Aromatic congeners of bilirubin: Synthesis, stereochemistry, glucuronidation and hepatic transport

Article abstract of DOI:10.1016/S0040-4020(01)00773-6

A new synthetic analog (1) of the bile pigment bilirubin-IXα (bilirubin, Fig. 1) with phenyl groups replacing vinyl was prepared by a constitutional scrambling reaction of a mixture of two new, symmetric phenylrubin analogs (2 and 3) of bilirubin-XIIIα and IIIα. The former (2) with two endo-phenyls, and the latter (3) with two exo-phenyls were synthesized by condensation of a dipyrrylmethane dialdehyde with appropriate methylphenylpyrrolinones, which were prepared in several steps from 4-methyl-3-phenyl-2-(p-toluenesulfonyl)pyrrole, obtained by the Barton-Zard pyrrole synthesis. Nuclear Overhauser effect ¹H NMR studies of 1-3 confirm that, like their bilirubin equivalents, these yellow-orange pigments adopt an intramolecularly hydrogen-bonded ridge-tile conformation. Reverse phase HPLC and TLC suggest that 3 is less polar than 2, and that 1 has intermediate polarity. Large differences in the induced circular dichroism spectra of 1-3 were found in pH 7.4 aqueous buffered solutions of human serum albumin. Despite the presence of bulky, lipophilic phenyl groups, 1-3 are metabolized like natural bilirubin in rats and require glucuronidation by the same enzyme for canalicular secretion from the liver into bile. However, there are striking qualitative differences between the three pigments in the ratio of mono- to diglucuronides formed. Phenyl substituents at the exo positions of the lactam rings diminish the proportion of diglucuronide more than phenyl substituents at the endo positions.

Full text of DOI:10.1016/S0040-4020(01)00773-6

TETRAHEDRON
Pergamon
Tetrahedron 57 %2001) 7813±7827
Aromatic congeners of bilirubin: synthesis, stereochemistry,
glucuronidation and hepatic transport
Justin O. Brower,a David A. Lightnera,p and Antony F. McDonaghb,p
aDepartment of Chemistry, University of Nevada, Reno, NV 89557-0020, USA
bG.I. Unit and the Liver Center, University of California, San Francisco, CA 94143-0538, USA
Received 11 June 2001;accepted 24 July 2001
AbstractÐA new synthetic analog %1) of the bile pigment bilirubin-IXa %bilirubin, Fig. 1) with phenyl groups replacing vinyl was prepared
by a constitutional scrambling reaction of a mixture of two new, symmetric phenylrubin analogs %2 and 3) of bilirubin-XIIIa and IIIa. The
former %2) with two endo-phenyls, and the latter %3) with two exo-phenyls were synthesized by condensation of a dipyrrylmethane dialdehyde
with appropriate methylphenylpyrrolinones, which were prepared in several steps from 4-methyl-3-phenyl-2-%p-toluenesulfonyl)pyrrole,
obtained by the Barton±Zard pyrrole synthesis. Nuclear Overhauser effect 1H NMR studies of 1±3 con®rm that, like their bilirubin
equivalents, these yellow-orange pigments adopt an intramolecularly hydrogen-bonded ridge±tile conformation. Reverse phase HPLC
and TLC suggest that 3 is less polar than 2, and that 1 has intermediate polarity. Large differences in the induced circular dichroism spectra
of 1±3 were found in pH 7.4 aqueous buffered solutions of human serum albumin.
Despite the presence of bulky, lipophilic phenyl groups, 1±3 are metabolized like natural bilirubin in rats and require glucuronidation by
the same enzyme for canalicular secretion from the liver into bile. However, there are striking qualitative differences between the three
pigments in the ratio of mono- to diglucuronides formed. Phenyl substituents at the exo positions of the lactam rings diminish the proportion
of diglucuronide more than phenyl substituents at the endo positions. q 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
Bilirubin is formed in large amounts in mammals but is
not excreted in bile or urine to any signi®cant degree.4
For excretion, it is metabolized by a hepatic glucuronosyl
transferase enzyme to isomeric monoglucuronides that are
excreted into bile. Although there are several glucuronosyl
Humans become jaundiced because of an inability to
ef®ciently eliminate the yellow pigment, bilirubin %Fig.
1),1,2 which is cytotoxic and a powerful anti-oxidant.3
Figure 1. Enzymic conversion of %blue-green) biliverdin to %yellow-orange) bilirubin, and formation of a bilirubin mono-glucuronide. Formation of mono-
glucuronides %and the diglucuronide) is an essential step in the hepatic elimination of bilirubin. The structures are drawn in a porphyrin-like shape, except for
the inset, where bilirubin is shown in its most stable conformation: intramolecularly hydrogen-bonded and folded like a ridge±tile. Hydrogen bonds are shown
by dashed lines.
Keywords: pyrroles;conformation;hydrogen bonding;circular dichroism.
Corresponding authors. Fax: 11-775-784-6804;e-mail: lightner@scs.unr.edu;tonymcd@itsa.ucsf.edu
p
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020%01)00773-6

7814
J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
novel lipophilic bilirubins %1±3, Fig. 2) in which the
enhanced lipophilicity stems from the replacement of
vinyl groups with relatively bulky phenyl groups. These
are the ®rst bilirubins containing aromatic substituents.
Despite their relative bulk, the phenyl substituents are
not expected to alter the three-dimensional structure and
hydrogen bonding depicted in Fig. 1, but their presence
could hamper the access of pigment to the active site of
UGT1A1.
2. Results and discussion
2.1. Syntheses
Preparation of the unsymmetric phenylrubin %1), the phenyl
analog of the natural bilirubin %IXa, Fig. 2) required the
syntheses %Scheme 1) of two symmetric phenylrubins %2
and 3), themselves phenyl analogs of the symmetric bili-
rubins-XIIIa and IIIa. The key synthetic step in the syn-
thesis of 1 was an acid-catalyzed, constitutional isomeri-
zation scrambling reaction that had been developed and
used successfully in other work.8 For ease of chromato-
graphic separation, the scrambling reaction was designed
to mix a dimethyl ester and a diacid so as to produce the
hybrid mono-methyl ester, which would be predicted to
have polarity intermediate between the more polar diester
and the less polar diacid and would thus be easily separated
chromatographically. Consequently, dimethyl ester 5 was
mixed and scrambled using acid catalysis with diacid 3 to
produce mono-methyl ester 4 in a mixture that was easily
separated by radial chromatography on silica gel.
Figure 2. Linear representations of bilirubin IXa, IIIa and XIIIa %solid
lines) and their analogs %1±3) with phenyls superimposed %dashed lines).
transferases in the liver, with frequent cross-reactivity for
substrates, only a speci®c isozyme %UGT1A1) glucuroni-
dates bilirubin in vivo in rats and humans.1±3 This enzyme
also accepts the two bilirubin monoglucuronides as
substrates, so that bilirubin diglucuronide is excreted in
bile along with the monoglucuronides. Presently little is
known about the mechanism of glucuronidation or about
the protein±pigment interactions at the catalytic site of the
enzyme. The high selectivity of bilirubin for UGT1A1 and
its double conjugation to a diglucuronide are unusual, and
it is surprising that bilirubin, which is an organic anion
at physiologic pH, should require glucuronidation for
excretion into bile. Its metabolic precursor biliverdin,
which is only two hydrogens different and has similar
pKas, can be eliminated rapidly from the liver into bile intact
without undergoing conjugation.4 This metabolic difference
between the two pigments is apparently related to their
lipophilicities, which are dependent on their three-dimen-
sional structures. Biliverdin adopts lock±washer confor-
mations, but bilirubin and its anions fold into ridge±tile
shapes %Fig. 1) that are stabilized by a network of intra-
molecular hydrogen bonds.5,6
The precursor symmetric phenylrubins 2 and 5 were
prepared by piperidine-catalyzed condensation of 3-methyl-
4-phenylpyrrolin-2-one %6) with the known 5,50-diformyl-
dipyrrylmethane 13,7,9,10 to give 5, or with diacid 14
%derived from 13) to give 2. Similarly, symmetric phenyl-
rubin 3 was prepared by reaction of 4-methyl-3-phenyl-
pyrrolinone 7 with 14 to give 3. Curiously, although
condensation of 13 with 6 in re¯uxing methanolic sodium
hydroxide also gave rubin product 2 in ,35% yield, the
same reaction failed to convert a mixture of 13 and 7 to 3.
The two pyrrolinones %6 and 7) were synthesized from a
common monopyrrole precursor %10), which was prepared
in 65% yield by a Barton±Zard pyrrole-forming reaction
between p-toluenesulfonylmethyl isocyanide %TosMIC)10
and 2-nitro-1-phenylpropanol acetate %15),11 which was
prepared from commercially available, inexpensive benz-
aldehyde and nitroethane. A similar Barton±Zard reaction
to yield 11 directly from 3-nitro-3-phenyl-2-propanol
acetate was less attractive because synthesis of the latter
requires coupling sensitive a-nitrotoluene with acetalde-
hyde and because previous studies11 had shown that 10
could be converted nearly completely into 11 by TFA-cata-
lyzed isomerization. Isomerization of 10 smoothly led to an
equilibrium mixture that favored 11 %10:1) and from which
pure 11 could be isolated by fractional crystallization.
Tosylpyrroles 10 and 11 could be converted to the desired
pyrrolinones 6 and 7 by either of two routes. In one, pyrro-
linone 8 was prepared from 10, ®rst by bromination %to give
12), then by treatment with TFA and water. Applying this
In a previous paper,7 we reported the synthesis of lipophilic
bilirubins in which the increased lipophilicity derived from
n-butyl substituents. The presence of n-butyl groups at the
3 and 17 %endo) positions on the lactam rings did not have
a large effect on glucuronidation and excretion. Biliary
excretion was somewhat slower than for the corresponding
ethyl substituted compound and the proportion of mono to
diglucuronide formation was increased. n-Butyl groups at
the 2 and 19 %exo) positions, adjacent to the lactam carbo-
nyls, had a much larger effect. Biliary excretion was greatly
retarded and only the monoglucuronide was detectable in
bile. In the present work, we describe the syntheses of three

