Bilirubin photo-isomers: Regiospecific acyl glucuronidation in vivo

Article abstract of DOI:10.1007/s00706-013-1076-6

(4Z,15Z)-Bilirubin-IXα, the end product of heme catabolism, requires uridine glucuronosyl transferase 1A1 (UGT1A1)-catalyzed glucuronidation for elimination in bile, where it appears as two isomeric monoglucuronides and a diglucuronide. When people are exposed to light, endogenous bilirubin is converted partly to photo-isomers that are produced in greater abundance during treatment of jaundiced babies with phototherapy. Little is known about the metabolism of the photo-isomers, other than that they appear not to require glucuronidation for elimination in bile. Studies have been hampered by their unavailability and instability, as well as confusion about the identity, structures, preparation, and purity of bilirubin photoproducts. This paper outlines methods for preparing photo-isomers of bilirubins in sufficient quantity and purity for metabolic studies in rats and reappraises the composition of some previous preparations. The studies show that (Z,E)-isomers of bilirubins and the structural isomer (Z)-lumirubin undergo glucuronidation in the rat, but unlike (4Z,15Z)-bilirubin, form only monoglucuronides. Moreover, glucuronidation is regiospecific for just one of the two propionic acid groups, the one attached to the isomerized half of the molecule. This unusual stereoselectivity appears to be dictated by intramolecular hydrogen bonding. Formation of hydroxylated bilirubins was not detected. During phototherapy, photo-isomers will compete with endogenous (4Z,15Z)-bilirubin for glucuronidation by nascent hepatic enzyme UGT1A1. Graphical abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s00706-013-1076-6

Monatsh Chem (2014) 145:465–482
DOI 10.1007/s00706-013-1076-6
ORIGINAL PAPER
Bilirubin photo-isomers: regiospecific acyl glucuronidation in vivo
Antony F. McDonagh
Received: 27 June 2013 / Accepted: 5 August 2013 / Published online: 14 December 2013
Ó Springer-Verlag Wien 2013
Abstract (4Z,15Z)-Bilirubin-IXa, the end product of
heme catabolism, requires uridine glucuronosyl transferase
appears to be dictated by intramolecular hydrogen bonding.
Formation of hydroxylated bilirubins was not detected.
During phototherapy, photo-isomers will compete with
endogenous (4Z,15Z)-bilirubin for glucuronidation by
nascent hepatic enzyme UGT1A1.
1
A1 (UGT1A1)-catalyzed glucuronidation for elimination
in bile, where it appears as two isomeric monoglucuronides
and a diglucuronide. When people are exposed to light,
endogenous bilirubin is converted partly to photo-isomers
that are produced in greater abundance during treatment of
jaundiced babies with phototherapy. Little is known about
the metabolism of the photo-isomers, other than that they
appear not to require glucuronidation for elimination in
bile. Studies have been hampered by their unavailability
and instability, as well as confusion about the identity,
structures, preparation, and purity of bilirubin photoprod-
ucts. This paper outlines methods for preparing photo-
isomers of bilirubins in sufficient quantity and purity for
metabolic studies in rats and reappraises the composition of
some previous preparations. The studies show that (Z,E)-
isomers of bilirubins and the structural isomer (Z)-lu-
mirubin undergo glucuronidation in the rat, but unlike
Keywords Bile pigments Á Enzymes Á Photochemistry Á
Jaundice
Introduction
Bilirubin-IXa (BR) is a yellow-orange bichromophoric
tetrapyrrole present in blood as a reversible non-covalent
complex with serum albumin. Cleared efficiently from the
circulation by the liver, it is eliminated in bile as two iso-
meric monoglucuronides and a diglucuronide whose
formation is catalyzed by a specific hepatic glucuronosyl
transferase isozyme, UGT1A1 [1]. The need for diglucu-
ronidation, when monoglucuronidation would suffice, has
long been a puzzle [2]. When the activity of UGT1A1 is
impaired, as in certain genetic disorders, BR accumulates in
the circulation, resulting in hyperbilirubinemia, which may
be manifested as jaundice. In newborn infants, UGT1A1 is
developmentally deficient while BR production is increased
(
4Z,15Z)-bilirubin, form only monoglucuronides. More-
over, glucuronidation is regiospecific for just one of the
two propionic acid groups, the one attached to the isom-
erized half of the molecule. This unusual stereoselectivity
Antony F. McDonagh: deceased October 22, 2012.
Correspondence: David A. Lightner, Department of Chemistry,
University of Nevada, Reno, Nevada 89557-0216, USA. e-mail:
lightner@unr.edu.
(relative to adults) and this combination can result in neo-
natal jaundice, which is common and which in severe cases
can cause irreversible brain damage [3, 4]. Phototherapy
with visible (blue) light is the most widely used first-line
therapeutic defense against severe neonatal jaundice [5].
BR absorbs visible light over a broad band of wave-
lengths centered around 460 nm and loses photoexcitation
energy primarily by radiationless processes, including
photo-isomerization reactions. When people are exposed to
light, a fraction of BR in the circulation is converted to two
Electronic supplementary material The online version of this
article (doi:10.1007/s00706-013-1076-6) contains supplementary
material, which is available to authorized users.
A. F. McDonagh (&)
Department of Chemistry, University of Nevada,
Reno, NV 89557-0216, USA
e-mail: lightner@unr.edu
123

4
66
A. F. McDonagh
HOOC
COOH
HOOC
COOH
8
12
H
N
5
15
1
0
O
E
Z
O
O
O
O
Z
Z
O
N
N
H
N
H
N
H
N
H
N
H
N
H
H
(
4Z,15Z )-Bilirubin
hν
hν
HOOC
COOH
HOOC
COOH
8
12
5
15
10
H
O
Z
Z
O
*
N
H
N
H
N
H
N
H
*
N
H
Z
O
N
N
(
N
H
H
(
4Z,15Z )-Bilirubin
Z )-Lumirubin
hν
hν
HOOC
COOH
HOOC
COOH
H
N
H
O
H
O
Z
E
*
N
O
N
H
N
H
N
H
*
E
N
N
N
H
H
(
E)-Lumirubin
Fig. 1 Configurational photo-isomerization of (4Z,15Z)-bilirubin to
its (4Z,15E)- and (4E,15Z)-isomers. A third photo-isomer, (4E,15E),
is not shown
Fig. 2 Structural photo-isomerization of (4Z,15Z)-bilirubin to (Z)-
lumirubin and configurational isomerization of the latter to (E)-
lumirubin. *Stereogenic centers
families of isomers, namely configurational isomers and
structural isomers [6–8]. The former result from reversible
twisting of the molecule about one or both of the unsym-
reversible, a photostationary state mixture of all four iso-
mers is reached when BR is irradiated in vitro, the
composition of which depends on the solvent and wave-
length of irradiation but in which the (4Z,15Z)-isomer
invariably predominates. [A claim that (4E,15E)-BR can be
separated chromatographically into two stable atropisomers
[13, 14] is not plausible (see ‘‘Results and discussion’’)].
Photo-isomerization reactions of BR in humans are
normally inconsequential but they play an essential role in
phototherapy of neonatal jaundice because the photo-iso-
mers, unlike the parent (4Z,15Z)-isomer, do not appear to
require glucuronidation for excretion into bile [5, 7, 8].
Thus, their formation facilitates the biliary elimination of
bilirubin pigments at a time when the normal pathway of
hepatic glucuronidation is inadequate. Formation of photo-
isomers during phototherapy also may act as a detoxifica-
tion process, rapidly reducing the concentration of the more
lipophilic (4Z,15Z), and presumably most toxic, isomer in
the circulation together with the attendant risk that it will
enter the brain and cause encephalopathy [15–17].
metrically substituted double bonds at C(4) and C(15)
1
Fig. 1); the latter from irreversible intramolecular cycli-
(
zation involving the endo vinyl group (Fig. 2). Of the two
competing processes, configurational isomerization is the
fastest, quantum yields for (Z ? E) isomerization of BR
bound to human serum albumin (HSA) being *0.1, some
5
0 times greater than for intramolecular cyclization [9–12].
Because configurational isomerization is photochemically
1
Lumirubin can be partially reconverted back to (4Z,15Z)-BR on
irradiation with UV light or prolonged irradiation with blue light [13,
2
6], but its formation is irreversible on the time scale required to
effect complete equilibration of (Z)-lumirubin and (E)-lumirubin.
Thus, irradiation of (Z)-lumirubin with blue light leads rapidly to a
photoequilibrium mixture of (Z)- and (E)-isomers, but not to a
photoequilibrium mixture with (4Z,15Z)-BR which appears only
along with overall pigment decomposition (McDonagh AF, unpub-
lished observations). Therefore on the time scales of configurational
isomerization and structural isomerization it is not inappropriate to
describe lumirubin formation as irreversible.
1
23

