Vol. 268, No. 35,Issue of December 15. PP.

26099-26106,1993
Printed in U.S.A .

THEJOURNAL
OF BIOLOGICAL
CHEMISTRY
8 1993 by The American Society for Bioehemiatry and Molecular Biology, Inc.

Inactivation of Phytochrome- andPhycobiliprotein-Chromophore

Precursors by Rat Liver Biliverdin Reductase*
(Received for publication, June 28, 1993, and in revised form, August 4, 1993)

Matthew J. Terry$, Mahin D. Mainesi, andJ. Clark LagariasSIl

From the $Section of Moleculur and Cellular Biology, Uniuersity of California, Davis California 95616 and the Department of
Biophysics, University of Rochester School of Medicine, Rochester,New York 14642

The phytochrome chromophore precursor, 3E-phy- Maines, 1990; Maines, 1990). Despite its heterogeneity at the
tochromobilin, and the phycobiliprotein chromophore protein level, BVR appears to be encoded by a single gene in
precursors,
3E-phycocyanobilin
and
3E-phycoeryrats (Fakhrai and Maines, 1992). Unique among eukaryotic
throbilin, are enzymatically converted to novel rubi- enzymes in having two distinct pH optima, BVR prefers
noid products by purified rat liver biliverdin reduc- NADH at neutral pH, while NADPH is preferred at pH 8.7
tase. Phytochromobilin and phycocyanobilin are par- (Kutty and Maines, 1981; Fakhrai and Maines, 1992). BVR
ticularly good substrates for
biliverdin reductase with also exhibits a broad bilin substrate specificity. Experiments
K,,,
and V,, values very similar to those
of the natural using a large number of structurally modified biliverdins have
substrate,biliverdin I X a . Phycoerythrobilin is the established that themajor requirement for reduction is a pair
least preferred of the three bilin substrates. H NMR
spectroscopy of phycocyanorubin, the productof phy- of propionic acid substituents (Awruch et al., 1984; Frydman
et al., 1987).
cocyanobilin catalysis by biliverdin reductase,and
In this study, we have examined the naturally occurring
comparison of absorption spectra
of all three rubinoid
products reveal that the C10 methine bridge is selec- linear tetrapyrroles, phycocyanobilin (PCB), phytochromotively reduced bybiliverdin reductase withoutaltering bilin (P@B), andphycoerythrobilin (PEB), aspotential BVR
In vitro phyto- substrates (see Fig. 1 for structures). P a B serves as the
theA-ringethylidenesubstituent.
immediate precursor of the chromophore of the plant photochromeassemblyexperimentsdemonstratethatthe
phytorubin products do not form photoactive adducts receptor phytochrome (Terry and Lagarias, 1991; Cornejo et
with recombinant apophytochrome. These
results sug- al., 1992; Terry et al., 1993), while PCB and PEB are chrogest that ectopic expression of biliverdin reductase inmophore precursors of the phycobiliprotein antennae complants will prevent assembly of the functional photo- plexes of algae and cyanobacteria (Glazer, 1989; Beale,1993).
receptor and thuswill potentially alter light-mediated Recent studies demonstrate that all three bilins can form
plant growth and development.
covalent adducts with apophytochrome in vitro (Li and La-

Biliverdin reductase (BVR) catalyzes the reduction of biIiverdin IXa (BV) to bilirubin (BR) in thefinal step of heme
catabolism in mammals and some species of fish (Singleton
and Laster, 1965). A cytosolic enzyme with a molecular mass
of approximately 33,000, BVR utilizes bothNADPHand
NADH for the reduction of BV (Kutty and Maines, 1981). At
the protein level, BVRexhibits extensive microheterogeneity.
At least 10 variants with different molecular mass and/or
charge have been reported in mammals (Huang et al., 1989a).
The pattern of expression of an individual BVR variant has
been shown to be tissue dependent and can be modified by
xenobiotic compounds (Huang et al., 1989b; Huangand
*This workwas supported by grants from the United States
Department of Agriculture AMD-9203377 (to J. C. L.) and National
Institutes of Health Grant ES04066 (to M.D. M.). The costs of
publication of this article were defrayed in part by the payment of
page charges. This article must therefore be hereby marked aduertisement in accordance with 18U.S.C. Section 1734 solelyto indicate
this fact.
ll To whom correspondence should be addressed Section of Molecular and Cellular Biology, University of California, Davis CA 95616.
Tel.: 916-752-1865; Fax: 916-752-3085.
* The abbreviations used are: BVR, biliverdin reductase; BV, biliverdin IXa; BR, bilirubin IXa; MBR, mesobilirubin IXa; PCB,
phycocyanobilin; PCB DME, phycocyanobilin dimethyl ester; PCR,
phycocyanorubin; PEB, phycoerythrobilin; PER, phycoerythrorubin;
PcPB, phytochromobilin; PaR, phytochromorubin; HPLC, high performance liquid chromatography; BSA, bovine serum albumin; ppm,
partfmillion.

