Squaric Acid N-Hydroxylamides: Synthesis, Structure, and Properties of Vinylogous Hydroxamic Acid Analogues

Article abstract of DOI:10.1021/jo035175g

The synthesis of squaric acid N-hydroxylamide esters 5 and amides 6 from dimethyl squarate 2a is described. These derivatives are analogues of the naturally occurring iron(III) chelator hydroxamic acid. On the basis of a comparative reactivity study, a co

Full text of DOI:10.1021/jo035175g

Squ a r ic Acid N-Hyd r oxyla m id es: Syn th esis, Str u ctu r e, a n d
P r op er ties of Vin ylogou s Hyd r oxa m ic Acid An a logu es
Nathaniel C. Lim,† Martha D. Morton,† Hilary A. J enkins,‡ and Christian Bru¨ckner*,†
Department of Chemistry, University of Connecticut, Storrs, Connecticut 06269-3060, and
Department of Chemistry, X-ray Crystallography Laboratory, Saint Mary's University,
Halifax, Nova Scotia B3H 3C3, Canada
c.bruckner@chem.uconn.edu
Received August 11, 2003
The synthesis of squaric acid N-hydroxylamide esters 5 and amides 6 from dimethyl squarate 2a
is described. These derivatives are analogues of the naturally occurring iron(III) chelator hydroxamic
acid. On the basis of a comparative reactivity study, a concerted retro-Cope mechanism for the
formation of the N-hydroxylamide esters 5 by reaction of dimethyl squarate with hydroxylamines
is proposed. A preliminary iron(III) binding study of these hydroxamic acid analogues is presented,
demonstrating binding of iron(III) to amides 6 in aqueous solutions, while the esters 5 did not
show any sign of metal ion binding. 13C NMR spectroscopic data (chemical shift and spin-lattice
relaxation time determination) of these and related derivatives delineate the resonance structures
predominant in these molecules. The resonance structures of the derivatives rationalize their
spectroscopic data, chemical reactivity, and iron(III) binding properties. Single-crystal X-ray
structure analyses of squaric acid N-hydroxylamide ester 5b and squaric acid N-hydroxylamide
amide 6c confirm their connectivity and provide structural evidence supporting the spectroscopically
derived conclusions. The squaric acid N-hydroxylamides are potentially useful in the construction
of chemosensors for iron(III).
SCHEME 1
In tr od u ction
Since the first synthesis of squaric acid (1) in 1959,1 a
range of squaric acid derivatives have been reported.2-5
In particular, the chemistry of squaric acid diesters 2,
monoamide monoesters 3 and diamides 4 is well devel-
oped (Scheme 1).5 Nucleophilic substitution of the di-
esters 2 with 1 equiv of amine generates the monoamide
monoesters 3.5 Further reaction of this half-ester with a
second equivalent amine forms the corresponding dia-
mides 4.5 This controlled, stepwise nucleophilic substitu-
tion reaction forms the basis for the use of diethyl
squarate as a coupling reagent to, for instance, conjugate
oligosaccharides to proteins or polyazamacrocycles.6
Squarate amides have also found use as anion recognition
systems.7 A number of pharmaceutical applications of
* Corresponding author. Phone: (860) 486-2743. Fax: (860) 486-
2981.
† University of Connecticut.
‡ Saint Mary’s University.
(1) Cohen, S.; Lacher, J . R.; Park, J . D. J . Am. Chem. Soc. 1959,
81, 3480.
(2) (a) Schmidt, A. H.; Ried, W. Synthesis 1978, 1-22. (b) Knorr,
H.; Ried, W. Synthesis 1978, 649.
squaric acid diesters are known. For instance, the di-n-
butylester is a potent allergen and has been used in the
treatment of alopecia areata, a nonscarring form of hair
(3) Schmidt, A. H.; Ried, W. Synthesis 1978, 869-880.
(4) West, R. Oxocarbons; Academic Press, Inc.: New York, 1980.
(5) Schmidt, A. H. Synthesis 1980, 961-994.
(6) (a) Vermeer, H. J .; Halkes, K. M.; van Kuik, A.; Kamerling, J .
P.; Vliegenthart, J . F. G. J . Chem. Soc., Perkin Trans. 1 2000, 2249-
2263. (b) Corsi, D. M.; Elst, L. V.; Muller, R. N.; van Bekkum, H.;
Peters, J . A. Chem. Eur. J . 2001, 7, 64-71. (c) Kitov, P. I.; Bundle, D.
R. J . Chem. Soc., Perkin Trans. 1 2001, 838-853. (d) Chernyak, A.;
Karavanov, A.; Ogawa, Y.; Kovac, P. Carbohydr. Res. 2001, 330, 479-
486.
loss, and in immunotherapy for warts in children.8,9
A
number of other squaric acid amide-based pharmaceuti-
cally active compounds are being studied.10 Squarate
diamides have also been used in the construction of chiral
auxilaries.11
10.1021/jo035175g CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/25/2003
J . Org. Chem. 2003, 68, 9233-9241
9233

Lim et al.
Our interest lies in the design of chemosensors for
metals of biological interest.12 Chemosensors are mol-
ecules that translate chemical information, such as the
presence of a certain metal ion, into an analytically useful
signal.13 Thus, a molecular recognition group is coupled
to a reporter group, and an efficient signal transduction
between these two moieties has to take place. In squaric
acid 1, two vinylogous acid moieties are “cross-conju-
gated”, enabling a tight electronic coupling between them.
This feature makes the squaric acid moiety attractive as
a linker between a metal recognition site and the reporter
group in the construction of chemosensors. A further
advantage is conceivably derived from involving the
squarate oxygens in the metal binding. Early reports on
the coordination chemistry of the squarate dianion
proposed an η2-coordination mode (1M);14 however, later
papers showed that the bite angle imposed by the
cyclobutene framework is too large to sustain this
coordination mode with first and second row transition
metals (bite distance of 3.3 vs 2.6 Å for the chelating
anion oxalate). Instead, only η1-coordination such as the
1,2-bismonodentate coordination mode (1M2) is ob-
served.15 Monodentate binding is less strong and specific
than chelate formation, and therefore to design a metal-
specific sensor incorporating a squaric acid moiety, we
had to redesign the metal binding motif.
The anions of hydroxamic acids are known to form
chelates with iron(III) with high specificity,16 and these
iron(III) hydroxamato complexes can possess great ther-
modynamic stability. This property is utilized by many
microorganisms to sequester iron by secreting hydrox-
amic acid-based siderophores.17,18 The squaric acid N-
hydroxylamide ester derivative 5 is a vinylogous hydrox-
amic acid (Scheme 1). Molecular modeling suggests that
the anion of this hydroxamic acid analogue possesses the
proper metric requirements for chelation, and it forms a
six-membered metallacycle 5M (bite distance between the
oxygens 2.8 Å).19 A range of N-hydroxylamide methyl-
esters of type 5 were first reported by Zinner and
Gru¨nefeld and tested for their iron(III)-binding proper-
ties.20 Surprisingly, however, chelation was only observed
under anhydrous conditions in aprotic solvents, suggest-
ing hydrolytic lability (i.e., low stability) for these com-
plexes. The authors also prepared the corresponding
N-hydroxylamide amides 6 but did not report on their
chelation behavior, and they reported an unexpected
inability to prepare squarate bis-N-hydroxylamides 7.
In our search for novel ways to bring a hydroxamate-
type chelating moiety into close electronic contact with,
for instance, a fluorophore for the construction of
chemosensors for iron(III), we decided to reinvestigate
the synthesis and chelation properties of the N-hydroxyl-
amide esters and amides. In this process, we fine-tuned
and expanded the synthetic methodology toward 5 and
6 and related compounds. We unveiled the reason bishy-
droxamates 7 could not be prepared by nucleophilic
displacement of diester 2. 13C NMR spectroscopy of the
squaric acid derivatives gains a particularly important
role in their characterization because of the paucity of
nonexchangeable or diagnostic protons in these struc-
tures, but the carbons of the cyclobutenedione framework
of many derivatives were characterized by unusually long
13C relaxation times. We determined the spin-lattice
relaxation time T1 of the squarate derivatives. This
investigation also provided spectroscopic evidence for the
degree by which resonance structures determine their
differing coordination properties. The results of the
synthesis and spectroscopic and structural investigations
of N-alkylhydroxylamide squarate esters 5, amides 6, and
related compounds are reported here.
