Nucleoside analogs. 14. The synthesis and antitumor activity in mice of molecular combinations of 5-fluorouracil and N-(2-chloroethyl)-N-nitrosourea moieties separated by a three-carbon chain

Article abstract of DOI:10.1021/jm9507237

5-Fluorouracil (5-FU) seco-nucleosides having as the "sugar" moiety a two-carbon (C₂) side chain carrying a N-(2-chloroethyl)-N-nitrosourea group were designed as molecular combinations of antimetabolite and alkylating agent, but hydrolytic release of free 5-FU was not fast enough for significant contribution to the high activity they showed against colon and breast tumors in mice. In the present study of the synthesis of the more reactive C₃ seco-nucleosides, it emerged that, of various groups attached to the aldehydic center in the precursor phthalimides, only the alkoxy/uracil-1-yl type was conveniently obtained by the standard method. The methylthio/uracil-1-yl analog required relatively large amounts of reagent methanethiol, and exploration of alternatives involving α-chlorination of alkyl methyl sulfide or Pummerer rearrangement of its S-oxide, or successive hydrolysis and methylation of isothiouronium bromide, gave disappointing yields. For successful preparation of the alkoxy/ uracil-3-yl compounds, the route used for C₂ homologs required considerable experimental modification. In addition to these O,N- and S,N-acetals, some N,N-acetals bearing two 5-FU residues were prepared. The new drugs have been tested against a panel of experimental tumors in mice. Although it is evident from a parallel study that even these C₈ seco-nucleosides release free 5-FU too slowly in vivo, several of them have shown impressive anticancer activity. Reviewing their performance in comparison with earlier molecular combinations, a short list of seven [B.4152 (6), B.4015 (5), B.4030 (10), B.3999 (4), B.3995 (2), B.4083 (3), and B.3996 (the N³-substituted analog of 1)] should be investigated further. This is particularly appropriate in light of the present understanding of the mode of action of chloroethylating agents. Following a prolonged period of clinical impatience with nitrosoureas because of limited selectivity of action, a new era is confidently anticipated as these powerful drugs are increasingly studied in combination with O⁶-benzylguanine and other more efficient inhibitors of repair enzymes like O⁶-alkylguanine-DNA-alkyltransferase now being developed.

Full text of DOI:10.1021/jm9507237

J . Med. Chem. 1996, 39, 1403-1412
1403
Nu cleosid e An a logs. 14. Th e Syn th esis a n d An titu m or Activity in Mice of
Molecu la r Com bin a tion s of 5-F lu or ou r a cil a n d N-(2-Ch lor oeth yl)-N-n itr osou r ea
Moieties Sep a r a ted by a Th r ee-Ca r bon Ch a in
R. Stanley McElhinney,*,† J oan E. McCormick,† Mike C. Bibby,‡ J ohn A. Double,‡ Marco Radacic,§ and
Patrick Dumont
Screening and Pharmacology Group, EORTC, and University Chemical Laboratory, Trinity College, Dublin 2, Ireland,
Clinical Oncology Unit, University of Bradford, Bradford, West Yorkshire, BD7 1DP, U.K., Ruder Bosˇkovic Institute,
Bijenicka 54, 41001 Zagreb, Croatia, and Institut J ules Bordet, Rue He´ger-Bordet, 1000 Brussels, Belgium
Received September 29, 1995X
5-Fluorouracil (5-FU) seco-nucleosides having as the “sugar” moiety a two-carbon (C2) side
chain carrying a N-(2-chloroethyl)-N-nitrosourea group were designed as molecular combina-
tions of antimetabolite and alkylating agent, but hydrolytic release of free 5-FU was not fast
enough for significant contribution to the high activity they showed against colon and breast
tumors in mice. In the present study of the synthesis of the more reactive C3 seco-nucleosides,
it emerged that, of various groups attached to the aldehydic center in the precursor
phthalimides, only the alkoxy/uracil-1-yl type was conveniently obtained by the standard
method. The methylthio/uracil-1-yl analog required relatively large amounts of reagent
methanethiol, and exploration of alternatives involving R-chlorination of alkyl methyl sulfide
or Pummerer rearrangement of its S-oxide, or successive hydrolysis and methylation of
isothiouronium bromide, gave disappointing yields. For successful preparation of the alkoxy/
uracil-3-yl compounds, the route used for C2 homologs required considerable experimental
modification. In addition to these O,N- and S,N-acetals, some N,N-acetals bearing two 5-FU
residues were prepared. The new drugs have been tested against a panel of experimental
tumors in mice. Although it is evident from a parallel study that even these C3 seco-nucleosides
release free 5-FU too slowly in vivo, several of them have shown impressive anticancer activity.
Reviewing their performance in comparison with earlier molecular combinations, a short list
of seven [B.4152 (6), B.4015 (5), B.4030 (10), B.3999 (4), B.3995 (2), B.4083 (3), and B.3996
(the N3-substituted analog of 1)] should be investigated further. This is particularly appropriate
in light of the present understanding of the mode of action of chloroethylating agents. Following
a prolonged period of clinical impatience with nitrosoureas because of limited selectivity of
action, a new era is confidently anticipated as these powerful drugs are increasingly studied
in combination with O6-benzylguanine and other more efficient inhibitors of repair enzymes
like O6-alkylguanine-DNA-alkyltransferase now being developed.
Alkylation was one of the first principles successfully
exploited in cancer chemotherapy, and the value of
nitrogen mustards like Melphalan and Cyclophospha-
mide and sulfonates (Busulfan) has endured1 to the
present day. N-(2-Chloroethyl)-N-nitrosoureas (CNUs)
such as Carmustine (BCNU) and Chlorozotocin are also
of this 30-40-year-old vintage.2 Although very many
congeners were made and tested, the clinical usefulness
of these agents reached a plateau because of limited
selectivity between cancer cells and bone marrow.
During the past decade improved experimental tech-
niques have led to a greater understanding of the
mechanism of action of alkylating drugs on biological
macromolecules, particularly DNA. The site of alkyla-
tion by CNUs and the function and significance of O6-
alkylguanine-DNA-alkyltransferase (ATase) have been
intensively studied.3 It can be claimed with some
confidence that a new era in the history of alkylating
agents is opening up, and not only will the performance
of early CNUs be greatly enhanced but the availability
of several specially-designed nitrosoureas4 may prove
something of an Aladdin’s cave, with the prospect of
tailor-made treatment for individual patients.
One of these classes has been the molecular combina-
tions incorporating the antimetabolite 5-fluorouracil
(5-FU) and the CNU alkylating function.4-6 Our de-
velopment of this program was facilitated by a novel
synthetic route devised for seco-nucleosides7 (nucleoside
analogs in which the “sugar” is linear instead of the
normal cyclic), enabling a pyrimidine to be readily
linked to a nitrosourea moiety. It was hoped that the
administered molecular combination would gradually
release 5-FU to make its own antitumor contribution
superimposed on that of the nitrosourea. This would
differ subtly from a physical combination (mixture) of
two drugs in that the single molecule of the mole-
cular combination is transported in vivo initially as
such, presenting a challenge for pharmacokinetic
study. By chemically adjusting the rate of release of
5-FU, in principle the overall antitumor effect could be
varied.
* Corresponding author.
† Trinity College, Dublin.
‡ University of Bradford.
§ Ruder Bosˇkovic Institute, Zagreb.
Institut J ules Bordet, Brussels. Present address: CNRS, Labora-
toire de Pharmacologie et de Toxicologie Fundamentales, 205, Route
de Narbonne, 31077 Toulouse Cedex, France.
X Abstract published in Advance ACS Abstracts, February 15, 1996.
The seco-nucleoside molecular combinations tested up
to the present had n ) 1 in the general structure
0022-2623/96/1839-1403$12.00/0 © 1996 American Chemical Society

1404 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7
McElhinney et al.
Sch em e 1
where R ) small alkyl group, X ) O (as in natural
sugars and nucleosides) or S, and the pyrimidine ring
was linked by N1 (as in natural nucleosides) or N3
(designated 5-FU and 5-FU3 respectively in the formulas
in the schemes). When n ) 1, 5-FU and CNU are
separated by a two-carbon chain, with the CNU sub-
stituent attached at the position corresponding to 2′ of
a furanose ring. Many of these C2 drugs [examples are
B.3839 (1), B.3995 (2), and B.4083 (3)]8 strongly inhib-
ited the growth of certain colon tumors in mice (the
MAC series), well-established as models for clinical
disease.
Acid-catalyzed hydrolysis releasing 5-FU from these
seco-nucleosides proceeds in the order of reactivity MeO
> MeS and 5-FU3 > 5-FU, thus B.4083 (3) is hydrolyzed
much faster than B.3839 (1)6,9 (see Scheme 1). However
the hydrolysis rates are greatly decreased at physiologi-
cal pH, and it became likely that even B.4083 (3) may
not release 5-FU at a rate sufficient to contribute
significantly to the antitumor activity. This would
indicate that all these C2 drugs act principally as
alkylating agents, in many cases having particularly
effective carriers. Such a conclusion was already im-
plied by the activity of the corresponding uracil (U) seco-
nucleosides which was often similar to that observed
in the 5-FU drugs at a high level.6,10
Since 5-FU seco-nucleosides with nitrogen or oxygen
functional groups separated from the pyrimidine by a
three-carbon chain, i.e., at the position corresponding
to 3′ of a furanose ring, are more readily hydrolyzed
than their homologs with a two-carbon chain,11 we
undertook a final attempt to prepare a molecular
combination whose antitumor action would conclusively
involve a contribution from released antimetabolite.
Recognizing the other factors governing reactivity, the
C3 analog 6 of B.4083 (3) would be an ideal target. To
complete the picture in this series it was also desirable
to assess the antitumor activity of the analogs 4 and 5
of B.3839 (1) and B.3995 (2), together with some water-
soluble drugs 8 and 9 and more hydrophobic compounds
such as the tert-butyl ether (10) and its C4 isomer (11).
We also chose analogs of 6 with a uracil (7) or a second
5-FU residue (12).
