11332 J. Am. Chem. Soc., Vol. 122, No. 46, 2000
BoiadjieV and Lightner
35 in 67% isolated yield. And, like the monopyrroles 38 and
40, averaged coupling constants (3J) in the propionate chain of
35 (as well as 28) indicated conformational mobility.
the presence of HMPA at -78 °C and allowed to reach ambient
temperature. The conditions were optimized by using excess
enolate, as any unreacted 44 present during the work up
produces a dipyrrylmethane (see above) that is removed only
by a tedious chromatography. The 2-carboethoxy group of 41
was saponified in a refluxing solution of NaOH in aqueous
ethanol, and the product was reacted directly with 46 in refluxing
methanol to give dipyrrinone 32 in 80% yield. Both monopyrrole
41 and dipyrrinone 32 exhibited restricted motion in the methyl
propionate chain, as evidenced by its 1H NMR vicinal coupling
constants: 3J ) 2.9 and 12.2 Hz in 41 and 2.8 and 12.0 Hz in
32.
Alkylation of 43 with isopropyl iodide in THF, as above,
failed to give the expected isopropyl pyrrole 39. However, with
added HMPA cosolvent, which had been shown to promote
alkylations with secondary alkyl halides,19 a 83% yield of pure
39 was achieved. Conversion of 39 to 30, as for 28 and 35,
proceeded uneventfully in 48% yield. Unlike 38 and 40, analysis
of the 1H{1H} coupling constants of the propionate ester chain
of 39 revealed a strong conformational preference. One of the
diastereotopic â-methylene protons of 39 showed 3J ) 4.7 Hz,
indicative of gauche vicinal local stereochemistry. The other
showed 3J ) 10.5 Hz, characteristic of an antiperiplanar
orientation. The same sort of conformation restriction was
observed in the propionate chain of dipyrrinone 30 but not in
dipyrrinones 28 and 35.
The procedure used above to introduce the propionate R-alkyl
groups in 38-40 could not be used to introduce the tert-butyl
and phenyl substituents. The preparation of 41 and 42 required
alternative methods. In one approach toward converting 43
to 42, we investigated a Friedel-Crafts alkylation of methyl
2,4-dimethyl-3-(2-bromo-2-carbomethoxyethyl)pyrrolecar-
boxylate18 with benzene in the presence of anhydrous AlBr3.
While the desired substitution of phenyl for bromine was
accomplished, the AlBr3 also apparently cleaved the pyrrole
2-carbomethoxy group to give the 2-carboxylic acid. The yield
of desired product was too low (39%), and its purification was
too tedious for our purposes; so, an alternative procedure was
developed for preparing both 42 and 41.
Rather than attempting to alkylate or arylate a pyrrole
3-propionate ester, the successful alternative pathway involved
reaction of chloromethylpyrrole 44 with the ester R-anions
prepared by treating either methyl neohexanoate20 (to give 41)
or methyl phenylacetate (to give 42) with LDA. We were
pleased to find that these reactions proceeded smoothly and
afforded satisfactory isolated yields (45-51%) of 41 and 42
despite severe steric hindrance at the chloromethyl group from
the adjacent ring methyls s a situation akin to the ortho steric
effect in 2,6-dimethylbenzene compounds21 where reactions at
carbon attached to C(1) are extremely difficult. (Chloromethyl)-
pyrrole 4422 is very reactive and tricky to prepare. Its synthesis
had been reported earlier only once22 and involved reaction of
the readily available 4-H pyrrole 4523 with paraformaldehyde
in glacial acetic acid saturated with HCl gas. The reaction is
capricious and must be controlled carefully, as the product (44)
is very sensitive toward solvolysis. Failure to control the reaction
properly leads to formation of bis(5-carboethoxy-2,4-dimethyl-
3-pyrryl)methane as the major product, a substance first
synthesized long ago by Fischer and Nenitzescu.24
Introduction of an R-phenyl group in 42 was achieved as
above for tert-butyl. Thus, the enolate formed by reaction of
methyl phenylacetate with LDA in THF at -50 °C was reacted
with â-(chloromethyl)pyrrole 44 in THF-HMPA at -78 to -40
°C to give a 51% of 42. After saponification, 42 was condensed
with 46 in refluxing methanol to afford pure dipyrrinone 34 in
1
58% isolated yield. The H NMR spectra of both 42 and 34
gave no evidence for restricted motion in the propionic ester
chain, as had been observed for the isopropyl and tert-butyl
analogues, 39, 41, 30, and 32.