(7) (a) Tomas, S.; Rotger, M. C.; Gonzalez, J . F.; Deya, P. M.;
Ballester, P.; Costa, A. Tetrahedron Lett. 1995, 36, 2523-2526. (b)
Prohens, R.; Tomas, S.; Morey, J .; Deya, P. M.; Ballester, P.; Costa, A.
Tetrahedron Lett. 1998, 39, 1063-1066. (c) Tomas, S.; Prohens, R.;
Deslongchamps, G.; Ballester, P.; Costa, A. Angew. Chem., Int. Ed.
1999, 38, 2208-2211. (d) Prohens, R.; Martorell, G.; Ballester, P.;
Costa, A. Chem. Commun. 2001, 1456-1457. (e) Prohens, R.; Rotger,
M. C.; Pin˜a, M. N.; Deya, P. M.; Morey, J .; Ballester, P.; Costa, A.
Tetrahedron Lett. 2001, 42, 4933-4936.
(8) Tietze, L. F.; Arlt, M.; Beller, M.; Glusenkamp, K. H.; J ahde, E.;
Rajewsky, M. F. Chem. Ber. 1991, 124, 1215-1221.
(9) (a) Gardner, S. H.; Freyschmidt-paul, P.; Hoffman, R.; Sundberg,
J . P.; Happle, R.; Nigel, J .; Tobin, D. J . Eur. J . Dermatol. 2000, 10,
443-450. (b) Freyschmidt-paul, P.; Sundberg, J . P.; Happle, R.;
McElwee, K. J .; Metz, S.; Boggess, D.; Hoffman, R. J . Investigative
Dermatol. 1999, 113, 61-68. (c) Kazumasa, M.; Motonobu, N.; Miyako,
N.; Tatsuo, K.; Yoshiki, M. J . Dermatol. 2002, 29, 661-664. (d)
Silverberg, N. B.; Lim, J . K.; Paller, A. S.; Mancini, A. J . J . Am. Acad.
Dermatol. 2000, 42, 803-808.
(10) (a) Bang-Andersen, B.; Ahmadian, H.; Lenz, S. M.; Stensbøl,
T. B.; Madsen, U.; Bøgesø, K. P.; Krogsgaard-Larsen, P. J . Med. Chem.
2000, 43, 4910-4918. (b) Butera, J . A.; Antane, M. M.; Antane, S. A.;
Angentieri, T. M.; Freeden, C.; Graceffa, R. F.; Hirth, B. H.; J enkins,
D.; Lennox, J . R.; Matelan, E.; Norton, N. W.; Quagliato, D.; Sheldon,
J . H.; Spinelli, W.; Warga, D.; Wojdan, A.; Woods, M. J . Med. Chem.
2000, 43, 1187-1202. (c) Gilbert, A. M.; Antane, M. M.; Angentieri, T.
M.; Butera, J . A.; Francisco, G. D.; Freeden, C.; Gundersen, E. G.;
Graceffa, R. F.; Herbst, D.; Hirth, B. H.; Lennox, J . R.; McFarlane, G.;
Norton, N. W.; Quagliato, D.; Sheldon, J . H.; Warga, D.; Wojdan, A.;
Woods, M. J . Med. Chem. 2000, 43, 1203-1214.
Resu lts a n d Discu sssion
Syn th esis of Squ a r ic Acid Hyd r oxa m a te Ester s
5 a n d Am id es 6. The ultimate starting material for all
our syntheses was 3,4-dimethoxycyclobut-3-ene-1,2-dione
(2a ), prepared from disilver(I) squarate and methyl iodide
following a procedure similar to a Williamson ether
synthesis.21 We chose the solid dimethyl ester 2a over
(11) Zhang, J .; Zhou, H.-B.; Lu¨, S.-M.; Luo, M.-M.; Xie, R. G.; Choi,
M. C. K.; Zhou, Z.-Y.; Chan, S. C.; Yang, T.-K. Tetrahedron: Asymmetry
2001, 12, 1907-1912.
(12) Lim, N. C.; Yao, L.; Freake, H. C.; Bru¨ckner, C. Bioorg. Med.
Chem. Lett. 2003, 13/ 14, 2251-2254.
(13) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J . M.; McCoy, C. P.; Rademacher, J . T.; Rice, T. E. Chem. Rev. 1997,
97, 1515-1566.
(16) Mehrota, R. C. In Comprehensive Coordination Chemistry;
Wilkinson, G. G., R. D., McCleverty, J . A., Eds.; Pergamon Press:
Oxford, 1987; Vol. 2, pp 505-514.
(17) Raymond, K. N.; Mu¨ller, G.; Matzanke, B. F. Topics Curr. Chem.
1984, 123, 49-102.
(18) Miller, M. J . Chem. Rev. 1989, 89, 1563-1579.
(19) Quantum CAChe 4.9, MM3 force field; Fujitsu Limited, 2002.
(20) Zinner, G.; Gru¨nefeld, J . Arch. Pharm. (Weinheim, Ger.) 1985,
318, 977-983.
(21) (a) Cohen, S.; Cohen, S. G. J . Am. Chem. Soc. 1966, 88, 1533-
1536. (b) Alternative methods for the synthesis of squarate esters are
available; see e.g.: Liu, H.; Tomooka, C. S.; Moore, H. W. Synth.
Commun. 1997, 27, 2177-2180.
(14) (a) West, R.; Niu, H. Y. J . Am. Chem. Soc. 1963, 85, 2589-
2590. (b) Ludi, A.; Schindler, P. Angew. Chem., Int. Ed. Engl. 1968, 7,
638.
(15) See e.g.: (a) Habenschuss, M.; Gerstein, B. C. J . Chem. Phys.
1973, 61, 852-860. (b) Weiss, A.; Riegler, E.; Robl, C. Z. Naturforsch.,
B: Chem. Sci. 1986, 41, 1329-1332. (c) Weiss, A.; Riegler, E.; Robl,
C. Z. Naturforsch., B: Chem. Sci. 1986, 41, 1332-1336. (d) Solans, X.;
Aguilo, M.; Gleizes, A.; Faus, J .; J ulve, M.; Verdaguer, M. Inorg. Chem.
1990, 29, 775-784. (e) Khan, M. I.; Chang, Y.-D.; Chen, Q.; Salta, J .;
Lee, Y.-S.; O’Connor, C. J .; Zubieta, J . Inorg. Chem. 1994, 33, 6340-
6350. (f) Hilbers, M.; Meiwald, M.; Mattes, R. Z. Naturforsch., B: Chem.
Sci. 1996, 51, 57-67.
9234 J . Org. Chem., Vol. 68, No. 24, 2003

Squaric Acid N-Hydroxylamides
SCHEME 2a
the more commonly used liquid diethyl ester because it
is readily prepared in high purity and is much less prone
to hydrolytic degradation. Parallel to the reactivity of the
diethyl ester, dimethyl ester 2a reacted with a slight
excess (∼1.1 equiv) of N-alkylamine in MeOH under
reflux conditions to afford the corresponding N-alkyl-
amide methylesters squarates 3a -c.5 Reaction of these
N-alkylamide methylesters with a molar excess of N-
alkylamine produced the bis-N-alkylamides with identi-
cal (4a , 4c) or mixed (4b, 4d ) N-alkyl substituents.8 The
preparation of the known (3a ,22 4a , 4c23) or novel (3b,
3c, 4b, 4d ) compounds serves as a reference point for
the synthesis of the N-hydroxylamide derivatives.
In analogy to the preparation of the monoamide
monoesters 3 upon treatment of dimethyl squarate 2a
with 1.1 equiv of free base N-alkylhydroxylamines in
MeOH at ambient temperature, the corresponding N-
alkylhydroxylamide methylester squarates 5a -c were
produced.20 Conversion of 5a -c to the corresponding
N-alkylhydroxylamide N-morpholinamide squarates 6a-c
was achieved by addition of 2 equiv of morpholine, also
in MeOH and at ambient temperature. The inverse
reaction, i.e., the preparation of the compounds 6a -c by
reaction of morpholinamide methylester squarate (3b)
with an N-alkylhydroxylamine was not possible, even
when using a large excess of hydroxylamine under forcing
conditions (reflux in MeOH or EtOH for 48 h, with or
without the addition of KOH or Et3N). The starting
materials were recovered quantitatively. Our results are
in accordance with the report by Zinner and Gru¨nefeld.20
This finding suggests a different mechanism for the
formation of the N-alkylhydroxylamide ester squarates
from squarate esters and hydroxylamines than implied
by the overall nucleophilic substitution reaction. A pos-
sible mechanism for this reaction is discussed below.
