Reaction of 2,3-butanedione with propylenediamine observed by 1H-NMR spectroscopy

Article abstract of DOI:10.3987/COM-01-9385

2,3-Butanedione reacted with propylenediamine to form an initial intermediate: acetylated methyl-N-(3-aminopropyl) imine, which was a geometrical mixture of syn- and anti- isomers. The intermediate isomerized to anti- or syn-form via pyrimidine ring flip on heating. A steric factor of the isomers affected the formation of the final products. The mechanism based on NMR spectroscopy and molecular orbital calculation is proposed.

Full text of DOI:10.3987/COM-01-9385

HETEROCYCLES, Vol. 57, No. 1, 2002, pp. 11 - 16, Received, 5th November, 2001
REACTION OF 2,3-BUTANEDIONE WITH PROPYLENEDIAMINE
1
OBSERVED BY H-NMR SPECTROSCOPY
Tadatoshi Yamaguchi,a* Shigeru Ito,b Nobuko Mibu,c and Kunihiro Sumotoc
aDepartment of Hygiene, Miyazaki Medical College, Kiyotake-cho, Miyazaki 889-
1692, Japan, bInstitute of Biomaterials and Bioengineering, Tokyo Medical and Dental
c
University, Tokyo 101-0062, Japan, Faculty of Pharmaceutical Sciences, Fukuoka
University,Fukuoka 814-0180, Japan
Abstract - 2,3-Butanedione reacted with propylenediamine to form an initial
intermediate : acetylated methyl-N-(3-aminopropyl) imine, which was a geometrical
mixture of syn - and anti- isomers. The intermediate isomerized to anti- or syn-form
via pyrimidine ring flip on heating. A steric factor of the isomers affected the
formation of the final products. The mechanism based on NMR spectroscopy and
molecular orbital calculation is proposed.
Some dihydropyrazines (DHPs) have revealed DNA strand-breakage activity.1 In the course of our attempt to
synthesize new DHPs having this activity, a few interesting facts also have become apparent.2,3 The new
reaction products and the pathway of 2,3-butanedione (BD) with ethylenediamine (ED) were already described
in our previous paper.4 All products showed DNA strand-breakage activity1,4 by generating hydroxyl and
carbon-centered radicals.5 Therefore, we attempted to obtain certain products expected to be DNA strand-
breakable compounds from a reaction of BD with propylenediamine (PD) instead of ED. Although the reaction
of diketones with diamines is well known6 and a number of products is obtained,7 there has been no
discussion about an imine derivative in anti/syn-form as an intermediate as will be reported herein. In this
communication, we would like to show several products from a reaction of BD with PD and the reaction
pathways based on NMR spectroscopy and molecular orbital calculation.
It was interesting that 2-acetyl-2-methylhexahydropyrimidine (1) was one of the products from the reaction of
BD with PD,4 because 2-acetyl-2-methylimidazolidine, similar to 1, could not be isolated in a reaction with
ED, indicating the high reactivity. The other products except for 1 were investigated in the present study. The
successive procedures of the reaction are, first, a mixing of the starting substances in a solvent under cooling,
followed by stirring at room temperature, refluxing and then post-treatment. Under the several reaction
conditions (reaction temperature, reaction time, solvent, post-treatment, etc.), the main product changed into an
alternate compound. The high reactivity of the products should be predictable similarly to the case with ED.
* E-mail address: yamaguti@post1.miyazaki-med.ac.jp Fax +81-985-85-5177