J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
7815
Scheme 1. aNaOH/CH3OH±H2O/re¯., then HOAc; bDMSO/HCl; cpiperidine; d6 M aq KOH±CH3OH, then HCl; eNaBH4/EtOH; fTFA/CH2Cl2 %1:9); gH2O2±
j
HOAc; hTFA±H2O %5:1); iPhNMe31Br32/CH2Cl2; Guanidine/THF±iPrOH.
method for converting 11 to 9, however, led to highly
variable %1±50%) yields of 7. In an alternate route, direct
oxidation of 11 to 9 in 25% yield was achieved with H2O2 in
HOAc. Removal of the tosyl group from either 8 or 9
proceeded smoothly and in high yield by reaction with
NaBH4 in CH3OH to afford the desired pyrrolinones, 6
and 7, respectively.
2.2. Constitutional structure and conformations
Although the structures of 2 and 3 follow logically from
their syntheses from well-characterized smaller molecules,
and the structure of 1 follows from the structures of 2 and 3
and the well-studied constitutional scrambling reaction,8 the
structures were con®rmed from their 13C NMR spectra in
%CD3)2SO %Table 1). For comparison purposes, we also
show the collected 13C NMR chemical shifts and assign-
ments for bilirubins-IIIa, IXa and XIIIa. As might be
expected, the resonances of the symmetric IIIa and XIIIa
structures correlate with those observed in the IXa struc-
ture. There is also a good correlation between the chemical
The central core dipyrrylmethane dialdehyde %13) had been
prepared earlier7,10 in three steps from methyl 3-[2-%carbo-
tert-butoxy)-3-methyl-5-acetoxymethyl-1H-pyrrolyl]pro-
pionate.10 Saponi®cation of diester 13 afforded diacid 14
in 90% yield.

7816
J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
Table 1. Comparison of 13C NMR chemical shift assignments of phenyl-
bilirubin analogs 1 %R1ˆR3ˆMe, R2ˆR4ˆPh), 2 %R1ˆR4ˆMe, R2ˆR3ˆPh)
dicular to the lactam ring, thus exposing the neighboring
21/181 methyls to the %shielding) face of the aromatic ring.
On the other hand, the exo-phenyls of 3 are more likely to lie
in the plane of the lactam, thereby exposing the neighboring
31/171-CH3 to the edge of the %deshielding) aromatic ring.
and
3
%R1ˆR4ˆPh, R2ˆR3ˆMe) with bilirubin-XIIIa %R1ˆR4ˆMe,
R2ˆR3ˆVn), bilirubin-IXa %R1ˆR3ˆMe, R2ˆR4ˆVn) and bilirubin-IIIa
%R1ˆR4ˆVn, R2ˆR3ˆMe)
Consistent with the 13C NMR results, in the 1H NMR run in
CDCl3 %Table 2), the 31/171-CH3 groups are relatively more
deshielded in 3 than in bilirubin-IIIa or 2, and the 21/
181-CH3 groups in 2 are relatively more shielded than in
bilirubin-XIIIa or 3. Consistent with an endo-phenyl
conformation with the aromatic ring perpendicular to the
lactam, the nearby 7/13-CH3 of 2 is more shielded than
that of 3. In addition, the 5/15 vCH of 2 is more shielded
than that of bilirubin-XIIIa or 3, and the 5/15 vCH of 3 is
more deshielded than that of bilirubin-IIIa. For most other
chemical shifts, there are strong similarities, except for the
obvious differences between phenyl and vinyl groups. Just
as the signals for bilirubin-IXa are a composite of those
from bilirubins-IIIa and XIIIa, so are the chemical shifts
found in 1 a composite of those from 2 and 3.
Bis-phenyl bilirubin
Bilirubin
Position
XIIIa %2) IXa %1) IIIa %3) XIIIa
IXa
IIIa
COOH
83/123
CO
1/19
2/18
173.93
171.31
123.42
140.36
127.43
99.13
173.89
173.89 173.91
171.30
170.36 170.38
123.31
122.47 122.38
140.41
141.84 141.96
127.54
173.91 173.94
173.95 174.04
171.09 171.09
171.22 170.22
124.00 124.00
125.98 125.95
145.13 145.13
141.88 141.89
128.04 128.06
128.49 128.55
101.11 101.09
99.89 99.89
122.24 122.19
122.06 122.20
119.55 119.58
119.74 119.78
123.24 123.25
123.92 123.97
130.81 130.81
3/17
4/16
5/15
6/14
7/13
8/12
9/11
128.31 128.21
99.12
The COOH and NH chemical shifts have been shown to be
characteristic of intramolecular hydrogen bonding of the
type shown in Fig. 1 for natural bilirubin and its symmetric
isomers. Those of phenylrubins 1±3 are very similar to
those found in bilirubin-IIIa, IXa and XIIIa, which
suggests intramolecular hydrogen bonding in similar
ridge±tile conformations. In such conformations, the
propionic acid segment is held in a ®xed staggered geometry
and exhibits a characteristic ABCX splitting pattern
%CHAHX±CHBHC±CO2H) rather than the A2B2 splitting
pattern typically found where greater segmental motion is
possible in the chain, as in %CD3)2SO solvent.7,12 The
coupling constants observed %Table 3) con®rm the ®xed
staggered geometry in the symmetric IIIa and XIIIa
isomers. The coupling constant analysis in the IXa isomer
is complicated by overlapping signals from the non-
equivalent C%8) and C%12) propionic acids, but the patterns
are clearly of the ABCX type.
99.96 100.00
122.12
122.31 122.23
119.59
119.79 119.79
123.31
124.01 124.11
130.63
131.39 131.34
122.04
119.52
123.21
130.79
132.26 132.26
23.74 23.65 23.74
10-CH2
CH3
23.58
9.47
23.60
9.36
9.18
23.67
9.25
9.04
9.06
10.68 10.74
21/31 %181/171)
31/21 %171/181) 127.36
127.34
132.16 132.17
131.24 131.29
129.44 129.43
129.19 129.18
128.47 128.6
127.97 128.00
128.29 128.30
127.03 127.03
9.02
9.15
19.19 19.23
19.23 19.29
34.17 34.20
34.20 34.26
127.03 127.08
121.95
117.12 117.12
22/32 %172/182) 122.04
21/31 %181/171)
24/34 %184/174)
±
±
±
±
±
±
±
±
±
±
±
9.12
±
CH3
71/131
CH2
9.10
9.10
19.24
19.24
34.28
34.28
9.00
9.13
19.21
34.17
9.20
Further con®rmation of the intramolecularly hydrogen-
bonded ridge±tile conformations of 1±3 may be found in
NOE studies. Signi®cant NOEs between the pyrrole and
lactam NHs, and between the C%5/15)-H and the C%7/13)-
CH3 groups con®rm the syn-Z dipyrrinone geometry %Fig.
3). Weaker NOEs between the COOH and lactam NH show
a close proximity, characteristic of that found in Figs. 1 and
5, and in the crystal structure of bilirubin-IXa. As with the
ordinary bilirubins and mesobilirubins with propionic acids
at C%8) and C%12), molecular dynamic calculations reaf®rm
the intramolecularly hydrogen-bonded ridge±tile confor-
mation %Fig. 4) as the global energy-minimum.13
19.22
34.21
81/121
CH2
82/122
Chemical shifts in d %ppm) down®eld from %CH3)4Si for spectra run in
%CD3)2SO solvent. Assignments are made from multiple-bond H±C
COSY experiments.
shifts of the phenylbilirubins and those of their vinyl
analogs, especially the propionic acid carbons, the lactam
CO, the methyl substituents and most of the ring carbons.
Noticeable differences are found in C-2/18, C-9/11 and
C-31/171 of 3 and bilirubin-XIIIa, and in C-3/17 and C-5/
15 of 2 and mesobilirubin-XIIIa, which has ethyl groups in
place of the phenyls of 2. Corresponding differences in
shielding are found in the IXa structures. It is interesting
to note that while the 21/181-CH3 groups of the XIIIa
isomers differ by only ,0.5 ppm, with those of 2 being
more shielded, the 31/171-CH3 groups of the IIIa isomers
differ by ,1.5 ppm, with 3 being more deshielded. These
data suggest that the endo-phenyl rings of 2 lie perpen-
2.3. UV±vis spectra
In bilirubin UV±vis spectra %Table 4) one of the features of
lmax for the long wavelength absorption is that endo-vinyls
cause a hypsochromic shift;whereas exo-vinyls cause a
bathochromic shift. This is clearly seen in comparing
bilirubin-XIIIa and IIIa over a wide range of solvents and
polarity. This same type of difference may also be seen for 2
and 3, with endo-phenyls and exo-phenyls, respectively.