Upon stirring of the ester hydroxamate 5 in water at
room temperature, the slow formation of the expected
hydrolysis products 8 was observed by +ESI-MS (e.g.,
for 8b, m/z ) 172, corresponding to MH+). However, even
after 1 week of reaction time, only partial conversion of
the starting material took place, despite the fact that an
autocatalytic reaction is expected to occur, as acids 8 are
expected to be strong acids (cf. to the pK1 and pK2 of 1 of
0.5 and 3.5, respectively).4 Moreover, the N-dehydroxy-
lated species 11 was formed over time (e.g., for 11b, m/z
) 156, corresponding to MH+). Hydrolysis of 5 under
acidic (1.2 M aq HCl) or basic (1 M aq KOH) conditions
resulted in the formation of 11 in quantitative yields. The
elimination of the hydroxyl group from hydroxamic acids
has been previously observed.24 Thus, we were not able
to prepare the N-hydroxylamide acids 8 in pure form. The
amide acids 11 could also be prepared independently by
hydrolysis of the ester amides 3. For instance, acid-
catalyzed hydrolysis of ester amide 3c provided 11d in
good yields.
a
Reaction Conditions: (i) 1.0 equiv of HNR2/MeOH, ∆;5 (ii)
molar excess of HNR2/MeOH, ∆;23 (iii) 1.7 equiv of HNROMe/
MeOH, 25 °C;20 (iv) 1.1 equiv of HNROH/MeOH, 25 °C;20 (v) 2
equiv morpholine/MeOH, 25 °C; (vi) 5 equiv of morpholine/MeOH,
25 °C; (vii) CH2N2/Et2O, 25 °C; (viii) H2O, 25 °C; (ix) 1.2 M aq
HCl or 1.0 M aq KOH.
The O-alkylated species 9 and 10 are expected to be
incapable of binding iron in the hydroxamate O,O-
bidentate fashion. They are thus intended to serve in
detailed binding studies as the nonspecific metal binding
analogues to hydroxamates 5 and 6. N-Methylhydroxy-
lamide ester 5a can be alkylated with diazomethane to
provide N,O-dimethylhydroxylamide methylester 9a . A
more convenient way to prepare this material, however,
is by reaction of dimethylsquarate 2a with O-alkylhy-
droxylamines.20 The O-methylhydroxylamide methylester
9a can be converted in good yield to the corresponding
O-methylhydroxylamide N-morpholinamide 10 by reac-
tion with a 5-fold molar excess of morpholine.
Mech a n ism of F or m a tion of th e Squ a r ic Acid
Hyd r oxa m a tes. As shown, squarate dimethyl ester 2a
readily reacts with N-alkylhydroxylamines to provide the
corresponding N-alkylhydroxylamide methylester squar-
ates 5a -c. Treatment of 5a -c with N-alkylhydroxyl-
amines, however, does not lead to the formation of the
bis-N-alkylhydroxylamides 7 under any condition tested
(Scheme 2).20 On the other hand, reaction of 5 with 1-3
equiv N-alkylamines affords the corresponding N-alkyl-
hydroxylamide alkylamide squarates 6 in good yields at
ambient temperature within 1 h. This is surprising in
light of the reported higher nucleophilicity of N-hydroxyl-
amines toward sp2 centers as compared to N-alkyl-
(22) Hall, L. A.; Williams, D. J .; Menzer, S.; White, A. J . P. Inorg.
Chem. 1997, 36, 3096-3101.
(23) Thorpe, J . E. J . Chem. Soc. B 1968, 435-436.
(24) For examples of the dehydroxylation of hydroxamates under a
variety of reaction conditions, see e.g.: (a) Surman, M. D.; Mulvihill,
M. J .; Miller, M. J . Tetrahedron Lett. 2002, 43, 1131-1134. (b)
Quadrelli, P.; Campari, G.; Mella, M.; Caramella, P. Tetrahedron Lett.
2000, 41, 2019-2022. (c) Ayyangar, N. R.; Kalkote, U. R.; Nikrad, P.
V. Indian J . Chem., Sect. B 1983, 22, 872-877. (d) Mattingly, P. G.;
Miller, M. J . J . Org. Chem. 1980, 45, 410-415. (e) Birchall, J . D.;
Glidewell, C. J . Chem. Soc., Dalton Trans. 1978, 604-607.
J . Org. Chem, Vol. 68, No. 24, 2003 9235

Lim et al.
SCHEME 326
SCHEME 5
SCHEME 4
transition state provides IIB. Formal N-to-O proton
transfer and subsequent anti elimination of methanol
(IIIB) leads to the formation of N-alkylhydroxylamide
methylester squarate 5. All steps involved are highly
suggestive of a concerted mechanism, making a collapse
of IB and direct formation of the final product likely.
Why, then, is the failure to form bis-N-hydroxylamide
squarates observed? Considering the amide character of
5 with its canonical structures 5A and 5B (Scheme 5),
nucleophilic attack at C3 and subsequent formation of
the five-membered transition state is unfavorable due to
the mismatched polarity of 5A and 5B and the nucleo-
phile. Resonance structure 5C is favored by the expres-
sion of a H-bond between the hydroxylamide hydrogen
and the adjacent carbonyl group, and the limiting reso-
nance structure is represented by 5C. Though this
structure polarizes C3 toward nucleophilic attack, the
C3-C4 bond no longer possesses any double-bond char-
acter. However, a double bond at this position is required
for the 3 + 2-type mechanism to take place. On the other
hand, the polarity of C3 allows a simple nucleophilic
substitution mechanism by an alkylamine to occur. This
hypothesis is consistent with the observed reactivity of
5. Further spectroscopic and structural evidence for the
polarity and limiting resonance structures evoked are
presented below.
Sp ectr oscop ic Com p a r ison of th e Hyd r oxa m a te
Squ a r a tes a n d Hyd r oxa m ic Acid s. Formally, squarate
esters 2 are not esters, i.e., they do not contain an ester
moiety. Likewise, hydroxamate squarates 5 and 6 are not
true hydroxamic acids. They are best described as viny-
logous hydroxamic acids. Furthermore, and in stark
contrast to hydroxamic acids, the N-hydroxyl group is not
in conjugation with its adjacent carbonyl group but is
instead in conjugation with the carbonyl group opposite
to it. This peculiarity has consequences when considering
the bond arrangements upon binding of a metal as
compared to those found in a hydroxamato metal complex
(Figure 1). The O,O-donor set in an [h yd r oxa m a to]M
complex is provided by a deprotonated N-hydroxide and
an amide carbonyl (I). The limiting resonance structure
II expresses the amide character of this carbonyl group;
the existence of resonance structures is responsible for
the tight binding of the metal. To illustrate the profound
differences in the bonding arrangement found in a
hydroxamato squarate metal complex, consider the reso-
nance structures I and II of the chelate 11M. Although
the O,O-donor set for the metal is also provided by a
N-hydroxide and its adjacent carbonyl group, the carbo-
nyl group is not in conjugation with the amide nitrogen
amines.25,26 Similarly, formation of the squarate diamides
4 by nucleophilic substitution of the squarate ester
amides 3 is possible, while the analogous reaction of 3
with N-hydroxylamines fails to provide any product.20
These findings suggest that N-alkylhydroxylamines
react with squarate esters by a different mechanism than
the expected nucleophilic substitution pathway of regular
N-alkylamines. Precedents for the varying reactivity of
the two nitrogen-based nucleophiles are known. For
instance, conjugate addition of hydroxylamines to R,â-
usaturated esters proceeds smoothly, but no reaction with
the corresponding O-alkylhydroxylamines and alkyl-
amines is observed.27 To explain this finding, a concerted
mechanism for the addition was proposed, assigning a
key role to the hydroxyl hydrogen (Scheme 3).26,27 Nu-
cleophilic attack by the nitrogen lone pair to the alk-
enenone IA generates the cyclic five-membered ring
transition state IIA. This transition state resembles that
of a [3 + 2] dipolar cycloaddition reaction or retro-Cope
elimination reaction. Its collapse generates the species
IIIA, and a rapid N-to-O proton shift then provides the
final product.
A similar mechanism can be hypothesized to explain
the formation of N-alkylhydroxylamide methylesters 5
by reaction of a squarate diester 2 with a N-alkylhy-
droxylamine (Scheme 4). Upon nucleophilic attack of
N-alkylhydroxylamine to dimethyl squarate 2a , the five-
membered transition state IB is formed. Collapse of this
(25) Collis, M. P.; Hockless, D. C. R.; Perlmutter, P. Tetrahedron
Lett. 1995, 36, 7133-7136.
