Triazinines: Synthesis and proteolytic decomposition of a new class of cyclic triazenes

Article abstract of DOI:10.1021/jo970721p

The reaction of 1-azido-3-chloropropane with various Grignard reagents and subsequent treatment with anhydrous isopropylamine results in the formation of the corresponding azimine. If the initial magnesium-triazene complex if first hydrolyzed with Dowex resin and then concentrated, the resultant linear triazene begins self-catalyzed cyclization to form the six- membered-ring triazenes as the major product, with HCl as the byproduct. Addition of an amine, at reduced temperature, allows for the neutralization of the byproduct, HC1, which would otherwise react with the linear triazene and the cyclic six-membered-ring triazene to form hydrolysis products. We have assigned the trivial name of triazinines to this new class of cyclic triazenes. The hydrolytic decomposition of these compounds in mixed acetonitrile-aqueous buffers predominantly forms 3-(alkylamino)-1-propanol and lesser amounts of the rearranged alcohol 1-(alkylamino)-2-propanol and N- alkyl-2-propenamine. The rate of hydrolysis of 1-alkyltriazinines is approximately equal to that of the analogous 1,3,3-trialkyltriazenes, about three times slower than that of the analogous 1-alkyltriazolines, and varies in the order ethyl > butyl > 3,3-diethoxypropyl > benzyl. As was true for other triazenes, the mechanism of the decomposition was found to be specific acid-catalyzed (A1), involving rapid reversible protonation followed by rate- limiting formation of a 3-(alkylamino)propyldiazonium ion. The slopes of the log k(obs) versus pH plots were near -1.0 The solvent deuterium isotope effect, k(H)₂O/k(D)₂O, was in all cases <1.0 and ranges from 0.82 for 1- benzyltriazinine to 0.89 for 1-ethyltriazinine. The activation parameters of the proteolytic decomposition of a series, 1-ethyltriazinine, 1- ethyltriazoline, 1,3,3-triethyltriazene, and 1-ethyl-3-methyltriazene, had similar values for ΔH(+) (+9 → 12 kcal/mol) and ΔS(+) (+7 → 15 eu), respectively.

Full text of DOI:10.1021/jo970721p

8660
J . Org. Chem. 1997, 62, 8660-8665
Tr ia zin in es: Syn th esis a n d P r oteolytic Decom p osition of a New
Cla ss of Cyclic Tr ia zen es
Brigitte F. Schmidt,† Emily J . Snyder,† Robin M. Carroll,‡ David W. Farnsworth,*,†
Christopher J . Michejda,† and Richard H. Smith, J r.†,‡
Molecular Aspects of Drug Design Section, ABL-Basic Research Program, NCI-Frederick Cancer Research
and Development Center, Frederick, Maryland 21702, and Department of Chemistry,
Western Maryland College, Westminster, Maryland 21157
Received April 22, 1997X
The reaction of 1-azido-3-chloropropane with various Grignard reagents and subsequent treatment
with anhydrous isopropylamine results in the formation of the corresponding azimine. If the initial
magnesium-triazene complex is first hydrolyzed with Dowex resin and then concentrated, the
resultant linear triazene begins self-catalyzed cyclization to form the six-membered-ring triazenes
as the major product, with HCl as the byproduct. Addition of an amine, at reduced temperature,
allows for the neutralization of the byproduct, HCl, which would otherwise react with the linear
triazene and the cyclic six-membered-ring triazene to form hydrolysis products. We have assigned
the trivial name of triazinines to this new class of cyclic triazenes. The hydrolytic decomposition
of these compounds in mixed acetonitrile-aqueous buffers predominantly forms 3-(alkylamino)-
1-propanol and lesser amounts of the rearranged alcohol 1-(alkylamino)-2-propanol and N-alkyl-
2-propenamine. The rate of hydrolysis of 1-alkyltriazinines is approximately equal to that of the
analogous 1,3,3-trialkyltriazenes, about three times slower than that of the analogous 1-alkyltria-
zolines, and varies in the order ethyl > butyl > 3,3-diethoxypropyl > benzyl. As was true for other
triazenes, the mechanism of the decomposition was found to be specific acid-catalyzed (A1), involving
rapid reversible protonation followed by rate-limiting formation of a 3-(alkylamino)propyldiazonium
ion. The slopes of the log kobs versus pH plots were near -1.0. The solvent deuterium isotope
effect, kH O/kD O, was in all cases <1.0 and ranges from 0.82 for 1-benzyltriazinine to 0.89 for
2
2
1-ethyltriazinine. The activation parameters of the proteolytic decomposition of a series, 1-ethyl-
triazinine, 1-ethyltriazoline, 1,3,3-triethyltriazene, and 1-ethyl-3-methyltriazene, had similar values
for ∆Hq (+9 f 12 kcal/mol) and ∆Sq (+7 f 15 eu), respectively.
Sch em e 1
In tr od u ction
We have been interested in the chemistry and biologi-
cal properties of triazenes for over a decade. The
chemistry of acyclic 1,3-dialkyl- and 1,3,3-trialkyltriaz-
enes has been studied extensively. Efficient methods
were developed for their synthesis from alkyl azides,1,2
and their hydrolytic decomposition in aqueous buffers
was shown to be specific acid-catalyzed.3,4 In general,
the mechanism of proteolytic decomposition of all of the
triazenes studied involves a rapid and reversible proto-
nation at N(3) followed by N(2)-N(3) heterolysis which
leads to the formation of a diazonium ion5 intermediate,
as indicated in Scheme 1.