J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
7817
1
Table 2. Comparison of the H NMR chemical shift assignments of phenyl rubins 1 %R1,R3ˆMe, R2,R4ˆPh), 2 %R1,R4ˆMe, R2,R3ˆPh) and 3 %R1,R4ˆPh,
R3,R2ˆMe) with bilirubin-IXa %BR-IX, R1,R3ˆMe, R2,R4ˆVn), bilirubin-XIIIa %BR-XIII, R1,R4ˆMe, R2,R3ˆVn) and bilirubin-IIIa %BR-III, R1,R4ˆVn,
R2,R3ˆMe) in CDCl3
Position
Bis-phenyl bilirubin
Bilirubin
XIIIa %2)
13.69
10.90
9.28
IXa %1)
IIIa %3)
XIIIa
13.67
10.79
9.26
IXa
IIIa
83/123-COOH
21/24-NHCO
22/23-NH
13.74
13.62
10.92
10.84
9.30
9.35
2.02
2.27
13.68
13.68
10.80
10.69
9.27
9.29
1.99
2.18
6.61
13.66
10.85
9.36
13.71
10.70
9.30
2/3, 17/18-CH3
21/31, 171/181
22/32 172/182
23/33, 173/183
24/34, 174/184
5,15-CHv
7,13-CH3
2.03
1.99
2.27
±
2.18
±
6.61a
6.49
6.50a
5.37c
5.35d
±
7.29b
7.46e
7.43f
5.97
7.29
7.47
7.46
7.41
7.43
7.30
5.97
6.23
5.62c
5.59c
±
5.62/5.37
5.59/5.35
±
7.47b
7.41e
7.30f
6.22
±
±
±
6.21
2.16
6.21
6.13
2.16
2.15
3.02g
2.58g
2.90g
2.90g
4.08
6.14
1.91
1.92
2.21
2.20
2.15
81,121-CH2
82,122-CH2
10-CH2
3.01g,h
2.56g,h
2.89g,h
2.79g,h
4.08
3.01/3.04g
2.56/2.61g
2.89/2.93g
279/2.82g
4.10
3.04g,h
2.60g,h
2.93g,h
2.82g,h
4.11
3.02g,h
2.58g,h
2.90g,h
2.82g,h
4.08
3.01g,h
2.58g,h
2.90g,h
2.81g,h
4.08
Chemical shifts in d %ppm) down®eld from %CH3)4Si.
Jˆ2.0, 11.5 Hz.
a
b
d, Jˆ7.5 Hz.
c
dd, Jˆ2.0, 17.5 Hz.
d
Jˆ14.0 Hz.
e
t, Jˆ7.5 Hz.
f
d, Jˆ7.5 Hz.
g
Multiplets.
See Table 3 for coupling constants.
h
Table 3. 1H NMR chemical shifts and coupling constants for the propionic
acid ±CbHAHX±CaHBHC±CO2H segments of phenyl rubins 2 and 3 and
bilirubin-XIIIa %BR-XIII) and bilirubin IIIa %BR-III) in CDCl3 at 228C
C%8,12)
b-CH2
a-CH2
HX
HA
HB
HC
d
2.56
3.01
2.89
2.79
2
JBXˆ1.5
JCXˆ4.0
JAXˆ214.5
2.6
JABˆ15.0
JACˆ3.0
JAXˆ214.5
3.04
JABˆ13.0
JBXˆ1.5
JBCˆ219.5
2.93
JACˆ3.0
JCXˆ4.0
JBCˆ219.5
2.82
d
3
JBXˆ1.5
JCXˆ4.5
JAXˆ214.5
2.58
JBXˆ2.0
JCXˆ4.0
JAXˆ214.0
2.58
JBXˆ2.0
JCXˆ3.5
JAXˆ213.5
JABˆ13.5
JACˆ3.0
JAXˆ214.5
3.02
JABˆ14.0
JACˆ2.5
JAXˆ214.0
3.01
JABˆ14.5
JACˆ2.5
JAXˆ213.5
JABˆ13.5
JBXˆ1.5
JBCˆ215.8
2.90
JABˆ14.0
JBXˆ2.0
JBCˆ218.8
2.90
JABˆ14.5
JBXˆ22.0
JBCˆ217.0
JACˆ3.0
JCXˆ4.5
JBCˆ215.8
2.82
JAXˆ2.5
JCXˆ4.0
JBCˆ218.5
2.81
JACˆ2.5
JCXˆ3.5
JBCˆ217.0
d
BR-XIII
d
BR-III
Figure 3. Selected nuclear Overhauser effects %NOEs) found in phenylrubin
1. Weaker NOEs are noted by dashed arrows. NOEs similar to those of the
left half of 1 are found in 2, whereas NOEs similar to those of the right half
of 1 are found in 3.
d, ppm down®eld from %CH3)4Si; J in Hz from 500 MHz spectra.

7818
J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
Figure 4. Ball and Stick drawings of 2 and 3 in their energy-minimized conformation, as determined by molecular mechanics calculations using Sybyl %Ref.
13).
Table 4. Solvent-dependence of UV±vis spectral data of phenyl rubins 1±3 and Bilirubins-IXa, XIIIa and IIIa %BR-IX, BR-XIII and BR-III)
Cpd
D1max%lmax, nm)
%CH3)2CO
Benzene
CHCl3
CH3OH
CH3CN
%CH3)2SO
1
2
3
61,300 %452)
50,300 %423)a
56,300 %448)
53,700 %436)a
65,500 %465)
56,000 %423)a
53,300 %455)
46,700 %433)a
45,580 %447)
63,300 %451)
52,400 %423)a
55,500 %446)
53,800 %435)a
66,000 %454)
61,900 %438)a
55,800 %453)
49,500 %432)a
46,920 %450)
60,600 %446)
50,300 %419)a
52,400 %440)
49,000 %424)a
64,300 %449)
60,700 %435)a
56,400 %449)
49,900 %429)a
46,600 %447)
34,600 %415)a
62,400 %451)
50,900 %424)a
61,400 %445)
47,700 %410)a
55,100 %440)
49,400 %420)a
61,200 %446)
54,000 %426)a
54,400 %449)
44,600 %417)a
45,300 %446)
37,500 %416)a
60,700 %451)
49,100 %417)a
59,200 %443)
48,600 %417)a
53,100 %439)
49,500 %423)a
62,300 %440)
58,500 %431)a
54,000 %447)
45,000 %421)a
47,000 %445)
39,400 %417)a
62,400 %451)
51,200 %423)a
64,200 %447)
46,900 %409)a
60,800 %441)
53,300 %421)a
67,300 %453)
54,900 %422)a
58,800 %455)
43,000 %415)a
47,800 %446)
37,700 %415)a
67,800 %459)
48,800 %416)a
BR
IX
BR
XIII
BR
III
61,700 %460)
48,100 %426)a
64,600 %456)
49,600 %424)a
At 228C in concentrations ,1.4£1025 M.
a
Shoulders %or) in¯ections were determined by ®rst and second derivative spectra.
The IX isomers have lmax lying between those of XIII and
III, and the 1 values are noticeably larger for the phenyl
analogs of bilirubins-IIIa, IXa and XIIIa. Whether phenyl
or vinyl, however, the long wavelength absorption is
characteristically broadened or dimpled due to exciton
coupling between the two component dipyrrinone chromo-
phores of each rubin. Such exciton coupling may be seen
even more clearly by circular dichroism spectroscopy.
Table 5. Comparison of circular dichroism and UV±visible spectroscopic
data for phenyl rubins 1±3 and bilirubins-IXa, XIIIa and IIIa %BR-IX, BR-
XIII and BR-III) in CHCl3 solutions containing quinine
Pigment
CD
UV
D1max%l1)
l
2 at D1%0)
D1max%l3)
1max
l %nm)
1
2
3
129 %400)
137 %396)
133 %402)
139 %405)
127 %402)
158 %408)
419
415
423
425
424
427
252 %446) 62,300
262 %443) 56,700
255 %452) 64,100
264 %453) 56,800
246 %454) 45,900
289 %456) 63,300
451
446
454
455
450
458
BR-IX
BR-XIII
BR-III
2.4. Circular dichroism spectra
Earlier studies have shown that bilirubin-IXa and related
rubins exhibit an induced circular dichroism %CD) in CHCl3
solvent in the presence of quinine and other optically active
Pigment conc. ,1.4£1025 M;quinine conc. ,4.2£1023 M;pigment:
quinine molar ratioˆ,1:300.