Bilirubin photo-isomers
467
Although hepatic UGT1A1 activity is relatively low at
birth and shortly thereafter, the enzyme is not totally
deficient and its activity increases during the first *10
postnatal days [3, 4]. Whether photo-isomers of BR are
substrates for UGT1A1 or other hepatic enzymes is
unknown and has been difficult to determine because of
their instability and unavailability, as well as confusion in
the literature regarding their structures, preparation, and
identification. However, as will be shown, their instability
can be turned to advantage in the identification of their
metabolites. This paper investigates the glucuronidation
and phase II metabolism of BR photo-isomers in the rat. It
outlines simple methods for preparing BR photo-isomers in
sufficient purity and quantity for in vivo studies and
reappraises earlier structural assignments and methods for
the preparation of ‘‘purified’’ photo-isomers and BR
O
O
COOH
H
O
O
N
H
N
N
H
N
H
Fig. 3 Spirolactone structure assigned to photobilirubin II and
cyclobilirubin [23, 26]
compound. In addition to their relevance to neonatal
jaundice and phototherapy, the studies provide an expla-
nation for the formation of BR diglucuronide in normal
metabolism and question the structures and importance of
various hydroxylated bilirubins [18–21] proposed as
products of BR metabolism and phototherapy.
2
photoproducts. The studies show that the principal BR
photo-isomers are substrates for a uridine glucuronosyl
transferase (UGT)1 enzyme in vivo but, in striking contrast
to the parent (4Z,15Z)-isomer, undergo highly regioselec-
tive monoglucuronidation, with no diglucuronidation. The
stereoselectivity of glucuronidation seems to be dictated by
the ability to form intramolecular hydrogen bonds, as
demonstrated by studies on a simple stable model
Results and discussion
For metabolism studies, solutions of pure individual photo-
isomers or, at least, solutions highly enriched in the content
of specific isomers are essential. In preliminary work the
methods of Stoll and colleagues for preparing ‘‘purified’’
photo-isomers of BR-IXa were examined, beginning with
photobilirubins IA and IB, which had been assigned
structures (4Z,15E)-BR and (4E,15Z)-BR, respectively
2
The nomenclature and identification of BR photoisomers in the
literature is confusing. The photoisomer structures shown in Figs. 1
and 2 were established unambiguously by NMR and other spectro-
scopic and chemical methods published in 1982 and, for simplicity,
the trivial name ‘‘lumirubin’’ (CAN 83664-21-5 and 83729-98-0) was
assigned to structural isomers with the novel cycloheptadienyl ring
[
13], followed by photobilirubin II, originally identified as
a separable mixture of two stable conformational isomers
of (4E,15E)-BR but later assigned other structures,
including that of (Z)-lumirubin [23].
(
Fig. 2) [7, 8]. Previously the term ‘‘photobilirubin’’ had been used to
describe early photoproducts of BR [22], but that term became
redundant once the individual photoisomer chemical structures had
been established. Yet the term was not completely abandoned,
appearing in the literature as photobilirubins IA and IB, photoprod-
ucts isolated by Stoll et al. who thought them to be (4E,15Z) and
Photobilirubins IA and IB
Samples were prepared by irradiation of BR in CHCl with
3
(
4Z,15E)-BR, respectively [13, 14]. However, as shown in this paper
a mercury lamp [13]. The final solution after irradiation
and the crude product obtained after extraction of photo-
isomers into acetone were dark green, indicating sub-
stantial by-product formation. Reversed-phase high-
performance liquid chromatography (HPLC) of the acetone
extract (Fig. 4) showed a complex mixture and revealed
that partial metathesis of the starting material had occurred.
Photobilirubin IA, isolated by thin-layer chromatography
those structure assignments are incorrect. Another photoproduct,
photobilirubin II, initially thought to contain two stable atropisomers
of (4E,15E)-BR [13, 14] was subsequently assigned either the
lumirubin structure (Fig. 2) or spiro-lactone structure (Fig. 3) [23],
and finally the lumirubin structure [24]. The lumirubin structure in
Fig. 2 is sometimes called (E,Z)-cyclobilirubin. Cyclobilirubin,
originally called ‘‘unknown pigment’’, was initially assigned the
(
[
4E,15E)-bilirubin structure [25], then the spiro-lactone structure
26], and eventually, in 1984 [27], the lumirubin structure elucidated
earlier [7]. The name is confusing because it implies that a (Z,Z)-
cyclobilirubin isomer could exist; however, such a structure is
stereochemically impossible. In 1987, Bonnett and Ioannou published
a table of structure/name correlations for the photoproducts isolated
by different investigators [28]. Unfortunately, there are errors in that
table. Adding further confusion, structures for several photoisomers
and for bilirubin glucuronides depicted in a more recent review [29]
are incorrect. The current paper uses unambiguous chemical nomen-
clature for configurational isomers and the trivial name lumirubin for
isomers with a cycloheptadienyl ring system linking two adjacent
pyrrolic rings formed by intramolecular cyclization of an endo vinyl
group. There is no longer a need for the ambiguous and confusing
photobilirubin or cyclobilirubin nomenclature.
(
(
TLC) as described [13], which had been assigned the
4Z,15E) structure, was found by HPLC to be a mixture
containing none of that isomer and with the approximate
bilirubin composition (based on relative peak areas) of 8 %
(
4E,15Z)-BR-XIIIa, 10 % (4E,15Z)-BR-IXa, 23 %
4Z,15Z)-BR-XIIIa, 47 % (4Z,15Z)-BR-IXa, and 11 %
(
(4Z,15Z)-BR-IIIa. Photobilirubin IB, which had been
assigned the (4E,15Z)-BR-IXa structure [13], contained
very small amounts of both (Z,E)- and (E,Z)-isomers of BR
and had the approximate composition (by HPLC) of 3 %
123

4
68
A. F. McDonagh
Fig. 4 Left HPLC
HOOC
COOH
chromatogram of partially
purified photobilirubins IA and
IB. Right Constitutional
structures of (4Z,15Z)-
bilirubins-XIIIa and -IIIa
O
O
N
H
N
H
N
H
N
H
(
Z,Z )-XIII
HOOC
COOH
O
O
N
H
N
H
N
H
N
H
(Z,Z )-IIIα
Repeated preparative TLC of this mixture yielded (Z)-lu-
mirubin-IXa in exceedingly low yield.
‘‘Photobilirubin’’ preparations
Careful repetition of the published procedures [13, 23, 30]
showed them to be unsatisfactory for the practical prepa-
ration of BR photo-isomers. Yields were exceedingly low
and products impure. This can be attributed to over-irra-
diation of solutions, use of unsuitable solvents, and an
inappropriate intense UV–Vis mercury light source with
strong emission within the main absorption band of BR in
the visible. Under these conditions BR-IXa undergoes
dismutation to -IIIa and -XIIIa isomers and isomerization
of the exo vinyl group to a highly reactive photo-isomer
Fig. 5 Partially purified product from the preparation of photobiliru-
bin II as described by Stoll et al. [13, 23]. The inset shows a
chromatogram of the final dark green (CH
irradiation of BR and before CH OH extraction of photo-isomers. The
chromatograms show only pigments with absorption at 450 nm
3 2
) SO solution after
3
[
31], as well as free radical reactions to verdinoid pig-
ments. Stoll et al. [13] originally ‘‘purified’’ photobilirubins
IA and IB [(Z,E)-isomers] by TLC on silica, which, as
noted earlier, catalyzes (E ? Z) reversion. Consequently,
the isolated products invariably contained (4Z,15Z)-BR
which was interpreted as evidence that ‘‘photobilirubins IA
and IB exist in solution as complexes with bilirubin’’ [13],
a speculation for which there is no experimental support.
(
4E,15Z)-BR-XIIIa, 2 % (4E,15Z)-BR-IXa, 3 % (4Z,15E)-
BR-IXa, 3 % (4Z,15E)-BR-IIIa, 8 % (4Z,15Z)-BR-XIIIa,
8 % (4Z,15Z)-BR-IXa, 25 % (4Z,15Z)-BR-IIIa, and 17 %
3
unidentified peaks. Thus, aside from questions of purity,
the structures originally assigned to photobilirubins IA and
IB [13] are incorrect. Contrary to the supposition of Stoll
et al. [13], the (4E,15Z)-isomer of BR-IXa precedes the
CHCl was used as solvent for the preparation of photo-
3
(
4Z,15E)-isomer on silica TLC, which is consistent with
bilirubins IA and IB, and (CH ) SO for preparation of
3
2
their relative mobilities on HPLC [8].
photobilirubin II. The choice of (CH ) SO was based on
3 2
the unfounded speculation that the solvent would break
hydrogen bonds in BR and favor formation of the
Photobilirubin II
(4E,15E)-photo-isomer [13]. Comparative solvent studies
Samples were prepared by irradiation of BR in (CH ) SO
3
showed this not to be the case and that CHCl3 and
(CH ) SO are poor solvents for preparing photo-isomers,
2
with a mercury lamp as described [13, 23]. The solution
was dark green at the end of the reaction indicating
extensive verdinoid by-product formation. HPLC revealed
a complex mixture containing less than 3 % lumirubin with
extensive disproportionation of the starting material to BR-
IIIa and -XIIIa (Fig. 5, inset). HPLC of the crude product,
3
2
as borne out in part by quantum yield measurements [10].
Ostrow et al. reported separation of (4E,15E)-BR by
TLC on silica into two stable ‘‘rotamers’’ termed photo-
bilirubins IIA and IIB [13, 14]. These were depicted as
conformational isomers (atropoisomers) that differed only
by rotation of the end rings about single bonds and were
supposedly stabilized by intramolecular hydrogen bonding.
Considering the low energy barrier for conformational
after CH OH extraction and before TLC purification
3
(
Fig. 5) showed numerous components, with (Z)-lumiru-
bin-IXa (photobilirubin II) as a minor constituent.
1
23