With the required left-half dipyrrinones (28, 30, 32, 34, 35)
now available, we were ready to take them on the path to verdin
monoesters 13-18 by oxidative coupling with dipyrrinone acid
36 (Scheme 1). However, while we anticipated (and observed)
no difficulty in saponifying verdins 13, 14, and 18 to verdin
diacids (or rubin diacids), we were concerned that the saponi-
fication or hydrolysis of the more hindered verdin monoesters
(15, 16, 17) would prove resistant to or require conditions that
would destroy the sensitive verdin or rubin structure. (For
example, nucleophilic attack at C(10) in verdins is well
documented.15) In such cases, it seemed wiser to accomplish
the ester to acid conversion on the more robust dipyrrinones,
e.g., 30 f 29, 32 f 31 and 34 f 33 (Schemes 1 and 3) and
then oxidatively couple these left half dipyrrinone acids with
dipyrrinone ester 37. The resulting verdin monoesters (15, 16,
17) were expected (and found) to be easily saponified to the
corresponding verdin diacids, which were reduced by NaBH4
to form rubin diacids (Scheme 2).
As anticipated, saponification of 30, 32, and 34 proved
difficult or impossible. Using the same conditions as in the
conversion of 37 to 36, refluxing in a 10% solution of NaOH
in aqueous ethanol for 4 h, was insufficient for converting 30
to 29, leaving ∼20% unreacted 30, and requiring reflux for an
additional 8 h. Similarly, 34 was converted to 33, but 32 resisted
saponification. More than a dozen different experiments were
attempted to convert rather hydrophobic 32 to 31, including
saponification by aqueous NaOH in refluxing mixtures of
ethanol or pyridine, glyme, or diglyme. Heating 32 in DMSO-
aqueous NaOH at >100 °C achieved eventual saponification,
but the reaction conditions also led to copious quantities of
silicate formed by partially dissolving the reaction flask. And
since the sodium salt of 31 is insoluble in water, we could not
isolate 31 from the silicate. However, by using an SN2
displacement of the carboxylate anion from the methyl ester
by reaction of 32 with anhydrous LiI in refluxing (>170 °C)
s-collidine,14 we achieved complete demethylation of the ester
and isolated a 99% yield of acid 31 by partially removing the
solvent and diluting with aqueous HCl to precipitate the product.
To prepare 41 and 32, the ester enolate of methyl neohex-
anoate was generated in THF at -5 °C and reacted with 44 in
(19) MacPhee, J. A.; Dubois, J.-E. J. Chem. Soc., Perkin Trans. 1 1977,
694-696.
(20) Nilson, A.; Carlson, R. Acta Chem. Scand. 1980, B34, 621. (b) Bott,
K.; Hellmann, H. Angew. Chem., Int. Ed. Engl. 1966, 5, 870-874. (c)
Botteron, D. G.; Shulman, G. P. J. Org. Chem. 1962, 27, 1059-1061. (d)
Traynham, J. G.; Battiste, M. A. J. Org. Chem. 1957, 22, 1551-1553.
(21) Fuson, R. C.; Walker, J. T. J. Am. Chem. Soc. 1930, 52, 3269-
3275. (b) Newman, M. S.; Connor, H. E. J. Am. Chem. Soc. 1950, 72,
4002-4003.
(22) McDonald, S. F.; Markovac, A. Can. J. Chem. 1965, 43, 3247-
3252.
(23) Robinson, J. A.; McDonald, E.; Battersby, A. R. J. Chem. Soc.,
Perkin Trans. 1 1985, 1699-1709.
(24) Fischer, H.; Nenitzescu, C. Liebigs Ann. Chem. 1925, 443, 113-
129.
13C NMR Spectra and Molecular Structure. Pairs of meso-
bilirubin diastereomers involving isomerism based on substitu-
tion at the R-position of the C(8) propionic acid and the â′-
position of the C(12) propionic acid have been distinguished