J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
7819
Figure 5. Interconverting intramolecularly hydrogen-bonded enantiomeric conformers of mesobilirubins 1±4 and bilirubin %see Fig. 1). The double headed
arrows represent the dipyrrinone long-wavelength electric transition moments %dipoles). The relative helicities %M, minus, or P, plus) of the vectors are shown
%inset) for each enantiomer. Hydrogen bonds are shown by dashed lines.
amines.14,15 Phenylrubins 1±3 are no exception and exhibit
strong bisignate CD Cotton effects %Table 5) associated with
the two exciton transitions at long wavelengths %that were
overlapping and less distinct in the UV±vis spectra). The
two-dipyrrinone chromophores of each rubin have strongly
allowed long wavelength electronic transitions %Table 4),
making them excellent candidates for transition dipole±
dipole interaction %exciton coupling). Such intramolecular
exciton interaction or resonance splitting produces two long
wavelength transitions in the UV±vis spectrum and two
corresponding, but oppositely signed, bands in the CD
spectrum.14,16 According to exciton chirality theory, a
long-wavelength negative and short-wavelength positive
CD couplet exhibits negative exciton chirality and corre-
sponds to a negative helical disposition of the relevant
dipyrrinone long-wavelength transition dipoles. Usually
this corresponds to the M-helical enantiomeric type of
Fig. 5. But, as has been shown previously, opening of the
interplanar angle of either enantiomer without inverting
molecular helicity also can cause the transition moments
to invert helicity and invert the Cotton effect signs of the
CD couplet.17
When human serum albumin %HSA) is used as the chiral
complexation agent in aqueous %pH 7.4 Tris buffer)
solutions,18 all but 2 exhibit moderately intense positive
chirality bisignate Cotton effects %Fig. 6A). Phenylrubin 2
exhibits a negative-exciton chirality, and its counterpart,
bilirubin-XIIIa, exhibits the weakest of the ®ve positive
exciton chirality Cotton effects, suggesting weaker or even
reversed enantioselectivity by the protein. Interestingly,
both 3 and bilirubin-IIIa give the most intense Cotton
effects, suggesting that the exo substituent is important for
favorable binding to the chiral complexation agent.
Adding trace amounts of CHCl3 %mL/mL) to buffered
aqueous HSA solutions of the rubins %except for 2) leads
to inverted Cotton effect signs %Fig. 6B, Table 6), as noted
previously.17 Upon addition of 5 mL CHCl3, all but bili-
rubin-IXa and IIIa show a negative exciton chirality. The
Cotton effects of 1 and 2 become especially intense.
Addition of a second 5 mL aliquot of CHCl3 causes all
rubins to exhibit a negative exciton chirality and leads to
increased magnitudes. The Cotton effect magnitudes of 1
and 2 remain exceptionally strong relative to the other
rubins. At a total of 15 mL/5 mL, further Cotton effect
intensi®cation is seen for bilirubins IIIa, IXa and XIIIa,
while that of 1 is changed little, and solutions of 2 and 3
have exceeded the saturation limit. The initially large
Cotton effects of the symmetric exo-phenyl %3) and exo-
vinyl rubins on HSA and their sluggishness to invert with
added CHCl3 suggest that they are tightly bound to HSA in a
favorable binding site. Apparently it is the location %exo) of
the aryl and vinyl groups rather than their presence that is
responsible for the phenomenon.
In a nonpolar, aprotic solvent such as chloroform, the
conformational equilibrium between bilirubin enantiomers
%Fig. 5) can be displaced from 1:1 by adding a chiral recog-
nition agent such as quinine or other optically active
amines.14,15 This leads typically to an intense bisignate
induced CD Cotton effect, as has been noted previously
for bilirubin in aqueous buffers with added albumin. The
strong, negative exciton chirality CDs observed in the
presence of quinine %Table 5) is characteristic of an exciton
system in which the two dipyrrinone chromophores of the
bichromophoric rubin molecule interact by coupling locally
excited p±pp transitions %electric dipole transition moment
coupling). Interestingly, at the same molar ratio of alkaloid/
pigment of ,300:1,14 which typically gives optimal Cotton
effects, the intensities are rather similarÐexcept for bili-
rubin-IIIa, which shows nearly double the intensity of the
others.
2.5. Polarity and solubility
Fig. 7 shows HPLC chromatograms of mixtures of bili-
rubins IIIa, IXa and XIIIa and diphenylrubins 1±3. Their
relative retention factors are bilirubin-XIIIa, 1.00;bilirubin-
IXa, 1.13;bilirubin III a and 2, 1.28; 1, 1.51 and 3, 1.76,
indicating that 1±3 are somewhat more lipophilic than their

7820
J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
Figure 6. Circular dichroism spectra of 1.4£1025 M 1±3 and their bilirubin parents in CHCl3 in the presence of %A) ca. 4.2£1023 M quinine and %B) ca.
2.8£1025 M HSA in pH 7.4 Tris buffer at 228C. The compound numbers are indicated on each CD curve.
vinyl counterparts. Bilirubin IIIa and diphenylrubin 2, the
phenyl counterpart of bilirubin-XIIIa, had identical
retention times and were unresolved when chromatographed
together. Similarly, on silica TLC, 1±3 have an Rf,0.58,
whereas the vinyl analogs have Rf 0.53 %solvent 99:1
CH2Cl2/MeOH by vol). For bilirubins with natural side-
chains, lipophilicity increases in the order XIIIa, IXa,
IIIa, and the same relative order is maintained for the
corresponding phenyl substituted compounds 1±3. Taken
collectively, the polarity and solubility characteristics of
1±3 are consistent with the propionic acid side-chains
being involved in intramolecular hydrogen bonding and
unavailable for solvent interaction.
2.6. Hepatic metabolism
Bilirubin is cleared from the circulation in healthy humans,
rats and other mammals by the liver, where it undergoes
enzymic conversion to two isomeric monoglucuronides.
These are partly converted by the same enzyme to bilirubin
diglucuronide. The resulting mixture of three acyl glucuro-
nides is secreted from the liver into bile. Formation of the
glucuronides is catalyzed by a speci®c isoform, UGT1A1,
of the uridine diphosphoglucuronosyl transferase family.1±3
This particular isoform is not expressed in homozygotes of
the mutant strain of rats known as Gunn rats. In these
animals, bilirubin glucuronides are not formed and are not

J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
7821
Table 6. In¯uence of chloroform on the circular dichroism spectra for phenyl-rubins 1±3 and bilirubins-IXa, XIIIa and IIIa %BR-IX, BR-XIII and BR-III) in
pH 7.4 Tris buffer containing HSA
Pigment
CHCl3 %mL)
CD
UV
1max
D1max%l1)
l
2 at D1%0)
D1max%l3)
l %nm)
1
2
3
0
218 %382)
158 %402)
243 %399)
225 %398)
28 %393)
263 %408)
1108 %400)
1145 %399)
114 %403)
29 %395)
112 %404)
217 %397)
1129 %399)
1150 %398)
115 %402)
120 %404)
139 %399)
121 %406)
401
433
418
418
411
421
419
417
423
412
431
418
418
417
422
427
422
425
147 %429)
223 %458)
166 %447)
150 %449)
127 %447)
167 %449)
2179 %449)
2225 %446)
221 %453)
125 %455)
215 %460)
123 %443)
2214 %448)
2239 %446)
222 %452)
232 %456)
269 %452)
242 %455)
52,660
54,300
57,300
46,070
37,670
54,470
65,980
62,100
34,300
46,050
39,570
55,570
65,460
62,600
25,300
49,850
40,930
56,620
441
451
449
459
455
462
451
450
449
458
451
461
451
450
452
458
449
460
BR-IX
BR-XIII
BR-III
1
2
3
BR-IX
BR-XIII
BR-III
1
2
3
BR-IX
BR-XIII
BR-III
1
2
3
5
10
15
1134 %399)
418
2223 %448)
66,330
451
a
a
a
a
a
a
a
a
a
a
BR-IX
BR-XIII
BR-III
143 %403)
151 %400)
140 %406)
424
421
425
271 %455)
294 %452)
272 %454)
52,550
43,830
58,130
456
449
460
Chloroform was added to 5 mL solutions that were ,1.4£1025 M in pigment and 1:2 molar ratio of pigment to HSA %human serum albumin) at 228C.
a
Solubility limit of CHCl3 exceeded.
present in bile, and unconjugated bilirubin accumulates in
their circulation and extravascular tissues.
each compound being the diglucuronide and the least polar,
the monoglucuronide. The absorption spectra of the
metabolites %Fig. 9, Table 7) are consistent with that
assumption. In each case, glucuronidation led to the appear-
ance of a weak shoulder on the short wavelength side of the
main absorption band, which is more pronounced for the
diglucuronide than the monoglucuronide. Acyl glucuro-
nides of bilirubin, mesobilirubin and other bilirubins show
the same phenomenon. For compound 2, identi®cation of its
metabolites as acyl glucuronides was con®rmed by their
hydrolysis to the parent pigment with both b-glucuronidase
and dilute NaOH %Fig. 10). Only one monoglucuronide
HPLC peak was detected for the unsymmetrically substi-
tuted biphenyl compound 1, rather than the two diastereo-
mer peaks expected. Whether this re¯ects poor HPLC
resolution or regiospeci®c glucuronidation is unclear. The
peak was, however, very slightly less symmetrical than the
peaks of the monoglucuronides of the symmetrical isomers
2 and 3, suggesting that two similar components might have
been present.
When compounds 1±3 were injected intravenously as
0.25 mg bolus doses into homozygous Gunn rats they
were not excreted in bile, either as the intact pigments or
as metabolites. In contrast, in normal rats each of the three
was excreted in bile as a pair of metabolites after intra-
venous administration %Fig. 8). Since these metabolites
were observed only in normal rats and not in Gunn rats, it
is logical to assume that they are acyl glucuronides formed
by UGT1A1 catalysis, with the most polar metabolite from
These metabolic studies show that biphenyl compounds 1±3
are metabolized much like bilirubin, being converted to
more polar cholephilic mono and diglucuronides by
UGT1A1. Their rates of excretion into bile as glucuronides
were slower than for similar compounds without phenyl
substitution, as shown by the rather broad, tailing biliary
excretion curves %Fig. 8). This could be a consequence of
slower uptake, impaired glucuronidation or retarded ef¯ux
from the liver into bile. Our data is insuf®cient to distinguish
these possibilities. The data do show, however, that the ratio
of mono to diconjugates formed from each compound is
highly dependent on the positions of the phenyl groups.
The proportion of diglucuronide was highest for the least
Figure 7. Superimposed reverse phase HPLC chromatograms of mixtures
containing %lower) bilirubin XIIIa, IXa and IIIa and %upper) diphenyl
bilirubins XIIIa %2), IXa %1) and IIIa %3). The mobile phase was 0.1 M
di-n-octylamine acetate in 5% water/MeOH at a ¯ow-rate of 0.75 mL/min.