(26) Niu, D.; Zhao, K. J . Am. Chem. Soc. 1999, 121, 2456-2459.
(27) (a) Fountain, K. R.; Erwin, R.; Early, T.; Kehl, H. Tetrahedron
Lett. 1975, 35, 3027-3030. (b) Moglioni, A. G.; Muray, E.; Castillo, J .
A.; Alvarez-Larena, A.; Moltrasio, G. Y.; Branchadell, V.; Ortun˜o, R.
M. J . Org. Chem. 2002, 67, 2402-2410.
9236 J . Org. Chem., Vol. 68, No. 24, 2003

Squaric Acid N-Hydroxylamides
nance stabilization by forming hydrogen bonds, whereas
amides 6 are much better H-acceptors. If the hydrogen
bond donor group is methylated (9a from ester hydrox-
amate 5a or 10 from amide hydroxamate 6a ), the
relaxation times increase dramatically (cf. 9a (11-20 s)
to 5a (9-13 s); cf. 10 (9 f 20 s) to 6a (6-11 s)) and fall
into the class of the corresponding ester amides 3 or
diamides 4, respectively.
Why does a greater ability to delocalize charge ef-
ficiently reduce the spin-lattice relaxation times in these
small molecules? In spin-lattice relaxation, the excited
nuclei transfer their excitation energy to their environ-
ment. They do so via interaction of their magnetic vectors
with fluctuating local fields of sufficient strength and a
fluctuation frequency of the order of the Larmor fre-
quency of the nuclear spin type.28 The efficiency of this
process is reflected in the measured relaxation time T1.
A short T1 corresponds to an efficient relaxation or a long
correlation time between nuclei. Depending upon the
atomic and electronic environment of a nucleus in a
molecule and the motion of that molecule, there are four
potential mechanisms contributing to spin-lattice relax-
ation of the nucleus: chemical shift anisotropy (CSA),
scalar coupling (SC), spin rotation (SR), and dipole-
dipole interaction (DD).28 DD is unlikely to be the
dominant relaxation mechanism in the squarates, since
this relaxation mechanism is dominant in C-H bonds
and no such bonds exist in the squarate core. Likewise,
SR and SC are not the dominant mechanisms, as all
compounds are of comparable shape and size and have
no large J couplings. This excludes dramatically different
spinning rates, anisotropies in their spinning motions,
or different segmental motions. This leaves CSA, which
is the modulation of the anisotropic shielding by a nuclear
or electron tensor within the molecule. The delocalization
and charge separation in the cyclobutenedione framework
aids relaxation in the hydroxamate derivatives. Thus, we
believe it is the CSA term that is primarily responsible
for the observed modulation of the spin-lattice relaxation
times. The different degrees to which resonance struc-
tures are expressed in the squarate esters and amides
change the dipole moments and polarizability of all bonds
of the framework and those attached to it. This, in turn,
changes the number of CSA interactions, which can
contribute to an efficient relaxation of the 13C excitation
energy to their environment. Nonetheless, some influence
of the SC term to the T1 modulation cannot be ruled out
since the cyclobutenedione framework in the compounds
considered is decorated with either four oxygens, three
oxygens, and one nitrogen or two elements of each. This
alteration results in different, albeit small, scalar cou-
pling terms.
F IGURE 1. Resonance structures of metal hydroxamates as
compared to metal hydroxamate squarate 11M.
of the N-hydroxide. Therefore, the donor capability of this
carbonyl oxygen is largely determined by the nature of
the fourth ring substituent, here the group -R. In
hydroxamate ester squarates such as 5, this carbonyl has
ester carbonyl character, and consequently, it is a
comparably bad donor. In the hydroxamate amide
squarates 6, this carbonyl has amide character and,
hence, carries a much higher partial negative charge and,
thus, is a better donor. This interpretation is the key to
understanding the largely differing iron binding proper-
ties of the N-hydroxamate esters and amides. Further,
this interpretation also predicts that the N-hydroxamate
acids 8, which carry a large negative charge on the
oxygen, should be excellent chelators. We are currently
seeking an alternative pathway toward their synthesis.
13C NMR Sp ectr oscop y. Our initial efforts to char-
acterize the squaric acid derivatives by {1H}13C NMR
spectroscopy using standard 1-2 s delay times were
thwarted by the inability to detect all of the expected
cyclobutenedione framework carbons in many of the
derivatives studied. The problem was found to lie in long
spin-lattice relaxation times (T1). This led to a systematic
investigation of the 13C T1 of these derivatives. The
results are tabulated in Table 2.
Across the table, the chemical shifts of the carbons vary
only slightly and in predictive ways: each (vinylogous)
ester moiety is characterized by a carbonyl carbon shift
between ∼185-190 ppm and the corresponding â-carbon
at ∼175-177 ppm (2, 3, 5, and 9). By contrast, each
(vinylogous) amide moiety resonates at ∼179-184 and
164-170 ppm (4, 5, 6, 9, and 10), respectively. Each
(vinylogous) acid functionality exhibits a signal between
186 and 190 ppm (1 and 11). These shifts can be
interpreted in light of the predominant resonance struc-
tures of the vinylogous ester, amide, and acid function-
alities.
In DMSO-d6, the T1 values range from 0.6 s in squaric
acid to well over 20 s in 2a , 3a , 4d , 9a , and 10. Some
trends in the relaxation times can be traced. Each
compound class is characterized by a specific range of
T1. Thus, the compound classes can be ranked according
to increasing T1: squaric acid 1 (less than 1 s) <
hydroxamate amides 6 (2-3 s) < amide acids 11 (8-10
s) < hydroxamate esters 5 (8-12 s) < diamide 4 (10-15
s) < ester amide 3a (12-20 s) ≈ diester 2a (13-26 s).
Again, the ranking of these compound classes reflects
their ability to undergo resonance stabilization. Two
examples may serve to illustrate this. Both hydroxamate
amides 6 and hydroxamate esters 5 can undergo reso-
Qualitatively, this interpretation is supported by the
trends observed in the νO-H and νCdO stretching frequen-
cies in the vibrational spectra of the derivatives. For
instance, the νCdO band of the vinylogous ester moieties
is found at ∼1800 cm-1 throughout all compounds (see
2, 3, 5, and 9). As expected, the amide νCdO band is found
at lower wavenumbers (1700-1660 cm-1) whereby their
position is dependent on the particular compound class.
Ester amides such as 3, 5, and 9 are found at the upper
(28) Breitmaier, E.; Spohn, K.-H.; Berger, S. Angew. Chem., Int. Ed.
Engl. 1975, 14, 144-159.
J . Org. Chem, Vol. 68, No. 24, 2003 9237

Lim et al.
TABLE 1. 13C NMR Sh ifts (100 MHz, DMSO-d 6, 27 °C) a n d Associa ted Sp in -La ttice Rela xa tion Tim es T1 for th e
Ca r bon s of th e Cyclobu ten ed ion e F r a m ew or k of Select Squ a r a tesa
a
T1 data was acquired using a standard inversion recovery sequence with proton decoupling using a QNP probe; delay between scans
was set at approximately 10 s with variable delays ranging from 0.5 to 60 s. For each delay, 400-512 scans were acquired. For a full
b
listing of the 1H and 13NMR shifts, see Experimental Section. Insufficient number of data points collected to determine the T1 with
greater precision. c Broadened signals for 11b-c suggest exchange phenomena that may diminish the validity of the T1 determinations;
see Supporting Information for the 13C spectra of these novel squaric acid monoamides.
range, while those of the diamides 4 and 10 and the
hydroxylamide amides 6 are shifted down to 1670-1660
cm-1. In the N-hydroxylamide amide 6a , a νCdO band at
1660 cm-1 is observed. O-Methylation breaks the possible
H-bond and the νCdO band shifts to 1675 cm-1. However,
a comprehensive interpretation of the IR data is made
impossible by the inability to unambiguously assign all
peaks above 1500 cm-1
.
P r elim in a r y Meta l Bin d in g Stu d ies. The red-purple
iron(III) complexes of hydroxamates are characterized by
a relatively strong (ꢀ ≈ 3000) and broad (half-height
width of ∼200 nm) ligand-to-metal charge-transfer band
centered at around 500 nm.29 Figure 2 shows the UV-
vis spectra of aqueous solutions of 5b and 6b in the
presence of 1/3 mol ratio Fe3+ in comparison to the
spectrum of the tris[N-tolylbenzhydroxamato]iron(III)
complex. Consistent with the report by Zinner and
Gru¨nefeld,20 squaric acid N-hydroxylamide ester 5b
provides no indication for metal binding. However, the
corresponding amide 6b exhibits a UV-vis band upon
treatment with iron(III) (Figure 2). This indicates a
binding event but does not allow a clear interpretation
of the binding mode, especially as the green color pales
over time. A detailed investigation of the coordination
interaction is underway. This experiment, however,
F IGURE 2. UV-vis spectra of 5b (‚‚‚), 6b (- - -), and
N-tolylbenzhydroxamic acid (-), all 0.57 mM in 50% EtOH/
H2O containing 0.19 mM FeCl3.
underlines the limitations of these “hydroxamate ana-
logues”, as they may not possess hydroxamate-like bind-
ing properties.
