EDC-mediated condensations of 1-chloro-5-hydrazino-9,10-anthracenedione, 1-hydrazino-9,10-anthracenedione, and the corresponding anthrapyrazoles

Article abstract of DOI:10.1016/j.tetlet.2004.04.041

The EDC-mediated condensation of 1-chloro-5-hydrazino-9,10-anthracenedione afforded an N-1 acyl anthrapyrazole instead of the expected hydrazide. The regiochemistry of the N-acyl substituent was assigned on the basis of an extensive set of NMR experiments, and identification of this isomer suggests a reaction sequence based on initial acylation and subsequent cyclization. In contrast, the parallel reaction of 1-hydrazino-9,10-anthracenedione proceeded to afford the expected hydrazide.

Full text of DOI:10.1016/j.tetlet.2004.04.041

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 4977–4980
EDC-mediated condensations of 1-chloro-5-hydrazino-9,10-
anthracenedione, 1-hydrazino-9,10-anthracenedione, and the
corresponding anthrapyrazoles
MeeKyoung Kim and David F. Wiemer*
Department of Chemistry, University of Iowa, Iowa City, IA 52242-1294, USA
Received 22 March 2004; accepted 8 April 2004
Available online 6 May 2004
Abstract—The EDC-mediated condensation of 1-chloro-5-hydrazino-9,10-anthracenedione afforded an N-1 acyl anthrapyrazole
instead of the expected hydrazide. The regiochemistry of the N-acyl substituent was assigned on the basis of an extensive set of
NMR experiments, and identification of this isomer suggests a reaction sequence based on initial acylation and subsequent cycli-
zation. In contrast, the parallel reaction of 1-hydrazino-9,10-anthracenedione proceeded to afford the expected hydrazide.
Ó 2004 Elsevier Ltd. All rights reserved.
The neuronal ceroid lipofuscinosis (NCLs), also known
as Batten disease, are a group of rare but devastating
diseases characterized by progressive neurodegenera-
tion.1 One form, known as late infantile neuronal ceroid
lipofuscinosis (or LINCL), is caused by inactivating
mutations in the lysosomal enzyme tripeptidyl protease I
(TPP-I).2;3 Current therapy for LINCL can only relieve
the seizure disorder and cannot cure the underlying
cause or slow the progression of the disease.1;2 Because
TPP-I is a lysosomal enzyme that can be secreted from
cells where it is over expressed and taken up by cells
where it is deficient, LINCL may be amenable to
development of gene-based therapies. However, inves-
tigations of potential gene therapies for this disease are
critically dependent upon advances in histochemical
methods for detection of TPP-I activity in tissue sec-
tions. In this context, we recently reported4 the prepa-
ration and partial characterization of several tripeptides
prepared from anthraquinone derivatives. Because these
compounds have been shown to be useful as histo-
chemical probes for TPP-I activity,4 their complete
characterization has become important. In this Letter,
we report the regiochemistry observed in EDC-mediated
coupling reactions with these substrates, and the spec-
troscopic studies that allow complete structural assign-
ments to the products.
According to Dikov et al.,5 1-chloro-5-hydrazino-9,10-
anthracenedione (1) was converted to the alanine
derivative 2 through DCC-mediated coupling with Boc-
L
-alanine, and the product was converted to the tri-
peptide derivative 3 through a sequence of coupling and
deprotection steps (Scheme 1). In our hands, reaction of
1,5-dichloro-9,10-anthracenedione (4) with hydrazine
smoothly affords the hydrazine derivative 1 (Scheme
2).6–10 However, attempted coupling of compound 1
with Boc-L-alanine in a DCC-mediated reaction gave a
coupled product only in low yield. The same compound
was obtained in much better yield when 1-(3-dimethyl-
aminopropyl)-3-ethylcarbodiimide hydrochloride (EDC)
was employed as the coupling agent. The carbodiimide
procedure for formation of the amide bond has been well
established,11 and DCC frequently has been used as a
coupling reagent. However, the resulting urea and
acylurea by-products may have solubility properties
similar to the peptides, so the desired products can be
hard to isolate.12 With the water soluble EDC reagent,
the by-products are readily removed by extraction with
water, which may explain the improved yield.13
The EDC-mediated condensation of the substituted
hydrazine 1 with Boc-L-alanine gave a single N-acylated
product as a dark red solid in 65% yield.7;9 The product
initially was assumed to be the amide 2 as previously
described.5 However careful analysis of the NMR
spectra did not support that assignment. Of special
significance is the 13C NMR data.14 The symmetrical
anthraquinone 4 shows a single resonance at 180.9 ppm
Keywords: Acylation; Anthrapyrazole; Hydrazide.
* Corresponding author. Tel.: +1-319-335-1365; fax: +1-319-335-1270;
e-mail: david-wiemer@uiowa.edu
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.04.041