7822
J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
Figure 8. Metabolism of diphenyl bilirubins XIIIa %2), IXa %1) and IIIa %3) in the rat. Panels a, c and d show HPLC chromatograms of bile before %tˆ0) and
after %tˆ21 or 27 min) injecting 0.25 mg of 2, 1 and 3, respectively, intravenously into adult male Sprague-Dawley rats. The tˆ0 chromatograms show the
presence of bilirubin diglucuronide %near 8 min) and its two diastereomeric monoglucuronides %near 14 min). MG and DG indicate, respectively, the mono and
diglucuronide of the injected pigment. Panels b, d and f show biliary excretion curves for the mono and glucuronides of 2, 1 and 3, respectively, derived from
the peak areas of the corresponding peaks in HPLC chromatograms of bile.
lipophilic isomer 2, in which the two phenyl groups are at
the endo positions of the lactam rings, and lowest for the
isomer 3 with phenyl groups at the two exo positions. The
unsymmetrical isomer 1, with one endo and one exo phenyl
group, yielded an intermediate proportion of diglucuronide.
We observed the same relationship between lipophilicity
and the proportion of diglucuronide previously for
lipophilic rubins substituted with n-butyl groups. How-
ever, we have also observed exclusive monoglucuronide
Table 7. Absorption spectra of diphenyl bilirubins 1±3 and their mono-
%MG) and di- %DG) glucuronides. Obtained from diode array spectra of
HPLC eluates in 0.1 M di-n-octylamine acetate in MeOH containing 8%
water %sh indicates shoulder)
Pigment
l %nm)
Parent
MG
DG
1 %IXa)
2 %XIIIa)
3 %IIIa)
450, 420 %sh)
444, 416 %sh)
454, 422 %sh)
448, 418 %sh)
444, 410 %sh)
452, 418 %sh)
448, 412 %sh)
444, 408
454, 414 %sh)
Figure 9. Normalized absorbance spectra of diphenyl bilirubin XIIIa %2)
and its mono and diglucuronides obtained by diode array scanning of the
corresponding HPLC chromatogram peaks.

J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
7823
tography was carried out on J.T. Baker silica gel IB-F plates
%125 m layers). HPLC analyses were carried out on a
Perkin±Elmer Series 4 high performance liquid chromato-
graph with an LC-95 UV±visible spectrophotometric
detector %set at 410 nm) equipped with a Beckman-Altex
ultrasphere-IP 5 mm C-18 ODS column %25£0.46 cm) and
a Beckman ODS precolumn %4.5£0.46 cm). For HPLC
analyses of metabolites in bile, detection %HP 1100 or
1050 DAD) was in the range of ,454 nm, and the column
was a Beckman-Altex ultrasphere-IP 5 mm C-18 ODS
column %25£0.46 cm) ®tted with a similarly packed pre-
column %4.5£0.46 cm). The ¯ow rate was 0.75 mL/min,
the elution solvent was 0.1 M di-n-octylamine acetate in
8% aqueous methanol, and the column temperature
,348C. Di-n-octylamine was obtained from Aldrich;
HPLC grade MeOH from Fisher; b-glucuronidase %E. coli
Type VII-A, 1000 units/vial), phosphatidyl choline %Type
XV-E), cholesterol, taurine, sodium cholate and human
serum albumin %defatted) from Sigma. All other solvents
and reagents, unless otherwise speci®ed, were purchased
from Acros, Aldrich or Fisher and used as received. Spectral
data were obtained in spectral grade solvents %Aldrich or
Fisher).
Figure 10. Hydrolysis of bile from a rat injected with diphenyl bilirubin
XIIIa %2). HPLC chromatogram a is of bile collected 39 min after injection
of pigment. Chromatogram b is of bile collected 45 min after injection and
treated with aqueous b-glucuronidase. Chromatogram c is of bile collected
24 min after injection and treated with dilute aqueous NaOH. B, BMGs and
BDG indicate bilirubin and its mono and diglucuronides, respectively. DG
and MG indicate mono and diglucuronide, respectively.
formation for bilirubins substituted with two methyl groups
at C10, which are much less lipophilic than bilirubin.19 Not
knowing the extinction coef®cients for the glucuronides,
we were unable to calculate the exact ratio of mono to
diglucuronidation for each compound. However, since the
aglycones and their glucuronides have very similar absorp-
tion spectra %Fig. 10), it is likely that mono to diglucuronide
ratios computed from the corresponding peak areas are close
to the true values.
The Ball and Stick drawings were created from the atomic
coordinates of the molecular dynamics structures using
the MuÈller and Falk `Ball and Stick' program %Cherwell
Scienti®c, Oxford, UK) for Macintosh. The molecular
dynamics calculations were carried out on a Silicon
Graphics workstation using Sybyl 6.6 with Gasteiger±
HuÈckel calculations of electrostatic charge.
Jaundiced homozygous male Gunn rats, weighing 300±
400 g, were obtained from our own colony and normal
Sprague-Dawley male rats, weight ,300 g, were obtained
from local commercial vendors. In vivo studies were done in
a windowless room under safe lights, as described else-
where.19,22 Each compound was studied in a single Gunn
rat and in at least two normal rats. Qualitatively similar
biliary excretion curves to those depicted in Fig. 8 were
obtained in replicate experiments.
These studies show that bilirubins containing bulky phenyl
groups at the endo or exo positions have similar three-
dimensional structures to natural bilirubin, but are more
lipophilic. In the rat, they are metabolized, like bilirubin,
to mono and diglucuronides that are excreted in bile.
However, the proportion of diglucuronide formation is
much less than for bilirubin and the overall rate of excretion
of the glucuronides into bile is slower than for bilirubin
glucuronides with simple alkyl groups %Me, Et or vinyl) in
place of the phenyl groups. Despite having relatively bulky
phenyl groups on the lactam rings, compounds 1±3 are
substrates of UGT1A1 in vivo.
3.1. CD and UV measurements
A stock solution of rubin %,7.0£1024 M) was prepared by
dissolving an appropriate amount of the desired rubin
compound in 2 mL of DMSO. Next, 100 mL of the stock
solution was diluted to 5 mL %volumetric ¯ask) with an HSA
solution ,2.8£1025 M in pH 7.4 Tris buffer. The ®nal
concentration of the solution was ,1.4£1025 M in pigment.
Up to four 5 mL solutions of each pigment were prepared, as
needed, in 5 mL volumetric ¯asks. To each ¯ask was added
the indicated volume %mL) of CHCl3. Gentle shaking was
occasionally required to dissolve all of the CHCl3 in the
aqueous solution, at which point CD and UV±vis measure-
ments were recorded.
3. Experimental
All ultraviolet±visible %UV±vis) spectra were recorded on a
Perkin±Elmer l-12 spectrophotometer, and all circular
dichroism %CD) spectra were recorded on a Jasco J-600
instrument. Nuclear magnetic resonance %NMR) spectra
were obtained on GE QE-300 spectrometer operating at
300 MHz, or on a Varian Unity Plus 500 MHz spectrometer
in CDC13 solvent %unless otherwise speci®ed). Chemical
shifts were reported in d ppm referenced to the residual
CHCl3 1H signal at 7.26 ppm and 13C signal at 77.23 ppm.
A J-modulated spin-echo experiment %Attached Proton
Test) was used to assign 13C NMR spectra. Melting points
were taken on a MelTemp capillary apparatus and are
uncorrected. Combustion analyses were carried out by
Desert Analytics, Tucson, AZ. Analytical thin layer chroma-
3.1.1. 4-Methyl-3-phenyl-2-p-toluenesulfonyl-1H-pyrrole
310). p-Toluenesulfonylmethyl isocyanide %TosMIC)10
%6.08 g, 31.2 mmol) and 1,1,3,3-tetramethyl guanidine
%7.60 g, 65.5 mmol) were dissolved in 36 mL of a 1:1
mixture of THF±2-propanol and stirred magnetically at
room temperature. A solution of 1-phenyl-2-nitro-1-propanol