Ch em istr y
Synthesis of these nitrosoureas was achieved by the
general route for C2 homologs: low-temperature hy-
drazinolysis of the corresponding phthalimides (thereby
avoiding ring closure of the liberated amine to a bicyclic
isomer), followed by reaction with 2,4,5-trichlorophenyl
N-(2-chloroethyl)-N-nitrosocarbamate (leading exclu-
sively to the desired N-nitroso isomer of the generated
urea).14,15
However, preparation of the precursor phthalimides
in the C3 series required a modified approach. Thus,
the precursors 22a and 19 of B.3839 (1) and B.3995 (2),
respectively, were readily obtained from the sulfoxide
(20a ) via Pummerer rearrangement which, influenced
by the â-phthalimido group, gave the required acetate
21a and the isomer 17a in the ratio 65:355,16 (see
Scheme 2). Rearrangement of the homolog 20b with
acetic anhydride yielded (60:35:5, respectively) mainly
the “normal” acetoxymethyl isomer 17b as anticipated,
together with unsaturated product 18 and only traces
of the acetate 21b necessary for the synthesis of 22b
and ultimately B.3999 (4). Trifluoroacetic anhydride
and 20b afforded only the 17b type of isomer and 18
(80:20).
Comparative data for earlier drugs with a three-
carbon chain resembling that in 6 are included in this
study. B.3971 (13)9 is isomeric with 9, and the phthal-
imide B.3957 (14)12 has a similar molecular framework.
It is noteworthy that the glycoside nitrosourea Acomus-
tine (15),13 with the 2-deoxy-3-CNU fragment charac-
teristic of all these C3 seco-nucleosides, has properties
superior to earlier hexopyranose analogs bearing a
2-CNU group such as Chlorozotocin.
Alternative approaches to the seco-nucleoside 22b
involved R-chlorination studies17-19,21 (see the Experi-
mental Section), which proved very unsatisfactory when
applied to 16b. Further, we investigated the adaptation
of an established route to thioglycosides containing the
fragment OCHSMe via sugar bromides (OCHBr) and
their reaction products with thiourea, isothiouronium
bromides. These salts are hydrolyzed to thiols (OCHSH)
which can be alkylated without isolation.20 We found
In the present paper we describe the synthesis of the
proposed C3 drugs and compare their activity against a
panel of experimental tumors in mice with that of the
C2 drugs and of standard nitrosoureas like Semustine
(MeCCNU) and Tauromustine (TCNU) in the light of
recent significant developments in the use of chloroeth-
ylating agents.3,4

Nucleoside Analogs
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7 1405
Sch em e 2
Sch em e 4
Sch em e 3
The alkoxyphthalimides 24a and 24b described ear-
lier have now been obtained on a preparative scale from
the chloride 28a , and 30 from the bromide 31; these
ethers are the precursors respectively of B.4015 (5),14
B.4024 (9), and B.4030 (10). The C4 homolog (27b) was
made, from N-(4-bromobutyl)phthalimide, in the same
manner11 as 27a and converted via the chloride into the
isopropoxyphthalimide (24d ) precursor of B.4029 (11).
The N3-substituted compound 42b needed for B.4152
(6) required a different approach from the isomer 24a ,
since reaction of silylated 5-FU with precursors such
as 36 affords (at least in practice) only N1-isomer 27.11
The C2 homolog of 42b was eventually obtained very
satisfactorily15 by a route based on reaction of the useful
reagent N1-(octylthio)carbonyl-5-FU22b with the chloro
ether 34 derived from the O,S-acetal 33a (see Scheme
4). Application to the C3 series would indicate 37 f 38
f 39b f 42b. Attempted preparation of 37 revealed
differences in the reactivity of C2 and C3 aldehyde
derivatives in addition to those already observed.11,25
Compound 33a is obtained (86% yield) from the
Pummerer acetate 21a in refluxing methanol with acid
catalysis.15 In the C3 series the S-(p-tolyl)acetal 37 is
necessary because the S-methyl compound 21b is so
inaccessible. The very convenient Pummerer rear-
rangement of the appropriate sulfoxide in trifluoroacetic
anhydride leads to 36, but reaction of this with refluxing
methanol afforded 37 in low and unreproducible yields.
However, room temperature gave a very satisfactory
result in this evidently more reactive system.26
The next steps (37 f 38 f 39b) also proved more
difficult in the C3 series. While chlorinolysis of O,S-
acetals in general proceeds smoothly, for 33a (C2) the
byproduct methanesulfenyl chloride is relatively volatile
and easily removed. The chloro ether 34 is stable and
reaction with the protected 5-FU (18 h at room temper-
ature) straightforward and high-yielding.15 For 37 the
(characteristically bright yellow) arenesulfenyl chloride
had to be extracted into petroleum ether, and attempted
reaction of the chloro ether 38 with N1-[(octylthio)-
carbonyl]-5-FU at room temperature usually afforded
no 39b or on one occasion ca. 20% yield, and the reagent
could be recovered (60-70%) as such or as free 5-FU.
b
that this sequence could be applied to seco-nucleoside
bromides (NCHBr) such as 31 (see below), and the
sulfide (NCHSMe, 22b) was obtained from the salt 32
(formed in 89% yield) via the thiol 29 (see Scheme 3).
However the sensitivity of the phthalimide group in
alkali is a complication, and some nuclear (N3) methy-
lation leading to 26 also occurs. After much experi-
mentation the best yield of sulfide 22b from the salt 32
was 17%.
Accordingly we reverted, reluctantly, to introducing
the methylthio group directly at the final stage in the
synthesis of 22b. We had earlier obtained C3 ethers (24,
n ) 2) from the chloride 28a , derived like the bromide
31 from the aryl sulfide 27a .11 This was available
without complication by a route corresponding to 16 f
20 f 21 f 22, since rearrangement of aryl sulfoxides
can produce only a single isomer.21 However, thiols do
not react satisfactorily with the chloride 28a in the form
of their sodio derivatives, a great disadvantage for very
volatile compounds. We had obtained15 the sulfide 25b
in 58% yield using neat mercaptoacetic acid, and we now
found that simpler thiols (e.g., propane-2-thiol giving
25a ) also reacted well, in nitromethane solvent, in this
acid-catalyzed transformation, but temperatures lower
than 90-100 °C were unsatisfactory. The sulfide 22b
was finally obtained successfully in 82% yield, although
considerable amounts of methanethiol (bp 6 °C) are
needed to maintain the concentration necessary for
efficient reaction.

1406 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7
McElhinney et al.
In case 38 was unstable27 even at room temperature
and/or very sensitive to generated HCl, we carried out
the condensation at 0 °C and at -15 °C, using 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU) to remove acid,
and obtained 39b in yields respectively of 28% and 55%
(based on 37).
Ta ble 1. Effects against MAC 13 Tumors
dose, no. deaths/ T/C,
a
compound
vehicle
mg/kg
treated
%
B.3839 (1) arachis oil
B.3999 (4) arachis oil
B.3995 (2) arachis oil
B.4015 (5) arachis oil
B.4083 (3) arachis oil
B.4152 (6) 10% DMSO/arachis oil 100
B.4184 (7) 10% DMSO/arachis oil 100
B.4146 (12) 10% DMSO/arachis oil 100
B.4084 (8) arachis oil
B.4024 (9) 10% ethanol/arachis oil 100
B.3971 (13) arachis oil
B.3957 (14) arachis oil
B.4030 (10) arachis oil
B.4029 (11) arachis oil
31.25
135
50
150
50
0/10
0/5
0/5
0/5
0/5
0/5
0/5
0/5
0/5
0/5
0/5
0/10
0/5
0/5
0/5
0/5
122
0
7
0.8
11
1
43
87
31
23
27
103
1
33
8
1
It was convenient or even desirable to avoid the
extraction of sulfenyl chloride because of the consider-
able solubility of 38, and a technique used in the
synthesis of unsaturated chloromethyl ethers from the
corresponding formaldehyde O,S-acetals suggested it-
self. Cyclohexene as scavenger rapidly formed a trans-
chlorocyclohexyl sulfide adduct which did not interfere
with subsequent nucleophilic reaction with the chlo-
romethyl ether.28 We found that by inclusion of an
equivalent of cyclohexene during the chlorinolysis, the
product 38 subsequently afforded 39b in 76% (overall)
yield in 1 h at -15 °C. This demonstrates the reactivity
of the C3 derivative 38 relative to 34,15 although in the
presence of cyclohexene reaction temperature was less
significant (63% yield of 39b at room temperature). The
deprotected phthalimide 42b was then obtained very
smoothly using isopropylamine in ether.
Reaction of the chloro ether 38 with N1-[(octylthio)-
carbonyl]uracil readily provided 42a and in turn B.4184
(7), underlining the great simplicity of seco-nucleoside
synthesis with these reagents.22b,29 An alternative route
to 42a seemed promising since, unlike 5-FU, uracil
tends to give mainly N3-seco-nucleosides, e.g., the
transformation corresponding to 36 f 40, using the
silylated reagent and tin(IV) chloride, is accomplished
in the C2 series in 60-70% yield.30 However the C3
ester 36 gave an inseparable mixture of approximately
equal amounts of 40 and its N1-isomer.
The versatility of the N1-(octylthio)carbonyl reagents
also suggested the synthesis of a “double-headed” seco-
nucleoside bearing two pyrimidine residues. Reaction
with the chloride 28a was slower than with the ether
38 but satisfactorily yielded the phthalimide 35, the
precursor of B.4146 (12). We have been unable to find
any references in the literature to double-headed nu-
cleoside analogs with geminal pyrimidinyl substituents
(N,N-acetals), although a dialdehydic xylopyranose
derivative31 has purine and pyrimidine residues at-
tached to the respective “anomeric” carbons (by analogy,
an N,O,N-hemialdal). N,N-Acetals have a long pedi-
gree,32 and compounds bearing geminal hydantoin rings
have been described.33
200
100
200
200
200
MeCCNU
TCNU
10% ethanol/arachis oil 20
saline 30
a
T represents mean weight of treated tumors and C represents
mean weight of control tumors.