Sin gle-Cr ysta l X-r a y Str u ctu r e of 5b a n d 6c. The
single-crystal X-ray structures of the hydroxamate ester
5b and amide 6c were determined. The structures prove
the connectivity of the squarate hydroxamates. However,
the conformations of both compounds are of note. Instead
of the initially expected intramolecular H-bond between
the N-OH group and an adjacent carbonyl group of the
putative metal binding site, the compounds form an
intermolecular H-bond in the crystal. The ball-and-stick
(29) Konetschny-Rapp, S.; J ung, G.; Raymond, K. N.; Meiwes, J .;
Za¨hner, H. J . Am. Chem. Soc. 1992, 114, 2224-2230.
9238 J . Org. Chem., Vol. 68, No. 24, 2003

Squaric Acid N-Hydroxylamides
excluded that the observation of the intermolecular
H-bonds are a crystal packing effect.
Con clu sion
In conclusion, squarate hydroxamate esters 5 and
hydroxamate amides 6 were prepared and fully charac-
terized. While esters 5 do not bind iron(III) in aqueous
systems, amides 6 do. The difference in their coordination
behavior can be rationalized on the basis of the variation
of the resonance stabilization of these molecules. The
dominating resonance structures also determine the
mechanism of formation of the squarate hydroxamates.
An investigation of the spin-lattice relaxation time T1
of a range of derivatives allowed us to define guidelines
for the reliable NMR spectroscopic characterization of
squarate derivatives. The spectroscopically derived con-
clusions regarding the prevailing resonance structures
in these molecules were also seen in the solid-state
structure of the hydroxamate ester and the amide. In all,
the hydroxamate squarates differ in key properties from
hydroxamic acids.
A metal coordinated to the O,O-chelate of the hydrox-
amic acid-like moiety of the hydroxamate amides 6 is
electronically coupled with the amide nitrogen on the
opposite side. This coupling makes them attractive
platforms for the synthesis of molecular chemosensors.
Thus, the understanding of the bond arrangements
controlling the metal binding in these novel iron(III)-
chelating moieties forms the basis for their use in the
construction of molecular sensors, a goal we are currently
pursuing.31
F IGURE 3. Ball-and-stick representation and numbering
scheme used for a H-bonded dimer found in the solid-state
structure of 6c; temperature of data collection ) 198 K.
Selected bond distances: C(1)-C(2) ) 1.470(2) Å, C(2)-C(3)
) 1.489(2) Å, C(3)-C(4) ) 1.466(2) Å; C(4)-C(1) ) 1.415(2)
Å, C(2)-O(1) ) 1.214(2) Å, C(3)-O(2) ) 1.232(2) Å, C(1)-N(3)
) 1.333(2) Å, N(3)-O(5) ) 1.413(2) Å, C(4)-N(4) ) 1.326(2)
Å, O(5)-O(2′) ) 2.657(2) Å; N(3)-O(5)-O(2′) angle ) 99.4°;
rms of cyclobutenedione framework ) 0.059 Å.
model showing one H-bond expressed by hydroxyl amide
squarate 6c in the solid state is shown in Figure 3.
In light of the deduced electronic and observed metal
binding differences between the ester and the amide
hydroxamates, the observed conformation may not be
entirely unexpected for the hydroxamate ester 5b, but it
is certainly unexpected for the hydroxamate amide 6c.
Nonetheless, much of the spectroscopic differences be-
tween the esters and the amides can also be seen in the
crystal structures. For instance, through comparison of
the relative trends in the bond lengths of equivalent bond
distances found in 5b and 6c, the differing electronic
influences of the ester and amide functionality on the
electron-density (or bond order) of the carbonyl oxygen
adjacent to the N-hydroxylamide functionality can be
traced. They are predicted on the basis of the spectro-
scopic evidence. The two carbonyl C-O bond distances
in the hydroxamate ester 5b (1.178 and 1.209 Å) are
much shorter than the two carbonyl C-O bond distances
in the hydroxamate amide 6c (1.214 and 1.232 Å). In
either case, the longer of the two bond lengths is found
for the C-O bond diagonally across the hydroxamate
moiety, providing evidence for the strongly electron-
donating nature of the N-OH moiety. The large difference
between the two bond lengths in the ester versus the
relatively equal bond distances in the amide highlights
the electronic difference between the ester and the amide
moiety. The C-C bond length differences observed in
these compounds also parallel the inferred resonance
pathways. Thus, the crystal structures confirm the
spectroscopically derived conclusions and provide a struc-
tural explanation for the differing metal binding proper-
ties of these compounds.
Exp er im en ta l Section
WARNING. Liquid, lipophilic squaric acid diesters are
known to be potent topical allergens.8 Two of us contracted a
contact dermititis on exposed areas on hands, arms, and face
that required medical attention after having worked with solid
amide esters such as 3 (exposure through dusts generated
while scraping the compounds from a round-bottom flask) and
during the preparation of diethylsquarate from halogenated
intermediates.3,32 We did not experience any problems when
working with the N-hydroxylamide derivatives. However, we
did not test the compounds and habitually exercised caution
when working with these compounds (gloves, transfer of solids
only under a well-ventilated fumehood).
Ma ter ia ls. 3,4-Dimethoxycyclobut-3-ene-1,2-dione (2a , di-
methyl squarate), Rf ) 0.89 (silica-CH3CN),21 3-methoxy-4-
N-diethylaminocyclobut-3-ene-1,2-dione (3a ),22 3,4-bis(diethyl-
amino)cyclobut-3-ene-1,2-dione (4a ),23 and 3,4-bis(dibenzyl-
amino)cyclobut-3-ene-1,2-dione (4c)23 were prepared according
to literature procedures. Spectroscopic and analytical data for
the known compounds are provided for comparison and
because some data differ from literature-reported data and no
13C NMR data were reported for any of these compounds.
3-Met h oxy-4-N-m or p h olin ylcyclob u t -3-en e-1,2-d ion e
(3b). 3,4-Dimethoxycyclobut-3-ene-1,2-dione (2a ) (500 mg, 3.52
mmol) was dissolved in MeOH (4 mL), and morpholine (307
µL, 3.52 mmol) was added; the mixture was heated to reflux.
The formation of a white precipitate was observed. Once all
The structures further highlight the fundamental
differences between hydroxamate squarates and hydrox-
amic acids, although the existence of intermolecular
H-bonds in 5b or 6c in solution cannot be excluded. The
molecules are arranged in the lattice in an extensive
network of intermolecular H-bonds but it cannot be
(30) Detailed description of the H-bond networks expressed by these
and related compounds will be reported in a separate manuscript.
(31) Lim, N. C.; Yao, L.; Freake, H. C.; Bru¨ckner, C. Abstract of
Papers, 225th National Meeting of the American Chemical Society,
New Orleans, LA, Mar. 27-23, 2003; American Chemical Society:
Washington, DC, 2003; INORG 884.
(32) Maahs, G. Angew. Chem., Int. Ed. Engl. 1963, 2, 690-691.