Recently we reported on the chemistry of simple
triazolines6,7 and a related class of high nitrogen com-
pounds, the azimines.8 Preparation of triazolines can be
accomplished by either base-catalyzed decomposition of
1,3-dialkyl-3-acyltriazenes6 or by a slight modification of
a previously reported synthesis involving the reaction of
alkyl azides with dimethylsulfoxonium methylide.7 The
reaction of 1-azido-3-chloropropane followed by treatment
with anhydrous isopropylamine results in the smooth
formation of azimines.8
Numerous cyclic triazenes, where the ring system
contains a carbonyl or imine group and is fused to an
aromatic moiety, have been reported.9 Chemical and
biological studies10 of these triazinones lead to derivatives
containing an imidazole ring.11-13 These compounds have
been studied as prodrugs which upon hydrolysis ring
open to linear triazenes14 analogous to the antitumor
drug DTIC.15
Small to medium-sized cyclic triazenes constrain the
triazene group to a Z-configuration (as opposed to the
X Abstract published in Advance ACS Abstracts, December 1, 1997.
(1) Sieh,D. H.; Wilbur, D. J .; Michejda, C. J . J . Am. Chem. Soc. 1980,
102, 3883-3887.
(2) Smith, R. H., J r.; Michejda, C. J . Synthesis 1983, 476-477.
(3) Smith, R. H., J r.; Denlinger, C. L.; Kupper, R.; Koepke, S. R.;
Michejda, C. J . J . Am. Chem. Soc. 1984, 106, 1056-1059.
(4) Smith, R. H., J r.; Denlinger, C. L.; Kupper, R.; Mehl, A. F.;
Michejda, C. J . J . Am. Chem. Soc. 1986, 108, 3726-3730.
(5) Smith, R. H., J r.; Wladkowski, B. D.; Mehl, A.F.; Cleveland, M.
J .; Rudrow, E. A.; Chmurny, G. N.; Michejda, C. J . J . Org. Chem. 1989,
54, 1036-1042.
(6) Smith, R. H., J r.; Wladkowski, B. D.; Herling, J . A.; Pfaltzgraff,
T.D.; Taylor, J . E.; Thompson, E.J .; Pruski, B.; Klose, J . R.; Michejda,
C. J . J . Org. Chem. 1992, 57, 6448-6454.
(7) Smith, R. H., J r.; Wladkowski, B. D.; Taylor, J . E.; Thompson,
E.J .; Pruski, B.; Klose, J . R.; Andrews, A. W.; Michejda, C. J . J . Org.
Chem.. 1993, 58, 2097-2103.
(8) Smith, R. H., J r.; Pruski, B.; Day, C. S.; Pfaltzgraff, T. D.;
Michejda, C. J . Tetrahedron Lett. 1992, 33, 4683-4686.
(9) Shealy, Y. F. Triazenylimidazoles, Other Triazenes, and Imidazo-
[5,1-d]-1,2,3,5-tetrazin-4(3H)-ones In Cancer Chemotherapeutic Agents;
Foye, W. O., Ed.; ACS Professional Reference Book: Washington, D.
C., 1995; pp 159-186.
(10) Stevens, M. F. G. J . Chem. Soc., Perkin Trans. 1 1974, 615-
620.
(11) Stevens, M. F. G.; Hickman, J . A.; Stone, R.; Gibson, N. W.;
Baig, G. U.; Lunt, E.; Newton, C. G. J . Med. Chem. 1984, 27, 196-
201.
(12) Ege, G.; Gilbert, K. Tetrahedron Lett. 1979, 4253-4256.
(13) Threadgill, M. D. In The Chemistry of Antitumor Agents;
Wilman, D. E. V., Ed.; Blackie and Son: Bishopbriggs, Glasgow,
Scotland, 1990; pp 187-201.
S0022-3263(97)00721-4 CCC: $14.00 © 1997 American Chemical Society

Triazinines
J . Org. Chem., Vol. 62, No. 25, 1997 8661
Sch em e 2
Ta ble 1. Str u ctu r e a n d Abbr evia tion s of Com p ou n d s
Stu d ied
E-form found in acyclic triazenes), and therefore cyclic
triazenes present an opportunity to explore the chemistry
of the Z-triazene moiety.6,7 Triazolines, 4,5-dihydro-1,2,3-
triazoles, are perhaps the best known simple cyclic
triazenes. Their chemistry has been dominated by the
study of thermolysis reactions16 with several reports of
acid-induced decompositions, generally involving bi- and
tricyclic-1-aryltriazolines.17 The hydrolysis of triazolines
involves the protonation of the triazoline N(1),18 followed
by heterolytic cleavage of the N(1)-N(2) bond to generate
the 2-(methylamino)ethyldiazonium ion, which undergoes
further transformations to the final products, Scheme 2.
In the present work, we describe the synthesis of a new
class of cyclic triazenes,1-alkyltriazinines. In addition,
we report the results of an investigation of the products
and mechanism of their acid-catalyzed decomposition in
aqueous buffers. The compounds discussed in this paper
are listed in Table 1.