Ta ble 2. Effects against MAC 15A Tumors
compound
vehicle
arachis oil
arachis oil
arachis oil
arachis oil
arachis oil
10% DMSO/arachis oil
10% DMSO/arachis oil
dose, mg/kg T/C, %a
B.3839 (1)
B.3999 (4)
B.3995 (2)
B.4015 (5)
B.4083 (3)
B.4152 (6)
B.4184 (7)
B.4146 (12) 10% DMSO/arachis oil
B.4084 (8)
B.4024 (9)
B.3971 (13) arachis oil
B.3957 (14) arachis oil
B.4030 (10) arachis oil
B.4029 (11) arachis oil
MeCCNU
TCNU
62.5
90
156
200
200
233
>850
360
200
182
275
233
136
148
280
250
154
210
50
100
100
100
100
100
200
100
100
200
300
300
20
10% DMSO/arachis oil
10% ethanol/arachis oil
saline
saline
30
a
T represents median survival times of treated animals and C
represents median survival times of control animals.
Ta ble 3. Effects against MCa Mammary Carcinoma
compound
vehicle
dose, mg/kg LT Sa T/C, %b
B.3839 (1)
B.3999 (4)
B.3995 (2)
B.4015 (5)
B.4083 (3)
B.4024c (9)
10% DMSO/water
10% DMSO/water
10% DMSO/water
10% DMSO/water
10% DMSO/water
10% DMSO/water
60
80
100
100
50
150
100
100
1/6
0/8
3/6
6/7
5/6
0/8
5/6
0/6
189
195
311
168
121
B.4030c (10) 10% DMSO/water
B.3957 (14) 10% DMSO/water
a
b
Long term survival. T represents median survival times of
treated animals and C represents median survival times of control
animals. c Treated on day 1.
Resu lts a n d Discu ssion
The effects of drugs against the various tumors
following ip administration at maximum tolerated doses
are presented in Tables 1-4. Earlier results6 for the
C2 drugs B.3839 (1), B.3995 (2), and B.4083 (3) are
included to make comparisons easier. We have at-
tempted to correlate the observed activity of the new
C3 drugs against the panel of tumors with chemical
reactivity and physicochemical properties. Details of
pharmacokinetics of B.4152 (6)34 and other very active
compounds are reported elsewhere, following those35 of
B.3839 (1), the original 5-FU/CNU molecular combina-
tion.
Ta ble 4. Effects against Colon 38 Adenocarcinoma
compound
vehicle
dose, mg/kg
T/C %a
B.3839 (1)
B.3999 (4)
B.3995 (2)
B.4015 (5)
B.4083 (3)
B.4084 (8)
B.4030 (10)
B.3971 (13)
saline
saline
saline
saline
saline
saline/Tween 80 (9/1)
saline/Tween 80 (9/1)
saline/Tween 80 (9/1)
100
100
100
100
50
100
100
50
0
63
38
47
17
0
42
68
a
T represents median tumor volume of treated animals and C
represents median tumor volume of controls.
It transpired that even B.4152 was not cleaved in vivo
to pyrimidine to a significant extent.34 Accordingly,
reaction of the nitrosourea moiety is the first and
principal event with all these seco-nucleoside drugs, and
any occurrence of high antitumor activity in compounds
which readily release 5-FU at lower pH’s must be
regarded as coincidental. Most of the trends discernible
for each tumor in the panel are probably due to

Nucleoside Analogs
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7 1407
Ta ble 5. N-(2-Chloroethyl)-N-nitrosoureas (CNUs)
b
starting phthalimide
CNU
yield, %
49
mp, °Ca
UV λmax (nm)
IR υmax (cm-1
)
formulac
C11H15ClFN5O5
24a
5, B.4015d
129.5-130
264
(262)
264
1536
1488
3520
1530
1488
1530
1489
1530
1495
1534
1492
24b
9,e B.4024
31
63-64.5f
C12H17ClFN5O6‚0.7EtOAc
(263)
30
10, B.4030
11, B.4029
4, B.3999
35
31
34
52
18
45
118-119
266
(265)
266
(266)
269
(270)
259
(294)
266
(305)
267
C
C
C
14H21ClFN5O5
14H21ClFN5O5
11H15ClFN5O4Sg
24d
22b
42a
42b
35
121-121.5
120-121
7, B.4184h
6,j B.4152h
12,k B.4146
121.5-122
122.5-123
190-210
C11H16ClN5O5
1528
1490
1535
1495
C11H15ClFN5O5
C14H14ClF2N7O6‚0.75EtOH
(270, 305)
a
b
With effervescence. In parentheses: those in the presence of NaOH. c Anal. C, H, N in all cases except for 9 (N: calcd, 15.8; found,
15.3) and 12 (C: calcd, 38.45; found, 37.7). Reference 14. e 1H NMR δ 11.7 (d, J ) 5, ring 3-H), 8.75 (t, J ) 6.5, CH2NH), 7.94 (d, J )
d
7, ring 6-H), 5.71 (t, J ) 6.5, OCHN), and 1.99 (s, CH3CO2Et; solvent content, 0.7 mol/mol CNU). f Effervescence at 70 °C. S: calcd, 8.7;
g
found, 9.0. Reference 34. j 1H NMR δ 10.94 (bs, ring 1-H) 8.78 (t, J ) 5.6, CH2NH), 7.82 (d, J ) 5.5, ring 6-H), 5.87 (t, J ) 6.3, OCHN),
h
and 3.22 (s, OMe); k 1H NMR δ 11.83 (bs, 3-H in 5-FU), 11.12 (bs, 1-H in 5-FU3), 8.80 (t, J ) 5.3, CH2NH), 8.27 (d, J ) 8.0, 6-H in 5-FU),
7.83 (d, J ) 5.5, 6-H in 5-FU3), 6.87 (t, J ) 6.5, NCHN), and 1.06 (t, CH3CH2OH; solvent content, 0.5-0.75 mol/mol CNU).
physicochemical factors. There are consistent patterns,
and some anomalies, in each table, but from a survey
of all the results a choice of several outstanding drugs
can be made for more detailed study.
(2) and B.4083 (3) are considerably more active than
B.3839 (1), as was observed for MAC 13. B.4152 was
not available for comparison against this tumor, and
while the methoxy C3 drug B.4015 (5) maintained good
activity with the characteristic high proportion of long-
term survivors, the methylthio-C3 B.3999 (4) was much
less active, resembling B.3839. Water-solubility [B.4024
(9)] was a disadvantage, but again the tert-butoxy
compound B.4030 (10) showed excellent activity. The
drug B.3957 (14) was once more inactive.
Against Colon 38 adenocarcinoma the methylthio C2
compound B.3839 (1) had high activity, more so than
the methoxy drugs B.3995 (2) and B.4083 (3). The C3
homologs B.3999 (4) and B.4015 (5) were less active,
especially the former. All these drugs were adminis-
tered in saline, which may influence bioavailability.
Preliminary studies with B.3839 suspended in Klucel
(hydroxypropylcellulose) resulted in increased toxicity.
Addition of some Tween 80 to the vehicle gave mixed
results: B.3971 (13) was inactive, B.4030 (10) barely
so, and B.4084 (8) very active, with no tumor measur-
able on day 20 as in the case of B.3839.
MAC 13 is a subcutaneous tumor to which drugs
administered intraperitoneally have to be transported
in order to exert any effect, enabling bioavailability to
be assessed. The methoxy C2 drugs B.3995 (2) and
B.4083 (3) are much more active against this tumor
than the methylthio analog B.3839 (1) of B.3995 (al-
though B.3839 is effective if administered orally).
However, all three C3 homologs B.3999 (4), B.4015 (5),
and B.4152 (6) are equally effective. Dose levels are
higher than for the C2 drugs, and the effect on the tumor
is marginally greater, almost eliminating it after 21
days. More water-soluble drugs like B.4084 (8), B.4024
(9), and B.3971 (13) are less active, as is the uracil-
containing B.4184 (7) and the C4 compound B.4029 (11).
On the other hand, the tert-butoxy C3 drug B.4030 (10)
isomeric with B.4029 is highly active. The relatively
insoluble (water and lipid) compounds B.3957 (14) and
B.4146 (12), of higher than average molecular weight,
are inactive against this tumor.
Thus it appears that the principal target drug B.4152
(6) has impressive activity against the MAC tumors,
significantly higher than its uracil analog B.4184 (7) and
than standard nitrosoureas like CCNU12 and PCNU,
MeCCNU, TCNU,6,16 which show activity in these
tumors only close to the maximum tolerated dose and
accompanied by normal tissue toxicities. Additional
evidence34 makes it a prime candidate for further
evaluation, together with three other C3 drugs which
have emerged from the present study: B.4015 (5),
B.4030 (10), and B.3999 (4), all outstandingly active
against the subcutaneous MAC 13 (colon) tumor when
administered ip in an oily vehicle. Treatment of the
MCa (mammary) carcinoma with B.4015 and B.4030 in
aqueous dimethyl sulfoxide leads to long-term survival
in 85% of the test animals, and the relative inactivity
of B.3999 in this system poses an interesting question.
All three drugs are relatively ineffective against the
Colon 38 tumor when administered in saline [but the
behavior of B.4084 (8) under these conditions would
indicate that this drug too should be short-listed].
In the ascitic tumor MAC 15A, tumor cells are located
within the peritoneal cavity into which the drugs are
injected, so the transport factor characteristic of MAC
13 is eliminated and drugs have merely to penetrate
the cell membrane. Of the C2 drugs, B.4083 (3) is
outstandingly more active than B.3839 (1) and B.3995
(2). Incidentally, its water-solubility (4 mg/mL) is
higher than the less active drugs (both 1 mg/mL). The
C3 homologs follow a similar though less marked trend,
but again B.4152 (6) is very active [and its uracil analog
B.4184 (7) less so]. Water-soluble drugs B.4084 (8) and
B.4024 (9) have good activity, but B.3971 (13) has not.
The lipid-soluble isomers B.4029 (11) and B.4030 (10)
also show useful activity, and as against MAC 13 the
insoluble drugs B.3957 (14) and B.4146 (12) are rela-
tively ineffective.