Bilirubin photo-isomers
469
interconversion in bilirubins [32–34], and the instability of
isomers and has the advantage that it is used isocratically
and can be recycled. Unfortunately, it is not useful for
preparative purification of photo-isomers because of the di-
n-octylamine acetate content.
(
E)-isomers during chromatography [35], that claim is
chemically implausible. The HPLC analyses in the present
paper show that all of the photo-isomer structural assign-
ments in the paper by Stoll et al. [13] are incorrect.
Consequently, the absorbance data assigned to particular
photo-isomers in the original paper are unreliable. (Con-
fusingly, several photo-isomer structures and the structure
of bilirubin diglucuronide in a more recent review [29] are
also incorrect.) In follow-up studies, Stoll et al. [30]
Injectate solutions highly enriched in (4Z,15E)-BR, the
isomer formed fastest during phototherapy of neonatal
jaundice [17], were prepared via photo-isomerization of
BR bound to HSA, which is highly regioselective for the
(4Z,15E)-isomer [8].
Preparation of (Z)-lumirubin was achieved by irradiation
of BR in CHCl /Et N beyond the initial photostationary
1
4
investigated the excretion of C-radiolabelled photobi-
lirubins I and II in Gunn rats. The composition of injectates
was not analyzed. The present studies, and data in the
published studies, show that the ‘‘purified’’ photo-isomers
injected were not, in fact, pure and were undoubtedly
complex mixtures which undermines the scientific basis for
the original conclusions.
3
3
state, followed by brief treatment with acid to revert all
(E)-isomers to the corresponding (Z)-isomers, CH OH
3
extraction, and TLC purification. Some decomposition of
lumirubin occurs during application of the pigment to the
TLC plate, but this can be minimized by working rapidly.
Much greater yields of lumirubin were obtained by irra-
diation of BR bound to HSA or rabbit serum albumin.
Since lumirubin formation is irreversible under the irradi-
ation conditions used in these studies, it should be possible
in theory to convert BR quantitatively to lumirubin. In
practice, however, unwanted secondary reactions set in
before complete conversion occurs, and optimal irradiation
times need to be determined in preliminary HPLC/UV–Vis
experiments. Since bilirubin photo-isomers are yellow,
final solutions after irradiation should also be yellow.
Green solutions indicate over-irradiation and formation of
unwanted by-products. No advantage was found to adding
EDTA to solutions [13].
Preparation of photo-isomers
Although photo-isomers of BR have yet to be isolated in
pure crystalline form, the present work shows that they can
be easily prepared photochemically in adequate quantity
and purity for metabolic studies in the rat by using iso-
merically pure starting material, visible light of appropriate
wavelength, a suitable deoxygenated solvent, and exclu-
sion of light during product work-up. In preparing
configurational isomers, solutions should be irradiated only
until a photostationary mixture of photo-isomers has
formed, as determined by preliminary difference spectra or
HPLC measurements, to avoid formation of by-products.
Yields are limited because the parent (4Z,15Z)-isomer
invariably predominates at photoequilibrium. CHCl /Et N
Exclusion of a spirolactone structure for lumirubin
A spirolactone structure (Fig. 3), with just one free car-
boxyl group, has at times been suggested for a photo-
isomer of BR [23, 26, 37]. This structure had been ruled out
on the basis of NMR studies [7], but because the present
studies were investigating glucuronidation, it was impor-
tant to definitively rule out the lactone structure by an
alternative method. Photochemical formation of the lactone
ring would be impossible if the C(8)-propionic acid were
methylated and, therefore, irradiation of BR-dimethyl ester
should not lead to photocyclization, whereas formation of
lumirubin dimethyl ester would rule out the cyclic struc-
ture. Accordingly, BR-dimethyl ester was irradiated in 1 %
NH OH/CH OH solution and the crude product treated
3
3
was used as solvent for preparing (Z,E)-isomers of BR-
IXa, BR-IIIa, BR-XIIIa, and dihydrobilirubin-VIIIa, this
solvent being chosen because it readily dissolves BR, is
easily removed under vacuum at room temperature,
provides a satisfactory quantum yield for (E ? Z) isomer-
ization, and inhibits acid-catalyzed reversion of (E)- to (Z)-
isomers. Separation of photo-isomers from starting mate-
rial relied on the differential solubility of BR and its photo-
isomers in CH OH [8, 35] and serum, which do not dis-
3
solve solid BR. Configurational isomers of BR undergo
slow spontaneous thermal (E ? Z) reversion in CH OH at
3
room temperature, so that CH OH extractions must be
3
4
3
done at low temperature and rapidly. Similar to the expe-
rience of others [35, 36], we found it impossible to purify
briefly with trifluoroacetic acid (TFA) to convert (E)-iso-
mers to (Z)-isomers. TLC showed the presence of only one
main product, which, after purification by TLC, showed an
(
E)-isomers by TLC on silica as reported by Stoll et al. [13,
0], or by HPLC on normal phase columns because of
E ? Z) reversion. However, the photo-isomers separate
3
absorbance spectrum (kmax = 429 nm) in CH OH similar
3
(
to that of lumirubin. This presumptive lumirubin dimethyl
ester was not characterized further, but on saponification
with NaOH was converted to lumirubin, which was iden-
tified by TLC comparison with authentic material using
well on reversed-phase HPLC with the methanolic di-n-
octylamine acetate solvent used in these studies. This sol-
vent inhibits acid-catalyzed reversion of configurational
123

4
70
A. F. McDonagh
three different solvent systems and by HPLC. Thus,
methylation of the COOH side chains of BR does not
prevent intramolecular photocyclization, ruling out the
spirolactone structure (Fig. 3) for lumirubin.
proportion of (Z)-lumirubin is small if solutions are not
over-irradiated.
(Z)-Lumirubin injectates
(
Z,E)-Isomer injectates
Irradiation of BR in CHCl /Et N, followed by brief treat-
3
3
ment with TFA to revert all (E)-configuration isomers to
the corresponding (Z)-isomers and preparative TLC,
afforded homogeneous (Z)-lumirubin, but yields were low
(*2 %) because of the low quantum yield for photocyc-
lization in that solvent [10]. Longer irradiation led to
unwanted side-reactions and products. However, quantum
yields for photocyclization of BR in the presence of excess
human or rabbit serum albumin in aqueous buffer are much
higher [11, 12, 38] and irradiation of BR in those solvents
led to much higher yields (*30–50 %) of (Z)-lumirubin,
albeit at the cost of a slightly more complicated work-up.
Millimolar extinction coefficients for (Z)-lumirubin were
Highly enriched solutions of (Z,E)-isomers of bilirubins-
IIIa, -IXa, and -IIIa in rat serum were easily prepared.
Irradiation of the corresponding (Z,Z)-isomers in CHCl3/
Et N to near-photoequilibrium was followed by removal of
3
solvent and extraction of (Z,E)-isomers into CH OH and
3
then rat serum. The time to reach photoequilibrium was
determined by preliminary absorbance difference mea-
surements. Final solutions were yellow, indicating no
formation of verdinoid by-products. (Looking ahead, typ-
ical photo-isomer enrichments can be seen by HPLC of the
injectates of Figs. 8 (inset), 12 (inset), 14a, and 16a.)
Product yields were limited by the isomer composition at
photoequilibrium, in which the (Z,Z)-isomer invariably
predominates. Common impurities in the preparations were
the parent (Z,Z)-isomer and, for pigments containing endo-
vinyl groups, lumirubin isomers. The proportion of (Z,Z)-
isomer in final product depends to a large degree on the
speed of work-up of the crude product, particularly during
found to be 33.0 in 1 % ammonia/CH OH (k
=
max
3
437 nm) and HSA solution (pH 7.4, kmax = 453 nm) and
39.0 ± 2.6 in the 0.1 M methanolic di-n-octylamine ace-
tate HPLC solvent (kmax = 440 nm); corresponding values
for (E)-lumirubin (emM/kmax) were 24.4/446–450, 24.1/
446–456, and 23.3/449, respectively.
the CH OH extraction step and evaporation steps. For
3
Excretion of photo-isomers in homozygous Gunn rats
glucuronidation studies, the presence of a small proportion
of the (Z,Z)-isomer in (Z,E)-preparations proved advanta-
geous because it facilitated the eventual identification of
Homozygous Gunn rats lack UGT1 activity and do not
glucuronidate BR [1, 2]. When subjected to phototherapy
with blue light the concentration of pigment in bile
immediately begins to increase, reaching a steady-state
level after about 2–3 h. Figure 6 shows typical HPLC
chromatograms (450 nm), on the same scale, of equal
volumes of Gunn rat bile collected before and after 90 min
of phototherapy. Before phototherapy the bile shows only
very minor peaks. One corresponds to a trace of (4Z,15Z)-
BR. The other two more polar peaks are acid labile and
correspond to (4E,15Z)- and (4Z,15E)-BR, probably gen-
erated during surgical preparation of the animal. During
irradiation of the animal there was copious excretion of
(Z)-lumirubin, (4E,15Z)-BR, and (4Z,15E)-BR, along with
(
Z,E)-glucuronides in bile. In the preparation of BR-IXa
photo-isomers, it is essential to use isomerically pure
starting material free of BR-IIIa and -XIIIa to ensure that
irradiations are not prolonged beyond the photostationary
state and do not cause disproportionation of the starting
material, to use narrow-band blue light, and to rigorously
maintain safelight precautions during extractions and
manipulation of product solutions. Consistent with the
experience of others [35, 36], attempted purification of
CH OH-extracted products by TLC [13] failed because of
3
(
acid-catalyzed) reversion of (E)- to (Z)-isomers.
Irradiation of BR-IXa in organic solvents and isolation
of the initially formed photo-isomers yields a mixture
highly enriched in both possible (4Z,15E)- and (4E,15Z)-
isomers. In contrast, photo-isomerization of the BR–HSA
complex in aqueous buffer is highly regioselective for the
(
15Z) double bond of BR [8, 38]. Injectate solutions highly
enriched in only (4Z,15E)-BR, the most rapidly formed
isomer in humans, were readily prepared in good yield by
irradiation of (4Z,15Z)-BR in HSA solution followed by
removal of HSA and extraction with CH OH and serum.
3
Contaminants in the product were (Z)-lumirubin, (4E,15Z)-
BR, and (4Z,15Z)-BR. The proportion of (4Z,15Z)-BR is
minimized by expeditious work-up of the product, and the
Fig. 6 HPLC chromatograms of bile collected from a Gunn rat
before and 90 min after beginning phototherapy
1
23