Resu lts
P r od u ct Stu d ies. The decomposition reaction of each
triazinine was followed periodically by 1H NMR analysis,
typically at times of 0, 0.5, 5, 24, 48, and 72 h. Product
yield data are recorded in Table 2 for the earliest time
at which all of the triazinine had disappeared, for 1 this
was 5 h (no further change in product distribution was
observed up to 72 h) and 5 h for 3. Values are reported
as a percentage of the total N-ethyl- or N-benzyl-
containing products, respectively. Each triazinine com-
pound contained a small amount of azimine. Because
this compound is stable to the reaction conditions and
does not appear to influence the reaction, it served as a
convenient internal standard for product peak integra-
tion. The identity of each product was determined by
the addition of standard compounds or by comparison
with spectra of authentic material.
for the lines in Figure 1 are as follows: 1, -0.833, 6.06;
3, -0.840, 5.56. The coefficient of determination (r2) are
0.996 and 0.999, respectively.
Rela tive Ra tes of Decom p osition of 1-Alk yltr ia z-
in in es. The rate of decomposition of the series of
1-alkyltriazinines was determined in pH 10.00, 0.10 M
lysine buffer at 25 °C. Added to this study was an acyclic
triazene, 1,3,3-triethyltriazene 7, and two cyclic triaz-
enes, 1-ethyl- and 1-benzyltriazolines 5 and 6, for com-
parison purposes. The rate constants obtained from
these studies are included in Table 4, which shows that
1-alkyltriazinines are approximately as reactive as the
analogous trialkyltriazenes but less reactive than the
analogous 1-alkyltriazolines under the reaction condi-
tions.
Solven t Deu ter iu m Isotop e Effect. The rates of
decomposition of all of the simple 1-alkyltriazinines were
measured in D2O buffers of pD equivalent to the pH
(10.00) of the analogous H2O buffers used in the relative
rate studies. In preparing the D2O buffers, the nominal
pH readings were 9.60, corrected according to the rela-
p H Dep en d en ce of th e Ra te of Tr ia zin in e Decom -
p osition . The rates of decomposition of representative
triazinines, 1-ethyl- and 1-benzyltriazinine, were deter-
mined in buffers over a range of pH levels. The data are
presented in Table 3. Plots of log kobs versus pH for these
data are displayed in Figure 1. The slopes and intercepts
(14) Baig, G. U.; Stevens, M. F. G. J . Chem. Soc., Perkin Trans. 1
1987, 665-670.
(15) Comis, R. L. Cancer Treat. Rep. 1976, 60, 165-176.
(16) Scheiner, P. Triazoline Decomposition In Thyagarajan, B. S.
Selective Organic Transformations; Wiley-Interscience: New York,
1970; Vol. 1, pp 327-362.
(17) Benson, F. R. Triazolines and Related Compounds In The High
Nitrogen Compounds; Wiley-Interscience: New York, 1984; pp 57-
82 and references therein.
(18) The numbering for cyclic triazenes is different than that for
the acyclic. The saturated nitrogen in cyclics is designated as N(1),
while in acyclics it is designated as the N(3) nitrogen.

8662 J . Org. Chem., Vol. 62, No. 25, 1997
Schmidt et al.
Ta ble 4. Solven t Isotop e Effects on th e Ra tea of
Ta ble 2. P r od u cts a n d Yield sa fr om th e Hyd r olysis of
1-Eth yl- a n d 1-Ben zyltr ia zin in e a t 25 °C in Aqu eou s
Bu ffer sb
Decom p osition of 1-Alk yltr ia zin in esb in Aqu eou s Bu ffer sc
comps
relative rated
kH
kD O
kH O/kD O
2O
2
2
2
1-ethyltriazininec
% product [5 h]d
1
2
3
4
5
6
7
1.00
0.49
0.25
0.27
2.70
0.97
1.02
6.02
2.94
1.53
1.63
16.18
5.86
6.13
6.74
3.38
1.86
1.92
27.90
0.89
0.87
0.82
0.85
0.58
3-(ethylamino)-n-propanol
N-ethyl allylamine
3-(ethylamino)-2-propanol
propionaldehydee
44.7
11.1
11.0
9.2
ethylamine
19.0
1-benzyltriazininec
% product [5 h]d
The rate constants (kobsd, s-1 × 103) are an average of at least
a
b
3-(benzylamino)-n-propanol
N-benzyl allylamine
3-(benzylamino)-2-propanol
propionaldehydee
56.2
13.3
16.8
12.3
24.3
two independent runs varying no more than (3%. Triazinine
initial concentration 3.0 × 10-5 M. c 0.1 M lysine buffer, pH 10.00,
0.25 M ionic strength (maintained with Na2SO4). Nominal pH
d
reading of 9.60 for D2O buffer. Rates based on the lead com-
benzylamine
pound, 1-ethyltriazinine, normalized to 1.00.
a
Yields were determined by 1H NMR analysis and are based
Ta ble 5. Tem p er a tu r e P r ofile for Ra tea of Hyd r olysis of
Cyclicb a n d Acyclic Tr ia zen esb in 0.1 M Lysin e Bu ffer sc
a n d Th eir Associa ted Activa tion P a r a m eter sd
on the percent of total N-Et or N-benzyl signals present in the
spectra as compared to the internal standard. The buffer was
b
25% (v/v) CD3CN in D2O containing 0.05 M sodium phosphate.
c The initial triazinine concentration was 0.5 M for each study.
compd
d
The value in the brackets represents the time at which all
temperature (°C)
1
5
7
8
triazinine was first noted to be absent from the 1H NMR and also
represents the time at which product yields were determined.
e Reported propionaldehyde yield is the total combined yield of
propionaldehyde and its hydrate, present in approximately equal
amounts.
