2-Methyl-(1Z,3E)-butadiene-1,3,4-tricarboxylic acid, " isoprenetricarboxylic acid"

Article abstract of DOI:10.1021/jo0507892

The synthesis of the title compound 7 from ethyl glyoxylate and dimethyl and diethyl β-methylglutaconate is described along with its physical properties that suggest its inability to assume a cis-dienoid structure due to steric hindrance between the methyl and carboxyl groups.

Full text of DOI:10.1021/jo0507892

SCHEME 1
2-Methyl-(1Z,3E)-butadiene-1,3,4-tricarboxylic
Acid, “Isoprenetricarboxylic Acid”
Mayer B. Goren,† Edward A. Sokoloski,‡ and
Henry M. Fales*,‡
6790 East Cedar Avenue, Denver, Colorado, and Laboratory
of Applied Mass Spectrometry, National Heart, Lung, and
Blood Institute, Bethesda, Maryland 20892
hmfales@helix.nih.gov
Received April 18, 2005
the ethyl ester of ethane-1,1,1-triacetic acid (2) were not
successful,5 nor did any of the reaction products bear any
similarity to synthetic tricarboxylic acid 7, and the
pursuit of 4 by this route was abandoned. Interestingly,
however, 7 failed to exhibit the typical cis-dienoid ultra-
violet absorption seen in, for example, cis,cis-â-methyl-
muconic acid (10)3 obtained by the peracid oxidation of
p-cresol.6 Examination of Fisher-Taylor-Hirschfelder
models strongly suggested that steric repulsion of the
2-methyl and 3-carboxyl groups was responsible (see 7b
below), but Professor Woodward was uneasy about this
explanation. To clarify this matter, and in view of the
more recent work on the stereochemistry of the lacton-
ization of 10,7a,b we have repeated the preparation of 7
and determined its properties using modern techniques,
ultimately reaffirming its structure as reported in the
M.B.G. thesis.
First attempts to prepare 7 depended on ring-opening
reactions of the type used successfully by Pauly et al. in
the conversion of 3-nitro-p-cresol to cis,cis-â-methyl-
muconic acid (10) with sulfuric acid4b or peracetic acid
ring opening as described by Bo¨eseken and Metz.9 This
would provide the advantage of ensuring fixed stereo-
chemistry of the double bonds. However, when the
peracid reaction was applied to 2-methyl-3-hydroxyben-
zoic acid (cresotinic acid), or even 2-methyl-3,4-dihy-
droxybenzoic acid, no such product was obtained, likely
due to the deactivating effect of their carboxyl groups.
Many other approaches were also attempted, and the
reader is referred to the original thesis3 for details.
Attention was then turned to base-catalyzed condensa-
tion reactions of glutaconic esters and aldehydes first
discussed by Feist andBeyer.10 Thus, diethyl (or here,
dimethyl) â-methylglutaconate (5) was treated with ethyl
glyoxylate (6) in methanolic KOH to produce an ether-
extractable, somewhat unstable acid (Scheme 2). There
The synthesis of the title compound 7 from ethyl glyoxylate
and dimethyl and diethyl â-methylglutaconate is described
along with its physical properties that suggest its inability
to assume a cis-dienoid structure due to steric hindrance
between the methyl and carboxyl groups.
In the late 1940s, Professor Robert B. Woodward was
concerned with possible synthetic approaches to tetra-
hedrane, a compound still awaiting synthesis although
its tetrakis tert-butyl derivative has been prepared.1
At that time, the structures of the cage compound 4
(Scheme 1) reported by Beesley and Thorpe2 appeared
to contain this ring system,and Woodward assigned one
of his students (Harold O. Larson) to attempt repetition
of its synthesis (1 > 4, in brackets). However, Professor
Woodward clearly doubted the validity of the intermedi-
ate 3. In its place, he postulated3 that alkali fusion of 2
had first led to a bromocyclopropane intermediate that
then lost the second bromine, rearranging to form three
isomeric acids of 2-methyl-1,3-butadiene-1,3,4-tricarboxy-
lic acid (7, Scheme 2), referred to at the time as “isopren-
etricarboxylic acid”, following earlier nomenclature.4a,b
The geometrical isomers of 7 were expected to have
properties consistent with those described for the three
isolated isomers of 3, and so he assigned a second
student, one of us (M.B.G.), to attempt a more straight-
forward synthesis of at least one of these compounds.