Bilirubin photo-isomers
471
lesser amounts of (E)-lumirubin, (4E,15E)-BR, and
intravenously, they do not appear to undergo significant
phase I or II metabolism in the absence of UGT1 enzymes.
(
4Z,15Z)-BR. Most of the (4Z,15Z)-BR is an artifact gen-
erated by thermal reversion of (E)-isomers as they exit
slowly through the biliary cannula; its proportion increased
with increasing length of the cannula.
Metabolism of (Z,E)-isomers in wild-type rats
Previous investigators have reported that Gunn rat bile
contains substantial quantities of poorly characterized
hydroxylated bile pigments whose concentration increases
during phototherapy or on induction of hepatic enzymes
Initial studies were done with solutions of the (Z,E)-iso-
mers of the symmetrically substituted BR analogues
(4Z,15Z)-BR-IIIa and (4Z,15Z)-BR-XIIIa in rat serum. On
intravenous (i.v.) administration to wild-type Sprague–
Dawley (SD) rats, each is rapidly eliminated in bile as a
mono- and a diglucuronide (MG and DG) whose HPLC
and spectral characteristics were determined in preliminary
experiments (not shown). The BR-IIIa photo-isomer in-
jectate contained, in addition to (Z,E)-BR-IIIa, relatively
small amounts of the (4E,15E)-isomer and the parent
(4Z,15Z)-isomer (Fig. 8, inset) which facilitated sub-
sequent metabolite identification. Figure 8 shows HPLC
chromatograms of SD bile before and 9 min after admin-
istration of the injectate. Comparison of the chromatograms
reveals that (Z,E)-BR-IIIa was excreted partly unchanged
but largely as a more polar metabolite. On treatment of bile
samples containing the metabolite with acid or on brief
exposure to blue light (Fig. 9), to convert (E)-isomers to
the corresponding (Z)-isomers, the metabolite peak and the
(Z,E)-BR-IIIa peak disappeared and the (Z,Z)-BR-IIIa
monoglucuronide and (Z,Z)-BR-IIIa peaks increased with
no increase in the (Z,Z)-BR-IIIa diglucuronide peak. This
identified the metabolite peak as a monoglucuronide of
(Z,E)-BR-IIIa. Consistent with this, on treatment of the bile
with b-glucuronidase all glucuronide peaks and the
metabolite peak disappeared, and a relatively large (Z,Z)-
BR-IIIa peak appeared (Fig. 10). During glucuronidase
treatment (E)-isomers revert to (Z)-isomers and glucuron-
ides are hydrolyzed to the parent aglycones.
[
18–21]. Extensive HPLC and TLC of concentrated
extracts of Gunn rat bile and of extracts treated with
alkaline methanolysis failed to reveal the presence of sig-
nificant quantities of such pigments in bile collected before
or during phototherapy. Trace amounts of bilirubins IX-b,
-c, and -d [39] were detected in bile collected in the dark,
consistent with the findings of Blanckaert et al. [40], along
with two pigments with bilirubinoid absorbance spectra
(
kmax = 443 and 449 nm, respectively). These may corre-
spond to the side chain hydroxylated BR derivatives found
by Blanckaert et al. [40] which are structurally different
from the hydroxylated compounds proposed by Ostrow
et al. [18–21].
When (Z)-lumirubin-IXa, (Z)-lumirubin-XIIIa, and the
(
Z,E)-isomers of BR-IIIa, -IXa and -XIIIa were injected
separately in Gunn rats the individual pigments were
excreted rapidly in bile, with no HPLC evidence of
metabolism to other rubins. To illustrate relative rates of
excretion, a synthetic mixture of (Z)-lumirubin, (4E,15Z)-
BR, and (4Z,15E)-BR-IXa in rat serum showing roughly
similar HPLC peak heights for each isomer was injected
intravenously as a bolus in a Gunn rat in the dark and bile,
collected at frequent intervals, was chromatographed
(
Fig. 7). The chromatograms confirm that all three photo-
isomers are excreted unchanged in bile and show that their
relative rates of excretion correspond to their order of
elution on reversed-phase HPLC. Thus, whether photo-
isomers are generated endogenously or are injected
Fig. 8 Metabolism of (4Z,15E)-BR-IIIa in a wild-type rat. The inset
shows an HPLC chromatogram of the highly enriched injectate and
the main panel shows chromatograms of bile taken just before and
9 min after injection. Most of the injected photo-isomer was excreted
in bile as a monoglucuronide metabolite (*), the remainder being
excreted unchanged
Fig. 7 Successive HPLC chromatograms of bile excreted by a Gunn
rat kept in the dark at various times after intravenous injection of the
pigment mixture shown in the inset. 1 (Z)-lumirubin, 2 (4E,15Z)-BR,
3
(4Z,15E)-BR, 4 (4Z,15Z)-BR
123

4
72
A. F. McDonagh
Fig. 11 Normalized absorbance spectra of
(
a
dashed line) and its monoglucuronide (solid line) and of
(4Z,15E)-BR-IIIa
b (4Z,15E)-BR-XIIIa (dashed line) and its monoglucuronide (solid
line)
Fig. 9 HPLC chromatograms of bile containing the metabolite (*) of
(
(
4Z,15E)-BR-BR-IIIa (top) after brief exposure to blue light and
bottom) after treatment with acid. Light exposure led to loss of the
metabolite peak and growth in (4Z,15Z)-BR-IIIa monoglucuronide,
along with the expected loss of (4Z,15E)-BR-IIIa and growth of
(
4Z,15Z)-BR-IIIa. Acid treatment resulted in loss of the metabolite
peak, a BR-IIIa monoglucuronide peak, and conversion of unchanged
photo-isomer to the parent isomer
Fig. 12 HPLC chromatograms of bile from a wild-type rat before
and 12 min after injecting a solution of (4Z,15E)-BR-XIIIa (inset)
intravenously as a bolus. The photo-isomer was excreted partly
unchanged and partly as a metabolite (*)
exposure to light or acid (Fig. 13) the metabolite peak
disappeared, a peak corresponding to (Z,Z)-BR-XIIIa
appeared, and there remained a strong peak at the retention
time of (Z,E)-BR-XIIIa. However, the absorption spectrum
of this strong peak was different from that of the (Z,E)-BR-
XIIIa photo-isomer and identical to that of a (Z,Z)-BR-
XIIIa monoglucuronide standard. In the HPLC system
used, the (Z,E)-BR-XIIIa photo-isomer and (Z,Z)-BR-
XIIIa monoglucuronide have similar retention times and
do not separate.
Fig. 10 Effect of b-glucuronidase on bile collected 15 min after
injecting enriched (4Z,15E)-BR-IIIa into a UGT1A1-competent rat.
All peaks disappeared and were replaced by (4Z,15Z)-BR-IXa and -
IIIa. The two minor peaks running close to the (4Z,15E)-BR-IIIa
metabolite (*) are the diglucuronides of (4Z,15Z)-BR-IXa and -IIIa
These observations provided strong evidence that the
(Z,E)-photo-isomers of both BR-IIIa and BR-XIIIa are
excreted partly unchanged and partly as monoglucuronides
in the rat, but they did not show which of the two non-
equivalent propionic acid groups in each photo-isomer had
undergone glucuronidation, i.e., whether just monoglucu-
ronide was formed. That question was answered by
studying the metabolism of (4Z,15E)-BR-IXa, the most
rapidly formed photo-isomer in humans.
Thus, (Z,E)-BR-IIIa is metabolized in part by a UGT1
enzyme to a monoglucuronide ((Z,E)-BR-IIIa monoglu-
curonide) that elutes close to endogenous (4Z,15Z)-BR-IXa
diglucuronide. This (Z,E)-monoglucuronide has a spectrum
(
kmax = 425 nm) different from that of the injected (Z,E)-
BR-IIIa isomer (Fig. 11) and rather similar to those of the
mono- and diglucuronides of (4Z,15Z)-BR-IIIa (not
shown).
Similarly to (E,Z)-BR-IIIa, (E,Z)-BR-XIIIa was also
excreted partly unchanged and partly as a more polar
metabolite with kmax = 424 nm (Figs. 11, 12). On
The (4Z,15E)-BR-IXa injectate was contaminated with
unchanged (4Z,15Z)-BR-IXa and (Z)-lumirubin, which
proved difficult to avoid (Fig. 14). However, as shown in
1
23

Bilirubin photo-isomers
473
Fig. 13 Chromatograms of bile containing the metabolite (*) of
(
(
4Z,15E)-BR XIIIa (top) after brief exposure to blue light and
bottom) after treatment with acid. Light exposure led to loss of the
Fig. 14 a (4Z,15E)-BR injectate. b HPLC chromatograms of bile
collected before and at 6 min and 21 min after administration of the
(4Z,15E)-BR injectate as a bolus in a wild-type rat. c HPLC of bile
collected 18 min after injection of (4Z,15E)-BR (top) and of bile
collected 10 min after injection and then subjected to b-glucuronidase
hydrolysis. d Normalized absorbance spectra of (4Z,15E)-BR and its
(monoglucuronide) metabolite. *Indicates the major (monoglucuro-
nide) metabolite
metabolite peak and growth in (4Z,15Z)-BR-XIIIa monoglucuronide
which has the same retention time as (4Z,15E)-BR-XIIIa but is
readily distinguished from it by its absorption spectrum. Acid
treatment resulted in loss of the metabolite peak and a marked
increase in the (4Z,15Z)-XIIIa monoglucuronide
Fig. 7, (Z)-lumirubin is excreted much faster than
(
4Z,15E)-BR. HPLC of bile collected after the more rap-
CHCl /Et N to a photostationary state followed by
3
3
idly excreted lumirubin had disappeared showed
unchanged (4Z,15E)-BR and a large metabolite peak with a
retention time close to that of BR diglucuronide (Fig. 14b)
and (kmax = 426–428 nm, Fig. 14d). On treatment of a bile
sample with b-glucuronidase all peaks (except for a small
lumirubin peak) disappeared and there was a marked
increase in the (4Z,15Z)-BR peak (Fig. 14c). This, along
with the fact that no metabolite is formed in Gunn rats, is
consistent with a glucuronide structure for the metabolite.
On brief treatment of bile samples with TFA or on brief
exposure to blue light the metabolic peak partially disap-
peared along with growth of the (4Z,15Z)-BR C(12)-
monoglucuronide peak, loss of the (4Z,15E)-BR peak, and
growth of the (4Z,15Z)-BR peak (Fig. 15). From these
observations it can be concluded that (4Z,15E)-BR is
metabolized in the rat to a monoglucuronide and that
glucuronidaton is regioselective for the C(12)-propionic
acid side chain.
extraction into CH OH and then rat serum (Fig. 16a). After
3
injection of this mixture of photo-isomers two overlapping
metabolite peaks, slightly more polar than BR diglucuro-
nide, were observed in bile (Fig. 16b). The ratio of these
two metabolites changed with time; the least polar, longer
retention time peak appeared first, eventually giving way to
the most polar (Fig. 16c). This allowed absorption spectra
for each component to be determined (Fig. 16d). These
were similar, but not identical. The least polar metabolite
had kmax = 422–426 nm; the most polar, kmax
=
426–430 nm; with a spectrum that closely resembled that
of (4Z,15E)-C(12)-glucuronide determined in the previous
experiment. This suggested that the least polar, but most
rapidly excreted, pigment might be (4E,15Z)-C(8)-mono-
glucuronide. That was confirmed by brief exposure of an
early bile sample to blue light to convert (E)-isomers to
(Z)-isomers. Exposure resulted in a marked increase of
(4Z,15Z)-C(8)-monoglucuronide (Fig. 17), as well as an
increase in (4Z,15Z)-BR.
These conclusions were supported by similar experi-
ments with injectates containing both (4Z,15E)-BR and
These experiments show that (Z,E)-isomers of BR and
other rubins are substrates for a UGT1 (presumably
UGT1A1) enzyme in the rat that catalyzes the coupling of
glucuronic acid stereoselectively to the propionic acid side
(
4E,15Z)-BR. lnjectates highly enriched in those two
photo-isomers, without significant lumirubin contamina-
tion, were easily prepared by irradiation of (4Z,15Z)-BR in
123