20.0
22.5
25.0
27.5
30.0
3.82
4.51
5.36
6.04
7.03
11.12
12.93
14.67e
16.59e
18.46e
3.81
4.71
5.65
6.46e
7.55
0.11e
0.14e
0.17
0.20
0.24
Ta ble 3. p H P r ofile for Ra tea of Decom p osition of
1-Alk yltr ia zin in esb in 0.1 M Lysin e Bu ffer sc a t 25° C
activation parameters
∆Hq, kcal/mol
∆Sq, cal/°C mol
∆G,q kcal/molf
10.10
10.64
6.93
8.32
6.72
6.32
11.28
14.71
6.90
12.47
11.73
8.97
compd
pH
1
3
10.00
10.25
10.50
10.75
11.00
5.25
3.34
2.02
1.37
0.745
1.45
a
kobs (s-1) × 103 average of two independent runs (<3%
0.875
0.529
0.346
0.206
variability). Initial concentration was 3 × 10-5 M. c 0.25 M ionic
b
d
strength, maintained with added Na2SO4. Calculated by using
values of kobs × [H3O+]-1; [H3O+] ) 1.0 × 10-10. Least-squares
analysis of each line gave a correlation coefficient of a least 0.993.
e Average of three independent runs (<3% variability). f Calcula-
tions performed at 25 °C.
a
k (s-1) × 103, average of two independent runs (<3% vari-
ablity). Triazinine initial concentration was 3 × 10-5 M. c 0.25
b
M ionic strength, maintained with added Na2SO4.
somewhat with triazinine structure, but in all cases kH O
/
2
kD O was < 1.0. Ethyltriazoline 5, included in the study
2
as a control, produced a value, 0.58, comparable to that
previously reported, 0.62.7
Activa tion P a r a m eter s of 1-Alk yltr ia zen in es. The
dependence of the rate of decomposition upon tempera-
ture was determined for 1-ethyltriazinine 1 and, for
comparison purposes, three additional triazenes, 1-eth-
yltriazoline, 5, 1,3,3-triethyltriazene, 7, and 1-ethyl-3-
methyltriazene 8 [which exists as a mixture of two
tautomers5]. The data, shown in Table 5, were deter-
mined in a pH 10.0, 0.10 M lysine buffer. Values for kobs
were obtained at five different temperatures ranging
from 20 to 30 °C. Duplicate and, in some cases, triplicate
runs were performed. Adjustments for pH were made
by dividing the averaged kobs by the [H3O+] to give kspec
.
Plots of log [kspec/T (K)] versus 1/T (K) gave slopes and
intercepts used to calculate the activation parameters,
∆Hq (kcal/mol) and ∆Sq (cal/deg mol eu). The values
calculated are for 1, [10.10, 10.64]; for 5, [8.32, 6.72]; for
7, [11.28, 14.71]; and for 8, [12.47, 11.73], respectively.
The values for 8 are in good agreement with those
previously reported [12.67, 11.36].5 The ∆Hq values favor
the reactivity order: 5 > 1 > 7 > 8, but the ∆Sq values
appear to predict a different order.
F igu r e 1. Plots of log kobs vs pH for the decomposition of
1-alkyltriazinines in 0.1 M lysine buffers at 25 °C.
Discu ssion
tionship pD ) pHnominal - 0.4.19 The measured rates of
decomposition in D2O buffers and the calculated values
for the solvent deuterium isotope effect, kH O/kD O, are
We report herein the preparation of a new class of
cyclic triazenes, one in which the triazene moiety is
incorporated in a six-membered ring. By analogy with
the five-membered analogues, triazolines, we have chosen
the trivial name triazinine for the members of this class.
2
2
recorded in Table 4. The solvent isotope effects varied
(19) Glasoe, P. K; Long, F. A. Phys. Chem. 1960, 102, 188-190.

Triazinines
J . Org. Chem., Vol. 62, No. 25, 1997 8663
Sch em e 3
of cyclization and decomposition.] To study the mecha-
nism of cyclization, aliquots of the linear triazene solution
were concentrated and immediately taken up in various
amines and then cooled to -20 °C. These were compared
with concentrated linear triazene, which was cooled but
had no amine added. The reactions were followed by
TLC, which after several days indicated that all of the
amine-containing reactions formed substantial amounts
of the cyclic triazinine. For several amines, a small
amount of the corresponding azimine was observed.
These experiments showed that (1) all amines tested
[isopropyl, diisopropyl, diethyl, triethyl, and pyridine]
allowed the formation of triazinines, with minimal
amounts of decomposition, as compared to the reaction
with no amine present, and (2) concentration alone
caused initial cyclization, with subsequent decomposition
of the product triazinine and the starting linear triazene
to give final products which contain only a small amount
of the azimine and mainly non-triazene-containing hy-
drolysis products.