Consequently, isolation of each product from the reaction mixture was expected to be very difficult, and
accordingly, NMR spectral analyses were attempted. Each product shown in Scheme 1 was independently
obtained from a separate reaction. The isolated yields of the products (1), 6,7-dihydro-2,3-dimethyl-5H-1,4-
diazepine8 (2) and 2,3,9,10-tetramethyl-1,4,8,11-tetraazacyclotetradeca-1,3,8,10-tetraene9 (3) were 89, 20,
and 6%, respectively. It is easily predicted that 2 is formed as the main product. However, the reaction mixture
in ether used as a solvent at the first trial gave 1 in high yield as the sole product. Since the preparation of 2
has not previously been described in the literature, it was particularly interesting to learn why 2 was not
obtained in this case. When the reaction mixture was refluxed in CHCl3 for prolonged time, the appearance of
2 increased sluggishly. Although the existence of 2 in more than 95% yield was detected in an NMR spectrum
of the reaction mixture which was refluxed for 7 days in CHCl3, the isolated yield (20%) was very low . One
reason is that significant amounts of the product were lost during distillation under high reduced pressure (1
mmHg) in the post-treatment.
HN
Me
Me
Me
O
Me
N
H2N
Me
Me
Me
Me
N
N
N
N
O
+
+
+
C
HN
H2N
N
Me
O
Me
1
2
PD
3
BD
Scheme 1
In the NMR spectra of the reaction mixture after addition of BD into a chloroform-d3 solution of PD under
cooling by ice bath, the mixture showed the signals due to the starting substances : BD and PD only, and then
began to represent signals of a product : acetyl, methyl, N-(3-aminopropyl)imine (st.1) as the sole product
under stirring for 10 min after removing the ice bath. The mixture at room temperature proceeded to give a
product (1), and the production of 1 further progressed under heating at 60 °C. Thus, the heating of the
reaction mixture for a short time affected the yield of 1. However, prolonged heating of the mixture showed
furthermore a conversion in NMR signals and, an additional product (2) appeared as shown in Scheme 5
below.
Me
N
Thus, the remarkable fact became
Me
H N
N
Me
c
heat
heat
clear that 1 changed into 2 when
isolated 1 was refluxed for a few
hours in CHCl3 ; in other words, a
O
C
H
2 N
C
H N
N
Me
Me
O
Me
2
anti/syn-st.1.
1
ring-expansion reaction proceeded
on heating (Scheme 2).
Scheme 2
A detailed study of the ring-expansion reaction showed that the initial transformation was ring-opening of 1 to
st.1 before the formation of 2, that is, st.1 was first produced. This indicated a route whereby 2 was
reconstructed via st.1 from 1 as shown in Scheme 2. It is very attractive that the proportion of anti- and syn-
forms10 in the reproduced st.1 was 1 : 1. On the other hand, st.1 resulting in the first stage (step 1) of the
reaction was shown in a proportion of anti- : syn-forms = 10 : 1 at 10 min when stirring of the reaction mixture

was started at room temperature after mixing of the starting substances under ice cooling. In another case
which was stirred for a few hours under continuous ice cooling, the syn-form was not detected. The
proportion underwent various changes in the progress of the reaction, while it is assumed that the steric factor
of the anti- and syn-forms affected the formation of alternate products. St.1 at syn-form can produce 1 and 2
without steric hindrance. Contrary to the syn-form, although the anti-form of st.1 is advantageous for
producing 1 to the syn-form of st.1 on inspection of the PM3 heats of formation (∆Hf) by energy difference
(49.2 - 48.0 = 1.2 kcal/mol) as mentioned below, st.1 at anti-form can not yield 2 since the N-atom of the
amino group can not achieve the distance to enable binding with the C-atom of the carbonyl group. Therefore,
it is inferred that the transformation of 1 to 2 under the heating condition requires a conversion of the anti-
form to syn-form of st.1. As an analogous case to st.1, acetyl, methyl, N-(3-dimethylaminopropyl)imine
(st.1’), which can not cyclize into a ring-closing compound such as 1 and 2, was examined to see if a mutual
conversion of anti- and syn-forms occurred by heating without conversion via a pyrimidine ring such as 1. It
was confirmed by NMR analysis that the anti-form of the imine (st.1’) resulting from a reaction of BD with
N,N-dimethyl-propylenediamine could not convert to syn-form by heating as shown in Scheme 3. This result
indicated that the conversion of the anti-form to syn-form in st.1 was impossible.
Me
O
N
Me
Me
Me
Me
O
N
N
N
X
N
Me
BD +
Me
Me
Me
NH
2
Me
st.1’ (anti-form)
st.1’ (syn-form)
Scheme 3
The slow-paced appearance of the syn-form of st.1 in the reaction mixture was recognized by detection of the
NMR signals and the process was postulated by PM3 calculation as mentioned below. The heating reaction of
1 in CHCl3 for prolonged time gave the ring-opened st.1, which contained nearly the same proportion of
--
anti- and syn-forms. This fact explained that the ring flip (1a < > 1b) of the pyrimidine ring in 1 achieved
the conversion of the anti-form to syn-form, which could be convertible to 2. The reaction system such as {
---
---
---
---
st.1 (anti-form) < >[ 1a < > 1b ] < > st.1 (syn-form) < > st.2} approached the reaction end to
yield 2 by the existence of an unreversible stage ( st. 2 ---> 2 ) as shown in Scheme 4. The PM3-calculated
heats of reaction (∆E) and heats of formation (∆Hf) support the preceding assumption. The ∆E of the
formation of each step is depicted in Scheme 4. It is estimated that the insignificant difference (0.6 kcal/mol)
between 56.5 kcal/mol for the formation st.1 (syn-form) from 1b and 57.1 kcal/mol for st.1 (anti-form)
from 1a, required a suitably prolonged heating. Also, the PM3 calculation indicated that the formation of 3
(∆Hf = 28.129 kcal/mol) from the syn-form of st.1 was more stable than that (∆Hf = 32.52 kcal/mol) from
the anti-form of st.1.
The total reaction pathway is expressed as shown in Scheme 5 involving the compound (2’), which was
identified by NMR spectra. The reaction proceeded at cool or room temperature, and on further refluxing for a