J . Org. Chem, Vol. 68, No. 24, 2003 9239

Lim et al.
the starting material was consumed (after 36 h, TLC control),
Et2O (1 mL) was added to the hot solution, which then was
cooled to 0 °C. Crystals were formed overnight and filtered
off using a glass frit (M). Washing of the filter cake with cold
Et2O and drying under vacuum at ambient temperature
afforded 3b (0.52 g, 75% yield) as analytically pure white
crystals: Rf ) 0.29 (silica-CH2Cl2/20% CH3CN); mp ) 152-
154 °C (uncorrected); 1H NMR (DMSO-d6, 400 MHz, 75 °C, δ)
4.29 (s, 3H), 3.72-3.70 (m, 8H) ppm; 13C NMR (DMSO-d6, 100
MHz, D1 ) 5 s, 75 °C, δ) 189.3, 182.5, 177.0, 170.3, 66.3, 60.8,
47.2 ppm; IR (KBr) νmax 1798 (w, CdO), 1705 (m, CdO), 1627
(s), 1496 (m), 1407 (m), 1277 (m), 1110 (m) cm-1; MS (FAB+,
NBA) m/z 198 (MH+, 76%); HR-MS (FAB+ of MH+, NBA) m/z
calcd for C9H12NO4 198.0766, found 198.0746.
a white precipitate resulted. Stirring at ambient temperature
was continued for ∼1 h (monitored by TLC). The mixture was
then filtered; the filter cake was washed with cold MeOH, and
the combined filtrates were taken to dryness by rotary
evaporation. Purification of 5a was achieved by column chro-
matography (silica-20% CH3CN/CH2Cl2) to yield a white solid
(4.9 g, 63% yield): Rf ) 0.45 (silica-20% CH3CN/CH2Cl2); mp
) 135-140 °C dec (uncorrected) (lit. 177-178 °C);20 1H NMR
(DMSO-d6, 400 MHz, δ) 4.28 (s, 3H) 3.38 (s, 3H) ppm; 13C NMR
(DMSO-d6, 100 MHz, D1 ) 5 s, LB ) 5 Hz, δ) 185.2, 179.3,
174.5, 167.8, 60.3, 40.7 ppm; IR (KBr) νmax 3087 (br), 2796 (br,
OH), 1806 (s, CdO), 1709 (s, CdO), 1626 (s), 1508 (m), 1489
(m), 1450 (m), 1390 (s), 1257 (m), 1202 (s), 1166 (s), 1122 (m),
1074 (s), 1040 (m), 944 (s), 885 (m), 834 (s), 782 (m), 769 (m)
cm-1; MS (ES-) m/z ) 156 (M-); MS (FAB+, NBA) m/z ) 158
(MH+, 83%); HR-MS (FAB+ of MH+, NBA) m/z calcd for C6H8-
NO4 158.0453, found 158.0448. Anal. Calcd for C6H7NO4: C,
45.86; H, 4.49; N, 8.91. Found: C, 45.92; H, 4.60; N, 9.11.
3-Meth oxy-4-N-isopr opylh ydr oxylam in ocyclobu t-3-en e-
1,2-d ion e (5b). Prepared as a white solid according to the
procedure described for 5a (92% yield, 3.59 g): Rf ) 0.65 (silica-
40% CH3CN/CH2Cl2); mp ) 148-150 °C dec (uncorrected) (lit.
141-142 °C);20 1H NMR (DMSO-d6, 400 MHz, δ) 10.53 (br s,
1H), 4.34 (br s, 1H), 4.29 (s, 3H), 1.19 (d, J ) 6.6 Hz, 6H) ppm;
13C NMR (DMSO-d6, 100 MHz, D1 ) 5 s, LB ) 5 Hz, δ) 185.4,
179.5, 175.0, 167.7, 60.3, 53.5, 19.2 ppm; IR (KBr) νmax 3096
(br), 2823 (br, OH), 1796 (s, CdO), 1710 (s, CdO), 1602 (s),
0.1496 (s), 1467 (s), 1402 (s), 1367 (s), 1193 (s), 1173 (s), 1122
(s), 1052 (s), 1028 (s), 941 (s), 904 (s), 850 (s), 752 (s) cm-1; MS
(ES-, 100% CH3CN, 30 V) m/z ) 184 (M-); MS (FAB+, NBA)
m/z ) 186 (MH+, 100%); HR-MS (FAB+ of MH+, NBA) m/z
calcd for C8H12NO4 186.0766, found 186.0743. Anal. Calcd for
C8H11NO4; C, 51.89; H, 5.99; N, 7.56. Found: C, 51.59; H, 5.61;
N, 7.92.
3-Ben zylam in e-4-m eth oxycyclobu t-3-en e-1,2-dion e (3c).
3,4-Dimethoxycyclobut-3-ene-1,2-dione (2a ) (4.0 g, 28.1 mmol)
was dissolved in MeOH (100 mL). Benzylamine (3.25 mL, 29.2
mmol) was added, and the mixture was stirred at ambient
temperature for ∼1 h. The solution was then taken to dryness
by rotary evaporation. Column chromatography (silica-20%
CH3CN/CH2Cl2) was used to isolate and purify 3c as a white
solid (5.1 g, 83% yield): Rf ) 0.60 (silica-CH2Cl2/20% CH3CN);
mp ) 122-124 °C (uncorrected); 1H NMR (DMSO-d6, 400
MHz, δ) 7.38-7.29 (m, 5H), 4.67 (s, 1H), 4.56 (s, 1H), 4.29 (s,
3H) ppm; 13C NMR (DMSO-d6, 100 MHz, δ) 189.8, 183.3, 178.0,
173.1, 138.7, 129.0, 127.9, 127.8, 60.3, 47.8 ppm; IR (KBr) νmax
3206 (br, NH), 2952 (m), 2925 (m), 1795 (s, CdO), 1705 (s,
CdO), 1594 (s), 1508, (s), 1435 (s), 1408 (s), 1355 (s), 1049 (m),
921 (s), 831 (m), 741 (s), 693 (s) cm-1; MS (ES+, 100% CH3-
CN, 30 V) m/z ) 240 (MNa+); MS (FAB+, NBA) m/z ) 218
(MH+, 51%); HR-MS (FAB+ of MH+, NBA) m/z calcd for
C
12H12NO3 218.0817, found 218.0809.
3-Diet h yla m in o-4-m or p h olin ocyclob u t -3-en e-1,2-d i-
on e (4b). Squarate amide ester 3a (228 mg, 1.24 mmol) was
dissolved in EtOH (6 mL) and reacted with morpholine (1.63
mL, 18.7 mmol) under reflux conditions. Upon consumption
of 3a as monitored by TLC (∼24 h), the solution was evapo-
rated in vacuo to yield a yellow solid residue. This solid was
recrystallized from hot, anhydrous EtOH, to yield light yellow
crystals of 4b (0.28 g, 94% yield): Rf ) 0.50 (silica-CH3CN);
mp ) 102 °C (uncorrected); 1H NMR (DMSO-d6, 400 MHz, δ)
3.69 (br t, J ) 5.0 Hz, 4H), 3.59 (br t, J ) 5.0 Hz, 4H), 3.51 (q,
J ) 7.2 Hz, 4H), 1.16 (t, J ) 7.2 Hz, 6H) ppm; 13C NMR
(DMSO-d6, 100 MHz, δ) 183.9, 183.1, 168.6, 167.5, 66.1, 48.3,
43.9, 13.8 ppm; IR (KBr) νmax 3427 (br), 2969 (m), 2926 (m),
2869 (m), 1778 (s, CdO), 1663 (s, CdO), 1590 (s), 1519 (s),
1455 (s), 1305 (m), 1274 (s), 1218 (m), 1167 (m), 1112 (s), 1065
(m), 1023 (m), 991 (m), 883 (m), 789 (m) cm-1; MS (ES+, 100%
CH3CN, 30 V) m/z ) 261 (MNa+); MS (FAB+, NBA) m/z )
239 (MH+, 96%); HR-MS (FAB+ of MH+, NBA) m/z calcd for
3-Met h oxy-4-N-cycloh exylh yd r oxyla m in ocyclob u t -3-
en e-1,2-d ion e (5c). Prepared as a white solid from 2a and
cyclohexylhydroxylamine according to the procedure described
for 5a (2.56 g, 81% yield): Rf ) 0.77 (silica-20% CH3CN/CH2-
Cl2); mp ) 158-159 °C dec (uncorrected) (lit. 149-151 °C);20
1H NMR (DMSO-d6, 400 MHz, δ) 10.56 (br s, 1H), 4.29 (s, 3H),
3.93 (br s, 1H), 1.79-1.06 (m, 10H) ppm; 13C NMR (DMSO-
d6, 100 MHz, D1 ) 3 s, LB ) 5 Hz, δ) 186.0, 180.1, 175.7, 168.3,
61.8, 60.9, 29.8, 25.1, 25.0 ppm; IR (KBr) νmax 2933 (s), 2856
(m, OH), 1787 (m, CdO), 1695 (m, CdO), 1637 (s), 1495 (s),
1381 (s), 1058 (m), 760 (m) cm-1; MS (ES-, 100% CH3CN, 30
V) m/z ) 224 (M-); MS (FAB+, NBA) m/z ) 226 (MH+, 28%),
264 (MK+, 99%), 302 (M - H+‚2K+, 66%); HR-MS (FAB+ of
MH+, NBA) m/z calcd for C11H16NO4 226.1079, found 226.1059.