In principle, it might be expected that the chemistry
of 1-alkyltriazinines would be similar to that of the five-
membered analogues, 1-alkyltriazolines. Previously re-
ported work7 on the hydrolytic decomposition of triazo-
lines in the presence of acids found the major product to
be an N-alkylaziridine, the result of acid-catalyzed ring
opening followed by loss of molecular nitrogen and
reclosure of the ring, as shown in Scheme 2. Lesser
amounts of 2-(alkylamino)ethanols, acetaldehyde, and
alkylamines were also observed.
We had previously reported that the reaction of 1-azido-
3-chloropropane with alkyl Grignard reagents, followed
by the addition of isopropylamine, leads to the formation
of N-alkylazimines.8 In that publication we noted that
hydrolysis of the initial magnesium salt with aqueous
NH4OH/NH4Cl buffers permits the isolation of the linear
triazene, 1-alkyl-3-(3-chloropropyl)triazene.
We now report that, in the workup, replacement of
isopropylamine with Dowex or NH4OH/NH4Cl buffer at
0 °C leads to the immediate formation of the 1-alkyl-3-
(3-chloropropyl)triazenes. Upon concentration this linear
triazene undergoes self-catalyzed cyclization to triazinine,
with small amounts of azimine and much decomposition.
If, on the other hand, concentration is followed im-
mediately by the addition with cooling of an amine, a less
rapid, more controlled formation of the triazinine occurs
with little decomposition. Presumably, the amine sta-
bilizes the formation of 1-alkyltriazinines by neutralizing
the byproduct of the displacement reaction, HCl. In the
procedure described herein, the linear triazene was not
purified, but simply formed through Grignard hydrolysis
with Dowex-resin or NH4OH/NH4Cl buffers and im-
mediately reacted with isopropylamine. The presumed
mechanism for this reaction involves intramolecular
displacement of the chloro group by the more distant of
the three nitrogens of the triazene moiety (Scheme 3).
In this mechanism we have chosen to use the 3-alkyl-1-
(3-chloropropyl) tautomer of the linear triazene as the
reactant and its N3 as the nucleophile. It is also possible
that the reaction could occur by N1 of the 1-alkyl-3-(3-
chloropropyl) tautomer functioning as the nucleophile.
As mentioned above, for the purposes of preparing the
compounds, we did not isolate the linear triazene inter-
mediate but rather carried out the cyclization step as
rapidly as possible (to minimize degradation of the
intermediate and product triazenes by the HCl formed).
We did, however, want to better understand the mech-
anism for triazinine formation. This study required the
formation of the linear triazene, which was prepared as
described above but not concentrated. The resulting
solution was dried and kept cold. Determinations by
NMR and thin-layer chromatography showed the linear
triazene to be stable in solution for as long as the study
was carried out, ∼3 weeks. [Note: when kept in ether
or THF, the linear triazene was only stable when the
solution was very dry; absorbed moisture caused bubbling
and the appearance of a white precipatate, indications
The data presented in Table 2 show that the decom-
position of a representative 1-alkyltriazinine, 1, in 25%
acetonitrile-75% aqueous phosphate buffer leads to a
mixture of ring-opened products. At no time during the
decomposition did product studies show the presence of
either an azetidine or an aziridine. It had been shown
previously7 that under the same reaction conditions the
aziridine would be quite stable. As little as ∼2.5% yield
of aziridine could be detected by NMR. These facts
appear to rule out any aziridines as possible products of,
or intermediates in, the decomposition of 1-alkyltriaz-
inines. The major product, formed in 45% yield, is
3-(ethylamino)-1-propanol. In addition, lesser amounts
of 3-(ethylamino)-2-propanol, N-ethyl-2-propenamine, eth-
ylamine, and propionaldehyde are also formed. The
latter two products presumably result from the initial
formation and subsequent hydrolysis of N-propylidene-
ethanamine. A proposed mechanism to explain the
formation of the above products is shown in Scheme 4.
By analogy with 1-alkyltriazolines, protonation at N(1)
is the likely step that initiates decomposition. Heteroly-
sis of the N(1)-N(2) bond generates 3-(alkylamino)-
propyldiazonium ion, species I in Scheme 4, a common
intermediate for the formation of all of the observed
products. Nucleophilic displacement of molecular nitro-
gen followed by transfer of a proton to solvent would
account for the major product, 3-(alkylamino)-1-propanol.
Removal of a proton from carbon-2, concerted with loss
of nitrogen, leads to the formation of the elimination
product N-alkyl-2-propenamine. As has been observed
for the proteolysis of 1-alkyltriazolines7 and acyclic 1,3-
dialkyltriazenes,4 loss of nitrogen can also be accompa-
nied by a 1,2-hydride shift to form a secondary carboca-
tion, II. Reaction with water produces 3-(alkylamino)-
2-propanol. Alternatively, a second 1,2-hydride shift
would produce the stabilized cation III. Reaction of this
species with water would then account for the formation
of alkylamine and propionaldehyde. The yields of prod-

8664 J . Org. Chem., Vol. 62, No. 25, 1997
Schmidt et al.
Sch em e 4
benzyl. It is reasonable to assume that N(1) substituents
such as benzyl (and 3,3-diethoxypropyl) exert a rate-
retarding effect through inductive electron withdrawl,
which destabilizes the conjugate acid form of the sub-
strate. Given a specific acid-catalyzed mechanism, the
overall rate for substrate disappearance is (1/Ka)kdissoc
,
where Ka is the acidity of the conjugate acid form of the
neutral triazene and kdissoc is the rate of N(1)-N(2)
heterolysis. When the basicity of the neutral substrate
is reduced, these N(1) substituents retard the overall rate
of triazinine proteolysis. The order ethyl > butyl is
counter to the inductive effect and suggests that steric
inhibition reduces the ability of the solvent to stabilize
the conjugate acid of the substrate.