The MCa mammary carcinoma like MAC 13 requires
transport of the drugs as administered, and the DMSO-
containing aqueous vehicle used may exert an influence
on bioavailability. Again the methoxy C2 drugs B.3995

1408 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7
McElhinney et al.
from the sulfoxide using the crude oily Pummerer trifluoro-
acetate directly (38% yield based on crude sulfoxide). Hy-
drazinolysis in MeOCH2CH2OH (3 h, 100 °C) of phthalimide
27b yielded (16%) a sample of the parent amine hydrochloride,
mp 257-258 °C dec (EtOH); UV λmax (nm) 270, 258 (sh). Anal.
(C15H18FN3O2S‚HCl) C, H, N, S.
Finally 27b (6.80 g, 15 mmol) was chlorinolyzed using SO2-
Cl2 (1.34 mL, 16.5 mmol) in CH2Cl2 (360 mL). After stirring
for 15 min, the clear solution was evaporated, leaving the
chloride (28b) and p-toluenesulfenyl chloride. The mixture
was treated with hot 2-propanol (150 mL) and refluxed for 20
min, the solution evaporated, and the residue triturated with
Et2O (60 mL). The isopropyl ether 24d (4.50 g, 77%, suf-
ficiently pure for further reaction) had mp 175.5-177.5 °C (i-
PrOH); UV λmax (nm) 269 (271 in presence of NaOH). Anal.
(C19H20FN3O5) C, H, N. The methyl ether 24c prepared
similarly had mp 205-207 °C (MeOH); UV λmax (nm) 268 (267
in presence of NaOH). Anal. (C17H16FN3O5) H,N; C: calcd
56.5, found 55.7.
Finally, three of the earlier C2 drugs also have a range
of high activity6 which requires further investigation:
B.3995 (2), B.4083 (3), and B.3996, the 5-FU3 analog of
B.3839 (1).
It is important to study the effects of administration
of O6-benzylguanine, and other more powerful O6-
alkylguanine-alkyltransferase inhibitors now being de-
veloped, in conjunction with a spectrum of chloroethy-
lating agents with differing biological properties. Bio-
availability, selectivity for tumor over normal tissue,
administration vehicle and site, and scheduling of
dosage will all be relevant in potential clinical applica-
tion. While the usefulness of familiar if inefficient
nitrosoureas like BCNU will no doubt be improved in
what appears to be an exciting new phase in the study
of these alkylating agents, the availability of more
effective or selective analogs should prove a considerable
advantage.
N-[3-(5-Flu or ou r acil-1-yl)-3-(m eth ylth io)pr opyl]ph th al-
im id e (22b). a . Methanethiol (about 35 g was used) was
passed into a stirred slurry of chloride 28a (10.55 g, 30 mmol)
in MeNO2 (150 mL) for 18 min at room temperature and then
40 min at 100 °C. The resulting solution was evaporated,
finally with addition of benzene (100 mL). The crumbly solid
was stirred with boiling PhMe (170 mL) and the hot mixture
filtered from impurity (0.99 g). The filtrate deposited 22b (9.55
g, 82%), recrystallization from PhMe giving a solvate, mp 152-
155.5 °C [Anal. (C16H14FN3O4S‚0.25C7H8) C, H, N]. Com-
pound 22b had mp 157-158 °C (MeOH); UV λmax (nm) 274
Exp er im en ta l Section
UV spectra were measured in MeOH on a Unicam SP-800
spectrophotometer, IR spectra using KBr disks on a Unicam
SP1000 instrument, and NMR spectra (80 MHz) in (CD3)2SO
(J values are given in hertz) on a Bruker WP-80. Melting
points, uncorrected, were determined in capillaries. Petroleum
ether had bp 40-60 °C unless otherwise stated. Column
chromatography was performed on Merck silica gel 60 (35-
70 mesh ASTM, Art. 7733). Microanalyses of all compounds
were within (0.4% of theory unless otherwise indicated.
N-[3-Alk oxy-3-(5-flu or ou r a cil-1-yl)p r op yl]p h t h a lim -
id es. a . The sulfide 27a 11 (7.46 g, 17 mmol) and SO2Cl2 (1.51
mL, 18.7 mmol) were stirred for 2 h in CH2Cl2 (187 mL). The
chloride 28a (5.32 g, 89%) was filtered from the yellow solution
of p-toluenesulfenyl chloride and taken up in boiling MeOH
(250 mL). Concentration (to 175 mL) and cooling yielded the
methyl ether 24a (4.63 g, 88%), mp 204-209 °C (lit.11 mp
202.5-203.5 °C).
b. The chloride 28a (6.83 g, 19.4 mmol) was stirred in
ethylene glycol (38.8 mL) for 15 min in a bath at 80 °C.
Cooling, addition of water (230 mL), and filtration gave the
2-hydroxyethyl ether 24b (7.23 g, 98%), showing a single spot
in TLC [CHCl3-MeOH (19:1)] and with the same IR spectrum
as that of an aliquot (mp 168-170 °C) recrystallized from
EtOAc (lit.11 mp 169-170 °C).
c. The sulfide 27a (4.39 g, 10 mmol) suspended in CH2Cl2
(200 mL) was treated with bromine (0.55 mL, 10.75 mmol) in
CH2Cl2 (40 mL). After being stirred for 2 h the bromide 31
(3.23 g, 81%) was filtered off and added to boiling tert-butyl
alcohol (800 mL). The solution was refluxed for 10 min, cooled,
and evaporated in vacuo. Trituration of the residue with
saturated aqueous NaHCO3 (240 mL) and filtration yielded
the product (2.96 g). Recrystallization from MeOH (200 f 100
mL) gave first some unsaturated material (104 mg, discarded)
and then the tert-butyl ether 30 (2.17 g, 70%), mp 205-209
°C (lit.11 mp 208-209 °C).
N-[4-(5-F lu or ou r a cil-1-yl)-4-isop r op oxyb u t yl]p h t h a l-
im id e (24d ). The intermediate N-[4-(5-flu or ou r a cil-1-yl)-
4-(p-tolylth io)bu tyl]p h th a lim id e (27b) was prepared in the
same way as the lower homologs.11 Thus, N-(4-bromobutyl)-
phthalimide afforded in turn N-[4-(p-tolylth io)bu tyl]p h th a l-
im id e (using p-toluenethiol and NaOMe), mp 54.5-56 °C
(MeOH) [Anal. (C19H19NO2S) C, H, N, S]; N-[4-(p-tolylsu lfi-
n yl)bu t yl]p h t h a lim id e [using 30% (w/v) aqueous H2O2 in
EtOH-AcOH], mp 85-87 °C (benzene-petroleum ether)
[Anal. (C19H19NO3S) C, H, N, S]; N-[4-a cet oxy-4-(p -t olyl-
th io)bu tyl]p h th a lim id e (by Pummerer rearrangement; 64%
yield based on starting bromide using crude sulfide and
sulfoxide both obtained in quantitative yield), mp 88-89 °C
(petroleum ether, bp 80-100 °C) [Anal. (C21H21NO4S) C, H,
N, S]; and 27b (using silylated 5-FU and SnCl4; 62% yield),
1
(274, decrease in intensity, in presence of NaOH); H NMR δ
11.83 (s, ring 3-H), 8.16 (d, J ) 7.0, ring 6-H), 5.60 (t, J ) 6.4,
SCHN) and 2.04 (s, MeS). Anal. (C16H14FN3O4S) H, N; C:
calcd 52.9, found 53.4.
b. Equimolar amounts of the bromide 31 and thiourea in
acetone (10 mL mmol-1) were refluxed for 1 h. The isothiou-
ronium bromide 32 (89%) had mp 197-198 °C (MeOH) [Anal.
(C16H15BrFN5O4S) C, H, N, S]. A mull of 32 (2.36 g, 5 mmol)
in dimethyl sulfoxide36 (2.5 mL) was chilled in ice-water while
2 M NaOH (12.5 mL) was added with stirring. After 15 min
at room temperature the solution of the sodium salt of the thiol
29 (phthalamate form) was treated with MeI (2.49 mL, 40
mmol) and acetone (12.5 mL) and stirred 30 min longer.
Acetone and excess MeI were evaporated, and the residual
cloudy solution (containing very little free 5-FU, from UV
spectrum) was brought to pH 4-5, with cooling, using 2 M
HCl. The S-methylated phthalamic acid was extracted into
CH2Cl2 (3 × 50 mL), washed with saturated brine (17 mL),
and dried (MgSO4). Evaporation gave a gum (2.14 g) which
was refluxed in AcOH (20 mL) for 2 h to reform the phthal-
imide ring. The brown gum remaining after evaporation was
triturated with water (5 mL), leaving an insoluble portion (869
mg) showing two main spots in TLC [CHCl3-MeOH (19:1)] of
which the slower running was due to the phthalimide 22b.
Its methanolic solution gradually deposited crystals (in all, 494
mg) at -18 °C. Recrystallization from PhMe yielded 22b (314
mg, 17%), mp 154.5-157 °C.
Application of this reaction requence to the isothiouronium
chloride made from chloride 28a gave 22b in 11% yield.
The water-insoluble gum (549 mg) from another experiment
on the bromide (5 mmol, as above) was carefully chromato-
graphed on a column of silica gel (27.4 g) using CH2Cl2
containing increasing amounts of MeOH. Only fractions
showing the two spots in TLC were obtained. However,
recrystallization of these from PhMe yielded 22b (165 mg, 9%),
mp 151-154 °C, and from the mother liquor the material (99
mg, 5%) faster moving in TLC. This was the N3-methyl-seco-
nucleoside 26, mp 145-146 °C (MeOH); UV λmax (nm) 274
(spectrum identical in presence of NaOH, characteristic of
N1,N3-dialkyl-5-FU). Anal. (C17H16FN3O4S) C, H, N, S.
N-[3-(5-F lu or ou r a cil-1-yl)-3-(isop r op ylt h io)p r op yl]-
p h th a lim id e (25a ). The chloride 28a (70 mg, 0.2 mmol) and
propane-2-thiol (0.37 mL, 4 mmol) were stirred for 2 h in
MeNO2 (1 mL) at 100 °C. After cooling, evaporation, and
addition of water (2.4 mL) the product (74 mg) was filtered
off. Recrystallization from MeOH gave 25a (39 mg, 50%), mp
mp 207-209.5 °C (MeCN); UV λmax (nm) 271 [Anal. (C23H20
-
FN3O4S) C, H, N, S], also obtained in the alternative way11

Nucleoside Analogs
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7 1409
192.5-194.5 °C; UV λmax (nm) 275 (274 in presence of NaOH).