garias, 1992). These bilins each contain the reduced A-ring

with an ethylidene substituent,astructural
feature which
appears to be required for proper assembly with their respective apoproteins (Arciero et al., 1988a, 1988b, 1988~;Li and
Lagarias, 1992). Until now, none of the bilin substrates for
BVR so far examined possess this ethylidene moiety. The
present studies demonstrate that BVR effectively converts
these ethylidene-containing bilins to a new class of rubinoid
pigments which cannot covalently assemble with apophytochrome in uitro. These resultssuggest a novel approach to the
regulation of plant and algal biliprotein function uia ectopic
expression of BVR.
EXPERIMENTALPROCEDURES

Biliuerdin Reductase Preparations-Biliverdin reductase was purified to homogeneity from rat liver essentially as described previously
(Kutty and Maines, 1981) using the modifications of Huang et ai.
(1989a). Protein concentrationwas measured by the method of Lowry
et al. (1951) using bovine serum albumin as a protein standard.
Bilin Preparations-Biliverdin IXa (BV) was prepared from bilirubin (Vega Biochemicals, Tucson, AZ) as described previously (Elich
et al., 1989). Mesobilirubin (MBR) was prepared from BR by catalytic
hydrogenation (Stoll and Gray, 1977). 3E-phytochromobilin (PaB)
and 3E-phycoerythrobilin (PEB) were isolated from solvent-treated
cells of Porphyridiurn cruenturn (La Jolla strain) using a modification
of the HgC1,-assisted cleavageprotocol described previously (Cornejo
et al., 1992). After cleavage and removal of HgC12, the crude bilin
mixture was diluted 10-fold with 0.1% (v/v) trifluoroacetic acid and
applied to a (21, Sep-Pak cartridge (Waters-Millipore Corp., Milford,
MA). The Sep-Pak cartridge was sequentially washed with 0.1%
(v/v) trifluoroacetic acid (2 X 3 ml) and acetonitrile/O.l% trifluoroacetic acid (2080; 2 X 2 ml) followed by elution of the crude bilin

26099

Biliverdin Reductase Substrate Specificity

26100
H-C

COzH

bilirubin (BR)
HO&

COaH

mesobilirubin (MBR)
C02H
H02C

phycocyanorubin (PCR)
HOX

COzH

phytochromorubin (PiPR)
H%C

C02H

phycoerythrorubin (PER)
FIG. 1.

study.

Structures of bilatriene substrates used in this

mixture with 3 ml of acetonitrile/O.l% trifluoroacetic acid (60:40)