1-Alkyltriazinines decompose about as fast as the
analogous 1,3,3-trialkyltriazenes; compare the rates for
1 and 7 in Table 4. It can also be seen (Table 4) that
1-alkyltriazinines decompose approximately three to four
times slower (1-ethyl, 6.02 × 10-3 s-1; 1-benzyl, 1.53 ×
10-3 s-1) than the analogous 1-alkyltriazolines (1-ethyl,
1.62 × 10-2 s-1; 1-benzyl, 5.86 × 10-3 s-1). We believe
these factors to be the result of a rather complex
combination of structural effects on both Ka and kdissoc
.
The dependence of the rate of decomposition upon
temperature was determined for 1-ethyltriazinine 1,
1-ethyltriazoline 5, 1,3,3-triethyltriazene 7, and 1-ethyl-
3-methyltriazene 8 (data shown in Table 5). Several
conclusions can be extracted from the results: (1) the
form of the triazene [cyclic or acyclic] seems to have little
effect on stability of the triazene moiety, (2) the small
positive values for the entropy of activation implies little
ordering in the transition state, and (3) the enthalpy
order 5 > 1 g 7 > 8 suggests that the higher reactivity
of the triazoline is reflected in the enthalpy of activation,
rather than in the entropy. The behavior of triazinines
is very similar to that of acyclic trialkyltriazenes.
In summary, 1-alkyltriazinines can be prepared in good
yield by an amine-mediated cyclization of linear (3-
chloropropyl)triazenes, though the latter substances can-
not be isolated in pure form due to spontaneous cycliza-
tion accompanied by decomposition. The general pattern
by which 1-alkyltriazinines undergo hydrolysis in aque-
ous buffers involves a specific acid-catalyzed (A1) mech-
anism, rapid reversible protonation of N(1) followed by
rate-limiting heterolysis of the N(1)-N(2) bond. This is
very similar to the mechanisms previously reported for
acyclic triazenes3,4 and 1-alkyltriazolines.7 The key
intermediate generated is the 3-(alkylamino)propyldia-
zonium ion, a species which is then partitioned among
several different pathways, with direct displacement of
molecular nitrogen dominating. The rates of reaction,
due to the alkyl substituents, follow that of their closely
related cyclic analogues, 1-alkyltriazolines, as opposed
to their linear analogues.
ucts derived from the unrearranged (I) and rearranged
(II) cationic species favor the former, a pattern reminis-
cent of that found both in 1-alkyltriazolines7 and acyclic
1,3-dialkyltriazenes.4
The products of the acid-mediated decomposition of 3
in mixed acetonitrile-buffer solutions are in general
similar to those observed from 1. 1-Benzyltriazinine
gives rise to 3-(benzylamino)-1-propanol, 3-(benzylamino)-
2-propanol, benzylamine, and propionaldehyde. The
latter two products again suggest the formation and
subsequent hydrolysis of N-propylidenebenzylamine.
Overall, the decomposition of simple 1-alkyltriazinines
in aqueous buffers involves the formation of a 3-(alkyl-
amino)propyldiazonium ion and yields the related 3-(alkyl-
amino)-1-propanol as the major product. The nature of
the 1-alkyl group had very little influence on the product
distribution.
Previous studies of acyclic triazenes3,4 and 1-alkyltria-
zolines7 have shown that the general mode of acid-
catalyzed hydrolysis in aqueous buffers follows a specific
acid-catalyzed, A1, mechanism. This involves rapid
reversible protonation followed by rate-determining for-
mation of an alkyldiazonium ion.
The mechanism of the hydrolytic decomposition of
triazinines is acid-catalyzed. The data in Table 3 and
Figure 1 show a direct and linear dependence of the rate
of decomposition of 1 and 3 on the hydronium ion
concentration of the buffer. Consequently, the slopes of
the log(kobs) versus pH plots are near -1.0. For all
1-alkyltriazinines studied, hydrolysis is more rapid in
D2O than in H2O buffers (Table 4). The calculated
solvent deuterium isotope effect, kH O/kD O, varies from
Exp er im en ta l Section
Sa fety Note. Because acyclic triazenes are potent biologi-
cal alkylating agents, it is only prudent to assume that
1-alkyltriazinines are also potentially toxic and carcinogenic.
At all times, efficient hoods and protective clothing should be
used in working with these substances. Further, alkyl azides
are treacherously explosive and should be treated with ex-
treme caution. Wherever possible, these compounds should
only be handled in solution.