4978
M. Kim, D. F. Wiemer / Tetrahedron Letters 45 (2004) 4977–4980
O
HBr.H2N
O
N
NH
NHBOC
O
O
HN
HN
O
NHNH2
O
NH
O
NH
(4 steps)
Cl
O
Cl
O
Cl
O
1
2
3
Scheme 1.
NHBOC
NHBOC
O
O
O
Cl
N
N
N
N
a
b
1
or
69%
65%
Cl
O
Cl
O
Cl
O
4
5
6
O
HBr.H2N
O
N
NH
O
N
N
Cl
Scheme 2. Reagents and conditions: (a) NH2NH2ÆH2O, Pyr, reflux, 3 h; (b) Boc-
O
7
L
-alanine, EDC, HOBt, NMM, THF.
that can be assigned to the two quinone carbonyl car-
bons. In the hydrazine derivative 1 this symmetry is lost,
and two distinct quinone resonances are observed at
183.8 and 182.5 ppm. In contrast, in the product of the
by condensation, but this would require addition of the
hydrazide nitrogen to the quinone carbonyl. On the
other hand, if cyclization precedes ring closure, forma-
tion of either isomer 5 or 6 may be possible.15
DCC and EDC condensations with Boc-L-alanine, only
one quinone-type resonance is apparent (181.1 ppm)
and, because the expected total of 16 aromatic or car-
bonyl resonances is observed, it is not possible that an
accidental overlap of two resonances has occurred. In-
stead it appears probable that cyclization to a pyrazole
derivative occurred under these conditions, as previ-
ously observed with alkyl-substituted hydrazines derived
from related anthraquinones.9 It was not immediately
clear whether this product was the N-1 acyl pyrazole 5
or the isomeric pyrazole 6 and there is precedent for
both regioisomers.15 Formation of compound 5 could be
envisioned through a reaction sequence where acylation
of the more nucleophilic terminal nitrogen was followed
The condensation product was assigned structure 5
through extensive analysis of its spectral data, and
especially the HMBC, HMQC, and NOESY data sum-
marized in Figure 1. A key HMBC correlation between
the carbonyl carbon observed at 181.1 ppm and the H-5
resonance at 8.04 ppm allowed assignment of hydrogens
and carbons on the C-ring through HMQC and COSY
experiments. Two carbons adjacent to the pyrazole
nitrogens were assigned by interpretation of a correla-
tion between the C-8 carbon resonance at 137.9 ppm and
the C-6 hydrogen observed at 7.77 ppm, and a second
correlation between the C-9 carbon at 144.1 ppm and
the C-1 hydrogen at 8.24 ppm.