4
74
A. F. McDonagh
Fig. 15 HPLC chromatograms of bile collected 12 or 9 min,
respectively, after injection of (4Z,15E)-BR into a wild-type rat
before and after brief exposure to blue light (top panel) or acid
Fig. 16 a HPLC chromatogram of injectate enriched in (4E,15Z)-
and (4Z,15E)-photo-isomers of BR. b Chromatograms of bile
collected before, and 12 and 24 min after injecting the mixture of
photo-isomers shown in panel a. c Time evolution of the metabolite
peaks following injection of a mixture of (4E,15Z)- and (4Z,15E)-
photo-isomers of BR into a wild-type rat. d Absorbance spectra of the
most and least polar components of the partially resolved metabolite
(bottom panel). *Indicates the major (monoglucuronide) metabolite of
4Z,15E)-BR
(
(monoglucuronide) peak shown in panel c. *Indicates two overlap-
ping metabolite peaks: less polar, more rapidly excreted (4E,15Z)-
chain attached to the dipyrrinone moiety containing the
isomerized (E) double bond.
C(8)-monoglucuronide and more polar, less rapidly excreted
(
4Z,15E)-C(12)-monoglucuronide
Metabolism of (Z)-lumirubin
lnjectate solutions of (Z)-lumirubin were isomerically
homogeneous and showed a single peak on HPLC
(
Fig. 18a). HPLC of bile samples collected after injection
of (Z)-lumirubin as a bolus in rat serum showed rapid
excretion of unchanged pigment, accompanied by a more
polar metabolite that eluted close to BR diglucuronide
(
Fig. 18a). The absorption spectrum of this metabolite
(
kmax = 435–441 nm) was identical in shape to that of the
parent (Z)-lumirubin but shifted by *2 nm to shorter
wavelength. On treatment of an early bile sample with b-
glucuronidase the metabolite peak disappeared and the
unchanged (Z)-lumirubin peak increased (Fig. 18b). On
treatment with NaOH the metabolite disappeared and the
Fig. 17 Effect of blue light exposure of bile collected 9 min after
injecting a mixture of (4E,15Z)- and (4Z,15E)-photo-isomers of BR as
a bolus in a wild-type rat. *Indicates unresolved mixture of
diastereoisomeric monoglucuronides containing predominantly the
least polar diastereoisomer. Light exposure led to a marked increase
in (4Z,15Z)-C(8)-monoglucuronide
(
Z)-lumirubin peak was retained, but underwent partial
decomposition (Fig. 18c). Both treatments resulted in
hydrolysis of endogenous BR glucuronide; and appearance
of (4Z,15Z)-BR as expected. Thus, (Z)-lumirubin is partly
metabolized in the rat to a glucuronide with an HPLC
retention time consistent with that of a monoglucuronide.
Similar results were obtained with (Z)-lumirubin-XIIIa.
The strategy used to confirm monoglucuronidation of
lumirubin and determine whether a particular one of the
two propionic acid side chains is selectively glucuronidated
is summarized in Fig. 19. The two isomeric C(8)- and
1
23

Bilirubin photo-isomers
475
monoglucuronide in which the glucuronic acid is conju-
gated with the propionic acid attached to the photocyclized
moiety of BR (Fig. 21).
Metabolism of dihydrobilirubin-VIIIa
Dihydrobilirubin-VIIIa (Fig. 22) is an analogue of
(4Z,15Z)-BR with the (C)12 propionic acid side chain
moved to C(13). When injected intravenously as a bolus in
a Gunn rat in the dark, unlike (4Z,15Z)-BR, the pigment
was excreted rapidly in bile (data not shown). In contrast,
in SD rats it was excreted partly in unchanged form, but
predominantly as a single more polar metabolite (Fig. 22a)
with an absorption spectrum that was consistent with a
glucuronide (Fig. 22b).
Metabolism of photo-isomers: summary
As illustrated in Fig. 7, (4Z,15E)-BR, (4E,15Z)-BR, and
(
Z)-lumirubin are excreted rapidly in bile without conju-
gation after intravenous injection in homozygous Gunn rats
deficient in UGT1 isozymes [7, 8]. In wild-type rats, the
three isomers, as well as (4Z,15E)-isomers of BR-IIIa and
BR-XIIIa, were excreted into bile partly in unchanged
form and partly as glucuronides, whose formation was
catalyzed presumably by UGT1A1. Strikingly, and in
contrast to the corresponding (4Z,15Z) parent isomers, the
photo-isomers formed only monoglucuronides. Further-
more, monoglucuronidation did not occur more or less
randomly on either of the two propionic acid side chains, as
with (4Z,15Z)-BR, but was specific for only one of the two
carboxyl groups, the one appended to the dipyrrinone
moiety in which photo-isomerization had occurred. In
linear molecular representations, as in Figs. 1 and 2, there
appears to be little difference between the two carboxyl
groups of each photo-isomer. However, 3D representations
of the photo-isomers (Fig. 23), based on the preferred
conformation of BR in organic solvents [41] or the crys-
talline state [42, 43], reveal that the carboxyl group that
becomes glucuronidated is the only one of the two that can
undergo intramolecular hydrogen bonding similar to that in
BR. Although the preferred conformations of BR photo-
isomers in vivo are not known, the observations suggest
that the catalytic site on UGT1A1 binds the photo-isomers
Fig. 18 a HPLC chromatograms of purified (Z)-lumirubin solution in
rat serum and chromatograms of bile collected before and 9 min after
injecting the solution as a bolus into a wild-type rat. b, c Effects of b-
glucuronidase and NaOH on bile samples collected 6 and 3 min,
respectively, after (Z)-lumirubin injection. *Indicates (monoglucuro-
nide) metabolite
C(12)-monomethyl esters of BR were prepared and irra-
diated to generate the corresponding isomeric lumirubin
monomethyl esters. After treatment with TFA to convert
(
E)-isomers to (Z)-isomers, each reaction product showed
two major peaks on HPLC (Fig. 20a)—a peak corre-
sponding to unchanged BR monomethyl ester and a peak
with a lumirubin absorption spectrum corresponding to one
of the two possible lumirubin monomethyl esters. These
two presumptive isomeric lumirubin monomethyl esters
were widely separated on HPLC as expected from 3D
models, the more polar being the lumirubin monomethyl
ester derived from BR C(8)-monomethyl ester. These were
compared by HPLC to a bile sample containing the
lumirubin metabolite that had been treated with alkaline
methanolysis to convert glucuronides to the corresponding
methyl esters (Fig. 20a). The lumirubin methyl ester from
methylated bile was indistinguishable on HPLC from the
lumirubin derived from BR C(8)-monomethyl ester.
Although the absorbance spectra of the two synthetic iso-
meric lumirubin monomethyl esters are almost identical
3
in a folded (P)- or (M)-helical conformation with the
almost planar, but chiral, intramolecularly hydrogen-bon-
ded moiety (Fig. 23) inserted into the active site rather than
the more exposed carboxyl group attached to the half that
(
Fig. 20b), they can be distinguished from each other, and
3
the absorbance spectrum of the lumirubin derived from rat
bile was identical to that of the synthetic lumirubin
monomethyl ester. Thus, (Z)-lumirubin, despite having two
(P) = plus, (M) = minus define the helical sense of the molecular
conformation and thus the chirality of the molecule, with reference to
the relative orientation of the component two dipyrrinone chromoph-
ores and their long wavelength electric transition dipole moments
lying along the long axis of each dipyrrinone [34].
carboxyl groups, is metabolized specifically to
a
123

4
76
A. F. McDonagh
Fig. 19 Strategy for
CH3OOC
8
COOH
12
HOOC
8
COOCH3
determining which propionic
acid side chain of (Z)-lumirubin
undergoes acyl glucuronidation
in vivo in the rat. Individual
diastereoisomers of (4Z,15Z)-
BR-IXa monomethyl ester were
converted photochemically to
the corresponding (Z)-lumirubin
monomethyl esters and these
were compared
12
O
O
O
O
N
H
N
H
N
H
N
H
N
H
N
H
N
N
H
H
1. hν
2. TFA
chromatographically to the
product obtained by alkaline
methanolysis of bile collected
from a wild-type rat after
injection of (Z)-lumirubin
CH3OOC
COOH
HOOC
COOCH3
H
H
O
O
O
O
N
H
N
N
H
N
H
N
H
N
N
H
N
H
KOH, MeOH
+
H , HPLC
Bile
HOOC
COOH
H
8
12
O
O
N
H
N
N
N
H
H
(
Z )-Lumirubin
UDPGA
UGT
HOOC
O
HO
HO
Fig. 20 a HPLC chromatograms of ꢀ product obtained by methan-
olysis of bile collected from a wild-type rat after intravenous injection
of (Z)-lumirubin, ` synthetic (Z)-lumirubin C(8)-monomethyl ester,
O
O
COOH
OH
´ synthetic (Z)-lumirubin C(12)-monomethyl ester. b Absorbance
spectra of (Z)-lumirubin C(8)- and C(12)-monomethyl esters
H
8
12
O
O
N
N
N
N
H
has not undergone photo-isomerization. If this is correct,
then (4Z,15Z)-bilirubin analogues bearing one propionic
acid located on C(8), as in BR itself, and a second propi-
onic acid located at a position other than C(8) or C(12)
which are sterically restricted from intramolecular hydro-
gen bonding should be excreted intact in Gunn rats and
excreted partly unchanged, and partly as a monoglucuro-
nide in wild-type rats. Studies on dihydrobilirubin-VIIIa
H
H
Fig. 21 Regioselective acyl glucuronidation of (Z)-lumirubin in vivo
the fact that it was not formed in Gunn rats, that it is a
monoglucuronide.
These results are relevant to neonatal jaundice and the
mechanism of phototherapy. There is no reason to believe
that glucuronidation of BR photo-isomers in humans and
rats will be qualitatively different. Since (4Z,15E)- and (Z)-
lumirubin undergo glucuronidation in vivo by the same
enzyme that catalyzes glucuronidation of the (4Z,15Z)-
isomer it is likely that they will compete with BR for
glucuronidation as the enzyme activity increases during
(
Fig. 22), along with previous studies on synthetic biliru-
bins [44], bear this out. In Gunn rats dihydrobilirubin-VIIIa
was excreted unchanged, but in wild-type rats it was
excreted partly as a metabolite. Although the metabolite
was not fully characterized, there is little doubt, on the
basis of its absorbance spectrum, HPLC retention time, and
1
23