Anal. (C18H18FN3O4S) C, H, N.
Reaction at 70 °C or the use of AcOH or DMF as solvents
gave unsatisfactory results.
was removed for 5 min and replaced again. Dry pyridine (5.6
mL) was added and the bath removed once more. After 20
min the solution was shaken successively with 1 M HCl (120
mL) and saturated aqueous NaHCO3 (120 mL), dried (MgSO4),
and evaporated. The resulting gummy trifluoroacetate 36
(13.0 g) in MeOH (120 mL) was treated with TsOH‚H2O (2.1
g). Next day the solution had deposited product (5.61 g), mp
62-65 °C. Evaporation of the filtrate, trituration of the solid
residue with aqueous NaHCO3, and filtration yielded a second
fraction (4.12 g) recrystallized from petroleum ether (bp 60-
80 °C) to give further methoxyphthalimide 37 (in all, 8.28 g,
81%). Another recrystallization did not raise the melting
point. Anal. (C19H19NO3S) C, H, N.
Meth yl 3-P h th a lim id op r op yl Su lfoxid e (20b). Meth-
anethiol (5.85 mL, 0.11 mol) was added to 1 M NaOMe (100
mL) and the solution treated with a warm (40 °C) solution of
N-(3-bromopropyl)phthalimide (26.8 g, 0.1 mol) in MeOH (100
mL). The next day MeOH was replaced by water (100 mL),
and methyl 3-phthalimidopropyl sulfide (16b), mp 54-57 °C
(lit.37 56 °C) was filtered off in quantitative yield. Oxidation
of a portion (11.75 g, 0.05 mol) in EtOH (120 mL) and AcOH
(60 mL) with aqueous H2O2 (30% w/v; 5.65 mL) for 3 days in
the usual way, finally drying (MgSO4) a CHCl3 solution of the
product, gave the sulfoxide 20b, again in quantitative yield:
mp 144-146 °C (EtOH). Anal. (C12H13NO3S) C, H, N, S.
P u m m er er Rea r r a n gem en t of Su lfoxid e 20b. a . The
sulfoxide (2.51 g, 10 mmol) and NaOAc (3.43 g) were stirred
for 3 h in refluxing Ac2O (46 mL). After cooling and evapora-
tion of solvent, the oily product (2.88 g) was isolated using CH2-
Cl2. No residual sulfoxide would be expected under these
reaction conditions, and the product consisted essentially of
the isomeric acetates 21b and 17b, together with the trans-
olefin 18: 1H NMR δ 7.65-7.92 (m, C6H4; 21b, 17b, and 18),
5.15 (s, OCH2S; 17b), 2.22 (s, MeS; 18), 2.19 (s, MeS; 21b),
2.09 (s, OCOCH3; 21b), and 2.08 (s, OCOCH3; 17b). Integra-
tion at the OCH2S signal shows the content of 17b to be about
60%; at MeS, 18 30-35% and 21b 5-10%; and at OCOCH3,
the percent values for 21b and 17b are confirmed.
A solution of the O-acetate11 (554 mg, 1.5 mmol) analogous
to 36 on similar treatment deposited the ether 37 (182 mg),
and workup gave a second fraction (total 295 mg, 58%).
Reaction of the trifluoroacetate 36 in refluxing MeOH (2 h),
or at room temperature omitting TsOH, gave products in lower
yield which were more difficult to purify.
N-[2-Meth oxy-2-(p-tolylth io)eth yl]p h th a lim id e (33b).
2-Phthalimidoethyl p-tolyl sulfoxide11 (313 mg, 1 mmol) was
converted into the trifluoroacetate (homolog of 36) as above.
This was treated with MeOH (4 mL) and TsOH‚H2O (70 mg).
Refluxing for 2 h, evaporation, and trituration of the solid
residue with aqueous NaHCO3 yielded a product (261 mg), mp
75-80 °C, showing three spots in TLC [C6H6-MeOH (4:1)].
The strongest corresponded to the O,S-acetal, with a slower
to the O,O-dimethyl acetal and the fastest to a byproduct such
as di(p-tolyl) disulfide. Recrystallization from MeOH gave the
main component 33b (103 mg, 32%): mp 123.5-125.5 °C
Treatment of the total product with silylated 5-iodouracil
in the usual way resulted merely in recovery of 5-iodouracil
(46%).
1
(benzene-petroleum ether); H NMR δ 7.63-7.90 [m, C6H4-
(CO)2], 7.06-7.46 (m, C6H4S), 5.01 (t, J ) 6.8, SCHO), 3.96(d,
J ) 7.1, CH2), 3.47 (s, MeO), and 2.33 (s, MeC). Anal. (C18H17
NO3S) C, H, N.
-
b. When the sulfoxide 20b (2.51 g) in CH2Cl2 (40 mL) was
treated with trifluoroacetic anhydride (1.82 mL) for 5 min,
followed by dry pyridine (1.85 mL) for 1.5 h, and the solution
washed successively with 1 M HCl and saturated aqueous
NaHCO3, dried (MgSO4), and evaporated, the product (3.04
g) consisted mainly (75-80%) of the linear trifluoroacetate
(17b, CF3CO for Ac): 1H NMR δ 7.65-7.92 (m, C6H4), 6.48 (s,
OCH2S), 2.15 and 2.14 (both s, MeS). The total content of MeS-
containing products (15-20%) is considerably less than in the
reaction product from Ac2O.
A similar reaction solution after 24 h at room temperature
yielded only a gummy mixture (305 mg) containing much
starting material. Repetition of the latter experiment with
the analogous O-acetate11 (355 mg, 1 mmol) showed a similar
pattern (TLC), and this solution was kept a further 24 h before
evaporation and washing in CH2Cl2 with aqueous NaHCO3.
The resulting gum (311 mg) consisted mainly of O,S-acetal and
O,O-acetal. Recrystallization from MeOH afforded only the
latter, N-(2,2-dimethoxyethyl)phthalimide (67 mg, 28%), mp
105-108 °C (MeOH) (lit.38 mp 103 °C); 1H NMR δ 7.6-7.9 [m,
C6H4(CO)2], 4.78 (t, J ) 5.7, OCHO), 3.83 (d, J ) 5.7, CH2)
and 3.38 (s, 2MeO). For C6H6-MeOH (4:1) using Merck
Reaction of the total product with silylated 5-FU gave a
crude solid product (502 mg), mp 115-125 °C, consisting
mainly of N3-substituted 5-FU derivative. Pure N-{3-[[(5-
flu or ou r a cil-3-yl)m eth yl]th io]p r op yl}p h th a lim id e (17b,
5-FU3 for AcO) (143 mg, 4%) has mp 145-147 °C (MeOH); UV
aluminum plates (Art. 5554) precoated with silica gel 60 F254
,
Rf values for the O,O-dimethyl acetal, starting S-p-tolyl
O-acetate, and O-methyl S-p-tolyl acetal are respectively 0.73,
0.80, and 0.84.
The S-p-tolyl O-acetate (355 mg) in refluxing MeOH was
comparable to the trifluoroacetate, yielding (from MeOH) O,S-
acetal (101 mg, 31%) followed by O,O-acetal (31 mg, 12%) from
the mother liquor.
λmax (nm) 270 (302 in presence of NaOH). Anal. (C16H14
FN3O4S) C, H, N, S.
-
Seco-Nu cleosid e F or m a tion fr om 3-P h th a lim id op r o-
p yl p-Tolyl Su lfid e via r-Ch lor in a tion . A mixture of this
sulfide11 (311 mg, 1 mmol) and N-chlorosuccinimide (NCS) (139
mg, 1.04 mmol) in Et2O (2.5 mL) and benzene (5 mL) was
stirred for 24 h. After filtration, solvents were replaced by
CH2Cl2, and a little more (total 56 mg) insoluble material
filtered off. Evaporation gave a semisolid residue which, in
CH2Cl2 (7 mL), was added to silylated 5-FU (from 130 mg, 1
mmol). The mixture was treated at -15 °C with SnCl4 (0.1
mL) and left overnight at room temperature. It was filtered
through Celite into 1 M HCl (4 mL). A solid (188 mg)
separated, mainly N1-substituted seco-nucleoside. Impurities
were extracted into boiling MeOH (50 mL), leaving compound
27a (143 mg, 33%), mp 243-248 °C (lit.11 mp 245-247 °C),
identified by UV spectral data and TLC.
When the sulfide (16b, TolS for MeS) (1 mmol) in CCl4 (3
mL) was treated with SO2Cl2 (0.081 mL, 1 mmol) in CCl4(1
mL) at 4 °C and stirring continued for 4 h at room tempera-
ture, the crystals initially separating mainly redissolved. The
solvent was replaced by CH2Cl2 (7 mL) and the reaction with
silylated 5-FU (1 mmol) carried out as above. The gummy
product (96 mg) gradually solidified and gave seco-nucleoside
27a (60 mg, 14%), mp 245-252 °C (MeOH).
R ea ct ion of t h e Tr iflu or oa cet a t e 36 w it h Silyla t ed
Ur a cil. The trifluoroacetate 36, made as above from the
sulfoxide (654 mg, 2 mmol), in CH2Cl2 (10 mL) was added to
silylated uracil (from 224 mg, 2 mmol) and the mixture treated
at -15 °C with SnCl4 (0.2 mL). The cooling bath was removed,
and next day shaking with 1 M HCl (8 mL) yielded a foam
(837 mg) from the organic layer. This crystallized very slowly
from MeOH, giving material (643 mg, ca. 75%) showing two
principal spots of equal intensity and very similar mobility in
TLC [C6H6-MeOH (4:1)]. The UV spectrum [λmax (nm) 261
(287 in presence of NaOH)] confirmed the presence of N3-
substituted seco-nucleoside 40 and N1-isomer (ca. 1:1). Re-
crystallization (MeOH, C6H6, or MeCN) did not change the
TLC pattern.