and drying in vacuo. BE-phycocyanobilin (PCB) was prepared from
lyophilized Spirulinaplotensis (Sigma) using a method similar to that
described by Beale and Cornejo (1991). Spirulina powder was rehydrated in deionized water (30 ml/g dry weight) for 10 min. The slurry
was centrifuged at 30,000 X g for 20 min, and the deep blue supernatant was decanted. Phycocyanin was precipitated from this supernatant with 1%(w/v) trichloroacetic acid by incubation for 1 h at
4 "C in the dark and then collected by centrifugation at 30,000 X g
for 20 min. After washing with methanol (2 X 20 ml/g Spirulinu
powder), the blue pellet was resuspended in methanol (2 ml/g Spirulina powder) containing 1 mg/ml HgCl,. Following incubation for
20 h a t 42 "C in darkness,the protein was removed bycentrifugation
at 10,000 X g for 10 min. 2-Mercaptoethanol(l pl/ml) was then added
to precipitate the dissolved mercuric ion, which was also removed by
centrifugation (30,000 X g for 10 min). The crude PCB solution was
then partially purified on a Cla Sep-Pak and dried in vacuo. The 3Eisomers of all three bilins were purified by CIS reverse-phase HPLC
using a Varian 5000 liquid chromatograph and a Beckman Ultrasphere ODS column (4.6 X 150 mm; 5-pm particle diameter). The
solvent system used was ethanol/acetone/water/acetic acid,
19:14661 (v/v/v/v) with a flow rate of 1.5 ml/min, and the column
eluate was monitored a t 370 nm. HPLC-purified bilins were concentrated using a CIa Sep-Pak as described above, dried in vacuo, and
stored at -20 "C. Before use, bilins were dissolved in dimethyl sulfoxide to aconcentration of1-1.5mM.An
aliquot of each stock
solution was diluted 200-fold into 2% HCl/methanol to estimate the
bilin concentration spectrophotometrically. The extinction coefficients of 66,200 M" cm" at 377 nm for BV (McDonagh and Palma,
1980), 64,570 M" cm" at 386 nm for 3E-PGB (Weller and Gossauer,
1980), 25,200 M" cm" at 594 nm for 3E-PEB (Chapman et al., 19671,
and 47,900 M" cm" at 374 nm for 3E-PCB (Cole et al., 1967) were
used for these determinations.
Apophytochrome Preparations-Recombinant
apophytochrome
was prepared from the yeast Saccharomycescerevisiae strain 29A that
contained the full-length oat phytochrome (phyA3) cDNA plasmid
pMphyA3 as described previously (Li and Lagarias, 1992). For assembly experiments (see below), apophytochrome was diluted from a
stock solution of127.8 pg/ml in 25 mM Tris-HC1 buffer, pH 7.8,
containing 25% (v/v) ethylene glycol, 1 mM dithiothreitol, 1 mM
EDTA, and 2 mM phenylmethylsulfonyl fluoride which was stored at
-80 "C.
Biliverdin Reductase Assays-NADPH-dependent BVR activity
was assayed in 1ml of 0.1 M Tris-HC1 buffer, pH 8.7, containing 0.1
mM NADP+, 1 mM glucose &phosphate, 0.1 units/ml glucose-6phosphate dehydrogenase (type XI1 from torula yeast), 1mg/ml BSA,
1-20 PM bilin substrate, and 200-800 ng/ml BVR. NADH-dependent
BVR activity was assayed in 1 mlof0.1
M potassium phosphate
buffer, pH 7.0, containing 0.2 mMNAD', 1 mM glucose 6-phosphate,
0.2 units/ml glucose-6-phosphate dehydrogenase (type XXIV from
Leuconostoc rnesenteroides), 1 mg/ml BSA, 1-20 pM bilin substrate,
and 200-800 ng/ml BVR. Assays were
performed a t room temperature
(25 "C), and all reactions were initiated by the addition of BVR. Bilin
reduction was monitored spectrophotometrically using a Hewlett
Packard 8450A UV visible spectrophotometer.
Extraction of Rubin Products for Absorption and 'H NMR Spectroscopy-Rubinswere
extracted into methylene chloride-ethanol
according to the method of Fevery et al. (1983) with the following
modifications. After the BVR reduction was judgedto be complete, 1
ml of 0.3 M NaCl, 5 ml of methylene chloride/ethanol (l:l, v/v), and
3.5 ml of0.4 M HCl (adjusted to pH 1.8 with glycine) were added
sequentially to 1-ml reaction mixtures. The mixture was shaken
vigorously, and the lower organic phase containing the rubins was
collected. After re-extraction of the upper phase with 2 ml of methylene chloride/ethanol ( L l , v/v), the lower phases were combined,
washed three timeswith an equal volume of water, and dried in vacuo.
For 'H NMR spectroscopy of phycocyanorubin (PCR), a 30-ml reaction mixture consisting of 0.1 M Tris-HC1 buffer, pH 8.7, containing
0.1 mM NADP', 1 mM glucose &phosphate, 0.1 units/ml glucose-6phosphate dehydrogenase (type XI1 from torula yeast),1mg/ml BSA,
30 p~ PCB (added in 300 p1 of distilled ethanol), and 800 ng/ml BVR
was incubated for 2 h at room temperature under argon. Following
completion of the reaction, PCR was extracted according to the
procedure described above with the addition of two extra steps.
Between the first and second water washes, the organic phase was
washed with freshly made 0.1 M NaHC03 to remove unmetabolized
and chemically oxidized bilins. After drying in vacuo, the sample was
dissolved in distilled methylene chloride, filtered through a small