7824
J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
acetate11 %15) %6.95 g, 31.2 mmol) was added dropwise, and
the mixture was stirred at room temperature overnight. It
was then diluted with 70 mL of CH2Cl2, washed with water
%2£70 mL) and brine %70 mL), and dried over anhydr.
MgSO4. The solvent was removed %rotary evaporator), and
the residue was crystallized from CH2Cl2±hexane to yield
5.14 g %53%) of a tan solid. An analytical sample was
prepared by recrystallization from CH2Cl2±hexane.
Colorless crystals of 10 were collected and dried over
P2O5 in a drying pistol overnight. Mp 202±2038C;IR
%KBr) n: 3380, 1596, 1313, 1301, 1211, 1143, 1081, 705,
129.05, 128.63, 77.72, 21.80, 10.13 ppm. Anal. calcd for
C18H17NO3S %327.4): C, 66.03;H, 5.24;N, 4.28. Found:
C, 65.76;H, 5.45;N, 4.60.
3.1.4. 3-Methyl-4-phenyl-3-pyrrolin-2-one 36). 3-Methyl-
4-phenyl-5-p-toluenesulfonyl-3-pyrrolin-2-one %8) %1.63 g,
5.01 mmol) and sodium borohydride %0.23 g, 6.01 mmol)
were stirred in 65 mL of absolute ethanol at room tempera-
ture for 10 min. The solution was then diluted with 65 mL of
CH2Cl2, washed with water %2£65 mL) and brine %65 mL),
and dried over anhydr. MgSO4. The solvent was removed
%rotary evaporator), and the residue was crystallized from
CH2Cl2±hexane to afford a light tan powder %0.76 g, 88%).
An analytical sample was prepared by precipitation of a
CH2Cl2 solution with hexane. Colorless crystals of 6 were
collected and dried over P2O5 in a drying pistol overnight.
Mp 155±1588C [lit.20 mp 155±1588C];IR %KBr) n: 3432,
1
589 cm21; H NMR d: 9.03 %1H, brs), 7.38±7.06 %5H, m),
6.79 %1H, d, Jˆ2 Hz), 2.32 %3H, s), 1.92 %3H, s) ppm; 13C
NMR d: 143.66, 139.53, 132.84, 130.75, 129.79, 129.44,
127.97, 127.56, 127.14, 125.01, 121.48, 120.22, 21.64,
10.53 ppm;MS: m/z %relative intensity) 379%12), 311%100),
231%30), 172%40), 117%20), 91%10) amu. Anal. calcd for
C18H17NO2S %311.4): C, 69.43;H, 5.50;N, 4.40%. Anal.
calcd for C18H17NO2S´1/4H2O %315.9): C, 68.44;H, 5.58;
N, 4.43%. Found: C, 68.42;H, 5.56;N, 4.68%.
1
1687, 1499, 1452, 1364, 766, 693, 614 cm21; H NMR d:
7.50±7.30 %5H, m), 7.05 %1H, brs), 4.23 %2H, d, Jˆ1 Hz),
2.10 %3H, s) ppm; 13C NMR d: 174.92, 151.21, 132.53,
131.83, 129.31, 128.44, 127.85, 50.46, 14.48 ppm.
3.1.2. 5-Bromo-4-methyl-3-phenyl-2-p-toluenesulfonyl-
1H-pyrrole 312). 4-Methyl-3-phenyl-2-p-toluenesulfonyl-
1H-pyrrole %10) %2.00 g, 6.43 mmol) was dissolved in
50 mL of CH2Cl2 and stirred in an ice bath magnetically.
A solution of phenyltrimethyl ammonium tribromide %PTT)
%5.07 g, 13.5 mmol) in 100 mL of CH2Cl2 was added
dropwise and the resulting yellow-orange solution was
stirred at 08C for 30 min. It was then washed with aq.
sodium bisul®te %3£150 mL) and brine %150 mL), then
dried over anhydr. MgSO4. The solvent was removed
using a rotary evaporator to yield a colorless solid %2.36 g,
94%). An analytical sample was prepared by the addition
of hexane to a CH2Cl2 solution of the pyrrole. Colorless
crystals of 12 were collected and dried over P2O5 in a drying
pistol overnight. Mp 172±1748C;IR %KBr) n: 3270, 1445,
3.1.5.
5,50-Diformyl-4,40-dimethyl-5,50-bis32-carboxy-
ethyl)-2,20-dipyrrylmethane 314). Dipyrrylmethane di-
methyl ester 137,9,10 %0.30 g, 0.75 mmol) was dissolved in
30 mL of methanol and 10 mL of 6 M aq. KOH. The solu-
tion was heated at re¯ux for 2 h, then cooled in an ice bath,
diluted with water, and acidi®ed with acetic acid to form a
tan precipitate. The solids were collected by vacuum ®l-
tration and air dried to afford diacid 14 %0.25 g, 90%),
which was used directly in the next step. Mp 248±2508C;
IR %KBr) n: 3248, 2919, 1713, 1619, 1449, 1384, 1207,
873 cm21; H NMR %DMSO-d6) d: 12.00 %2H, s), 11.48
1
%2H, s), 9.48 %2H, s), 3.91 %2H, s), 2.51 %4H, t, Jˆ8 Hz),
2.18 %6H, s), 2.06 %4H, t, Jˆ8 Hz) ppm; 13C NMR %DMSO-
d6) d: 176.71, 173.75, 134.57, 130.41, 128.01, 120.73,
34.21, 22.46, 18.63, 8.46 ppm.
1
1317, 1142, 1192, 1085, 772, 707, 590 cm21; H NMR d:
9.12 %1H, brs), 7.37±7.08 %5H, m), 2.33 %3H, s), 1.85 %3H,
s) ppm; 13C NMR d: 144.00, 139.14, 132.25, 130.65,
130.54, 129.76, 129.54, 128.11, 128.05, 127.32, 121.43,
104.45, 21.67, 10.50 ppm;MS: m/z %relative intensity)
391%100), 325%10), 245%32), 154%70), 127%22), 91%15)
amu. Anal calcd for C18H16BrNO2S %390.3): C, 55.39;H,
4.13;N, 3.59. Found: C, 55.10;H, 3.96;N;3.58.
3.1.6. 3,17-Des-vinyl-3,17-bis-phenyl-bilirubin-XIIIa 32).
In a 50 mL round bottomed ¯ask 3-methyl-4-phenyl-3-
pyrrolin-2-one %6) %0.253 g, 1.46 mmol) and diformyl-
dipyrrylmethane 14 %0.129 g, 0.341 mmol) were dissolved
in 55 mL of 2-propanol and 0.28 mL of piperidine. The
solution was heated at re¯ux overnight, during which time
the product precipitated. The mixture was chilled in an ice
bath, and the solids were collected by vacuum ®ltration then
washed with water %2£10 mL). The yellow-orange solid
was air dried to afford pure 2 %0.081 g, 37%). Mp 320±
3228C;IR %KBr) n: 3417, 2920, 1692, 1634, 1443, 1385,
3.1.3. 3-Methyl-4-phenyl-5-p-toluenesulfonyl-3-pyrrolin-
2-one 38). 5-Bromo-4-methyl-3-phenyl-2-p-toluenesul-
fonyl-1H-pyrrole %12) %2.00 g, 5.13 mmol) was dissolved
in 40 mL of TFA and 8 mL of water, then stirred at room
temperature overnight. The resulting dark solution was
diluted with 100 mL of CH2Cl2 and washed sequentially
with water %2£100 mL), aq. sodium bicarbonate %2£
100 mL) and brine %100 mL);then it was dried over anhydr.
MgSO4. The solvent was removed on a rotary evaporator to
yield 1.63 g %98%) of a light gray powder. Colorless crystals
were obtained by the addition of hexane to a CH2Cl2
solution of pyrrolinone 8. They were collected and dried
over P2O5 in a drying pistol overnight. Mp 80±828C;IR
%KBr) n: 3431, 1700, 1446, 1357, 1136, 1083, 816, 636,
1
1252, 1020, 700 cm21; H NMR d: 13.69 %2H, brs), 10.90
%2H, s), 9.28 %2H, s), 7.48±7.28 %10H, m), 5.97 %2H, s), 4.08
%2H, s), 3.01 %2H, ddd, Jˆ15.0, 3.0, 214.5 Hz), 2.89 %2H,
ddd, Jˆ13.0, 219.5, 1.5 Hz), 2.79 %2H, ddd, Jˆ3.0, 219.5,
4.0 Hz), 2.56 %2H, ddd, Jˆ214.5, 1.5, 4.0 Hz), 2.03 %6H, s),
1.91 %6H, s) ppm; 13C NMR d: 179.85, 174.26, 146.45,
133.86, 132.32, 129.75, 128.82, 128.69, 128.66, 124.97,
124.79, 124.53, 119.86, 104.56, 32.79, 22.49, 18.79,
10.32, 9.06 ppm. Anal. calcd for C41H40N4O6 %684.8): C,
71.91;H, 5.89;N, 8.18. Found: C, 71.95;H, 6.05;N, 8.08.
1
547 cm21; H NMR d: 7.46±7.23 %5H, m), 7.40 %2H, d,
Jˆ8 Hz), 7.22 %2H, d, Jˆ8 Hz), 6.85 %1H, s), 5.70 %1H, s),
2.42 %3H, s), 1.80 %3H, s) ppm; 13C NMR d: 173.17, 146.84,
145.89, 144.77, 134.06, 131.33, 129.84, 129.70, 129.58,
3.1.7. 3-Methyl-4-phenyl-2-p-toluenesulfonyl-1H-pyrrole
311). 4-Methyl-3-phenyl-2-p-toluene-sulfonyl-1H-pyrrole
%10) %6.00 g, 19.3 mmol) was dissolved in 19.2 mL of