N-[3-(5-F lu or ou r a cil-3-yl)-3-m eth oxyp r op yl]p h th a lim -
id e (42b). To a solution of the sulfide 37 (938 mg, 2.75 mmol)
and cyclohexene (0.30 mL, 3 mmol) in CH2Cl2 (12.5 mL) cooled
in ice-water was added dropwise, with hand-swirling, SO2-
Cl2 (0.24 mL, 3 mmol) in CH2Cl2 (2.5 mL) (internal tempera-
ture, 2 °C f 7 °C f 2 °C). After 30 min in the bath, the
solution (always virtually colourless) was evaporated, and the
oily chloro ether 38 in DMF (3.75 mL) at -15 °C treated all
N-[3-Meth oxy-3-(p-tolylth io)p r op yl]p h th a lim id e (37).
A solution of 3-phthalimidopropyl p-tolyl sulfoxide11 (9.81 g,
30 mmol) in CH2Cl2 (120 mL) cooled in ice-water was treated
with trifluoroacetic anhydride (5.4 mL, 39 mmol). The bath

1410 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7
McElhinney et al.
at once with a solution of N1-[(octylthio)carbonyl]-5FU22b (755
mg, 2.5 mmol) and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)
by Et2O (240 mL). The resulting gummy solid (4.91 g)
hardened and was recrystallized from MeOH (10 mL). The
N,N-acetal 35 (2.43 g, 48%) had mp 175-215 °C (MeOH); UV
(0.41 mL, 2.75 mmol) in DMF (2.5 mL) from
a funnel,
continuing to exclude moisture. Crystals soon separated, and
after 1 h at -15 °C as much solvent as conveniently possible
was removed below 40 °C in vacuo, and the residue triturated
with aqueous AcOH (5%; 25 mL). The resulting solid was
filtered off, washed with water, taken up in CH2Cl2, dried
(MgSO4), and isolated by evaporation. Trituration with Et2O
(25 mL) yielded the N1-(octylthio)carbonyl derivative 39b (982
mg, 76%), mp 151.5-152.5 °C (MeCN) [Anal. (C25H30FN3O6S)
C, H, N]; UV λmax (nm) 271, shifted immediately to 304 on
addition of NaOH, corrresponding to deprotection. In contrast,
the spectrum of the reagent N1-[(octylthio)carbonyl]-5-FU, with
UV λmax (nm) 267 (263, 280 sh in present of NaOH), shifts
completely to λmax (nm) 298 in alkali only after 20-30 min.
The derivative in MeCN (15 mL) was filtered from a trace
of insoluble material and treated at room temperature with
isopropylamine (0.18 mL, 2.1 mmol). After 30 min, evapora-
tion, trituration with Et2O (10 mL), and filtration yielded 42b
(550 mg, 84%), mp 182-184 °C (EtOAc); UV λmax 273 (303 in
presence of NaOH). Anal. (C16H14FN3O5) C, H, N.
λ
max (nm) 268 (274, 304 in presence of NaOH); 1H NMR δ 11.66
(bs, 3-H in 5-FU and 1-H in 5-FU3), 8.29 (d, J ) 7.9, 6-H in
5-FU), 7.83 (m, 6-H in 5-FU3 and C6H4) and 6.80 (t, J ) 8.1,
NCHN). Anal. (C19H13F2N5O6.0.5H2O) C, H, N.
N-(2-Ch lor oeth yl)-N-n itr osou r ea s (Ta ble 5). The start-
ing phthalimides were hydrazinolyzed either in DMF10,15 (1.5
mL mmol-1) at -15 °C overnight (24a ,b, 22b, 35; 30 at 3 °C),
or during 1 h in refluxing MeOH (8 mL mmol-1) for N3-
substituted seco-nucleosides (42a ,b) where formation of a
bicyclic isomer30 is not a potential problem. Hydrazine hydrate
(1.1 equiv) was used initially (24a ,b, 30, 22b), filtering off
phthalohydrazide (from water) after acidification and evapora-
tion of DMF. When N2H4‚H2O (2 equiv) was used (42a ,b, 35),
excess reagent was subsequently removed as benzalazine.15
The aqueous methanolic solution of amine hydrochloride was
treated10,15 with NaOMe and 2,4,5-trichlorophenyl N-(2-chlo-
roethyl)-N-nitrosocarbamate for 2 h at 0-5 °C. After quench-
ing and extraction into CH2Cl2 (5, 10, 4) or EtOAc, 2,4,5-
trichlorophenol was removed by trituration of total product
with Et2O. The CNUs crystallized except in the case of 9
(solvate from EtOAc) and 10 which required chromatography
on a column of silica gel [elution with CH2Cl2-MeOH (9:1)]
and crystallized from H2O-EtOH (6:4). The other compounds
were recrystallized from EtOH.
For the C4 seco-nucleoside 24d , bicyclic isomer30 formation
is also unlikely, and hydrazinolysis (using 1.1 equiv in
MeOCH2CH2OH during 1 h at 100 °C), followed by evapora-
tion, acidification with 0.25 N HCl and filtration, and re-
evaporation, left crude amine hydrochloride which with trif-
luoroacetic anhydride in EtOAc yielded (20-30%) the tri-
fluoroacetamide (24d , CF3CONH for PhtN), mp 150.5-152 °C
(EtOAc-petroleum ether); UV λmax (nm) 268 (267 in presence
of NaOH) [Anal. (C13H17F4N3O4) C, H, N]. The product of
hydrazinolysis of 24d at -15 °C in DMF was very impure, and
best results were obtained by warming 24d (3.50 g, 9 mmol)
in 0.1 N NaOMe (90 mL) and N2H4‚H2O (1.34 mL, 27 mmol)
for 30 min at 45 °C. Acidification, evaporation, and benzala-
zine workup and reaction with the nitrosocarbamate as before
gave a product which was extracted into CH2Cl2, chromato-
graphed on a silica gel column and crystallized from EtOH,
yielding 1.09 g (31%).
Biologica l Eva lu a tion . Tu m or System s: MAC Tu m or s
(Br a d for d ). The MAC system has been shown to be a good
model of human colorectal cancer in terms of its chemosensi-
tivity39 and has been shown to have the appropriate charac-
teristics to be used in the development of new combination
regimens.40 MAC 15A (1 × 106 cells) was injected intraperi-
toneally (ip) in male NMRI mice, and MAC 13 tumor frag-
ments were implanted subcutaneously (sc) in the flank of
female NMRI mice by means of a trocar.
Ma m m a r y Ca r cin om a (MCa ) (Za gr eb). A total of 1 ×
106 viable MCa tumor cells were injected intramuscularly (im)
into the right flank of inbred 10-14-week-old CBA/H mice at
the Ruder Bosˇkovic Institute, Zagreb.41
Colon 38 Ad en oca r cin om a (Br u ssels). The Colon 38
adenocarcinoma originated from the NCI Frederick Cancer
Research Facility (Frederick, MD) and was serially maintained
in C57B1/6 mice. Tumor fragments of 3 mm × 3 mm were
transplanted sc in B6D2F1 mice (weighing 19-22 g) on day
0. The B6D2F1 mice used for testing were purchased also from
Frederick.
Exp er im en ta l Ch em oth er a p y: MAC Tu m or s. (i) MAC
13. Two days after transplantation, when tumors were ∼2
mm3 in volume, mice were allocated into groups by restricted
randomization. Compounds for injection were made up freshly,
and treatment (0.1 mL/10 g of body weight) was ip. Chemosen-
sitivity was assessed by tumor weight inhibition 21 days later
and presented as T/C % where T represents mean weight of
treated tumors and C represents mean weight of control tumor.
Reaction of the chloro ether 38 for 1 h at room temperature
afforded 39b in 63% yield. In two other experiments on sulfide
37 (3.3 mmol) cyclohexene was omitted and the bright yellow
oily products extracted with petroleum ether (5 + 3 mL) to
remove much of the sulfenyl chloride; reaction (overnight) of
the relatively insoluble 38 at 0 °C and at -15 °C respectively
gave 39b in yields of 28% and 55%.
N1-[(Octylth io)ca r bon yl]u r a cil. No details about the
preparation or properties of this compound are available.29
Uracil (784 mg, 7 mmol) was dissolved in DMF (28 mL) at
110 °C and the solution cooled (just) to room temperature.
S-Octyl chlorothioformate (1.61 g, 7.7 mmol) in DMF (7 mL)
was added in one portion, followed by pyridine (5.6 mL). The
next day the solution was filtered from a trace of solid and
evaporated. Trituration of the crystalline residue with aque-
ous AcOH (5%; 42 mL) yielded crude product (1.40 g), reduced
(to 1.07 g) by washing with petroleum ether (12 mL). Treat-
ment with water (14 mL) at 100 °C removed a little uracil,
leaving N1-[(octylthio)carbonyl]uracil (947 mg, 48%), mp 163.5-
167 °C. The analytical sample (from MeCN) had mp 169-
171 °C [Anal. (C13H20N2O3S) C, H, N]; UV λmax (nm) 262 (261
in presence of NaOH; after 20-30 min, shifts completely to
287 indicating cleavage to uracil). A reaction time of 3 days
did not increase the yield, nor addition of various proportions
of MeCN as solvent.
N-[3-(Ur a cil-3-yl)-3-m eth oxyp r op yl]p h th a lim id e (42a ).
The sulfide 37 (2.73 g, 8 mmol) was converted into the chloride
38 as described before. This was treated in DMF (11 mL) with
a solution of N1-[(octylthio)carbonyl]uracil (2.27 g, 8 mmol) and
DBU (1.32 mL, 8.8 mmol) in DMF (27 mL) at room temper-
ature. After 1 h the solvent was evaporated, and workup gave
crude 39a (3.10 g), contaminated with some reagent. This was
treated in MeCN (51 mL) with isopropylamine (0.8 mL) for
30 min. Uracil (200 mg, 22%) was filtered off, and evaporation,
trituration with Et2O (30 mL), and filtration yielded 42a (1.65
g, 62% based on 37). Recrystallization from MeOH gave a
product (1.36 g), mp 159-172 °C, containing some solvent, and
the analysis sample (from MeCN; dried at 100 °C in vacuo)
had mp 171.5-174.5 °C; UV λmax (nm) 264, 298 sh (294 in
presence of NaOH). Anal. (C16H15N3O5) C, H, N.