Ma ter ia ls. All chemicals were reagent grade (Aldrich
Chemical Co.) and were used as purchased without further
purification. The synthesis of the alkyl azide used in the
preparation of 1-alkyltriazinines has been reported previ-
ously.2,4,5 The preparation of buffers and the UV determina-
2
2
0.89 for 1 to 0.82 for 3. Taken together, these data again
point to a specific acid-catalyzed mechanism, A1, in which
rapid reversible protonation of the substrate occurs prior
to the rate-limiting N(1)-N(2) heterolysis step. [Note
that the numbering of the triazene moiety in cyclic
triazenes is the reverse of that in acyclic triazenes.]18
Rates of decomposition for various simple 1-alkyltri-
azinines, Table 4, follow the order ethyl > butyl > 3,3-
diethoxypropyl > benzyl. This pattern is similar to that
seen with 1-alkyltriazolines,7 ethyl > methyl > propyl >

Triazinines
J . Org. Chem., Vol. 62, No. 25, 1997 8665
tions for both chemical analysis and kinetic studies have been
reported.4-7 NMR spectra were obtained on a Varian XL-200
spectrometer. All samples submitted for exact mass determi-
nation were shown to be of >92% purity by 1H NMR analysis.
The major impurity in these samples was found to be the
corresponding azimine. Removal of final traces of this impu-
rity is extremely difficult due to the extreme sensitivity of
1-alkyltriazinines to hydrolytic decomposition.
addition of this cosolvent, however, permits the use of more
highly concentrated, and thus more easily analyzed, reaction
solutions. In a typical experiment, buffer plus acetonitrile-d3
was added to a weighed amount of the compound. The sample
was mixed thoroughly to ensure homogeneity, and 0.5 mL was
placed in a capped NMR tube. The reaction course was
followed7 until all of the starting triazinine had been decom-
posed. The initial triazinine concentration in each reaction
was 0.050 M. It was shown that after the triazinine was
completely decomposed, the pH had changed by no more than
(0.4 pH unit. Assignment of the NMR peaks arising from
the various products was made by comparison with authentic
samples and by coincidence of peaks upon the addition of
authentic materials. Yields were determined by comparative
integration of the product peaks. In the case of 1-benzyltri-
azinine, several products were confirmed by reports in the
literature.20,21
Kin etic Stu d ies. The method employed for the preparation
of the buffers used in these kinetic studies has been reported
previously.4-7 Rates of triazinine decomposition in aqueous
solution were followed spectrophotometrically.4-7 The reaction
solutions were contained in thermostated 1-cm cells, and the
temperature was held constant to within (0.1 °C. The
disappearance of each triazinine was followed by monitoring
the change in absorbance at its highest wavelength λmax (see
Experimental Section). In a typical kinetic run, the reaction
cuvette was charged with 1.341 mL of a 0.1 M lysine buffer
(ionic strength ) 0.25 M maintained with Na2SO4) and the
reaction was initiated by the addition of 9 µL of a 4.5 × 10-3
M solution of the 1-alkyltriazinine in acetonitrile; the initial
triazinine concentration was thus 3.0 × 10-5 M. The reference
cuvette contained 1.341 mL of buffer (the addition of 9 µL of
acetonitrile proved unnecessary). A minimum of 100 absor-
bance vs time readings were obtained over 3.5 half-lives. The
first-order rate constants were calculated from these data by
means of a computer program based on the Guggenheim
approximation least-squares method.22 The rate constants
were an average of at least two separate runs with standard
errors of less than 3%. In several cases, 100 absorbance vs
time readings were unable to be obtained; in those experiments
a minimum of at least three independent runs were used with
standard errors of less than 3%. Temperature studies were
performed in a similar fashion, using 1-cm thermostated cells
connected to a circulating waterbath maintaining constant
temperature to within (0.1 °C. Buffers used in these studies
were pH adjusted at the temperature at which a particular
rate was to be determined.
Because of the above-stated reactivity of 1-alkyltriazinines,
we determined that combustion analysis is not an appropriate
method of establishing purity or identity. An alternative
method, high-resolution mass spectrometric molecular formula
determination, was instead chosen. Exact mass measure-
ments were determined on either a VG 70-250 mass spectrom-
eter using a peak matching technique or on a ZAB-2F, using
fast atom bombardment (FAB) and peak matching. (See
Supporting Information for availability of these spectra.)
1-Eth yltr ia zin in e (1). To a 2.4 g (20 mmol) solution of
chloropropyl azide in 30 mL of dry THF was added, dropwise,
21 mmol of ethylmagnesium chloride under N2 at -30 °C. Over
the next 30-45 min the reaction warmed to -10 °C and was
hydrolyzed with 5 g of DOWEX (RG 501-X8, 20-50 mesh, fully
generated). The magnesium salt was precipitated with 30 mL
of diethyl ether and filtered, and the filtrate was concentrated
in vacuo. The yellowish oil was immediately diluted with 8.5
mL of isopropylamine and refrigerated overnight. The pre-
cipitate was filtered off and the filtrate concentrated. The
resultant oil was purified by column chromatography, using
basic Al2O3 eluted with (90:10) diethyl ether/isopropylamine
[Rf ) 0.7] to yield 735.9 mg (31%): UV (CH3CN) λmax 243 nm
1
(log ꢀ 3.90); H NMR (CDCl3, Me4Si) δ 1.25 (3H, t, J ) 7.21
Hz), 1.79 (2H, dt, J ) 5.84 Hz, J ) 6.60 Hz), 3.22 (2H, dt, J )
6.60 Hz, J ) 1.54 Hz), 3.47 (2H, dt, J ) 5.84 Hz, J ) 1.54 Hz),
3.61 (2H, q, J ) 7.21 Hz); proton-decoupled 13C NMR (CDCl3,
Me4Si) δ 13.0, 15.7, 43.3, 47.0, 50.9; exact mass calcd m/z for
M+, C5H11N3 113.0953, found 113.0977 (by EI).