This synthesis was finally accomplished3 through the
strong base-catalyzed reaction of diethyl â-methylgluta-
conate (5) and ethyl glyoxylate (6) (Scheme 2). On the
other hand, many attempts by Larson to repeat the
Beesly and Thorpe work which depended on the prepara-
tion and alkaline fusion of the dibromination product of
(5) Larson, H. O. Ph.D. Thesis, Harvard University, Cambridge, MA,
May 1950.
† 6790 East Cedar Avenue, Denver, CO.
(6) Elvidge, J. A.; Linstead, R. P.; Sims, P. J. Chem. Soc. 1951, 3386.
(7) (a) Chen, B.; Kirby, G. W.; Rao, G.; Cain, R. B. J. Chem. Soc.,
Perkin Trans. 1 1996, 1153. (b) Freer, A. A.; Kirby, G. W.; Rao, G. V.;
Cain, R. B. J. Chem. Soc., Perkin Trans. 1 1996, 2111.
(8) Cawley, J. D. J. Chem. Soc. 1955, 77, 4125.
(9) Bo¨eseken, J.; Metz, C. F. Rec. Trav. Chim. Pays-Bas Belg. 1935,
54, 345.
‡ National Heart, Lung, and Blood Institute.
(1) Maier, G. Angew. Chem., Int. Ed. Engl. 1988, 27, 309-446.
(2) Beesley, R. M.; Thorpe, J. F. J. Chem. Soc. 1920, 117, 591.
(3) Goren, M. B. Ph.D. Thesis, Harvard University, Cambridge, MA,
September 1949.
(4) (a) Pauly, H.; Gilmour, R.; Will, G. Ann. Chem. 1914, 403, 138.
(b) Pauly, H.; Will, G. Ann. Chem. 1918, 416, 1
(10) Feist, F.; Beyer, O. Ann. 1906, 345, 117.
10.1021/jo0507892 CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/10/2005
J. Org. Chem. 2005, 70, 7429-7431
7429

SCHEME 2
SCHEME 3
seems little doubt that this is in fact the desired “iso-
prenetricarboxylic acid” 7 since it has the correct molec-
ular weight by negative-ion electrospray, losing succes-
sively three molecules of CO2 on MS/MS, and exhibits a
reasonable NMR (see below). On standing open to the
atmosphere, the dry acid decomposed with evolution of
gas (CO2) in a matter of days, or more rapidly on heating
in solution, to yield a product from which two lactones
could be isolated by trituration with ether. The least
soluble is the decarboxylation product, lactone 8, first
described by Pauly et al.,4a and as described in the M.B.G.
thesis, it was found to be identical with a sample of
authentic 8. The second, more soluble product is the
dicarboxylic acid lactone 9. Again, both of these lactones
show NMR and mass spectra that are in accord with their
proposed structures and recall the very similar lacton-
ization found to occur stereospecifically by Kirby et al.
when 3-methyl-cis,cis-muconic acid (10) is produced in
the presence of certain bacteria and fungi7a,b where these
workers were also able to isolate lactone 11, closely
related to 9.
The 1Z,3E orientation of the double bonds suggested
by its NMR and shown in 7a receives considerable
support from the easy formation of these lactones, i.e.,
the 1Z linkage remains in 8 after decarboxylation while
the 3E is apparent in 9. Thus, the C1 proton is differenti-
ated from the C4 proton by the presence of allylic coupling
to the methyl group in the latter, and its chemical shift
at δ 6.61 in both the acid and its methyl ester suggests
that it is deshielded by the adjacent cis carboxyl at C2 in
the 1E conformation.11 As expected for the 3Z conforma-
tion, the C2 methyl group does not appear to be similarly
deshielded at δ 2.05, but the chemical shift range
available in ref 8 is too wide to be convincing. Similarly,
allylic coupling constants between the H and CH3 groups
were found to be virtually identical in that reference (1.60
Hz).11
The apparent UV maximum of acid 7a is at 210 nm (ꢀ
15 900) like fumaric and crotonic acids rather than 265
nm (ꢀ 17 800) like cis,cis-â-methyl muconic acid 10,6 or
muconic acid (265 nm, 25 100) or cis,cis-â-carboxymuconic
acid12 (254 nm, 8200). This shows that there is, in fact,
considerable steric hindrance between the 2-methyl and
3-carboxy groups that prevents the double bonds from
attaining coplanarity in the cis-dienoid form 8a respon-
sible for absorption at that wavelength. The dimethyl
ester of lactone 9 could be easily opened with sodium
methoxide,3 and the resulting dimethyl ester acid 12
apparently retains the same stereochemistry as 7 since
it also shows a UV maximum at only 210 nm (ꢀ 15 900).