Bilirubin photo-isomers
477
because competition between photo-isomers and BR for
conjugation will favor the latter statistically because it
carries two carboxyl groups that undergo glucuronidation.
Second, the observations strongly suggest that (4E,15E)-
BR and (E)-lumirubin, which lack carboxyl groups that can
form the hydrogen-bonded motif shown in Fig. 23, will be
completely excreted in bile unchanged and will not
undergo conjugation to glucuronides in vivo. These two
photo-isomers, which do not accumulate in serum during
phototherapy, may make a greater contribution to the
effectiveness of the treatment than generally thought.
Third, the pigment composition of bile of infants under-
going phototherapy is likely to be more complex than
previously believed. Initially, the bile will contain princi-
pally photo-isomers of BR along with relatively low
concentrations of (4Z,15Z)-BR formed by thermal rever-
sion of its configurational isomers in bile. However, as the
glucuronidating system in the liver becomes active these
pigments will be accompanied, not only by mono- and
diglucuronides of BR, but by monoglucuronides of (Z)-
lumirubin, (4E,15Z)-BR, and (4Z,15E)-BR.
HOOC
COOH
O
N
O
N
H
N
N
H
H
H
Fig. 22 Metabolism of dihydrobilirubin-VIIIa in the rat. a HPLC
chromatograms of bile before and 9 min after intravenous injection.
b Normalized absorbance spectra of metabolite (solid line) and parent
pigment (dashed line). *Indicates (monoglucuronide) metabolite
The observations lead to a stereochemical explanation
for the formation of BR diglucuronide in normal BR
metabolism. Monoglucuronidation is adequate to ensure
rapid biliary elimination of BR and diglucuronidation
appears to be redundant and unnecessary. For a time it was
thought that the two processes, mono- and diglucuronida-
tion, were catalyzed by two different enzymes [45], a view
no longer current. In its preferred conformation, BR has
two intramolecularly hydrogen-bonded dipyrrinones
(Fig. 23). These are structurally similar, but not identical
because of differences in the vinyl substitution sequence on
the two lactam end rings. There is little selectivity in
binding of one or the other of these to the UGT1A1 active
site because the C(8)- and C(12)-monoglucuronides are
formed in similar amounts when BR undergoes glucuron-
idation. Glucuronidation converts BR into a product that,
like BR photo-isomers, still has a single dipyrrinone moiety
capable of intramolecular hydrogen bonding. Once glucu-
ronidation has occurred and the monoconjugate leaves the
active site it can be transported into bile or, maintaining the
same (M)- or (P)-chirality, it can undergo a rotation of 180°
around an axis bisecting the C(9)–C(10)–C(11) bond angle
and present the remaining dipyrrinone moiety to the active
site for conjugation and formation of the diglucuronide.
Thus, it is likely that BR monoglucuronides compete with
unconjugated BR for glucuronidation and formation of BR
diglucuronide can be seen to be an inevitable and acci-
dental consequence of the 3D structure and symmetry of
the BR molecule rather than a ‘‘purposeful’’ process. The
final ratio of BR diglucuronide to monoglucuronides
excreted into bile, which shows a wide species variation
[46], will depend on intracellular transport dynamics.
Fig. 23 Preferred conformation of (4Z,15Z)-BR stabilized by intra-
molecular hydrogen bonding and hydrogen-bonded conformations of
BR photo-isomers based on this. Note: (1) Each conformer can exist
as a pair of mirror-image (M)- or (P)-chirality enantiomers which
interconvert rapidly in solution [33, 34]. Only a single enantiomer is
shown for each structure. The structure depicted for (4Z,15Z)-BR is
an (M)-chirality enantiomer. The inset shows the structural motif that
seems to be preferred by UGT1A1. (2) If monoglucuronidation of BR
involves preferential binding of, say, a (P)-chirality conformer, then
the monoglucuronide (and the photo-isomers) will have to maintain
the same chirality in order to undergo further glucuronidation
developmental maturation of the neonatal liver. In this
way, phototherapy might retard the normal excretion of BR
as the liver matures. However, this is unlikely to be a major
effect because BR is always in excess in the circulation and
123

4
78
A. F. McDonagh
Alternate pathways of BR metabolism
their extraction, chromatography, and isolation. Until their
structures, mechanism of formation, and physiological
relevance have been substantiated they should be viewed
with skepticism.
There are several reports that Gunn rats metabolize BR to
hydroxylated pigments whose formation is augmented by
inducers of cytochrome P450, such as tetrachlorodibenzo-
dioxin, and phototherapy. Although often referred to as
having established structures, these hydroxylated pigments
have not been well characterized. Three pathways have
been proposed to explain their genesis (Fig. 24) [18–21].
One pathway (Fig. 24a), suggested as important in photo-
therapy and in the metabolism of BR in the Gunn rat in the
dark [18], involves initial ‘‘phototautomerization’’ of BR
followed by regioselective reaction with singlet oxygen to
give a dioxetane which undergoes rearrangement to a vic-
diol. ‘‘Phototautomerization’’ of the type proposed has not
been observed in BR photochemistry and is energetically
unlikely not least because it involves migration of a
hydrogen from one saturated carbon center to another. The
source of the singlet oxygen implicated in the second step
was not explained and the cleavage of the 1,2-dioxetane in
the final step, which involves dissociation of a hydrogen
from a saturated carbon center, is highly improbable. A
second pathway (Fig. 24b) invokes addition of water to a
bis-lactim tautomer of BR and also dehydrogenation to
biliverdin followed by addition of water. However, water
does not add to BR or biliverdin as suggested, nor does
dehydrogenation of BR to biliverdin occur under physio-
logical conditions. And if it did, the biliverdin would be
reduced back to BR by biliverdin reductase in vivo. The
third pathway (Fig. 24c) postulates addition of hydrogen
peroxide, of unexplained origin, to the C(4) (or C(15))
double bond of BR, yielding a vic-diol, which spontane-
ously loses water to yield a C(5)- (or C(15)-)hydroxy-BR.
However, hydrogen peroxide does not normally add to
double bonds to give vic-diols, except under UV light [47].
Thus, the mechanisms proposed for the formation of
hydroxylated bilirubins in homozygous Gunn rats are all
highly unlikely.
Experimental
In general, solvents and solutions were purged with Ar
([99.9 % purity) before use. Isomerically pure BR-IIIa,
-IXa, and -XIIIa, used for synthesis of the corresponding
photo-isomers, were isolated by preparative HPLC on sil-
ica, crystallized from CHCl /CH OH, and dried overnight
3
3
at 65 °C under vacuum. Commercial BR (Porphyrin
Products, Logan, UT, and Koch-Light Laboratories,
Colnbrook, Bucks., UK, containing 2 % BR-IIIa, 95 %
BR-IXa, and 3 % BR-XIIIa by HPLC) was used for
preparation of photobilirubins I and II. Rat serum was
isolated from fresh rat blood, collected by venisection
under ether anesthesia, briefly evaporated on a rotary
evaporator to remove ether anesthetic, and frozen at below
-50 °C until needed. HSA and rabbit serum albumin were
from Sigma-Aldrich. Blue light refers to Westinghouse
Special Blue fluorescent lamps (F20T12/BB); 20 W lamps
were used for in vitro irradiation of solutions and 40 W
lamps for irradiating Gunn rats. All work, except for irra-
diation experiments, was done under safelights in a
darkroom. Where indicated, solutions were evaporated
rapidly at room temperature on a rotary evaporator fitted
with a vacuum pump and the final residues evacuated to
less than 13 mbar before being stored under Ar at below
-20 °C. For flash-freezing, bile samples in 6 9 50 mm
glass Pasteur tubes were plunged into powdered dry-ice.
Preparative and analytical TLC were run on 20 9 20 cm 9
250 mm silica gel G (Analtech, Newark, DE, USA) or
homemade silica gel H (E. Merck Laboratories) plates
activated at 110 °C for 1 h and used after equilibration in
air at room temperature. Silica K6 and silica gel D-O TLC
plates, used as described [13], were also investigated for
the preparation and separation of photobilirubins I and II.
Developed plates were viewed under blue light to improve
visualization of yellow bands or spots. Essential details of
the reversed-phase HPLC analyses have been described
previously [44]. Eluted peaks were detected at 450 nm and
relative peak areas measured with Hewlett-Packard HP
ChemStation software. The isocratic HPLC eluent was
During the present studies and our extensive studies on
the excretion and metabolism of BR photo-isomers and BR
model compounds in wild-type and Gunn rats and on the
effects of phototherapy on biliary pigment excretion in
Gunn rats, we have observed no evidence for the excretion
in bile of significant quantities of hydroxylated pigments of
the type proposed by Ostrow and co-workers [18–21]. As
can be seen in Figs. 6 and 7, BR photo-isomers are
excreted unchanged in Gunn rats and are the only signifi-
cant products absorbing near 450 nm excreted in bile
during phototherapy. We have found no evidence that
0.1 M di-n-octylamine acetate in CH OH containing from
3
3
2 to 8 % water at a flow rate of 0.75–1.00 cm /min. Used
solvent can be exposed to blue light and recycled. In a few
early studies less than 0.1 M di-n-dodecylamine acetate
was used in place of or in admixture with di-n-octylamine
acetate to improve peak resolution. Metabolism studies
were done in male homozygous Gunn rats or wild-type SD
‘‘phototherapy simply accelerates the alternate path-
way(s) of bilirubin catabolism which exist normally in the
Gunn rat’’ [48]. Possibly the putative hydroxylated pig-
ments are non-physiological artifacts generated during
1
23