3-Mor p h olin o-4-N-m et h ylh yd r oxyla m in ocyclob u t -3-
en e-1,2-d ion e (6a ). Squarate hydroxamate ester 5a (600 mg,
3.8 mmol) was dissolved in MeOH and titrated with morpho-
line at ambient temperature (TLC control). Once all the
starting material was consumed, the solution was taken to
dryness by rotary evaporation yielding a yellow residue that
was recrystallized from hot water yielding 6a as a white solid
(0.50 g, 62% yield): Rf ) 0.23 (silica-CH3CN); mp ) 128-130
C
12H19N2O3 239.1396, found 239.1400.
3-Ben zyla m in o-4-d iet h yla m in ocyclob u t -3-en e-1,2-d i-
on e (4d ). Squarate amide ester 3c (220 mg, 1.0 mmol) was
dissolved in EtOH (4 mL) and reacted with diethylamine (1.6
mL, 15.5 mmol) under reflux conditions for 5 min. Workup
was performed as described for 3b (Et2O, 3 mL). The dried
white crystals formed were analytically pure (0.26 g, 99%
yield): Rf ) 0.24 (silica-20% CH3CN/CH2Cl2); mp ) 175-177
°C (uncorrected); 1H NMR (DMSO-d6, 400 MHz, δ) 7.38-7.27
(m, 5H), 4.78 (s, 2H), 3.54 (br s, 4H), 1.13 (t, J ) 7.1 Hz, 6H);
13C NMR (DMSO-d6, 100 MHz, D1 ) 4 s, LB ) 5 Hz, δ) 183.0,
182.2, 167.4, 167.0, 140.0, 128.9, 127.8, 127.6, 46.9, 43.9, 15.5
ppm; IR (KBr) νmax 3203 (br, NH), 2966 (m), 1789 (m, CdO),
1659 (s, CdO), 1570 (s), 1528 (s), 1443 (s), 1315 (m), 1075 (m),
696 (m) cm-1; MS (FAB+, NBA) m/z ) 258 (M+, 88%), 259
(MH+, 99%); HR-MS (FAB+ of MH+, NBA) m/z calcd for
1
°C dec (uncorrected); H NMR (DMSO-d6, 400 MHz, δ) 10.91
(br s, 1H), 3.85 (t, J ) 4.5 Hz, 4H), 3.68 (t, J ) 4.4, 4H), 3.47
(s, 3H) ppm; 13C NMR (DMSO-d6, 100 MHz, D1 ) 5 s, δ) 180.5,
178.6, 165.1, 164.4, 66.0, 48.7, 41.2 ppm; IR (KBr) νmax 3413
(br), 2662 (br, OH), 1783 (s, CdO), 1660 (s, CdO), 1588 (s),
1445 (s), 1406 (s), 1310 (s), 1292 (s), 1205 (s), 1109 (s), 1025
(s), 867 (m) cm-1; MS (ES+, 100% CH3OH, 30 V) m/z ) 235
(MNa+); MS (FAB+, NBA) m/z ) 212 (M+, 42%), 213 (MH+,
76%), 235 (MNa+, 15%); HR-MS (FAB+ of MH+, NBA) m/z
calcd for C9H13N2O4 213.0875, found 213.0855. Anal. Calcd for
C9H12N2O4; C, 50.94; H, 5.70; N, 13.20. Found: C, 50.76; H,
6.01; N, 13.26.
C
15H19N2O2 259.1447, found 259.1437.
3-Meth oxy-4-N-m eth ylh yd r oxyla m in ocyclobu t-3-en e-
1,2-d ion e (5a ). 3,4-Dimethoxy-3-cyclobutene-1,2-dione (2a )
(7.0 g, 49.3 mmol) was dissolved in MeOH (60 mL), and a slight
excess of N-methylhydroxylamine hydrochloride (4.62 g, 55.3
mmol) was added. Upon addition of KOH (3.35 g, 59.7 mmol),
3-Mor p h olin o-4-N-isop r op ylh yd r oxyla m in ocyclob u t -
3-en e-1,2-d ion e (6b). Prepared in 67% yield (131 mg) from
5b and morpholine as described for 6a : Rf ) 0.43 (silica-CH3-
1
CN); mp ) 148-150 °C dec (uncorrected); H NMR (DMSO-
9240 J . Org. Chem., Vol. 68, No. 24, 2003

Squaric Acid N-Hydroxylamides
d6, 400 MHz, δ) 10.28 (br s, 1H), 4.66 (heptet, J ) 6.5 Hz, 1H),
3.86 (t, J ) 4.5, 4H), 3.69 (t, J ) 4.5 4H), 1.18 (d, J ) 6.5 Hz,
6H) ppm; 13C NMR (DMSO-d6, 100 MHz, D1 ) 5 s, δ) 180.8,
178.7, 165.7, 164.7, 66.0, 53.4, 48.7, 19.2 ppm; IR (KBr) νmax
3435 (br), 2750 (br, OH), 1779 (s, CdO), 1662 (s, CdO), 1585
(s), 1497 (s), 1445 (s), 1364 (m), 1308 (s), 1281 (s), 1190 (s),
1121 (s), 1023 (m), 1006 (m), 883 (m), 826 (m) cm-1; MS (ES+,
100% CH3OH, 30 V) m/z ) 263 (MNa+); MS (FAB+, NBA) m/z
) 240 (M+, 41%), 241 (MH+, 74%); HR-MS (FAB+ of MH+,
NBA) m/z calcd for C11H17N2O4 241.1188, found 241.1178.
Anal. Calcd for C11H16N2O4; C, 54.99; H, 6.71; N, 11.66.
Found: C, 54.99; H, 6.47; N, 11.90.
3-Mor p h olin o-4-N-cycloh exylh yd r oxyla m in ocyclobu t-
3-en e-1,2-d ion e (6c). Prepared in 71% yield (360 mg) from
5c and morpholine as described for 6a : Rf ) 0.29 (silica gel-
CH3CN); mp ) 167-169 °C dec (uncorrected); 1H NMR
(DMSO-d6, 400 MHz, δ) 10.35 (br s, 1H), 4.23 (m, 1H), 3.85 (t,
J ) 4.4 Hz, 4H), 3.68 (t, J ) 4.4 Hz, 4H), 1.79-1.25 (m, 10H)
ppm; 13C NMR (DMSO-d6, 100 MHz, D1 ) 10 s, LB ) 5 Hz, δ)
180.8, 178.7, 165.7, 164.7, 66.1, 60.9, 48.7, 29.3, 24.8 (2
overlapping signals) ppm; IR (KBr) νmax 3119 (br), 2928 (m),
2855 (m, OH), 1784 (m, CdO), 1672 (s, CdO), 1575 (s), 1492
(s), 1442 (s), 1280 (s), 1116 (m) cm-1; MS (ES-, 100% CH3CN,
30 V) m/z ) 279 (M-); MS (FAB+, NBA) m/z ) 280 (M+, 40%),
281 (MH+, 65%); HR-MS (FAB+ of M+, NBA) m/z calcd for
C14H21N2O4 281.1501,found 281.1497.Anal.Calcd for C14H20N2O4;
C, 59.99; H, 7.19; N, 9.99. Found: C, 59.63; H, 7.62; N, 10.11.
3-Meth oxy-4-N,O-d im eth yla m in ocyclobu t-3-en e-1,2-d i-
on e (9a ). 3,4-Dimethoxycyclobut-3-ene-1,2-dione (2a ) (620 mg,
4.36 mmol) was dissolved in MeOH (2.5 mL) and free base
N,O-dimethylamine (1.7 equiv, 11.0 mL of a solution 0.7 M in
MeOH) was added. The methanolic free base N,O-dimethyl-
amine solution was prepared by addition of a methanolic KOH
solution (0.43 g in 4 mL MeOH) to a solution of N,O-
dimethylammonium hydrochloride in MeOH (0.75 g in 7 mL
MeOH), followed by filtration of the precipitated KCl. The
reaction mixture was stirred at ambient temperature until all
starting material was consumed (∼1 h, monitored by TLC).