Biliverdin Reductase Substrate Specificity

26101

of PCB and P+B reduction were based on the absorbance

changes shown in Fig. 2 and themolar absorption coefficients
of the respective rubinoid product. These coefficients were
determined from the known concentration of the bilin starting
material andthe empirically determined relationship between
substrate and product absorbances (see Table I). Owing to
the complexity of the PEB conversion described above, the
analysis of the initial rates of PEB reduction was based on
the disappearance of PEB substrate. All three bilin substrates
obeyed Michaelis-Menten kinetics and catalytic constants for
both NADPH- and NADH-dependent BVR activities shown
in Table I1 werecalculated from Lineweaver-Burk reciprocal
plots(datanot
shown). Under our assay conditions, the
M
NADPH and 4.7
measured K, values for BV (i.e. 3.7 ~ L with
p~ with NADH) proved to be very close to the published
values of 3.0 and 5.0 p~(Kutty andMaines, 1981). K,,, values
for both PCB andP@Bare only slightly larger than those for
BV while their V,, values are slightly lower those of the
natural substrate for both NADPH and NADH-dependent
BVR activities (Table 11). K , values for PEB are nearly 3fold larger than those for BV, although the Vma, values are
11).These analyses indicate
not substantially different (Table
that all three bilins are effective substrates for BVR, with
PCB and P@B having an affinity for the enzyme similar to
the natural substrate.
B VR Reduces the ClO Methine Bridge of the Plant andAlgal
RESULTS
Bilins-For further characterization, the tetrapyrrole prodThe Three Plant and Algal Biliprotein Chromophore Pre- ucts of the BVR-catalyzed reduction of P+B, PCB, and PEB
cursors Are BVR Substrates-The catalysis of 3E-PCB, 3E- were purified using a modified solvent extraction protocol
P+B, and 3E-PEBby both NADPH- and NADH-dependent (Fevery et al., 1983). With this method, BR can be recovered
activities of BVR was examined spectrophotometrically. For in nearly quantitative yield following BVR-catalyzed reduccomparative purposes, BV was analyzed under identical assay tion of BV (data not shown). Fig. 3 illustrates the absorption
conditions. Fig. 2 shows that the threebilins were substrates spectra of the three solvent-extracted tetrapyrrole products.
for both BVR activities. These experiments show that the For the PCB and P+B reduction products, two absorption
bilins PCBand P+Bare converted to productswith an maxima are observed as shown in Fig. 3 (panels a and b), one
absorption maximum between 400 and 450 nm (Fig. 2, panels in the blue region (i.e. at 416 nm and 442 nm, respectively)
C-F; Table I) thatis quite similar to theabsorption maximum and another in the near UV region (i.e. at 366 and 361 nm,
respectively). Only the shorter wavelength absorption maxiof BR (Fig. 2, panels A and B; Table I). In addition, both
reactions exhibit defined isosbestic points (see Fig. 2,panels mum at 358 nm is present in the absorption spectrum of the
C-F), which indicates that PCB and P+B are converted to PEB reduction product (Fig. 3, panel c). Since the absorption
single products. The spectrophotometric analysis of the BVR- spectra of rubin compounds reflect light absorption by two
dependent catalysis of PEB is also shown in Fig. 2 (panels G dipyrrin-1-ones, the spectra of the PCB and P@B reduction
and H).In contrastwith the othertwo bilins, this conversion products reveal the presence of two distinct dipyrrole chroappears to yield at least two products, one with an absorbance mophores. This contrasts with the absorption spectra of BR
maximum at about 360 nm (Fig. 2, panels G and H ) and a and mesobilirubin IXa (MBR) which each exhibit a single
second minor product that absorbs between 400 nm and 450 absorption maximum at 450 and 428 nm, respectively, in this
nm (most noticeable in panel G, Fig. 2). The lack of defined solvent (data not shown). As shown in Fig. 4,MBR and BR
isosbestic points is consistentwith this interpretation (Fig. 2, are comprised of two nearly identical dipyrrin-1-one chropanels G and H ) . These results suggest that theinitial product mophores which accounts for the single absorption band. The
of PEB reduction by BVR is either chemically unstable or similarity between the long wavelength absorption maxima of
further metabolized. However, two species of PEB appear to the PCBand P@B reduction products and the respective
be initially present in the reaction mixture, one with an absorption maxima of MBR and BR indicates that thedipyrabsorption maxima at about 540 nm and theother at 616 nm, rin-1-one chromophore represented by the C- and D-rings is
which may also contribute to the complexity of the spectro- present in the two bilin reduction products. The short wavephotometric analysis. For all of the above studies, control length absorption maximum, present in the spectra of all
experiments without BVR were also performed. Except for a three bilin products, supports the hypothesis that the ethylislight bleaching of the respective absorption maxima, little dene moiety is also retained during BVR-catalyzed reduction
change was detected in the absorption spectra of all three of the parent bilins. Indeed, the position of this absorption
bilins during a 30-min incubation period in the absence of band is reasonable for an ethylidene-containing dipyrrole
BVR (data not shown). Taken together, these spectrophoto- chromophore represented by the A- and B-rings. Taken tometric measurements establish that all three bilins are BVR gether, the absorption spectra of the three products are consubstrates.
sistent with the hypothesis that BVR catalyzes the reduction
The kinetics of the BVR-dependent catalysis of the three of the C10 methine bridge to yield the corresponding rubibilins were next compared with those of the natural substrate noids, phytochromorubin (P@R), phycocyanorubin (PCR),
BV. The initial rates of reaction for each bilin were deter- and phycoerythrorubin (PER), whose structures areproposed
mined at five different substrate concentrations. Initial rates in Fig. 4.