J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
7825
TFA and 172.8 mL of CH2Cl2 to make a 0.1 M solution. The
solution was stirred at room temperature for 4 days, during
which time it darkened. It was washed with water %2£
200 mL), aq. sodium bicarbonate %2£200 mL) and brine
%200 mL), then dried over anhydr. MgSO4. The solvent
was removed on a rotary evaporator, and the residue was
crystallized from ethyl acetate±hexane to afford 3.92 g
%65%) of a tan solid. Colorless crystals of 11 were obtained
by precipitation from ethyl acetate±hexane. They were
collected and dried over P2O5 in a drying pistol overnight.
Mp 117±1188C;IR %KBr) n: 3400, 1664, 1442, 1302, 1141,
1087, 706, 691, 599 cm21; 1H NMR d: 9.33 %1H, brs), 7.82
%2H, d, Jˆ8 Hz), 7.36±7.28 %5H, m), 7.29 %2H, d, Jˆ8 Hz),
7.00 %1H, d, Jˆ3 Hz), 2.41 %3H, s), 2.30 %3H, s) ppm; 13C
NMR d: 143.95, 139.97, 134.63, 130.02, 128.65, 128.53,
127.95, 126.95, 126.74, 125.02, 123.42, 120.93, 21.69,
dipyrrylmethane 14 %0.147 g, 0.393 mmol) were dissolved
in 60 mL of 2-propanol and 0.32 mL of piperidine. The
solution was heated at re¯ux overnight, during which time
the product precipitated. The mixture was chilled in an ice
bath, and the solids were collected by vacuum ®ltration then
washed with water %2£10 mL). The yellow-orange solid
was air dried to afford pure 3 %0.203 g, 75%). Mp 260±
2628C;IR %KBr) n: 3407, 2923, 1700, 1684, 1653, 1576,
1
1496, 1448, 1250, 667 cm21; H NMR d: 13.66 %3H, brs),
10.85 %2H, s), 9.36 %2H, s), 7.48±7.28 %10H, m), 6.22 %2H,
s), 4.11 %2H, s), 3.04 %2H, ddd, Jˆ13.5, 3.0, 214.5 Hz), 2.93
%2H, ddd, Jˆ13.5, 215.8, 1.5 Hz), 2.82 %2H, ddd, Jˆ3.0,
215.8, 4.5 Hz), 2.60 %2H, ddd, Jˆ214.5, 1.5, 4.5 Hz), 2.27
%6H, s), 2.20 %6H, s) ppm; 13C NMR d: 179.90, 173.58,
142.54, 134.13, 131.81, 129.44, 129.39, 128.57, 128.55,
127.75, 125.00, 124.64, 120.01, 102.35, 32.82, 22.52,
18.81, 11.02, 10.42 ppm. Anal. calcd for C41H40N4O6
%684.8): C, 71.91;H, 5.89;N, 8.18. Found: C, 72.22;H,
6.13;N, 8.12.
10.40 ppm;MS:
m/z %relative intensity) 311%100),
246%12), 172%10), 154%12), 128%10) amu. Anal. calcd for
C18H17NO2S %311.4): C, 69.43;H, 5.50;N, 4.40. Anal.
calcd for C18H17NO2S´1/4H2O %315.9): C, 68.44;H, 5.58;
N, 4.43. Found: C, 68.00;H, 5.33;N, 4.58.
3.1.11. 3,17-Des-vinyl-3,17-bis-phenyl-bilirubin-XIIIa-
dimethyl ester 35). In a 50 mL round bottomed ¯ask
3-methyl-4-phenyl-3-pyrrolin-2-one %6) %0.341 g, 1.97
mmol) and diformyldipyrrylmethane 137,9,10 %0.198 g,
0.493 mmol) was dissolved in 10 mL of methanol and
2.0 mL of piperidine. The solution was heated at re¯ux
overnight, during which time the product precipitated
from the solution. The mixture was chilled in an ice bath,
the solids collected by vacuum ®ltration, then washed with
water %2£10 mL). The yellow-orange solid was air dried to
afford pure dimethyl ester 5 %0.351 g, 56%). Mp 230±2348C
%dec);IR %KBr) n: 3338, 2918, 1738, 1658, 1633, 1464,
3.1.8. 4-Methyl-3-phenyl-5-p-toluenesulfonyl-3-pyrrolin-
2-one 39). 3-Methyl-4-phenyl-2-p-toluenesulfonyl-1H-
pyrrole %11) %1.00 g, 3.21 mmol) was dissolved in acetic
acid %40 mL) and 30% aq. hydrogen peroxide %0.55 mL,
4.8 mmol) and stirred at room temperature for 3 days. The
solution was diluted with CH2Cl2 %100 mL), washed with
water %2£100 mL) and brine %100 mL), and dried over
anhyd. MgSO4. Removal of the solvent by rotary evapo-
ration gave a crude solid that was puri®ed by radial
chromatography %2% methanol in CH2Cl2 eluent) to yield
0.28 g %25%) of pure product. Mp 133±1348C;IR %KBr) n:
1361, 1252, 1165, 1019, 700 cm21 1H NMR d: 11.27
;
1
3425, 1702, 1386, 1317, 1136, 1081, 700, 587 cm21; H
%2H, s), 10.42 %2H, s), 7.36±7.11 %10H, m), 5.91 %2H, s),
4.18 %2H, s), 3.64 %6H, s), 2.85 %4H, t, Jˆ7.4 Hz), 2.45 %4H,
t, Jˆ7.4 Hz), 1.95 %12H, s) ppm; 13C NMR d: 173.79,
173.69, 145.73, 132.84, 131.66, 130.02, 129.25, 128.38,
128.20, 124.75, 124.60, 124.32, 119.37, 104.57, 51.70,
35.68, 22.92, 20.30, 9.91, 8.73 ppm. Anal. calcd for
C43H44N4O6 %712.8): C, 72.45;H, 6.22;N, 7.86. Calcd for
C43H44N4O6´1/2H2O %721.9): C, 71.55;H, 6.28;N, 7.76.
Found: C, 71.97;H, 6.07;N, 7.65.
NMR d: 7.73 %2H, d, Jˆ8 Hz), 7.34±7.28 %5H, m), 7.15
%1H, s), 7.06 %2H, d, Jˆ8 Hz), 5.17 %1H, s), 2.34 %3H, s),
2.32 %3H, s) ppm; 13C NMR d: 172.02, 146.29, 144.69,
136.68, 130.70, 129.96, 129.63, 129.52, 129.03, 128.76,
128.47, 79.45, 21.77, 14.51 ppm. Anal. calcd for
C18H17NO3S %327.4): C, 66.03;H, 5.24;N, 4.28. Found:
C, 65.71;H, 5.05;N, 4.30.
3.1.9. 4-Methyl-3-phenyl-3-pyrrolin-2-one 37). 4-Methyl-
3-phenyl-5-p-toluenesulfonyl-3-pyrrolin-2-one %9) %0.63 g,
1.93 mmol) and sodium borohydride %0.09 g, 2.32 mmol)
were stirred in 25 mL of absolute ethanol at room tempera-
ture for 10 min. The solution was then diluted with 25 mL of
CH2Cl2, washed with water %2£25 mL) and brine %25 mL),
and dried over anhydr. MgSO4. The solvent was removed
%rotary evaporator) and the residue was crystallized from
CH2Cl2±hexane to afford a light tan powder %0.33 g,
quant.). An analytical sample of 7 was prepared by precipi-
tation of a CH2Cl2 solution with hexane. Colorless crystals
were collected and dried over P2O5 in a drying pistol over-
night. Mp 132±1338C [lit.21 mp 187±1948C, 198±2008C];
IR %KBr) n: 3424, 1677, 1497, 1444, 1387, 1082, 744, 700,
3.1.12. 2,17-Des-vinyl-2,17-bis-phenyl-bilirubin-IXa 12-
monomethyl ester 34). 2,18-Des-vinyl-2,18-bis-phenyl-
bilirubin-IIIa %3) %0.052 g, 0.073 mmol) and 3,17-des-
vinyl-3,17-bis-phenyl-bilirubin-XIIIa dimethyl ester %5)
%0.050 g, 0.073 mmol) were dissolved in DMSO %5 mL);
then conc. HCl %0.50 mL) was added with stirring. The
solution was stirred in the dark under a nitrogen atmosphere
for 1 min and was poured into water %75 mL). The resulting
solids were removed by centrifugation, and collected by
vacuum ®ltration. The solids were washed with water
%2£10 mL) and air dried to afford a 1:1:2 mixture of the
two starting materials and the desired product, respectively.
The rubins were separated and isolated by radial chroma-
tography %eluent: from 100% CH2Cl2 to 5% methanol in
CH2Cl2). The desired monomethyl ester %4) was obtained
as a bright yellow-orange solid %0.048 g, 93%). Mp 280±
2828C;IR %KBr) n: 3334, 2918, 1658, 1626, 1456, 1362,
1
612 cm21; H NMR d: 8.03 %1H, brs), 7.44±7.32 %5H, m),
3.92 %2H, s), 2.13 %3H, s) ppm; 13C NMR d: 174.92, 151.21,
132.53, 131.83, 129.31, 128.44, 127.85, 50.46, 14.48 ppm.
1
3.1.10. 2,18-Des-vinyl-2,18-bis-phenyl-bilirubin-IIIa 33).
In a 50 mL round bottomed ¯ask, 3-methyl-4-phenyl-3-
pyrrolin-2-one %7) %0.288 g, 1.57 mmol) and diformyl-
1249, 1166, 869, 700 cm21; H NMR d: 13.60 %1H, brs),
10.96 %1H, s), 9.39 %1H, s), 9.21 %1H, s), 8.56 %1H, s), 7.48±
7.28 %10H, m), 6.12 %1H, s), 5.99 %1H, s), 4.04 %2H, s), 2.90±