To further examine this matter, 3-methyl-4-phenylb-
utadiene-1,2-dicarboxylic acid (benzal-â-methylglutaconic
acid) (13a,b, Scheme 3), reported earlier by Feist and
Beyer,10 was prepared by strong base-catalyzed conden-
sation of benzaldehyde and diethyl â-methylglutaconate.3
Based on the known cis orientation of the aryl and R
groups obtained in these condensations as found by
Cawley,8 13 was expected to have the 1E,3Z stereochem-
istry shown in 13a and 13b. In this case, the cis-dienoid
rotamer 13b is hindered both by the interaction of the
C3 methyl and the C2 carboxyl mentioned above, and
(12) MacDonald, D. L.; Stanier, R. Y.; Ingraham, J. L. J. Biol. Chem.
1954, 210, #2, 809.
(11) Jackman, L. M.; Wiley: R. H.J. Chem. Soc. (London) 1960 2886.
7430 J. Org. Chem., Vol. 70, No. 18, 2005

MHz, THF-d): δ 24.7, 119.7, 125.9, 148.8, 152.6, 165.6, 166.1,
166.6. IR (Nujol) (cm-1) 1460, 1377. MS positive-ion electrospray
in water: m/z 223 (M + Na)+ and 239 (M + K)+; negative-ion
electrospray: m/z 199 (M - H)-. MS/MS: m/z 155, 111, and 66
(loss of 1-3 CO2). The acid was not particularly stable, and it is
essential that it be stored under vacuum to prevent decomposi-
tion. If it is merely air-dried and placed in a bottle, pressure
develops over a matter of days as the compound undergoes
decarboxylation, ultimately forming lactones 8 and 9 (see below).
The trimethyl ester of 7 was prepared with ethereal diaz-
omethane as an oil which was not crystallized but molecularly
distilled at 0.3 mm. IR (CHCl3, cm-1): νmax 1709. UV (EtOH):
214m (17 800). 1H NMR (200 MHz, CDCl3): δ 5.75 (br s, 1H),
6.61 (br s, 1H), 2.05 (br s, 1H), 3.78, 3.69, 3.58 (3s, 3 CH3O).
Anal. Calcd for C11H14O6: C, 54.51; H, 5.78. Found: C, 54.12;
H, 5.73.
more seriously by the C1 carboxyl and the benzene ring.
However, even the trans-dienoid form 13a is subject to
hindrance from the C2 carboxyl and the benzene ring,
preventing full planarity of the ring and trans-diene. In
fact, 13 showed only the two maxima at 220 nm(ꢀ 17 800)
and 274 nm (ꢀ 15 900) expected for trans-cinnamic acid.
(204-215 nm (ꢀ 15 160) and 273 nm (ꢀ 20 000).3 Simi-
larly, Cawley working on the chemistry of vitamin A,8
found that 2-methyl-4-phenyl-(1E,3E)-butadiene-1,3-di-
carboxylic acid (14)13 shown in its trans-dienoid form
showed only trans-cinnamic acid absorption (276 nm, ꢀ
17 600) rather than that expected for a phenylbutadi-
ene.13 This spectral difference was assigned to the steric
hindrance between the o-hydrogen on the phenyl ring and
the methyl group in the trans-dienoid rotamer 14, an
effect that might be expected to be smaller than in the
case of 13a by comparing methyl and carboxyl group
sizes.
Lactone 8. The residue from evaporation of the above
chloroform washings was triturated with ether to produce a
white solid that was recrystallized from the same solvent, mp
124-126 °C either alone or on admixture with an authentic
sample obtained from D. Taub prepared according to Pauly et
al.4a Anal. Calcd for C7H8O4: C, 53.85; H, 5.13. Found: C, 53.37;
1
H, 5.17. H NMR (200 MHz, D2O): δ 2.05 (d, J ) 1.4 Hz, 3H),
Experimental Section
2.29 (dd, J ) 9.0, 15.6 Hz 1H), 2.72 (dd, J ) 15.6, 4.3 Hz, 1H),
5.24 (dd, J ) 9.0, 4.3 Hz, 1H), 5.80 (dq, J ) 1.4, 1.4 Hz 1H). 13
C
The preparation of the title compound is a slight modification
of that described in the M.B.G. thesis.3 The available dimethyl
ester of â-methyl glutaconic acid was used in place of the diethyl
ester used in the thesis. Several other preparations, mp, and
data other than NMR are also taken from the thesis.
NMR (50.3 MHz, D2O): δ 14.2, 40.4, 85.0, 116.2, 173.9, 177.8.