column of Silica Gel G (type 60; 1-ml bed volume), and dried under
a stream of argon gas. 'H NMR spectroscopy of PCR dissolved in
pyridine-& (100.0 atom % D; Sigma catalog no. 23,699-3) was performed using a GEOmega300 spectrometer at the University of
California at Davis NMR Facility.
Holophytochrome Assembly-All assembly experiments were performed under green safelight. P+B (5 PM, final concentration) was
added to two separate reaction mixtures in 25 mM Tris-HC1 buffer,
pH 7.8, which contained 25% (v/v) ethylene glycol, 1 mg/ml BSA, 1
mM dithiothreitol, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride,
and 0.1 mM NADPH. BVR was addedto one of the reaction mixtures
to give a final concentration of 800 ng/ml, and both reactions were
incubated for 50 min at 25 "C. After the conversion of P+B was
judged to be complete in the +BVR sample by spectrophotometric
assay, apophytochrome was added to both reaction mixtures to yield
a final concentration of 10 rg/ml. Both mixtures were further incubated for 10 min at 4 "C,centrifuged at 200,000 X g for 15 min, and
phytochrome difference spectra were measured at 5 'C using an HP
8450A spectrophotometer as described previously (Lagarias et ul.,
1987; Elich and Lagarias, 1989). NADPH was subsequently oxidized
to NADP+ inboth reaction mixtures by addition of 0.75 mM oxidized
glutathione and 0.2 units/ml glutathione reductase. After incubation
at 25 "C for 10 min, a second phytochrome difference spectrum was
determined for both +BVR and -BVR (control) samples. An additional aliquot of 5 p~ P 9 B was then added to both samples and, after
incubating for 15 min at 4 "C, a third set of difference spectra were
measured. Western blots of samples taken at each stage of the
experiment were performed as described previously (Wahleithner et
UL, 1991).

Biliverdin Reductase Substrate Specificity

26102
0.30

BV, NAPDH ( p H 8 . 7 )

BV, NADH ( p H 7.0)

1

0.25

0.20

0.15
0.10

0.05
0.00
0.6

0.5

PCB, NADPH ( p H 8 . 7 )

11

PCB, NADH (pH 7.0)

0.4

0.3
0.2

0.1
0.0

0.8

P@B, NADPH (pH 8 . 7 )

P q B . NADH ( p H 7 . 0 )

0.8

0.4

0.2

0.0

PEB. NADPH (pH 8 . 7 )

0.30

0.25
0.20

0-16
0.10

0.05
0.00

wavelength (nm)
FIG.2. Spectrophotometric assay of the time course of the BVR-catalyzed reduction of BV, PCB, PaB, and PEB. BVRcatalyzed reduction of BV, PCB,PaB, and PEB
was followedspectrophotometrically using both NADPH (panelsA , C, E , and G ) and NADH
(panels B, D,F, and H ) assay protocols as described under Experimental Procedures. Mathematically smoothed spectra were obtained at
time zero and a t various intervals up to 45 min. The arrows indicate the direction of absorbance changes during the incubation. In punels A
and B , 5 PM BV and 200 ng/ml BVR mixtures are shown; last spectra correspond to 20-min ( A )and 45-min ( B )incubations. In panels C and
D, 10 FM PCB and 800 ng/ml BVR mixtures are shown; last spectra shown correspond to 10-min (C) and 24-min (D)incubations. In panels
E and F,17 p~ P@B and800 ng/ml mixtures are shown; last spectra shown correspond to 22 min ( E )and 25 min (F)incubations. In panels
G and H , 20 p~ PEB and800 ng/ml BVR mixtures are shown; last spectra shown correspond to 22-min ( G ) and 8-min ( H ) incubations.

