The use of synthetic linear tetrapyrroles to probe the
verdin sites of human biliverdin-IXa reductase and human
biliverdin-IXb reductase
Edward M. Franklin
1
, Seamus Browne
1
, Anne M. Horan
1
, Katsuhiko Inomata
2
, Mostafa A. S.
Hammam
3
, Hideki Kinoshita
2
, Tilman Lamparter
4
, Georgia Golfis
1
and Timothy J. Mantle
1
1 School of Biochemistry and Immunology, Trinity College, Dublin, Ireland
2 Division of Material Sciences, Graduate School of Natural Science and Technology, Kanazawa University, Ishikawa, Japan
3 Department of Chemistry, School of Science, Nagoya University, Aichi, Japan
4 Institut fu
¨
r Botanik I, Universita
¨
t Karlsruhe (TH), Germany
Introduction

and IXc isomers of biliverdin. Therefore, the activity of BVR-A can be
measured using biliverdin-IXa as a specific substrate. We now show that
the dimethyl esters of biliverdin-IXb and biliverdin-IXd are substrates for
BVR-B, but not for BVR-A. This provides a useful method for specifically
assaying the activity of both BVR-A and BVR-B in crude mixtures, using
biliverdin-IXa for BVR-A and the dimethyl ester of either biliverdin-IXb
or biliverdin-IXd for BVR-B. Human BVR-A has been suggested as a
pharmacological target for neonatal jaundice. Because of the absence of a
crystal structure with biliverdin bound to BVR-A, we have investigated
indirect ways of examining tetrapyrrole binding. In the present study, we
report that a number of sterically locked conformers of 18-ethylbiliverdin-
IXa are substrates for human BVR-A, and discuss the implications for the
biliverdin binding site. The oxidation of bilirubin-IXa ditaurate to biliver-
din-IXa ditaurate is also described. We show that biliverdin-IXa ditaurate
is a substrate for human BVR-A and discuss the possibility of using a com-
peting substrate, which is reduced to a water soluble and excretable rubin,
as a prototypic inhibitor of BVR-A.
Abbreviations
18EtBV, 18-ethylbiliverdin-IXa; BVR-A, biliverdin-IXa reductase; BVR-B, biliverdin-IXb reductase; DDQ, 2,3-dichloro-5,6-dicyanobenzoquinone;
hBVR-A, human biliverdin-IXa reductase; hBVR-B, human biliverdin-IXb reductase.
FEBS Journal 276 (2009) 4405–4413 ª 2009 The Authors Journal compilation ª 2009 FEBS 4405
cytoprotective levels of bilirubin-IXa [5]. This antioxi-
dant mechanism has been suggested to involve cycling
between biliverdin-IXa and bilirubin-IXa [5]. Intrigu-
ing evidence has been obtained from animal transplan-
tation studies [6–8] showing that the administration of
biliverdin-IXa is cytoprotective for the heart, colon
and liver. In a notable study, Yamashita et al. [8]
demonstrated that short-term treatment (3 weeks) with
biliverdin-IXa is sufficient to induce tolerance in a

dine nucleotide and biliverdin, we set out to further
probe the nature of the biliverdin binding pocket in a
significant extension to our earlier studies [2]. In the
present study, sterically-locked conformers of 18-ethyl-
biliverdin-IXa (18EtBV) are shown to be substrates for
human BVR-A (hBVR-A), and the implications for
the nature of the binding of biliverdin-IXa to BVR-A
are discussed.
We also show that biliverdin ditaurate is a substrate
for BVR-A and discuss the possibility of using a com-
peting substrate such as this to reduce the levels of the
lipophilic and potentially toxic bilirubin-IXa by pro-
ducing bilirubin ditaurate, which is water soluble and
readily excreted.
Additionally, by studying various biliverdin-IX iso-
mers, we present a method that specifically allows the
assay of both BVR-A and biliverdin-IXb reductase
(BVR-B) enzymes in crude preparations. Catalytically,
the major difference identified to date between the two
human enzymes is that BVR-B catalyses the reduction
of biliverdin-IXb, biliverdin-IXd and biliverdin-IXc,
but not biliverdin-IXa [2,10]. BVR-A prefers biliverdin-
IXa as substrate, but can reduce all three other isomers
[10]. Thus, although it is possible to specifically measure
the activity of BVR-A with biliverdin-IXa, it is not pos-
sible to specifically measure the activity of BVR-B with
any of the free acids of biliverdin (IXa,IXb,IXc or
IXd). We now show that BVR-B can reduce biliverdin-
IXb, biliverdin-IXd and biliverdin-IXc as their dimethyl
esters in clear distinction to BVR-A. This permits the