N-[3-(5-F lu or ou r a cil-1-yl)-3-(5-flu or ou r a cil-3-yl)p r op -
yl]p h th a lim id e (35). A mixture of the chloride 28a (4.39 g,
12.5 mmol) and DMF (25 mL) was treated dropwise over 10
min with a solution of N1-[(octylthio)carbonyl]-5-FU (3.40 g,
11.25 mmol) and NEt3 (1.74 mL, 12.5 mmol) in DMF (18 mL).
Stirring was continued overnight and the solvent then evapo-
rated. Addition of aqueous AcOH (5%; 112 mL) gave a
semisolid which was taken up in CH2Cl2 (125 mL). Extraction
of the aqueous layer with further CH2Cl2 (2 × 30 mL), drying
(MgSO4), and evaporation yielded the crude (octylthio)carbonyl
intermediate as a syrup which was dissolved in MeCN (150
mL) and clarified by filtration. It was stirred with isoprop-
ylamine (1.07 mL, 12.5 mmol) for 30 min (the solid which
initially separated redissolved) and the solvent then replaced
(ii) MAC 15A. Animals bearing MAC 15A tumor cells were
allocated into groups and treated (0.1 mL/10 g of body weight)
on day 2. Antitumor activity was determined by comparison

Nucleoside Analogs
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7 1411
(6) Bibby, M. C; Double, J . A.; McCormick, J . E.; McElhinney, R.
S.; Radacic, M.; Pratesi, G.; Dumont, P. Nucleoside Analogues.
13. The Effect on Anti-tumour Activity of Varying the Uracil
5-Substituent and the Point of Attachment (N1 or N3) of the
Uracil Moiety in Seco-nucleoside Nitrosoureas. Anti-Cancer Drug
Des. 1993, 8, 115-128.
(7) (a) McCormick, J . E.; McElhinney, R. S. Thio-sugars. Part 3.
4-Thiotetrofuranose Nucleosides. J . Chem. Soc. 1978, 500-505.
(b) McCormick, J . E.; McElhinney, R. S. Thio-sugars. Part 6.
Seco-nucleosides related to 4-Thio-DL-ribofuranose. J . Chem. Res.
(S) 1981, 12-13; (M) 256-281. (c) McElhinney, R. S.; McCor-
mick, J . E.; Lucey, C. M. Evolution of Molecular Combinations
Involving 5-Fluorouracil. Cancer Treat. Rev. 1988, 15, 73-81.
(8) McCormick, J . E.; McElhinney, R. S.; McMurry, T. B. H.;
Maxwell, R. J . Nucleoside Analogues. Part 12. The Anomalous
19F NMR Spectrum of B.3996, a Molecular Combination of
5-Fluorouracil and N-(2-Chloroethyl)-N-nitrosourea and Syn-
thesis of its N ′-Nitroso Isomer and Related Compounds. J .
Chem. Soc., Perkin Trans. 1 1991, 877-880.
(9) McElhinney, R. S.; McCormick, J . E.; Bibby, M. C.; Double, J .
A.; Atassi, G.; Dumont, P.; Pratesi, G.; Radacic, M. Nucleoside
Analogues: 7. Effect on Colon, Breast, and Lung Tumours in
Mice of 5-Fluorouracil/Nitrosourea Molecular Combinations
Incorporating Alkoxy and Oxidized Sulphur Functions. Anti-
Cancer Drug Des. 1989, 3, 255-269.
(10) McElhinney, R. S.; McCormick, J . E.; Bibby, M. C.; Double, J .
A.; Atassi, G.; Dumont, P.; Pratesi, G.; Radacic, M. Nucleoside
Analogues. 9. Seco-nucleoside Analogues of some 5-Fluorouracil/
Nitrosourea Molecular Combinations having Uracil as Base:
Synthesis and Anti-Tumour Activity. Anti-Cancer Drug Des.
1989, 4, 191-207.
(11) McCormick, J . E.; McElhinney, R. S. Nucleoside Analogues. Part
1. Some 5-Fluorouracil Seco-nucleosides and their Hydrolysis
by Dilute Acid. J . Chem. Res. (S) 1983, 176-177; (M) 1736-
1769.
(12) Double, J . A.; Bibby, M. C.; McCormick, J . E.; McElhinney, R.
S. Nucleoside Analogues. 5. Molecular Combination of Anti-
cancer Drugs: Activity of 5-Fluorouracil/Nitrosourea Combina-
tions against Mouse Colon Tumours. Anti-Cancer Drug Des.
1986, 1, 133-139.
(13) Roger, P.; Monneret, C.; Fournier, J .-P.; Choay, P.; Gagnet, R.;
Gosse, C.; Letourneux, Y.; Atassi, G.; Gouyette, A. Rationale for
the Synthesis and Preliminary Biological Evaluation of Highly
Active New Antitumour Nitrosoureido Sugars. J . Med. Chem.
1989, 32, 16-23.
(14) McCormick, J . E.; McElhinney, R. S. Nucleoside Analogues. Part
6. The Preparation by Dephthaloylation and Properties of N-(ω-
Aminoalkyl)uracils, Key Intermediates in the Synthesis of
Molecular Combinations of Certain Anti-tumour Agents. Proc.
R. Irish. Acad. 1989, 89B, 213-228.
of the lifespan of treated and control groups. Deaths were
recorded and the median survival time (MST) determined.
Ma m m a r y Ca r cin om a (MCa ). Drugs were administered
ip as a single dose at a constant volume of 0.02 mL/g of body
weight on day 1 after tumor cell implantation. Effects of
therapy were assessed by comparison of MST of treated and
control groups.
Colon 38 Ad en oca r cin om a . All compounds were pre-
pared freshly on the day of injection. B.3958, B.3970, and
B.3971 were dissolved in saline (0.9% NaCl) while the other
compounds were suspended in saline by adjunction of Tween
80 (0.1%). All treatments were intraperitoneally (ip) admin-
istered on days 2 and 9 as volumes of 0.1 mL/10 g of mouse
body weight and all individual animal body weights recorded
on days 2 and 20.
The tumor size measured on day 20 was converted into
tumor volume by using the formula P ) L(W 2/2) where P is
the tumor weight in mg when the length (L) and width (W)
are given in millimeters. The median tumor volume was
calculated for both treated and untreated control groups. The
T/C, i.e., the median tumor volume of the treated group divided
by that of the control group, indicated antitumor activity. A
T/C of <42% was considered necessary to demonstrate activity.
Ack n ow led gm en t. We are grateful to Rhoˆne-Pou-
lenc Ltd. (Dagenham) for microanalyses and IR spectra,
to Professor T. B. H. McMurry of the University Chemi-
cal Laboratory, Trinity College, Dublin, for NMR spec-
tra, and to Neil Lucey for preparation of intermediates.
We acknowledge with thanks financial support from the
Irish Cancer Society, Bradford’s War on Cancer, and the
Screening and Pharmacology Group of the EORTC. We
thank Mr. George Curran of BASF Ireland Ltd. for a
generous gift of S-octyl chlorothioformate.
Refer en ces
(1) Haddow, A.; Timmis, G. M. Myleran (GT41) in Chronic Myeloid
Leukaemia. Chemical Constitution and Biological Action. Lancet
1953, 264 (i), 207-208. Bergel, F.; Stock, J . A. Cyto-active
Amino-acid and Peptide Derivatives. Part I. Substituted Phe-
nylalanines. J . Chem. Soc. 1954, 2409-2417. Arnold, H.;
Bourseaux, F. Cytostatic Active Cyclic N′,O-Alkylene-N,N-bis-
(â-chloroethyl) Diamidophosphates. Angew. Chem. 1958, 70,
539-544.
(2) Schabel, F. M., J r.; J ohnston, T. P.; McCaleb, G. S.; Montgomery,
J . A.; Laster, W. R., J r; Skipper, H. E. Experimental Evaluation
of Potential Anticancer Agents. VIII. Effects of Certain Ni-
trosoureas on Intracerebral L1210 Leukaemia. Cancer Res. 1963,
23, 725-733. J ohnston, T. P.; McCaleb, G. S.; Montgomery, J .
A. Synthesis of Chlorozotocin, the 2-Chloroethyl Analog of the
Anticancer Antibiotic Streptozotocin. J . Med. Chem. 1975, 18,
104-106.
(15) Lucey, N. M.; McCormick, J . E.; McElhinney, R. S. Nucleoside
Analogues. Part 10. Synthesis of Three Unusual 5-Fluorouracil
Seco-nucleosides Incorporating N-(2-Chloroethyl)-N-nitrosourea
Residues. J . Chem. Soc., Perkin Trans. 1 1990, 795-802.
(16) McElhinney, R. S.; McCormick, J . E.; Bibby, M. C.; Double, J .
A.; Atassi, G.; Dumont, P.; Pratesi, G.; Radacic, M. Nucleoside
Analogues: 8. Some Isomers of B.3839, the Original 5-Fluoro-
uracil/Nitrosourea Molecular Combination, and Their Effect on
Colon, Breast and Lung Tumours in Mice. Anti-Cancer Drug Des.
1989, 4, 1-20.
(3) Baer, J . C.; Freeman, A. A.; Newlands, E. S.; Watson, A. J .;
Rafferty, J . A.; Margison, G. P. Depletion of O6-Alkylguanine-
DNA-Alkyltransferase Correlates with Potentiation of Temozo-
lomide and CCNU Toxicity in Human Tumour Cells. Br. J .
Cancer 1993, 67, 1299-1302. Arris, C. E.; Bleasdale, C.; Calvert,
A. H.; Curtin, N. J .; Dalby, C.; Golding, B. T.; Griffin, R. J .; Lunn,
J . M.; Major, G. N.; Newell, D. R. Probing the Active Site and
Mechanism of Action of O6-Methylguanine-DNA-Methyltrans-
ferase with Substrate Analogues (O6-Substituted Guanines).