Biliverdin
Specificity
Substrate
Reductase

26103

TABLEI
Absorption data for the products of BVR-catalyzed reduction of BV, PCB, POB, and PEB
Data were determined from the spectrophotometric analysis of bilin reduction by BVR shown in Fig. 2. Results for both NADPH and
NADH assay protocols are shown (see Experimental Procedures for details). Molar absorption coefficients at theabsorption maximum of
the rubin products ( t ) were calculated as described under Results.
Product ,A,

Isosbestic points (nm)

Substrate

NADPH

BV
PCB
P+B
PEB
a ND, not determined.

NADPH

NADH

Product e (M cm)

(nm)
NADH

NADPH

NADH
p H 7.0

55,800
39,600
42,800
ND

p H 8.7

p H 7.0

p H 8.7

p H 7.0

p H 8.7

407,526
394,480
413,514
ND

404,522
391,482
408,507
ND

466
422
450
361

458
428
448
356

64,100
36,900
42,300
ND

TABLEI1
Comparative kinetic constants for BVR-dependent catalysis of BV,
PCB, POB, and PEB
K,,,( p ~and
)
V,, (pmol rubin/mg BVR/min) values were determined from Lineweaver-Burk reciprocal plots using both NADPH
and NADH assay protocols (see Experimental Procedures for details). These analyses demonstrate that all reactions obey MichaelisMenten kinetics (data not shown).

v-

K,

Kl

p H 8.7

4.7

BV2.6
PCB
POB
1.4
PEB

3.7
3.9
3.8
9.8

4.8
4.8
14.4

PiPlR

PER

V0.00

p H 7.0

2.1
1.4
1.2

NADH

NADPH

Substrate

PCR

a,
2.0
3.5

H NMR spectroscopy was used to establish the structure

of PCR. Fig. 5 shows the 300 MHz H NMR spectrumof PCR
in pyridine-& Detailed assignments of this spectrum were
based on homonuclear spin decoupling experiments andcomparative analysis with the H NMR spectra of 3E-PCB dimethyl ester (BE-PCB DME) and synthetic MBR, an isomer
of PCR, in the same solvent (Fig. 5 and Table 111). These
analyses confirm the hypothesis that BVR catalyzes the exclusive reduction of the C10 methine bridge of PCB. This
conclusion is supported by the loss of the C10 methine singlet
at 6.99 ppm inthe H NMR spectrum of PCBandthe
appearance of a new methylene resonance at 4.19 ppm in the
H NMR spectrum of PCR (Fig. 5). The chemical shift
position of the latter is in good agreement with that of the
C10 methylene group of MBR at 4.25 ppm (Table 111). The
two high field methyl resonances in the spectrum of PCR also
provides strong evidence that the BE-ethylidene moiety is
retained upon reduction of PCB. In this regard, spin decoupling experiments show that themethyl group signals at 1.22
ppm and 1.56 ppm are coupled to multiplets at 3.11 and 5.9
ppm, respectively (Fig. 5, see decoupled resonances labeled ac). The relative chemical shift positions of these two multiplets me in good agreement with those assigned to 2-methine
and 3I-ethylidene protons of PCB at 3.39 and 6.33 ppm (Table
111). Both high methyl resonances and their corresponding
multiplets are not observed in the spectrum of MBR which
lacks the ethylidene moiety (Table 111). Integration of the
methyl resonances in the region between 1.9 and 2.2 ppm
indicates that PCR has three pyrrole methyl groups which
contrasts with four for MBR (data not shown). Based on
these analyses, the assignment of the structurefor PCRshown
in Fig. 4 has been made. Unfortunately, it was not possible to
obtain HNMR spectra for PQR and PERdue to insufficient
starting material. As described above, the features of the
absorption spectra of these compounds are fully consistent
with their structures as proposed in Fig. 4.

0.12

442

a
rQ

k
0
v1
rQ
rr(

0.00

ry

0.04

b
358
I

0.00

300

360

400

460

600

660

600

wavelength (nm)
FIG. 3. Absorption spectra of the products of PCB, P9B,
and PEB reduction by BVR. The absorption spectra of the solventextracted products of BVR-catalyzed reduction of PCB (panel a ) ,
POB (panel b), and PEB(panel c) are shown. Spectra were obtained
in methylene chloride and have been mathematically smoothed.