dependent decrease at 660 nm and an increase at
460 nm (Fig. 1B). Similar results were obtained with
the dimethyl ester of biliverdin-IXd (data not shown).
We did not have sufficient material for detailed studies
Probing the verdin site of BVR-A and -B E. M. Franklin et al.
4406 FEBS Journal 276 (2009) 4405–4413 ª 2009 The Authors Journal compilation ª 2009 FEBS
with the dimethyl ester of biliverdin-IXc. Initial rate
kinetics show that the dimethyl esters of the IXb and
IXd isomers all demonstrate substrate inhibition with
apparent substrate inhibitory K
i
values in the lm range
(data not shown).
Probing the biliverdin binding pocket of BVR-A
using synthetic biliverdin isomers
The substrate specificity of hBVR-A was examined
using a range of synthetic biliverdins locked in
various conformations. These biliverdin derivatives
have their C and D rings sterically locked by cycliz-
ing with an additional two- or three-carbon chain
[22] (and are termed 15Z-syn 18EtBV, 15Z-anti
18EtBV, 15E-syn 18EtBV and 15E-anti 18EtBV;
Fig. 2). All of the fixed isomers were substrates for
human BVR-A (Fig. 4). The spectra of all of the
fixed isomers exhibited the characteristic immediate
A
B
Fig. 1. Spectra of the hBVR-B catalysed reaction of the dimethyl
ester of biliverdin-IXb. (A) Biliverdin-IXb dimethyl ester (7 l
M)in

inhibition profile was evident (Fig. 5A–C). Although
we have not isolated or characterized the putative
bilirubin products from the locked isomer reactions,
the spectral changes that we observed are entirely
consistent with enzyme-catalysed formation of the
corresponding bilirubin. The possibility of a C-10
adduct (other than hydride from NADPH) is highly
unlikely. Thiol compounds will adduct to biliverdin
to give a yellow product; however, because neither
NADPH nor BVR-A alone were capable of initiating
an increase in A
460
, adduct formation of this type
can be ruled out. The fact that the linear increase
with time at A
460
required NADPH, BVR-A and the
locked biliverdin isomer is entirely consistent with
enzyme-catalysed hydride transfer from NADPH to
the various locked isomers. The demonstration that
locked isomers of biliverdin can be substrates for
BVR-A has been reported previously for the rat liver
enzyme [26], although the physiologically less rele-
vant IXc isomer and its locked variants were used in
that study.
Competitive substrates as inhibitors
As an alternative strategy to biliverdin reductase
inhibitors for human BVR-A, we were interested in
the possibility that competing substrates may provide
a means of inhibiting BVR-A in vivo. In the case of

enzymes of tetrapyrrole metabolism. The two down-
stream products of haem oxygenase activity, the linear
tetrapyrroles biliverdin-IXa and bilirubin-IXa, have
been identified as agents that improve the efficacy of
organ transplantation [8,27]. BVR-A has been identi-
fied as a pharmacological target for treating neonatal
jaundice [9]. Increasingly, there is a requirement for
straightforward assays that are specific for measuring
the activity of BVR-A and BVR-B in crude prepara-
tions. Until now, no method of specifically assaying
the activity of both enzymes in crude preparations with
biliverdin substrates has been proposed.
As described previously, the free acid of biliverdin-
IXa is not a substrate for hBVR-B [2], and the
dimethyl ester behaves similarly. The bridging propio-
nate side chains of biliverdin-IXa preclude access to
the substrate pocket of hBVR-B in a productive mode
and the tetrapyrrole rotates 90° compared to meso-
biliverdin-IVa [28]. The rotated configuration observed
with biliverdin-IXa bound to BVR-B is not consistent
with hydride transfer [28]. It is likely that a similar
binding orientation is adopted when the dimethyl ester
of biliverdin-IXa binds to hBVR-B. Biliverdin-IXa
dimethyl ester clearly does bind because it inhibits
hBVR-B activity with an apparent K
i
in the micro-
molar range. The dimethyl ester of biliverdin-IXa is
not a substrate for BVR-A (data not shown), which is
in agreement with previous findings [29]. Clearly, the