7826
J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
2.77 %8H, m), 2.27 %3H, s), 2.20 %3H, s), 2.00 %3H, s), 1.93
%3H, s) ppm; 13C NMR d: 179.85, 176.30, 174.34, 171.18,
146.39, 142.15, 133.35, 132.33, 132.27, 131.86, 130.35,
129.78, 129.71, 129.60, 129.01, 128.70, 128.65, 128.43,
128.11, 127.60, 124.77, 124.70, 124.65, 123.07, 120.27,
118.96, 104.78, 99.55, 53.14, 34.07, 32.83, 22.52, 19.57,
18.99, 11.12, 10.57, 9.99, 9.08 ppm. Anal. calcd for
C42H42N4O6 %698.8): C, 72.19;H, 6.06;N, 8.02. Calcd for
C42H42N4O6´H2O %716.8): C, 70.37;H, 6.19;N, 7.82. Found:
C, 70.70;H, 5.77;N, 7.57.
2708C in the dark until analyzed by HPLC. For HPLC,
samples %20 mL) were mixed with 80 mL of ice-cold
0.1 M methanolic di-n-octylamine acetate %prepared by
dissolving equimolar quantities of di-n-octylamine and
acetic acid in methanol), microfuged for 30 s and 20 mL
of the clear supernate was injected onto the column. Biliary
excretion curves were derived by plotting HPLC peak
areas, determined by integration, against time and are not
corrected for small variations in bile ¯ow-rate or differences
in the extinction coef®cients of the individual components.
For hydrolysis of glucuronides with b-glucuronidase, bile
%20 mL) was mixed with 40 mL of b-glucuronidase solution
%prepared by adding 1 mL of water to one vial of bacterial
b-glucuronidase on ice), the mixture incubated for 1 h at
378C in the dark and then mixed vigorously with 140 mL
of 0.1 M di-n-octylamine acetate in methanol. This mixture
was microfuged for 30±60 s and an aliquot %20 mL) of the
supernate was taken for HPLC. For base hydrolysis of
glucuronides, bile %20 mL) was mixed with 1 M NaOH
%10 mL). After 30 min at room temperature, 10 mL 1 M
HCl was added followed, after mixing, by 160 mL of
0.1 M di-n-octylamine acetate in methanol. The mixture
was vortexed, microfuged for 30±60 s and an aliquot
%20 mL) of the supernate was taken for HPLC.
3.1.13. 2,17-Des-vinyl-2,17-bis-phenyl-bilirubin-IXa 31).
2,17-Des-vinyl-2,17-bis-phenyl-bilirubin-IXa
12-mono-
methyl ester %4) %0.046 g, 0.068 mmol), disodium-EDTA
%2.0 mg), and ascorbic acid %50.0 mg) were dissolved
in 120 mL of methanol and 60 mL of 1 M aq. sodium
hydroxide. The solution was stirred under argon at 408C
for 30 min. After acidi®cation with acetic acid, the solution
was extracted with CH2Cl2 %3£50 mL). The combined
extracts were washed with water %150 mL) and brine
%50 mL), and dried over anhydr. MgSO4. The solvent was
removed on a rotary evaporator and the residue puri®ed by
radial chromatography %eluent: CH2Cl2) to afford 1 as a pure
orange-yellow product %0.025 g, 56%). Mp 315±3168C;IR
%KBr) n: 3408, 2912, 1693, 1681, 1573, 1440, 1404, 1249,
1
785 cm21; H NMR d: 13.74 %1H, s), 13.62 %1H, s), 10.92
%1H, s), 10.84 %1H, s), 9.35 %1H, s), 9.30 %1H, s), 7.48±7.28
%10H, m), 6.23%1H, s), 5.97 %1H, s), 4.10 %2H, s), 3.08±2.54
%8H, m), 2.27 %3H, s), 2.21 %3H, s), 2.02 %3H, s), 1.92 %3H,
s) ppm; 13C NMR d: 179.88, 174.29, 173.57, 146.47,
142.52, 134.20, 133.79, 132.32, 131.81, 129.75, 129.45,
129.40, 128.83, 128.69, 128.66, 128.56, 127.75, 127.23,
124.98, 124.80, 124.63, 124.53, 120.01, 119.88, 104.60,
102.30, 32.85, 32.76, 22.53, 18.82, 18.78, 11.04, 10.43,
10.32, 9.06 ppm. Anal. calcd for C41H40N4O6 %684.8): C,
71.91;H, 5.89;N, 8.18. Found: C, 71.97;H, 5.91;N, 8.08.
Acknowledgements
We thank the US National Institutes of Health %HD-17779,
DK 26307 and GM 36633) for generous support of
this work. We also thank Ms Wilma Norona for technical
assistance. Justin Brower is an R. C. Fuson Graduate
Fellow.
References
3.2. Animal studies
1. Bosma, P. J.;Seppen, J.;Goldhoorn, B.;Bakker, C.;Oude
Elferink, R. P.;Chowdhury, N. R.;Chowdhury, N. R.;Jansen,
P. L. J. Biol. Chem. 1994, 269, 17960±17964.
2. Coffman, B. L.;Green, M. D.;King, C. D.;Tephly, T. R. Mol.
Pharmacol. 1995, 47, 1101±1105.
Details of the experimental procedures used for studying the
hepatic metabolism and biliary excretion of 1±3 in rats have
been described in detail elsewhere.19,22 Brie¯y, a femoral
vein and the common bile duct were cannulated under
ketamine anesthesia and a liposomal solution7,10,19 con-
taining phosphatidylcholine %1.5 g), cholesterol %62 mg),
sodium cholate %12.95 g) and taurine %3.75 g) in 1 L of
water was infused %2 mL/hr) through the femoral catheter
to maintain hydration and ensure a steady bile ¯ow rate
%measured gravimetrically). Once the bile ¯ow rate and
body temperature %maintained with an infra-red heating
lamp) had stabilized %,30±60 min), a solution containing
0.25 mg pigment in 1 mL rat serum was infused via the
femoral vein as a bolus over a period of about 1 min. The
infusate was prepared by dissolving the pigment in 0.1 mL
DMSO, diluting this solution with 1 mL rat serum and
microfuging to remove any undissolved or particulate
material. Bile was collected in 20 mL aliquots from the tip
of the bile duct cannula immediately before injection of
pigment and at frequent intervals thereafter. The total length
of the biliary cannula was 3 in., of which ,1.25 in.
protruded to the outside through the skin. Collection of
each bile sample took about 20 s. Immediately after collec-
tion, samples were ¯ash frozen in dry-ice and stored at
3. Owens, I. S.;Ritter, J. K.;Yeatman, M. T.;Chen, F. J. Phar-
macokin. Biopharma. 1996, 24, 491±508.
4. Ramonas, L. M.;McDonagh, A. F.;Palma, L. A. J. Pharma-
col. Methods 1981, 5, 149±164.
5. Bonnett, R.;Davies, J. E.;Hursthouse, M. B.;Sheldrick, G. M.
Proc. R. Soc. Lond., Ser. B 1978, 202, 249±268.
6. Person, R. V.;Peterson, B. R.;Lightner, D. A. J. Am. Chem.
Soc. 1994, 116, 42±59.
7. Brower, J. O.;McDonagh, A. F.;Lightner, D. A. Tetrahedron
2000, 56, 7869±7883.
Â
8. Trull, F. R.;Rodrõ guez, M.;Lightner, D. A. Synth. Commun.
1993, 23, 2771±2783.
9. Mironov, A. F.;Evstigneeva, R. P.;Preobrazenskii, N. A.
J. Gen. Chem. USSR 1965, 35, 1938±1943.
10. Chen, Q.;Huggins, M. T.;Lightner, D. A.;Norona, W.;
McDonagh, A. F. J. Am. Chem. Soc. 1999, 121, 9253±9264.
In this reference may be found a simpli®ed high yield TosMIC
synthesis.
11. Kohori, K.;Kinoshita, H.;Inomata, K. Chem. Lett. 1995, 799±
800.
12. Kaplan, D.;Navon, G. Isrl. J. Chem. 1983, 23, 177±186.

J. O. Brower et al. / Tetrahedron 57 22001) 7813±7827
7827
13. Molecular mechanics calculations and molecular modeling
were carried out on an SGI Octane workstation using version
6.6 of Sybyl %Tripos Assoc., St Louis, MO) as described in
Ref. 6. The ball and stick drawings were created from the
atomic coordinates of the molecular dynamics structures
using MuÈller and Falk's Ball and Stick program for the
Macintosh.
17. Pu, Y.-M.;McDonagh, A. F.;Lightner, D. A. J. Am. Chem.
Soc. 1993, 115, 377±380.
18. Lightner, D. A.;Wijekoon, W. M. D.;Zhang, M.-H. J. Biol.
Chem. 1988, 263, 16669±16676.
19. McDonagh, A. F.;Lightner, D. A. Cell. Mol. Biol. 1994, 40,
965±974.
20. Eby, J. M.;Moore, J. A. J. Org. Chem. 1967, 32, 1346±1353.
21. Wineholt, R. L.;Wyss, E.;Moore, J. A. J. Org. Chem. 1966,
31, 48±52.
14. Lightner, D. A.;Gawronski, J. K.;Wijekoon, W. M. D. J. Am.
Chem. Soc. 1987, 109, 6354±6362.
22. McDonagh, A. F. The Porphyrins;Dolphin, D., Ed.;
Academic Press: New York, 1979;Vol. 6, pp 293±491.
15. Lightner, D. A.;An, J. Y.;Pu, Y. M. Arch. Biochem. Biophys.
1988, 262, 543±559.
16. Harada, N.;Nakanishi, K. Circular Dichroic SpectroscopyÐ
Exciton Coupling in Organic Stereochemistry;University
Science Books: Mill Valley, CA, 1983.