M. Kim, D. F. Wiemer / Tetrahedron Letters 45 (2004) 4977–4980
4979
ded only recovered starting material; none of the pyr-
azole 8 was detected under these conditions. Taken
together, these experiments suggest that compound 5 is
formed from the substituted hydrazine 1 through a
reaction sequence where N-acylation of the terminal
nitrogen is followed by formation of the pyrazole ring
system.
Parallel studies with the related anthraquinone 9 have
shown that relatively brief treatment with hydrazine
affords the aryl hydrazine derivative 10 (Scheme 4), and
longer reaction time gives the anthrapyrazole 13
(Scheme 5). Treatment of compound 10 with EDC and
Figure 1. (A) Key HMQC ( ) and HMBC ( ) correlations and (B)
NOESY correlations for compound 5.
Boc-L-alanine clearly gave compound 11 rather than a
pyrazole product, as evidenced by observation of two
quinone carbon resonances at 185.3 and 182.9 ppm.16
The parallel reaction with a protected serine provides
the analogous product 12.
The final assignment of the product as structure 5 also
was supported by a combination of COSY and NOESY
experiments. For example, compound 5 showed strong
NOESY correlations between both the a hydrogen of
the alanine moiety (5.74 ppm) and the alanine methyl
group (1.65 ppm) and the C-1 anthraquinone hydrogen
(8.24 ppm, Fig. 1). This strong correlation provides clear
support for assignment of the structure as presented in
Figure 1, especially when considered with the energy
minimized structure as shown (PM3 level).
N-Acylation of compound 13 under the reaction condi-
tions employed with compound 8 afforded the pyrazole
derivative 14 in 68% yield (Scheme 5). Assignment of a
pyrazole product was relatively straightforward based on
observation of a single quinone carbonyl resonance at
182.2 ppm. The regiochemistry of this reaction was
established through extensive analysis of the NMR spec-
tra. As shown in Figure 2, key HMBC correlations were
observed between the C-4 hydrogen at 8.32ppm, as well as
the C-5 hydrogen at 8.03ppm, and the carbonyl carbon at
182.2 ppm. These correlations allowed assignment of
hydrogens and carbons in the A- and C-rings through
HMQC and COSY data. The A-ring has four aromatic
hydrogens and the C-ring has just three, so COSY and
HMQC correlations readily allowed the assignment of all
hydrogens and carbons in these two ring systems. The two
aromatic carbons adjacent to the pyrazole nitrogens were
assigned by interpretation of a correlation between the C-
8 carbon at 138.0 ppm and the C-6 hydrogen at 7.72ppm
and a second correlation between the C-9 carbon at
144.4 ppm and the C-1 hydrogen at 8.08ppm.
Because Dikov’s report5 contained no spectral data and
only a very brief summary of the reactions employed, it
is not possible to establish whether their reported bio-
logical evaluations were based on a tripeptide derived
from a quinone or a pyrazole. In our study,4 the pyr-
azole derivative 5 was carried on to the corresponding
tripeptide 7 through standard methods (Scheme 2), and
shown to be a useful probe for TPP-I activity.4
Several experiments were conducted to illuminate the
reaction sequence that leads to compound 5. Extended
heating of compound 4 with hydrazine in the presence of
iPr2NEt leads to the pyrazole 8 (Scheme 3).7 When
compound 8 was treated with Boc-L-alanine under
The final structure assignment of compound 14 was
based on a combination of COSY and NOESY experi-
ments. Compound 14 showed strong NOESY correla-
tions between the a hydrogen at 5.78 ppm and one of the
–CH2O– hydrogens (at 4.20 ppm) and the C-1 hydrogen
of the aromatic ring (Fig. 2). These correlations clearly
support assignment as the N-1 derivative shown rather
than the N-2 isomer.
standard coupling conditions followed by an aqueous
work-up, the major product could be assigned as the
amide 2 given that two quinone carbonyl resonances
were observed in the 13C NMR spectrum (183.2 and
181.4 ppm). The minor product of this reaction proved
to be identical to the material prepared earlier (i.e.,
pyrazole 5). Compound 1 also was treated with a car-
bodiimide (DCC or EDC), N-methylmorpholine
(NMM), and 1-hydroxylbenzotriazole (HOBt) in THF
In summary, the EDC-mediated coupling of the chloro-
anthraquinone derivative 1 with Boc-L-alanine proceeds
in the absence of Boc-L-alanine. These reactions affor-
to afford the N-1 acyl pyrazole 5, while the parallel
EDC-mediated condensation of pyrazole 8 followed by
work-up with aqueous acid gives the ring-opened
product 2. In the absence of the chloro substituent,
different reactivity is observed. The EDC-mediated
condensation of hydrazine 10 proceeds without pyrazole
formation and gives the hydrazide 11, and condensation
of the pyrazole 13 gives the N-1 acyl pyrazole 14. While
determination of the ways that the presence or absence
of a remote chloro substituent influences these reactions
will require further investigation, this work establishes
N
NH
a
b
4
2
51%
5
2%
+
Cl
O
8
Scheme 3. Reagents and conditions: (a) NH2NH2ÆH2O, iPr2NEt,
reflux; (b) Boc- -alanine, EDC, HOBt, NMM, THF.
L