Bilirubin photo-isomers
479
HOOC
COOH
HOOC
COOH
a
hν
O
N
O
O
OH
N
H
N
H
N
H
N
N
H
N
H
N
H
H
H
H
H
H
1O2
COOH
HOOC
COOH
HOOC
O
OH
OH
O
O
N
OH
O
OH
N
H
N
N
H
N
N
H
N
N
H
H
H
H
HOOC
COOH
HOOC
COOH
b
-
H2
HO
N
OH
HO
OH
N
H
N
N
N
N
N
H
N
H
H
H
+H2O
+H2O
HOOC
COOH
HOOC
COOH
HO
OH
HO
N
OH
N
N
H
N
H
N
N
H
N
H
N
H
OH
OH
HOOC
COOH
HOOC
COOH
c
H2O2
O
N
O
O
O
OH
N
H
N
H
N
N
N
H
N
H
N
H
H
H
H OH
H
H
H
H
-H2O
HOOC
COOH
O
O
N
H
OH
N
H
N
H
N
H
H
H
Fig. 24 Mechanisms proposed for formation of hydroxylated bilirubins in vivo in homozygous Gunn rats during phototherapy and in the dark
18–21] (redrawn from original figures)
[
rats weighing more than 250 g fitted with a short
*7.5 mm) indwelling biliary cannula for bile collection
and a femoral venous line for administration of samples as
previously described [39, 44]. The external segment of
biliary cannula was painted black or red with nail varnish
and experiments were done under safelights in a darkroom.
Surgery was done in the dark or under safelights with only
the small area of the abdomen and abdominal cavity being
operated on illuminated with light from a fiber-optic hal-
ogen-lamp illuminator. For generation of photo-isomers
(
123

4
80
A. F. McDonagh
3
in vivo, homozygous Gunn rats, with dorsal hair removed
with a surgical depilatory cream, were exposed in restraining
cages to 40-W Special Blue lights for *4 h and bile was
collected.
UV–Vis analysis and 0.5- to 1.0-cm portions for i.v.
injections in metabolism studies. These injectates con-
tained predominantly (4Z,15E)-BR along with lesser
amounts of (4Z,15Z)-BR and (4E,15Z)-BR and (Z)-lumiru-
bin as seen in Fig. 14a. The final pigment concentration in
(
Z,E)-Bilirubin-IIIa, (Z,E)-bilirubin-XIIIa,
injectates, estimated spectroscopically, varied from 0.3 to
3
.0 mg/cm . Substitution of rabbit or pigeon serum albu-
and (4E,15Z)/(4Z,15E)-bilirubin IXa injectate
1
Bilirubin-IIIa, -XIIIa, or -IXa (1.3 mg) was dissolved in
min for HSA in the above preparation gives final solutions
containing both (4E,15Z)- and (4E,15Z)-stereoisomers of
BR, with the former predominating.
3
3
1
0 cm Ar-purged CHCl /Et N (1:1) in a 15-cm Erlen-
3 3
meyer flask and the solution irradiated from below with
blue light with continuous bubbling of Ar for 10 min
(
photostationary state). The solution was flash-evaporated
(Z)-Lumirubin-IXa injectate
A fresh solution of BR (2.5–2.7 mg) in 0.2 cm Ar-
3
under vacuum and the residue rapidly rinsed thrice with
3
ice-cold CH OH (2, 2, 1 cm ). The rinsings were decanted,
degassed 0.1 M NaOH was diluted into a solution of 0.28 g
3
rabbit serum albumin in 10 cm Ar-purged 0.1 M potas-
3
sium phosphate buffer, pH 7.4 in a 15-cm Erlenmeyer
3
briefly centrifuged, combined, and flash-evaporated under
3
vacuum in a 25-cm round-bottom flask, and stored until
required at below -50 °C under Ar. Residues from two
preparations were rinsed with 1.5 cm rat serum and the
flask. The solution was purged slowly with Ar for 10 min
in the dark and then, with continued Ar bubbling, irradiated
3
liquid phase decanted and centrifuged. Samples of the clear
from below with blue light for 4 h. The yellow solution
3
was then added slowly to 50 cm 0.1 M NH
supernate were taken for HPLC and UV–Vis analyses and
3
.5-cm portions were for i.v. injections in metabolism
4
OAc in
0
CH
OH and the mixture centrifuged. The supernate was
3
3
decanted into a 250-cm round-bottom flask and the residue
studies. Pigment concentrations in the injectates, deter-
mined spectroscopically using emax/kmax for BR in rat
serum (50,000/400 nm; A.F. McDonagh and L.A. Palma,
unpublished observations) were in the range 0.1–0.8 mg/
3
rinsed with 5 cm methanolic NH OAc. The combined
4
supernate and rinsings were evaporated to a moist residue
3
on a rotary evaporator. This was mixed with 10 cm
3
cm . Typical enrichments are shown in the insets of Figs. 8
CH Cl/CH OH (9:1, v/v) and the organic phase washed
3 3
3
with water (2 9 10 cm ), clarified by centrifugation, and
and 12 and in Fig. 16a. Injectate solutions contained
predominantly (Z,E)-photo-isomers along with a relatively
small proportion of the corresponding (Z,Z)- and (E,E)-
isomers.
3
transferred to a 25-cm round-bottom flask. To this solution
3
was added 0.020 cm TFA, to convert exocyclic (E) to
(Z) double bonds, and the solution flash-evaporated at once.
The residue was dried under high vacuum, rinsed with
3
–5 cm ice-cold Ar-purged CH OH in three portions, and
(
4Z,15E)-Bilirubin-IXa injectate
BR (2.5–2.7 mg) was dissolved in 0.2 cm Ar-degassed
.1 M NaOH and the solution added at once to 0.28 g HSA
2
3
3
3
the rinsings decanted into a 12-cm centrifuge tube and
centrifuged. The clear supernate was flash-evaporated to
dryness and the residue stored at below -50 °C under Ar.
0
3
in 10 cm Ar-purged 0.1 M potassium phosphate buffer,
3
pH 7.4 in a 15-cm Erlenmeyer flask. The solution was
3
For metabolism experiments this was dissolved in 1.5 cm
purged slowly with Ar gas (beware of frothing) for 10 min
and then, with continued Ar bubbling, irradiated from
of rat serum to give an absorbance of *2 in a 1-mm-
pathlength cell; 0.5-cm aliquots were used for injectates.
3
below with blue light for 10 min. The solution was shaken
3
with 50 cm 0.1 M ammonium acetate (NH OAc) in
HSA can be substituted for rabbit albumin, but yields are
somewhat lower. Yields were in the range 35–50 %,
measured spectrophotometrically in NH OH/CH OH
4
CH OH and centrifuged for 1 min. The supernate was
3
4
3
rapidly evaporated on a rotary evaporator to a moist residue
(1:99, v/v). (Z)-Lumirubin-XIIIa injectates were prepared
(
speed is essential at this step to avoid poor yields and
similarly.
3
enrichments). The residue was mixed with 10 cm CHCl /
3
CH OH (9:1, v/v) and the organic phase washed with water
3
The following simpler, lower yield, procedure not
3
3
2 9 10 cm ) using brief centrifugation to separate phases.
(
involving protein was also used. BR (2.5 mg) in 10 cm
3
The organic phase was evaporated to dryness on a rotary
Ar-purged CHCl /Et N (1:1) in a 15-cm Erlenmeyer was
3 3
evaporator and then on a high-vacuum line. The residue
3
was rinsed rapidly with 5 cm ice-cold Ar-purged CH OH
irradiated with blue light with continuous bubbling of Ar
for 4 h and the solution evaporated to dryness under vac-
3
uum. The residue was dissolved in 10 cm CHCl
3
in three portions and the combined rinsings were centri-
fuged. The supernate was flash-evaporated and the residue
further dried under vacuum. Three such preparations were
/CH
3
3
OH
3
(9:1, v/v), 0.020 cm TFA was added, and the solution
flash-evaporated. Crude products from two such prepara-
3
tions were mixed with 5 cm ice-cold Ar-purged CH
3
mixed with 1.5 cm rat serum and the solution centrifuged.
3
OH
Samples of the clear supernate were taken for HPLC and
and the mixtures decanted and centrifuged to remove
1
23