P r ep a r a tion of Oth er Tr ia zin in es. The above reaction
procedure was used in the synthesis of each of the following
triazinines.
1-Bu tyltr ia zin in e (2) was obtained in 10% yield following
purification by column chromatography, using neutral alumina
eluted with (9:1:0.5) pentane/diethyl ether/isopropylamine [Rf
1
) 0.4]: UV (CH3CN) λmax 243 nm (log ꢀ 3.83); H NMR (500
MHz, CDCl3, Me4Si) δ 0.95 (3H, t, J ) 7.38 Hz), 1.37 (2H, m),
1.63 (2H, m), 1.79 (2H, dq, J ) 6.55 Hz, 5.87 Hz), 3.22 (2H, tt,
J ) 6.55 Hz, 1.51 Hz), 3.37 (2H, dt, J ) 5.87 Hz, 1.51 Hz),
3.56 (2H, t, 7.3 Hz); proton-decoupled 13C NMR (CDCl3, Me4-
Si) δ 13.8, 15.7, 19.8, 30.0, 44.0, 46.9, 56.2; exact mass calcd
m/z for MH+ C7H16N3 142.1344, found 142.1347 (by FAB).
1-Ben zyltr ia zin in e (3) was obtained in 47% yield following
purification by column chromatography, using neutral alumina
eluted with (9:1:0.5) pentane/diethyl ether/isopropylamine [Rf
) 0.3]: UV (CH3CN) λmax 244 nm (log ꢀ 4.02); 1H NMR (CDCl3,
Me4Si) δ 1.73 (2H, m), 3.11 (2H, tt, J ) 6.50 Hz, J ) 1.56 Hz),
3.48 (2H, tt, J ) 7.86 Hz, J ) 1.56 Hz), 4.77 (2H,s), 7.32 (5H,
m); proton-decoupled 13C NMR (CDCl3, Me4Si) δ 15.5, 43.3,
46.9, 60.3, 126.9, 127.9, 128.6, 136.9; mass calcd m/z for M+,
C10H13N3 175.1205, found 175.1109 (by EI).
Ack n ow led gm en t. Research sponsored by the Na-
tional Cancer Institute, DHHS, under contract with
A.B.L. and (R.H.S., in part) by grants from the National
Science Foundation (CHE-8521385 and CHE-8910890).
The contents of this publication do not necessarily
reflect the views or policies of the Department of Health
and Human Services, nor does mention of trade names,
commercial products, or organizations imply endorse-
ment by the U.S. Government. We are grateful to Mr.
J ohn Roman for providing mass spectrometry data, Mr.
J ohn Klose and Mrs. Mary McGuire for providing NMR
spectrometry data, and Ms. Lisa Taneyhill for assistance
with NMR data.
(3,3-Dieth oxyp r op yl)tr ia zin in e (4) was obtained in 47%
yield following purification by column chromatography, using
neutral alumina eluted with (7:5:0.5) pentane/diethyl ether/
isopropylamine [Rf ) 0.5]: UV (CH3CN) λmax 243 nm (log ꢀ
3.84); 1H NMR (CDCl3, Me4Si) δ 1.21 (3H, t, J ) 7.05 Hz),
1.79 (2H, dt, J ) 5.95 Hz, J ) 6.60 Hz), 2.0 (2H, dt, J ) 5.67
Hz, J ) 7.33 Hz), 3.23 (2H, tt, J ) 6.60 Hz, J ) 1.56 Hz), 3.47
(2H, tt, J ) 5.95 Hz, J ) 1.56 Hz), 3.58 (2H, q, J ) 7.05 Hz),
3.64 (2H, t, J ) 7.33 Hz), 4.58 (1H, t, J ) 5.59 Hz); proton-
decoupled 13C NMR (CDCl3, Me4Si) δ 15.3, 15.8, 32.4, 44.4,
47.1, 52.4, 61.6, 101.1; mass calcd m/z for MH+, C10H22N3O2
216.1712, found (relative intensity) 216 (by FAB).
P r od u ct Stu d ies. The products of hydrolytic decomposi-
tion were determined for 1-ethyl- and 1-benzyltriazinine by
carrying out reactions in solutions containing 25% (v/v) ac-
etonitrile-d3 dissolved in 0.05 M buffers of sodium phosphate
in D2O adjusted to the appropriate pH with a D2O solution of
NaOD. A previous study7 determined that the formation of
products was unaffected by the presence of acetonitrile. The
Su p p or tin g In for m a tion Ava ila ble: Supporting 1H NMR
and 13C NMR spectra (12 pages). This material is contained
in libraries on microfiche, immediately follows this article in
the microfilm version of the journal, and can be ordered from
the ACS; see any current masthead page for ordering
information.
J O970721P
(20) Pouchert, C. J . and Campbell, J . R. Aromatic alcohols, mer-
captans, sulfides, amines, nitro and nitroso compounds In The Aldrich
Library of NMR Spectra, Aldrich Chemical Co., Inc., 1974; Vol. 5, pp
41, and 96.
(21) Saavedra, J . E. J . Org. Chem. 1985, 50, 2271-2273.
(22) Guggenheim, E. A. Philos. Magn. 1926, 2, 538-543.