the same such that the reaction centre for both BVR-B
and BVR-A is likely to involve the B-face hydrogen
atom attached to C4 of the nicotinamide ring and the
two bridging pyrroles (rings B and C) with the linking
C10 methene bridge as the electrophilic centre. Over-
lapping conformationally stable forms of the sterically-
locked conformers of 18EtBV, where the two C-10
bridging pyrroles (B and C) are ‘fixed’, reveals that the
outer pyrrole rings (A and D) can adopt a variety of
conformations in the verdin binding site of BVR-A.
There are two models that require consideration to
define biliverdin binding in light of the present obser-
vations. The first model assumes that all four pyrroles
are bound in the active site, which must be able to
accommodate various conformations of the tetrapyr-
role. If we assume a helical ‘lock washer’ conformation
for the ‘nonlocked’ physiological substrate biliverdin-
IXa, then, although the B and C rings may deviate
slightly from planarity, the deviation for the outer A
and D rings is sufficient to suggest that they may inter-
act with the ‘roof’ and ‘floor’ of a hypothetical ‘verdin
binding site’, with the nicotinamide ring forming part
of the floor. If the bound conformation is helical, there
would be a significant distance between the ‘roof’ and
the ‘floor’, allowing a variety of conformational states
to be accommodated, as appears to be the case for the
locked isomers shown to be substrates in the present
study. Such a model would also allow the binding of
both P- and M-helical configurations described by
Hayes & Mantle [31] for the wild-type and Y102A

cyanobacterial [31] forms of BVR-A. The second
model is clearly applicable for BVR-B, where the inner
pyrroles are in good density [28], although the outer
pyrroles show low occupancy and are modelled in the
solvent. The second model does not allow protein
binding to stabilize either P- or M-helical forms, which
we clearly see in the case of both human and cyano-
bacterial forms of BVR-A.
Clearly, we require the crystal structures of ternary
complexes with biliverdin-IX a to define the biliverdin
binding site. It is likely that conformational changes
may be needed to facilitate biliverdin binding once the
enzyme-pyridine nucleotide complex has formed. This
information will allow the development of potential
BVR-A inhibitors. In this regard, we have shown that
biliverdin-IXa sulfonate is a substrate for human
BVR-A, although its synthesis is not straightforward.
The rubin product is water soluble and readily
excreted, whereas bilirubin-IXa requires conjugation
with glucuronic acid prior to excretion. The concept of
a competing substrate as a BVR-A inhibitor, where
the product is readily excretable, is of some interest. In
the present study, we have also shown that the synthe-
sis of biliverdin ditaurate is relatively straighforward
and that it is a substrate for human BVR-A. Our preli-
minary investigations indicate that an in vivo evalua-
tion of this compound is warranted to determine
whether it can competitively reduce the levels of the
potentially toxic bilirubin-Ixa, whereas the bilirubin
ditaurate is predicted to retain the ability to be elimi-

BF
3
⁄ MeOH [17]. The synthesis of the sterically locked
bilins has been described previously [20,21] and the
recorded absorbance spectra are reported [22]. Biliverdin
ditaurate was synthesized by oxidation of bilirubin-IXa
ditaurate using DDQ. Briefly, bilirubin-IXa ditaurate
(53 mg) was dissolved in 20 mL of sterile distilled water in
a round bottomed flask, to which 30 mg of DDQ was
added and mixed. The solution was allowed to react for
5 min and then silica gel (3 g) was added to the flask. The
biliverdin ditaurate was dried onto silica gel by rotary evap-
oration. A silica gel 60G column of 20 mL bed volume was
equilibrated in ethyl acetate (100 mL). The dried biliverdin
ditaurate ⁄ silica gel material was loaded onto the column.
The column was washed with 300 mL of ethyl acetate to
remove the reduced quinone. The column was then washed
with 150 mL of ethyl acetate ⁄ methanol (4 : 1, v ⁄ v) to elute
the unreacted quinone and bilirubin ditaurate. Finally, the
biliverdin ditaurate was eluted with 100% methanol and
dried to a powder by rotary evaporation. This material was
homogenous by TLC and the NMR spectrum of the final
product was consistent with oxidation at C10. The extinc-
tion coefficient at 660 nm at pH 6.8 is 11.3 mm
)1
Æcm
)1
.
Biliverdin-IXa sulfonate was a generous gift from Professor
David Lightner (University of Reno, NV, USA).