4980
M. Kim, D. F. Wiemer / Tetrahedron Letters 45 (2004) 4977–4980
NHBOC
O
R
HN
O
Cl
O
NHNH2
NH
O
a
b or c
67%
O
O
O
9
10
11 R = CH3 (98%)
12 R = CH2OCH2C6H5 (94%)
Scheme 4. Reagents and conditions: (a) NH2NH2ÆH2O, Pyr, reflux, 3 h; (b) Boc-L-alanine, EDC, HOBt, NMM, THF; (c) Boc-O-benzyl-L-serine,
EDC, HOBt, NMM, THF.
2. Sleat, D. E.; Donnelly, R. J.; Lackland, H.; Liu, C.-G.;
Sohar, I.; Pullarkat, R. K.; Lobel, P. Science 1997, 277,
1802–1805.
3. Vines, D. J.; Warburton, M. J. FEBS Lett. 1999, 443, 131–
135.
NHBOC
O
N
NH
OBn
N
N
b
a
4. Kim, M. K.; Mao, Q.; Davidson, B. L.; Wiemer, D. F.
J. Med. Chem. 2003, 46, 1603–1608.
9
68%
99%
5. Dikov, A.; Dimitrova, M.; Ivanov, I.; Krieg, R.; Halbh-
uber, K.-J. J. Cell. Mol. Biol. 2000, 46, 1219–1225.
6. Mohlau, R. Chem. Ber. 1912, 45, 2233–2247.
7. Bradley, W.; Geddes, K. W. J. Chem. Soc. 1952, 1630–
1636.
O
13
O
14
Scheme 5. Reagents and conditions: (a) NH2NH2ÆH2O, Pyr, reflux,
9 h; (b) Boc-O-benzyl- -serine, EDC, HOBt, NMM, THF.
8. Showalter, H. D. H.; Johnson, J. L.; Hoftiezer, J. M.
J. Heterocyclic Chem. 1986, 23, 1491–1501.
L
9. Krapcho, A. P.; Avery, K. L., Jr. J. Org. Chem. 1988, 53,
5927–5929.
10. Zhang, L. H.; Meier, W. E.; Watson, E. J.; Gibson, E. P.
Tetrahedron Lett. 1994, 35, 3675–3678.
11. Goodman, M.; Kenner, G. W. In Advances in Protein
Chemistry; Academic: New York, 1957; Vol. 12, pp 488–
638.
12. Sheehan, J. C.; Cruickshank, P. A.; Boshart, G. L. J. Org.
Chem. 1961, 26, 2525–2528.
13. Xu, Y.; Miller, M. J. J. Org. Chem. 1998, 63, 4314–4322.
14. For compound 5: 1H NMR (CDCl3) d 8.39 (d, J ¼ 8:1 Hz,
1H), 8.24 (dd, J ¼ 6:6, 2.2 Hz, 1H), 8.04 (d, J ¼ 7:3 Hz,
1H), 7.77 (t, J ¼ 7:8 Hz, 1H), 7.64–7.57 (m, 2H), 5.76–5.66
(m, 1H), 5.44 (d, J ¼ 8:1 Hz, 1H), 1.65 (d, J ¼ 6:9 Hz, 3H),
1.49 (s, 9H); 13C NMR (CDCl3) d 181.1, 173.6, 155.1,
144.1, 137.9, 137.3, 135.0, 133.2, 132.2, 131.6, 129.6, 126.4,
124.6, 123.6, 123.1, 120.7, 80.1, 49.2, 28.2 (3C), 19.2;
HRMS (FAB) calcd for C22H21N3O4Cl (M+H)þ 426.1221,
found 426.1226.
Figure 2. (A) Key HMQC ( ) and HMBC ( ) correlations and (B)
NOESY correlations for compound 14.
the structures of the condensation products and will
advance their use as probes of TPP-1 distribution fol-
lowing gene-transfer experiments.
15. Oliva, A.; Ellis, M.; Fiocchi, L.; Menta, E.; Krapcho, A. P.
J. Heterocyclic Chem. 2000, 37, 47–55.
16. For compound 11: 1H NMR (CDCl3) d 10.54 (s, 1H,
exchanges with D2O), 8.31–8.21 (m, 3H), 7.78–7.71 (m,
3H), 7.42–7.30 (m, 5H), 7.23 (d, J ¼ 8:5 Hz, 1H), 5.46
(d, J ¼ 6:7 Hz, 1H), 4.64 (d, J ¼ 11:7 Hz, 1H), 4.59
(d, J ¼ 11:7 Hz, 1H), 4.51–4.48 (m, 1H), 4.00 (dd,
J ¼ 8:8, 3.6 Hz, 1H), 3.70 (dd, J ¼ 9:0, 6.4 Hz, 1H), 1.49
(s, 9H); 13C NMR (CDCl3) d 185.3, 182.9, 170.4, 155.5,
151.0, 137.1, 135.1, 134.1, 134.0, 133.8, 133.3, 132.6, 128.5
(2C), 128.0, 127.9 (2C), 126.8, 126.6, 118.4, 118.0, 114.4,
80.7, 73.7, 69.7, 53.4, 28.3 (3C); HRMS (FAB) calcd for
C29H30N3O6 (M+H)þ 516.2135, found 516.2129.
Acknowledgements
Financial support from the Batten Disease Support and
Research Association is gratefully acknowledged.
References and notes
1. Bennett, M. J.; Hofmann, S. L. J. Inherit. Metab. Dis.
1999, 22, 535.