BVR-dependent Catalysis of P 9 B and PCBPrecludes Their

Assembly with Apophytochrome-In an earlier study, we demonstrated thatapophytochrome will covalently assemble with
several different ethylidene containing bilins including the
phycobiliprotein chromophore precursors, PCB and PEB (Li
and Lagarias, 1992). Since the reduction of the plant and
algal bilins byBVRdoes not alter the ethylidene moiety,
experiments were undertaken to assess the effect of BVRmediated reduction of these compounds on their ability to

Biliverdin Reductase Substrate Specificity

26104
no2c

co2n

biliverdin (BV)

phycocyanobilin (PCB)

This treatment effectively oxidizes all of the NADPH within

a few minutes (data not shown). As shown in Fig. 6b (-POB
sample), a small amount of photoactive holophytochrome was
detected following thistreatment.Thisresultappearsto
reflect BVR-mediated reoxidation of P+R to POB in the
presence of NADP (i.e. the reverse enzymatic reaction). The
alternative explanation for this result, chemical oxidation of
either free P9R or apophytochrome-bound POR, should yield
additional holophytochrome during a subsequent incubation
period which was not observed (data not shown). As shown
in Fig. 6b (+P9B),
addition of a fresh aliquot of POB to the
NADPH-depleted, BVR-treated P+B-apophytochrome reaction mixture led to a significant increase in the amount of
photoactive holophytochrome almost to theamount obtained
in the original control sample (compare with Fig. 6a, -BVR
spectrum). The reduced amount of holophytochrome produced in this experiment appears to be due to phytochrome
denaturation which occurs under the incubation conditions
used for these experiments (data notshown). Taken together,
theseexperiments indicate that covalent adducts between
apophytochrome and P+R or PCR are notproduced. Hence,
phytochrome chromophore precursors are functionally inactivated by BVR catalysis.
DISCUSSION

These investigations show that all three plant and algal

bilins are recognized by purified rat liver BVR and reduced
to novel rubinoid products by both NADPH- and NADHdependent activities. The ability of BVR to metabolize PCB,
P+B, andPEB was expected based on the results of Frydman
and co-workers (Awruch et al., 1984; Frydman et al., 1987).
phytochromobilin (PQB)
By examining a large number of synthetic biliverdins, their
studies showed an absolute requirement for two propionic
acid side chains for reduction by BVR. All three bilins analyzed in the present study have propionate side chains at the
C8 and C12 positions, which is the substitution pattern found
in the natural substrateBV (see Fig. 1);thus it is reasonable
that these pigments are effectively recognized by BVR. Indeed, PCB and P+B areexcellent BVR substrates, exhibiting
K , and V,, values very closeto those of the natural substrate.
H
H
H
These resultsclearly indicate that theethylidene moiety does
not measurably influence bilin binding to the enzyme, nor
does it lead to the formation of a covalent suicide adduct
phycoerythrobilin (PEB)
with a nucleophilic residue within the active site of BVR. By
FIG. 4. Structures of BR, MBR, and the products of PCB,
contrast, the larger K , values for PEB suggest that structure
POB, and PEB catalysis by BVR, PCR, POR, and PER.
of this pigment is less suited to the bilin-binding site of the
enzyme.
The lack of the C15 double bond in PEB permits
assemble with apophytochrome. As shown in Fig. 6a, P+B
catalysis by BVR prevents assembly with recombinant apo- free rotation of the D-ring which probably accounts for its
phytochrome as revealed by the lack of the characteristic reduced binding affinity.
Based on H NMR spectroscopy of PCR, the product of the
holophytochrome difference spectrumin the BVR-treated
sample (+BVR) compared with the untreated control sample catalysis of PCB by BVR, and comparison of the absorption
(-BVR). The same experiment was performed using PCB as spectra of the major products of P+B and PEBreduction by
studies establish that BVR
a substrate. In this case, BVR treatment also prevented the BVR, P+R, and PER, the present
reduces
the
C10
methine
bridge
without
altering the A-ring
formation of photoactive holophytochrome (data notshown).
To test the possibility that the BVR enzyme itself inhibits ethylidene substituent. Recent investigations have shown that
phytochrome photoconversion directly, an additional control the ethylidene group is required for theformation of the
experiment was performed. This experiment showed that thioether linkage to apophytochrome in vitro (Li and Lagarholophytochrome photoreversibility is not affected by BVR ias, 1992). For this reason, we anticipated that therespective
either in the presence or in the absence of NADPH (data not rubin products would besubstrates for assembly with apophytochrome. The present experiments clearly demonstrate that
shown).
The lack of photoreversibility of the BVR-treated sample this is not the case. One explanation for this observation is
configurations of the phytorubins (i.e. PGR
shown in Fig. 6a could reflect the formation of a covalent that the structural
POR-apophytochrome adduct which is photochemically in- and PCR) are poorly suited to the chromophore pocket of
active. To test this
possibility, the BVR-treated P+B-apophy- apophytochrome. It is well established that BR adopts an
tochrome mixture shown in Fig. 6a (+BVR) was first incu- intramolecular hydrogen-bonded configuration which is conbated with oxidized glutathione and glutathione reductase. siderably distorted from the flat helical conformations of its