and spectra were recorded at intervals over a period of
20 min.
Acknowledgement
This work was supported by a grant from Science
Foundation Ireland.
References
1 Schluchter WM & Glazer AN (1997) Characterization
of cyanobacterial biliverdin reductase. Conversion of
biliverdin to bilirubin is important for normal phyco-
biliprotein biosynthesis. J Biol Chem 272, 13562–13569.
2 Cunningham O, Dunne A, Sabido P, Lightner D &
Mantle TJ (2000) Studies on the specificity of the tetra-
pyrrole substrate for human biliverdin-IXalpha reduc-
tase and biliverdin-IXbeta reductase. Structure-activity
relationships define models for both active sites. J Biol
Chem 275, 19009–19017.
3 Foresti R, Green CJ & Motterlini R (2004) Generation
of bile pigments by haem oxygenase: a refined cellular
strategy in response to stressful insults. Biochem Soc
Symp 71, 177–192.
4 Stocker R, Yamamoto Y, McDonagh AF, Glazer AN
& Ames BN (1987) Bilirubin is an antioxidant of possi-
ble physiological importance. Science 235, 1043–1046.
5 Baranano DE, Rao M, Ferris CD & Snyder SH (2002)
Biliverdin reductase: a major physiologic cytoprotec-
tant. Proc Natl Acad Sci USA 99, 16093–16098.
6 Nakao A, Otterbein LE, Overhaus M, Sarady JK,
Tsung A, Kimizuka K, Nalesnik MA, Kaizu T,
Uchiyama T, Liu F et al. (2004) Biliverdin protects the
functional integrity of a transplanted syngeneic small

(2001) Nuclear localization of biliverdin reductase in
the rat kidney: response to nephrotoxins that induce
heme oxygenase-1. J Pharmacol Exp Ther 296, 1091–
1097.
14 Trakshel GM, Rowley PT & Maines MD (1987) Regu-
lation of the activity of heme degradative enzymes in
K562 erythroleukemic cells: induction by thymidine.
Exp Hematol 15, 859–863.
15 Wheelock AM, Boland BC, Isbell M, Morin D, Weges-
ser TC, Plopper CG & Buckpitt AR (2005) In vivo
effects of ozone exposure on protein adduct formation
by 1-nitronaphthalene in rat lung. Am J Respir Cell
Mol Biol 33, 130–137.
16 Aldini G, Carini M, Vistoli G, Shibata T, Kusano Y,
Gamberoni L, Dalle-Donne I, Milzani A & Uchida K
(2007) Identification of actin as a 15-deoxy-delta12,14-
prostaglandin J2 target in neuroblastoma cells:
mass spectrometric, computational, and functional
approaches to investigate the effect on cytoskeletal
derangement. Biochemistry 46, 2707–2718.
17 Bonnett R & McDonagh AF (1973) The meso-reactivity
of porphyrins and related compounds. VI. Oxidative
cleavage of the haem system. The four isomeric biliver-
dins of the IX series. J Chem Soc [Perkin 1] 9
, 881–888.
18 Heirwegh KP, Blanckaert N & Van Hees G (1991)
Synthesis, chromatographic purification, and analysis of
isomers of biliverdin IX and bilirubin IX. Anal Biochem
195, 273–278.
19 McDonagh AF & Palma LA (1980) Preparation and

(1990) The enzymatic and chemical reduction of
extended biliverdins. Biochem Biophys Res Commun
171, 465–473.
27 Kato Y, Shimazu M, Kondo M, Uchida K, Kumamoto
Y, Wakabayashi G, Kitajima M & Suematsu M (2003)
Bilirubin rinse: a simple protectant against the rat liver
graft injury mimicking heme oxygenase-1 precondition-
ing. Hepatology 38, 364–373.
28 Pereira PJ, Macedo-Ribeiro S, Pa
´
rraga A, Pe
´
rez-Luque
R, Cunningham O, Darcy K, Mantle TJ & Coll M
(2001) Structure of human biliverdin IXbeta reductase,
an early fetal bilirubin IXbeta producing enzyme. Nat
Struct Biol 8, 215–220.
29 Colleran E & O’Carra P (1977) Enzymology and
comparative physiology of biliverdin reduction. In
Chemistry and Physiology of the Bile Pigments (Berk
PD & Berlin NI, eds), pp. 69–80, U.S. Dept. Health,
Education and Welfare, Washington, DC.
30 Smith LJ, Browne S, Mulholland AJ & Mantle TJ
(2008) Computational and experimental studies on
the catalytic mechanism of biliverdin-IXbeta reductase.
Biochem J 405, 61–67.
31 Hayes JM & Mantle TJ (2009) The effect of pH on the
initial rate kinetics of the dimeric biliverdin-IXa
reductase from the cyanobacterium Synechocystis
PCC6803. FEBS J, in press.