Ring inversion
H
N
H
N
Me
NH
O
Me
Me
NH
Me
O
DHf =-49.80 kcal/mol
DHf =-48.01 kcal/mol
1a
1b
56.5
48.0
57.1
49.2
Me
N
Me
N
NH2
X
H2N
BD + PD
O
Me
O
Me
DHf =-40.73 kcal/mol
DHf =-40.71 kcal/mol
st.1(anti-form)
st.1(syn-form)
48.23
X
45.8
N
Me
Me
Me
N
N
NH
Me
OH
st.2
2
DHf =-43.23 kcal/mol
Scheme 4
short time gave the sole product (1), and the reaction under refluxing for a prolonged time gave the plural
products, mainly 2 and small amounts of 2’ and 3. Using a solvent with higher boiling point increased the
formation of 3 and 2’.
Successive research for the synthesis of a new DNA strand-breakable compound is in progress. Compound
(2) was already recognized to be a DNA strand-breakage agent. The reactions of various diketones with PD
have been already performed and a number of novel products were obtained. The detailed data and biological
activity of the products will be described in a following paper.
ACKNOWLEDGMENT
This work was supported in part by a Grant-in-Aid for Scientific Research (B) from Japan Society for the
Promotion of Science (JSPS).

Me
Me
N
Me
Me
Me
Me
N
N
N
C
C
intramolecular
BD + PD
H2N
O
NH
OH
st.1.
2
st.2.
room
temperature
heat
Me
Me
Me
Me
Me
N
N
N
N
N
HN
Me
O
N
C
HN
H
Me
2'
3
1
Compounds (1, 2, and 3) were isolated. St.1 and 2' were identified by NMR spectra
Scheme 5
REFERENCES AND NOTES
1. T. Yamaguchi, N. Kashige, N. Mishiro, F. Miake, and K. Watanabe, Biol. Pharm. Bull., 1996, 19,
1261.
2. T. Yamaguchi, S. Ito. Y. Iwase, K.Watanabe, and K. Harano, Heterocycles, 1999, 51, 2305.
3. T. Yamaguchi, S. Ito. Y. Iwase, K.Watanabe, and K. Harano, Heterocycles, 2000, 53, 1677.
4. T. Yamaguchi, M. Eto, K. Harano, N. Kashige, K. Watanabe, and S. Ito, Tetrahedron, 1999, 55,
675.
5. T. Yamaguchi, S. Matsumoto, and K. Watanabe, Tetrahedron Lett., 1998, 39, 8311.
6. e.g. H. Stett, Chem. Ber., 1953, 86, 69; T. Ishiguro and M. Matsuura, Yakugaku Zasshi, 1958, 78, 229;
T. Ishiguro, M. Matsuura, and M. Awamura, Yakugaku Zasshi, 1958, 78, 571.
7. e.g. B. Fuches and A. Ellencweig, Recl. Trav. Chim. Pay-Bas, 1979, 98, 326; R. L. Willer and D. W.
Moor, J. Org. Chem., 1985, 50, 2365; R. L. Willer, D. W. Moor, and C. K. Lowe-Ma, J. Org. Chem.,
1985, 50, 2368.
8. bp 60-65°C (1 mmHg) by Kugel Role distillation Instrument equipped with a vacuum gauge : DVR2
(Vaccbrand).
Colorless odor oil. FAB-MS : m/z 249 (2M+H+, 24%), 125 (M+H+, 19%), 99 (100%). 1H (500 MHz) /
13C (125 MHz)-NMR (δ, CDCl3) : CH3- ; 2.08 / 22.90, -C-CH2-C- ; 2.18 / 32.14, =N-CH2- ; 3.21 /
47.91, >C=N- ; - / 168.60 ppm.
9. mp 66-67°C (ether). FAB-MS : m/z 249 (M+H+, 32%), 99 (100%). 1H (500 MHz) / 13C (125 MHz)-
NMR (δ, CDCl3) : CH3- ; 2.06 / 12.64, -C-CH2-C- ; 2.02-2.09 / 31.86, =N-CH2- ; 3.54 / 50.27, >C=N;
- / 168.15 ppm. The metal complex of 3 has been reported,11 however 3 as a metal-free compound was
not found in the literature.
10.1H-NMR (δ, 500 MHz, CDCl3) : anti-form [ 1.95 (d, J = 1.2 Hz, CH3-C=N-), 2.35 (s, CH3-CO-), 3.54

(dt, J =1.2, 6.7 Hz, >C=N-CH2-)] : syn-form [ 1.96 (d, J = 1.2 Hz, CH3-C=N-), 2.38 (s, CH3-CO-),
3.59 (dt, J = 1.2, 6.7 Hz, >C=N-CH2-)]. The NOE was observed between the methyl (CH3-C=N-) and
the methylene (>C=N-CH2-) signals of st.1 indicating an anti-form.
11. S. C. Jackels, K. Farmery, E. K. Barefield, N. J. Rose, and D. H. Busch, Inorg.Chem., 1972, 11,
2893.