Bilirubin photo-isomers
481
insoluble bilirubin. The CH OH solution was evaporated to
3
procedure [24] using ammoniacal CH
3
OH as solvent and
dryness and the residue purified by preparative TLC on
silica gel G. Pigment was applied to the plates in Ar-purged
CHCl /CH OH (1:1) and plates developed with 1 % acetic
blue fluorescent lights was not investigated because
preliminary studies indicated relatively low quantum yields
for lumirubin formation in that solvent.
3
3
acid (HOAc) in CHCl /CH OH (9:1). The isolumirubin
3
3
Isomeric (Z)-lumirubin-IXa monomethyl esters
and lumirubin bands were collected [7], combined, and
eluted with developing solvent and the solution evaporated.
Yield: 1.7 %. For animal studies the residue was rinsed
Isomeric C(8)- and C(12)-BR-monomethyl esters [50] were
separated and purified by preparative TLC on 250-lm
silica G plates using CHCl /CH OH/glacial acetic acid
3
3
3
with 1.5 cm rat serum and the liquid phase decanted and
(HOAc) (97:2:1, v/v/v) as developing solvent and eluent.
centrifuged. Samples of the clear supernate were taken for
3
HPLC and UV–Vis analysis and 0.5-cm portions were
The C(8)-isomer runs ahead of the C-12) isomer. The
identity of the products was confirmed by HPLC and TLC
comparison with authentic samples (kindly provided by Dr.
D.A. Lightner, University of Nevada, Reno) and literature
4
used for i.v. injections.
Photobilirubins IA and IB
The published procedure [13] was followed exactly on a
smaller scale, starting with 20 mg rather than 200 mg
commercial BR. Briefly, BR was irradiated in CHCl3,
data [50, 51]. Individual purified isomers were dissolved in
3
Ar-saturated CH
OH (*0.35 cm ) to give an absorbance
3
of *2 in a 1-mm-pathlength quartz cuvette and each
solution was irradiated with blue light through a Lucite
filter. Changes were followed by absorbance difference
under N for 2 min/mg BR with an unfiltered 100-W Hg
2
spotlight, CHCl3 was removed by evaporation and the
residue extracted with acetone. Evaporation of the acetone
extract gave crude product to which TLC on silica was
applied, using CHCl /CH OH/H O (40:9:1, v/v/v) as
3
measurements and 0.040-cm samples were taken at
intervals for HPLC. By 60 min most of the starting
material had been converted to photo-isomers. The solution
3
3
2
developing solvent. Bands corresponding to photobiliru-
remaining after 60 min irradiation was evaporated to
3
dryness, the residue dissolved in 0.2 cm CHCl
bins IA and IB (R values *0.42 and 0.52) were collected
3
, and
f
3
0.0020 cm TFA added. This solution was immediately
and pigments eluted.
3
evaporated to dryness and the residue taken in 0.080 cm
Photobilirubin II
CH OH for HPLC.
3
The published procedures [13, 23] were followed exactly
on a smaller scale. BR (20 mg) was irradiated in (CH ) SO
3
2
Preparation and hydrolysis of lumirubin dimethyl ester
containing ethylenediaminetetraacetic acid (EDTA) with a
00-W Hg spotlight as outlined above for the preparation
1
BR dimethyl ester (13 mg; Porphyrin Products/Frontier
3
Scientific, Logan, UT) in 100 cm Ar-purged 1 % NH OH/
of photobilirubins IA and IB. The solution was diluted with
CHCl and water and the CHCl extracts washed with
4
3
3
3
CH OH in a 500-cm Erlenmeyer flask was irradiated from
water, filtered, and evaporated. The residue was extracted
3
below with continuous Ar bubbling for 4 h. The solution
was evaporated to dryness and the residue dissolved in
with CH OH, the extracts evaporated, and the residue
3
applied to a TLC plate in CHCl . The plate was developed
3
3
0 cm CHCl . The solution was washed with 3 9 50 cm
3
5
and the photobilirubin II band(s) collected and eluted. Two
slightly different variations of the procedure have been
published. In the first, the UV light was unfiltered,
3
water, filtered through CHCl -moistened filter paper, sat-
3
3
urated with Ar, and 0.1 cm TFA added. The solution was
immediately evaporated and the residue chromatographed
(
CH ) SO was purged with N , CHCl /CH OH/formic
3 2 2 3 3
on four CH OH-washed silica gel G plates (analytical
3
acid (30:3:1) was used for TLC development, and eluted
pigments in CHCl were washed with 0.1 M NaHCO and
thickness) with CHCl /CH OH/HOAc (94:5:1, v/v/v) as
3
3
3
3
irrigant. The main yellow product band (which turned
brown if not eluted at once) was collected and eluted with
the developing solvent. The residue after evaporation of
solvent was rechromatographed on three plates and the
main yellow band collected and eluted. This showed a
single yellow spot on analytical TLC and one peak on
water to remove formic acid [13]. In the second, the
spotlight was filtered through a Wratten 2A filter (to cut out
Hg emission at 365 and 405 nm), (CH ) SO was purged
3
2
with Ar, and CHCl /CH OH/H O (40:9:1) used for TLC
3
3
2
development [23]. Both variants were investigated. A third
3
4
HPLC. The purified product was dissolved in 1 cm
3
degassed CH OH and 0.1 cm of this was diluted to 5 cm
(
Z)-Lumirubin has two stereogenic centers, denoted by *in Fig. 2.
3
Prepared by irradiation of BR-HSA, (Z)-lumirubin is not identical to
Z)-lumirubin prepared by irradiation of BR in CHCl /Et N. Because
3
(
3
3
for UV–Vis absorbance measurements (k
To the remaining 0.9 cm of solution was added 0.5 cm
= 429 nm).
max
of chiral induction by the protein the former is chiral and optically
active whereas the latter is a racemate and optically inactive [49].
Both preparations behaved identically in the in vivo studies and are
not distinguished in this paper.
3
3
1
5
M NaOH and the solution kept at 37 °C under Ar for
min, followed by addition of 5 cm water and adjustment
3
123

4
82
A. F. McDonagh
of the pH to 4 with HOAc. This solution was extracted with
3
22. Lightner DA, Wooldridge TA, McDonagh AF (1979) Proc Natl
Acad Sci U S A 76:9
3
5
cm CHCl and the extract washed with 5 cm water and
3
2
3. Stoll M, Vicker N, Gray CH, Bonnett R (1982) Biochem J
01:179
evaporated to dryness. The residue was analyzed by TLC
and HPLC.
2
24. Bonnett R, Buckley D, Hamzetash D, Hawkes G, Ioannou S, Stoll
M (1984) Biochem J 219:1053
Acknowledgments I thank Wilma Norona, Lucita A. Palma, and
Andrew Phimister for technical assistance and the National Institutes
of Health for partial financial support.
25. Isobe K, Onishi S (1981) Biochem J 193:1029
26. Onishi S, Itoh S, Isobe K, Sugiyama S (1981) Photomed Photo-
biol 3:59
2
7. Onishi S, Miura I, Isobe K, Itoh S, Ogino T, Yokoyama T,
Yamakawa T (1984) Biochem J 218:667
2
2
3
3
3
3
8. Bonnett R, Ioannou S (1987) Mol Aspects Med 9:457
9. Vitek L, Ostrow JD (2009) Curr Pharm Design 15:2869
0. Stoll M, Zenone E, Ostrow J (1981) J Clin Invest 68:134
1. Cheng L, Lightner D (1999) Photochem Photobiol 70:941
2. Manitto P, Monti D (1976) J Chem Soc Chem Commun 1976:122
3. Navon G, Frank S, Kaplan D (1984) J Chem Soc Perkin Trans
References
1
2
. Ikushiro S (2010) Drug Metab Rev 42:13
. Chowdhury R, Chowdhury NR, Gartner U, Wolkoff AW, Arias
IM (1982) J Clin Invest 69:595
3
. Hanchard NA, Skierka J, Weaver A, Karon BS, Matern D, Cook
W, O’Kane DJ (2011) BMC Med Genet 12:57
. Miyagi SJ, Collier AC (2011) Drug Metab Disp 39:912
. Maisels MJ, McDonagh AF (2008) New Engl J Med 358:920
. McDonagh AF (1986) N Engl J Med 314:121
2
:1145
4. Person R, Peterson BR, Lightner DA (1994) J Am Chem Soc
16:42
3
3
4
5
6
7
1
5. Falk H, M u¨ ller N, Ratzenhofer M, Winsauer K (1982) Monatsh
Chem 113:1421
6. Sloper R, Truscott T (1982) Photochem Photobiol 35:743
7. Onishi S, Itoh S, lsobe K, Togari H, Kitoh H, Nishimura Y (1982)
Pediatrics 69:273
8. Kanna Y, Arai T, Tokumaru K (1993) Bull Chem Soc Jpn
6
9. McDonagh AF (1979) Bile pigments. Biladienes and 5,15-bil-
atrienes. In: Dolphin D (ed) The porphyrins, vol VI. Academic,
New York, p 293
. McDonagh AF, Palma LA, Lightner DA (1982) J Am Chem Soc
04:6867
3
3
1
8
9
. McDonagh AF, Palma LA, Trull FR, Lightner DA (1982) J Am
Chem Soc 104:6865
. Agati G, Fusi F, Pratesi R, McDonagh A (1992) Photochem
Photobiol 55:185
3
3
6:1586
1
1
0. Kanna Y, Arai T, Sakuragi H, Tokumam K (1990) Chem Lett 631
1. Kanna Y, Arai T, Tokumaru K (1993) Bull Chem Soc Jpn
6
6:1482
4
0. Blanckaert N, Fevery J, Heirwegh KP, Compernolle F (1977)
Biochem J 164:237
1
1
2. McDonagh AF, Agati G, Fusi F, Pratesi R (1989) Photochem
Photobiol 50:305
3. Stoll MS, Zenone EA, Ostrow JD, Zarembo JE (1979) Biochem J
4
4
1. Dorner T, Knipp B, Lightner D (1997) Tetrahredron 53:2697
2. LeBas G, Allegret A, Mauguen Y, De Rango C, Bailly M (1980)
Acta Cryst B 36:3007
¨
1
83:139
1
1
1
1
4. Cohen AN, Ostrow JD (1980) Pediatrics 6:740
5. Hansen TWR (2010) Sem Perinatol 34:231
6. McDonagh AF, Palma LA, Lightner DA (1980) Science 208:145
7. Mreihil K, McDonagh AF, Nakstad B, Hansen TR (2010) Pediat
Res 67:656
4
3. Bonnett R, Davies JE, Hursthouse MB, Sheldrick GM (1978)
Proc Royal Soc Lond B 202:249
4. McDonagh A, Lightner DA (2007) J Med Chem 50:480
5. Jansen PL, Chowdhury JR, Fischberg EB, Arias IM (1977) J Biol
Chem 252:2710
4
4
1
1
8. Berry CS, Zarembo JE, Ostrow JD (1972) Biochem Biophys Res
Comm 49:1366
9. Cohen AN, Kapitulnik J, Ostrow JD, Webster CC (1986) Hepa-
tology 6:490
0. Ostrow JD (1972) Prog Liver Dis 4:447
1. Ostrow JD, Kapitulnik J (1986) Alternate pathways of heme and
bilirubin metabolism. In: Ostrow JD (ed) Bile pigments and
jaundice. Marcel Dekker, New York, p 421
4
4
4
4
6. Spivak W, Carey MC (1985) Biochem J 225:787
7. Milas N, Kurz P, Anslow W Jr (1937) J Am Chem Soc 59:543
8. Ostrow J, Branham R (1970) Birth Defects Orig Artic Ser 6:93
9. McDonagh AF, Lightner DA, Reisinger M, Palma LA (1986) J
Chem Soc Chem Commun 249
2
2
5
5
0. Lightner D, Trull F, Zhang M (1987) Tetrahedron Lett 28:1047
1. Blanckaert N (1980) Biochem J 185:115
1
23