178.0.
2-Methyl-1,3-butadiene-1,3,4-tricarboxylic Acid Mono-
lactone (9). The ether filtrates from the above crystallization
were evaporated and a few drops of methanol added. Addition
of chloroform caused the lactone to precipitate and the product
was recrystallized from the same solvent mixture to give
polyhedra. Mp: 151-153 °C. IR (CHCl3, cm-1): νmax 1709, 1751.
Anal. Calcd for C8H8O6: C, 48.01; H, 4.03. Found: C, 47.71; H,
3.91. Neutralization equivalent with NaOH: calcd 100.07, found
100.7.
2-Methyl-(1E,3Z)-butadiene-1,3,4-tricarboxylic acid, “Iso-
prenetricarboxylic acid” (7). Ten grams of dimethyl â-me-
thylglutaconate (5) and 14.8 g of 50% ethyl glyoxylate (6) in
toluene were added dropwise under nitrogen to 20 g of KOH
dissolved in 150 mL of cold (0 °C) anhydrous methanol over 45
min and the reaction mixture allowed to stand overnight at room
temperature. A precipitate formed in the bottom of the flask from
which the original solvents were decanted, and the residue was
washed with 20 mL of cold methanol. That was likewise
decanted, the solid was dried under vacuum and taken up in 12
mL water, 100 mL of ether was added, and the flask was cooled
in an ice bath. Concentrated HCl was added until the Congo
red point was obtained, the solution was extracted with six 100
mL portions of ether, and the extracts were dried over 30 g of
anhydrous sodium sulfate. The ether was evaporated at 30 °C
and the residue dried with an oil pump, yielding 4.00 g of white
crystals, mp 113-116 °C. This was washed with 20 mL of
chloroform to remove small amounts of lactones 8 and 9 (see
below) and the acid again pumped dry with an oil pump after
which it melted at 130-146 °C dec with foaming. Recrystalli-
zation was quite difficult, but fine needles could be prepared by
dissolving the acid in excess ether and reducing the volume.
Mp: 151-152 °C. The UV shows only the low wavelength
absorption expected for R,â-unsaturated acids at 214 m (ꢀ
17 800).; IR (ether, cm-1): νmax 1709. Anal. Calcd for C8H8O6:
C, 48.01; H, 4.03. Found: C, 47.96; H, 4.25. IR (ether): 1709
cm-1 (COOH); 1653, 1626 cm-1 (CdC). 1H NMR (200 MHz, THF-
d): δ 5.75 (br s 1H), 6.61 (s, 1H), 2.04 (br s 3H). 13C NMR (50.3
The dimethyl ester of 9 was prepared by treating the above
monolactone with either hydrogen chloride in methanol or excess
diazomethane; crystallized on trituration with ether. Mp: 44-
45 °C. IR (CHCl3, cm-1): νmax 1709, 1751. Anal. Calcd for
C
10H12O6: C, 52.63; H, 5.28. Found: C, 52.69; H, 5.29. 1H NMR
(200 MHz, CDCl3) δ 1.62 (s 3H), 2.92, (d, J ) 15.8 Hz, 1H), 3.04
(d, J ) 15.8 Hz, 1H) 3.52, (s, 3H), 3.82 (s, 3H). 13C NMR (50.3
MHz, CDCl3): δ 24.9, 40.9, 51.6, 52.6, 85.5, 126.3, 158.5, 160.9,
168.5, 169.3.
The dimethyl ester (200 mg) was allowed to stand with 23
mg of sodium dissolved in 10 mL of anhydrous methanol for 2
days, evaporated, and taken up in 5 mL of cold water. The
solution was extracted once with ether, acidified, and again
extracted with ether to produce the dimethyl ester acid 12
which was recrystallized from ether. Mp: 126-128 °C. Anal.
Calcd for C10H12O6: C, 52.63; H, 5.28. Found: C, 52.34; H, 5.26.
UV (dilute KOH): 214 m (ꢀ 17 800).
Benzal â-methylglutaconic acid (13) was prepared accord-
ing to ref 10 and recrystallized from aqueous ethanol. Mp: 186-
192° dec. NE C13H12O4: 117.9 (calcd 116.1). UV (dilute NaOH):
∼215 m (ꢀ 17 600), 274 m (ꢀ 17 400) [lit.8 UV 276 (ꢀ 17 600)].
(13) Braude, E. A. Ann. Rep. Chem. Soc. 1945, 45, 105.
JO0507892
J. Org. Chem, Vol. 70, No. 18, 2005 7431