Anti-Cancer Drug Des. 1994, 9, 401-408. Clark, A. S.; Deans,
B.; Stevens, M. F. G.; Tisdale, M. J .; Wheelhouse, R. T.; Denny,
B. J .; Hartley, J . A. Antitumor Imidazotetrazines. 32. Synthesis
of Novel Imidazotetrazinones and Related Bicyclic Heterocycles
to Probe the Mode of Action of the Antitumor Drug Temozolo-
mide. J . Med. Chem. 1995, 38, 1493-1504. Chae, M.-Y.; Swenn,
K.; Kanugula, S.; Dolan, M. E.; Pegg, A. E.; Moschel, R. C.
8-Substituted O6-Benzylguanine, Substituted 6(4)-(Benzyloxy)-
pyrimidine, and Related Derivatives as Inactivators of Human
O6-Alkylguanine-DNA-Alkyltransferase. J . Med. Chem. 1995, 38,
359-365.
(17) Dilworth, B. M.; McKervey, M. A. Organic Synthesis with
R-Chlorosulfides. Tetrahedron 1986, 42, 3731-3752.
(18) Tuleen, D. L.; Stephens, T. B. Chlorination of Unsymmetrical
Sulfides. J . Org. Chem. 1969, 34, 31-35.
(19) Meffre, P.; Lhermitte, H.; Vo-Quang, L.; Vo-Quang, Y.; Le Goffic,
F. A New Class of Unusual R-Aminoacids: R-Aminoacid â,γ-Enol
Thioethers. Tetrahedron Lett. 1991, 32, 4717-4720.
(20) Cerny, M.; Paca´k, J . Eine Neue Methode zur Darstellung
Acetylierter â-D-Thioglucopyranoside. (A New Method for the
Preparation of Acetylated â-D-Thioglucopyranosides.) Collect.
Czech. Chem. Commun. 1959, 24, 2566-2569. Durette, P. L.;
Shen, T. Y. Insulin-like, and Insulin-Antagonistic, Carbohydrate
Derivatives. The Synthesis of Aryl and Aralkyl D-Mannopyra-
nosides and 1-Thio-D-mannopyranosides. Carbohydr. Res. 1980,
81, 261-274.
(21) 27a is obtained (50-60% yield overall) from TolS(CH2)3NPht via
sulfoxide and Pummerer trifluoroacetate 36;11 via R-chlorination
using NCS the yield of 27a is 33%, using sulfuryl chloride 14%.
The NCS sequence on MeS(CH2)3NPht (16b) gave only a trace
(3% yield) of 22b, isolated with some difficulty. Reaction of
R-chloro sulfides with nitrogen nucleophiles has been infre-
quent,17,22 and their use in nucleoside synthesis23 has been
superseded.7a,24
(4) McCormick, J . E.; McElhinney, R. S. Nitrosoureas from Chemist
to Physician: Classification and Recent Approaches to Drug
Design. Eur. J . Cancer 1990, 26, 207-221.
(5) McCormick, J . E.; McElhinney, R. S. Thio-sugars. Part 7.
Secouridines with Amino-groups in the “Carbohydrate” Com-
ponent: Intramolecular Addition across the 5,6-Double Bond and
Molecular Combination of 5-Fluorouracil and N-(2-Chloroethyl)-
N-nitrosourea Residues. J . Chem. Res. (S) 1981, 310-311; (M)
3601-3641.
(22) (a) Worley, J . W. N-(R-Chloroalkyl)phthalimides. J . Org. Chem.
1979, 44, 1178-1180. (b) Ozaki, S.; Watanabe, Y.; Fujisawa, H.;
Hoshiko, T. 5-Fluorouracil Derivatives. VI. Alkylation of 5-Fluo-
rouracil Using the (Octylthio)carbonyl Group as a Protecting
Function. Heterocycles 1984, 22, 527-530.

1412 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 7
McElhinney et al.
(23) Bobek, M.; Block, A.; Parthasarathy, R.; Whistler, R. L. Synthesis
and Biological Activity of 5-Fluoro-4′-thiouridine and Some
Related Nucleosides. J . Med. Chem. 1975, 18, 784-787 and
references cited.
(24) Uenishi, J .; Motoyama, M.; Nishiyama, Y.; Wakabayashi, S.
Stereocontrolled Preparation of Cyclic Xanthate; a Novel Syn-
thetic Route to 4-Thiofuranose and 4′-Thionucleoside. J . Chem.
Soc., Chem. Commun. 1991, 1421-1422. Secrist, J . A., III; Riggs,
R. M.; Tiwari, K. N.; Montgomery, J . A. Synthesis and Anti-
HIV Activity of 4′-Thio-2′,3′-dideoxynucleosides. J . Med. Chem.
1992, 35, 533-538.
(25) Chloride in 28a (C3) is displaced in good yield by alkoxy by
refluxing briefly in lower alcohols, while compound 23 (C2) can
be recrystallized from methanol and only reacts after longer
heating in the presence of Ag+. In turn the ether 24a (C3) is
hydrolyzed by dilute acid very much faster than 19 (C2). The
deactivating effect of a 2-substituent has long been known in
sugar chemistry, confirmed by the great difference in rates of
acid hydrolysis of 5-fluorouridine and 5-fluoro-2′-deoxyuridine.11
The S-oxides of 27a (C3) and of 22a (C2) also differ greatly in
the rate of acid hydrolysis to 5-FU; further, the C3 sulfoxide
yields olefin by elimination of sulfenic acid smoothly in boiling
toluene, whereas the C2 sulfoxide requires fusion at 215 °C in
vacuo.
(31) Ferrier, R. J .; Tyler, P. C. Photobromination of Carbohydrate
Derivatives. Part 5. Preparation and Reactions of (5S)-1,2,3,4-
Tetra-O-acetyl-5-bromo-â-D-xylopyranose; a New Type of “Double-
Headed” Nucleoside. J . Chem. Soc., Perkin Trans. 1 1980, 2767-
2773.
(32) Stenhouse, J . Liebigs Ann. Chem. 1840, 33, 92. Zaugg, H. E.;
Martin, W. B. R-Amidoalkylations at Carbon. Org. React. 1965,
14, 52-269 (p 107).
(33) J ohnson, H. E.; Crosby, D. G. Reaction of Hydantoin with
Acetals. J . Org. Chem. 1962, 27, 2077-2080.
(34) Loadman, P. M.; Matthew, A. M.; McCormick, J . E.; McElhinney,
R. S.; Bibby, M. C. Pharmacokinetics and Anti-Tumour Activity
of B.4152, a Reactive Molecular Combination of 5-Fluorouracil
and N-(2-Chloroethyl)-N-nitrosourea, and its Uracil Analogue.
Anti-Cancer Drug Des., in the press.
(35) Loadman, P. M.; Bibby, M. C.; Double, J . A.; McElhinney, R. S.
Pharmacokinetics and Cytotoxicity of B.3839,
a Molecular
Combination of 5-Fluorouracil and N-(2-Chloroethyl)-N-
nitrosourea, in a Mouse Model. Invest. New Drugs 1992, 10,
149-158.
(36) Pan, H.-L.; Fletcher, T. L. A Facile Preparation of Sulphydryl
Compounds in Dimethyl Sulphoxide. Chem. Ind. (London) 1968,
546.
(37) Karrer, P.; Scheitlin, E.; Siegrist, H. U¨ ber Homologe des Sul-
foraphans und u¨ber ω-Aminoalkyl-sulfoxyde. (A Study of Ho-
mologs of Sulforaphan and of ω-Aminoalkyl Sulfoxides.) Helv.
Chim. Acta 1950, 33, 1237-1245.
(38) Za´bransky, J .; Houska, M.; Ka´lal, J . Aldehydes and Acetals
Based on Acryl- and Methacrylamides as Potential Monomers.
Makromol. Chem. 1985, 186, 223-229.
(39) Double, J . A.; Ball, C. R. Chemotherapy of Transplantable
Adenocarcinomas of the Colon in Mice. Cancer Chemother. Rep.
1975, 59, 1083-1089.
(40) Cowen, D. M.; Siegerstetter, J .; J anik, A. C.; Double, J . A. The
Assessment of Response of Murine Transplantable Colon Tu-
mours to Combination Chemotherapy. J . Cancer Res. Clin.
Oncol. 1982, 103, 119-126.
(26) In a direct comparison, the C2 homolog of the p-tolyl trifluoro-
acetate 36 was found to react very incompletely with methanol
at room temperature, forming several products; at reflux tem-
perature O,S-acetal 33b (32%) was obtained and some O,O-
dimethyl acetal identified. The corresponding C2 p-tolyl acetate11
(i.e., 21a with Tol for Me) had similar reactivity: at room
temperature displacement of arylthio was extensive and O,O-
dimethyl acetal (28% after 48 h) was the only product isolated
by recrystallization. The C3 p-tolyl acetate11 analogous to 36 was
only slightly less reactive than 36, giving 37 in 58% yield.
(27) Summers, L. The R-Haloalkyl Ethers. Chem. Rev. 1955, 55, 301-
353 (p 337).
(28) Antonsen, O.; Benneche, T.; Hagelin, G.; Undheim, K. Synthesis
of Unsaturated Chloromethyl Ethers. Acta Chem. Scand. 1989,
43, 56-61.
(29) Ozaki, S.; Watanabe, Y.; Hoshiko, T.; Fujisawa, H.; Uemura, A.;
Ohrai, K. 5-Fluorouracil Derivatives. VII. Tert-amine Promoted
N-Glycosidation of Pyrimidines. Tetrahedron Lett. 1984, 25,
5061-5062.
(41) Radacic, M.; Boranic, M.; Basic, I. A Mouse Mammary Carci-
noma as a Model for Evaluation of Drug Effects on Primary
Tumour Growth and Metastases. In Tumour Progression and
Markers; Lapis, K., J eney, A., Eds.; Kluger Publications: Am-
sterdam, 1982; pp 217-223.
(30) McCormick, J . E.; McElhinney, R. S. Nucleoside Analogues. Part
2. Further Molecular Combinations of (5-Substituted) Uracil and
N-(2-Chloroethyl)-N-nitrosourea Residues as Anticancer Agents.
J . Chem. Soc., Perkin Trans. 1 1985, 93-100.
J M9507237