4.63-4.55 (br m, 1H), 4.40 (dd, 1H, J ) 8, 6),
4.27 (dd, 1H, J ) 6), 3.76 (dd, 1H, J ) 10, 14), 3.62 (d, 1H, J
) 14), 3.46-3.31 (m, 2H), 3.25 (t, 1H, J ) 6), 3.04 (s, 6H),
2.09-1.97 (m, 1H), 1.97-1.85 (m, 1H); 13C NMR (100 MHz,
D2O) δ 177.1, 155.5, 152.8, 148.5, 140.2, 118.9, 89.4, 76.4, 72.1,
71.6, 65.6, 62.4, 52.6, 51.7, 51.4, 26.0; HRMS calcd for
C
16H26N7O5 396.1995, found 396.1998. This procedure could
be improved by treating the crude reaction product with base
(aqueous Bu4NOH followed by neutralization with AcOH)
before loading onto the ion-exchange column. This would
cleave the boron complex of 3, which seemed to have been
formed in the reaction.
Ack n ow led gm en t. We thank the Biotechnology and
Biological Sciences Research Council for financial sup-
port (Grant GR/F73328) and the Engineering and
Physical Sciences Research Council for a Research
Studentship (to M.J .T.). We also thank Dr. David
J akeman for assistance with NMR titration experi-
ments.
N-(2′,3′-Di-O-ter t-bu tyld im eth ylsilyla d en osin -5′-yl)-N-
[2-(3′-b en zyloxyca r b on yloxa zolid in -5-on e)-4-yl]-et h yl-
N,N-d im eth yla m m on iu m Iod id e 24. Amine 13 (252 mg,
0.48 mmol) was suspended in dry acetonitrile (7 mL) under
argon, and then dry DMF (3 mL) was added. Iodide 10 (281
mg, 0.75 mmol, 1.5 equiv) was then introduced and the mixture
heated at 57 °C for 2 d. Solvent was evaporated and the residue
purified by flash column chromatography (MeOH/DCM, 8:92)
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails and characterization data for compounds 3 (alternative
methods used), 7, 9, 12-14, 18, 19 (using methylamine), and
2′,3′-O,O-benzylidene adenosine. NMR spectra and/or ad-
ditional characterization data for compounds 2, 10, 11, 15,
17a ,b , and 21-24. X-ray structural data for compound 10,
including tables of atomic coordinates, bond lengths and bond
angles. This material is available free of charge via the
Internet at http://pubs.acs.org.
(36) Baldwin, J . E.; Miranda, T.; Moloney, M. Tetrahedron 1989,
45, 7459.
J O9907742