The solvent was removed, and the resulting yellow residue was
taken up in hot methanol and then cooled to 0 °C. Crystals
were formed and filtered off using a glass frit (M). The yellow
crystals were washed with cold petroleum ether 30-60 and
dried under vacuum at ambient temperature to afford 9a (0.68
g, 91% yield) as yellowish crystals: Rf ) 0.37 (silica-CH2Cl2);
mp ) 112-114 °C (uncorrected) (lit. 101-103 °C);20 1H NMR
(DMSO-d6, 400 MHz, δ) 4.31 (s, 3H), 3.75 (s, 3H), 3.38 (s, 3H)
ppm; 13C NMR (DMSO-d6, 100 MHz, δ) 185.3, 180.9, 175.9,
169.0, 62.7, 61.2, 37.8 ppm; IR (KBr) νmax 1799 (m, CdO), 1715
(s, CdO), 1625 (s), 1497 (s), 1398 (s) cm-1; MS (FAB+, NBA)
m/z 171 (M+, 27%), 172 (MH+, 100%), 210 (MK+, 15%); HR-
MS (FAB+ of M+, NBA) m/z calcd for C7H10NO4 172.0610,
found 172.0610.
3-Mor ph olin o-4-N,O-dim eth ylam in ocyclobu t-3-en e-1,2-
d ion e (10). Squarate ester hydroxamate 9a (160 mg, 0.94
mmol) was dissolved in MeOH (5 mL), and morpholine (0.5
mL, 5.7 mmol) was added at ambient temperature. Once all
the starting material was consumed (TLC control) and a white
precipitate had formed, the product was filtered off, yielding
10 as white solid (130 mg, 62% yield): Rf ) 0.74 (silica-CH3-
1
CN); mp ) 161-162 °C dec (uncorrected); H NMR (DMSO-
d6, 400 MHz, δ) 3.80 (t, J ) 4.3 Hz, 4H), 3.72-3.69 (m, 7H),
3.44 (s, 3H) ppm; 13C NMR (DMSO-d6, 100 MHz, δ) 182.4,
179.3, 166.7, 164.4, 66.5, 60.2, 48.9, 37.7 ppm; IR (KBr) νmax
1785 (m, CdO), 1675 (s, CdO), 1601 (s), 1499 (s), 1287 (m),
1107 (m) cm-1; MS (ES+, 100% CH3CN, 30 V) m/z ) 227 (M+);
MS (FAB+, NBA) m/z ) 226 (M+, 22%), 227 (MH+, 35%), 265
(MK+, 5%); HR-MS (FAB+ of M+, NBA) m/z calcd for
C
10H15N2O4 227.1037, found 227.1030.
3-H yd r oxy-4-N-isop r op yla m in ocyclob u t -3-en e-1,2-d i-
on e (11b). Squarate ester hydroxamate 5b (160 mg, 0.86
mmol) was dissolved in aq HCl (3.6%, 10 mL) and stirred at
ambient temperature. Once all the starting material was
consumed (TLC control, ∼2 days), the solution was taken to
dryness by rotary evaporation, yielding a yellowish residue.
The residue was taken up in aq HCl (1M, 10 mL) and extracted
with ether (3 × 20 mL). Isolation of the organic phase and
evaporation of solvent afforded the hydrolyzed product 11b
1
(0.10 g, 70% yield): mp ) 233-235 °C dec (uncorrected); H
NMR (DMSO-d6, 400 MHz, δ) 8.35 (d, J ) 7.6 Hz, 1H), 3.94
(m, 1H), 1.14 (t, J ) 7.2 Hz, 6H) ppm; 13C NMR (DMSO-d6,
100 MHz, D1 ) 14 s, LB ) 5 Hz, δ) 187.0 (v br), 186.1, 184.7
(v br), 174.8, 47.9, 25.1 ppm; IR (KBr) νmax 3225 (br), 2979 (br),
2638 (br), 1818 (m), 1689 (s, CdO), 1628 (s, CdO), 1535 (s),
1464 (s), 1425 (s), 1337 (s), 1127 (m) cm-1; MS (ES+, 100%
CH3CN, 30 V) m/z ) 156 (MH+). Anal. Calcd for C7H9NO3; C,
54.19; H, 5.85; N, 9.03. Found: C, 53.80; H, 5.61; N, 9.14.
3-H yd r oxy-4-N-cycloh exylh yd r oxyla m in ocyclob u t -3-
en e-1,2-d ion e (11c). Prepared in 60% yield (110 mg, 0.56
mmol) from 5c and aq HCl (3.6%, 10 mL) as described for 11b:
mp ) 233-235 °C dec (uncorrected); 1H NMR (DMSO-d6, 400
MHz, δ) 8.38 (d, J ) 7.7 Hz, 1H), 3.6 (s, 1H), 1.81-1.04 (m,
10H) ppm; 13C NMR (DMSO-d6, 100 MHz, D1 ) 10 s, δ) 187.0
(v br), 186.2, 184.7 (v br), 174.9, 55.0, 35.3, 26.8, 26.4 ppm; IR
(KBr) νmax 3248 (br), 2932 (br), 2659 (br), 1818 (m), 1677 (br,
s), 1534 (s), 1426 (s), 1365 (m), 1092 (m), 803 (m) cm-1; MS
(ES+, 100% CH3CN, 30 V) m/z ) 196 (MH+). Anal. Calcd for
C10H13NO3; C, 61.53; H, 6.71; N, 7.17. Found: C, 61.22; H, 6.94;
N, 6.95.
3-Hyd r oxy-4-N-b en zyla m in ocyclob u t -3-en e-1,2-d ion e
(11d ). Prepared in 96% yield from 3c and aq HCl (3.6%, 10
mL) as described for 11b (0.9 mmol scale): mp ) 205-210 °C
dec (uncorrected); 1H NMR (DMSO-d6, 400 MHz, δ) 8.83 (s,
1H), 7.35 (m, 5H), 4.58 (d, J ) 6.2 Hz, 2H) ppm; 13C NMR
(DMSO-d6, 100 MHz, D1 ) 10 s, δ) 184.7 (v br peak overlap-
ping sharp signal), 183.8 (v br), 173.7, 138.6, 128.5, 127.5, 46.9,
15.1 ppm; IR (KBr) νmax 3260 (br), 1810 (m), 1680 (s), 1567
(s), 1510 (s), 1341 (m), 1261 (m), 1082 (m), 912 (m), 694 (m)
cm-1; MS (ES+, 100% CH3CN, 30V) m/z ) 204 (MH+); MS
(ES-, 100% CH3CN, 30V) m/z ) 202 (M-).
3-Met h oxy-4-m et h oxya m in ocyclob u t -3-en e-1,2-d ion e
(9b). 3,4-Dimethoxycyclobut-3-ene-1,2-dione (2a ) (530 mg, 3.73
mmol) was dissolved in MeOH (3.5 mL) and free base meth-
oxyamine (3.0 mL of a solution 1.25 M in MeOH, prepared as
described for 9a from 0.21 g KOH in 1 mL of MeOH and 0.31
g methoxyammonium hydrochloride in 2 mL of MeOH) was
added. The mixture was stirred at ambient temperature until
all starting material was consumed (∼24 h, monitored by TLC).
The solution was taken to dryness by rotary evaporation.
Column chromatography (silica-CH2Cl2) was used to isolate
and purify 9b as a white solid (120 mg, 20% yield): Rf ) 0.24
(silica-CH2Cl2/20% CH3CN); mp ) 110-112 °C (uncorrected)
(lit. 103-104 °C);20 1H NMR (DMSO-d6, 400 MHz, δ) 4.28 (s,
3H), 3.70 (s, 3H) ppm; 13C NMR spectrum could not be recorded
because the compound decomposed over the course of 12 h in
DMSO-d6; IR (KBr) νmax 3167 (br, NH), 1803 (m, CdO), 1703
Ack n ow led gm en t. We thank the donors of the
Petroleum Research Fund (PRF), administered by the
American Chemical Society (ACS), the Research Foun-
dation at the University of Connecticut (C.B. and N.L.),
and Saint Mary’s University (H.J .) for financial support.
We thank Andreas Decken, University of New Bruns-
wick, for diffraction data collection for 6c.
Su p p or t in g In for m a t ion Ava ila b le: 13C and 1H NMR
spectra of 5a -c, 6a -c, and 11b-d and details of the X-ray
crystallographicdetermination (tabulated crystaldata,ORTEPS,
and a CIF file) of 5b and 6c. This material is available free of
charge via the Internet at http://pubs.acs.org.
(s, CdO), 1613 (s), 1514 (m), 1371 (s), 1011 (m), 938 (m) cm-1
;
MS (ES-, 100% CH3CN, 30 V) m/z ) 156 (M-); MS (FAB+,
NBA) m/z ) 158 (MH+, 100%); HR-MS (FAB+ of M+, NBA)
m/z calcd for C6H8NO4 158.0453, found 158.0456.
J O035175G
J . Org. Chem, Vol. 68, No. 24, 2003 9241