26105

Biliverdin Reductase Substrate Specificity

i

\
2'
FIG. 5. 'H NMR spectrum of 3EPCR obtained from incubating 3EPCB with rat liver BVR. The 300
MHz spectrum of 0.3 mM PCR dissolved
in pyridine-ds was recorded at 25 "C.The

following spectrometer parameters were

used (100, 90" pulses; 2.05-9 acquisition
time/pulse; 4000 Hz spectral width; presaturation of the H20signal at 5 ppm).
Resonance assignments are indicated on
the spectrum and are listed in Table 111.
Signals labeled with the letter i represent
unknown solvent and/or sample impurities. Decoupled resonances lubeled a-c
correspond to saturating irradiation of
coupled spins systems at 1.55, 1.22, and
5.9 ppm, respectively.

I 8*

15

IJ

10

...
"T

1
7

._,

,"

~
6

"
"

18

..

.1.

'

"

"

PPm

"

"

1
2

'

'

"

~
0

TABLE111
300 MHz 'H NMR spectra
of 3E-phycocyanorubin (3E-PCR), 3E-phycocyanobilin dimethyl ester (3E-PCB DME),
and mesobilirubin (MBR)
in pyridine-&
Assignment
3E-PCB

BE-PCR"

DMEb

MBR"

182-Ethyl
7.5)
(t, J = 7.5)
1.00 (9,J = 7.8)
1.06 (t, J =1.24
2l
(s)
1.22 (m)'
1.47 2.15
(d, J = 7.4)
3'-Ethyl
present Not
Not present
1.00 (4, J = 7.8)
Not present
3'-Ethylidene
1.56 (d, J = 7.3)
1.72
(d, J = 7.3)
7l, 13', 17l
1.94,
2.17
2.14,
(3s)
2.20
2.12,
2.03,
(3s)
2.15
1.92,
1.82,
(3s)
2.47 (9, J = 7.5)
2.30 (q, J = 7.8)
18'-Ethyl
2.32 (q, J = 7.5)
Not present
2.36 (4, J = 7.8)
Not present
3l-Ethyl
2.81 (br t)
2.68 (m)
8',12'
2.73, 2.80 (Pt, J = 7)
2.99 (m)
2.95 (m)
2.89 (m)
8', 12'
3.39 (m)
2
Not present
3.11 (4, J = 7.7)
8', 12'
present
Not
3.8
3.6,
(2s)
Not present
10-Methylene
4.19 (broad s)
Not present
4.25 (broad s)
Not present
3I-Ethylidene
5.9 (dq, J = 2.2,7.3)
6.33 (dq, J =7.2)
2.1,
5-Methine
5.85 (s)
6.18 (s)
5.97 ( 8 )
6.22 (s)
6.06 (s)
15-Methine
6.18 (a)
10-Methine
6.99 (s)
present
Not
Not present
This work.
From Gossauer and Hirsch (1974).
A doublet is expected for this methyl resonance, but it is obscured in part by an impurity in the sample. Spin decoupling experiments
show that theshape of this signal changes upon irradiation of the resonance at 3.11 ppm (data not shown).

oxidized bilatriene parent BV (Falk, 1989). The pronounced

solubility of the phytorubins inorganic solvents suggests that
these compounds adopt a similar conformation to BR. The
inability of the phytorubins to covalently attach with apophytochrome therefore suggests that thisH-bonded configuration
must be disrupted to facilitate interaction with the chromophore-binding site. Another possibility for the inefficient assembly of the phytorubins may be due to the fact that these
compounds are more electron rich than their parent bilins.
Indeed, the chromophore precursor presumably functions as

an electrophile in the assembly reaction (Terry et al., 1993),

therefore this property may also disfavor holophytochrome
assembly.
Since BVR inactivates the precursors of the plant and
algal
biliprotein chromophores by their reduction to phytorubins,
the potential for regulation of biliprotein assembly via ectopic
expression of BVR is an intriguing one. Our studies suggest
that transgenic plants andalgae which express the BVR gene
would not only be able to metabolize the phytobilins themselves but also BV, the common precursor of these compounds

26106
FIG. 6. Holophytochrome assembly of BVR-treated P*B. Panel a,
apophytochrome samples (10 pg/ml)
were incubated with SE-P@B (-BVR)
and the product of P@B reduction by
BVR, P@R (+BVR). Panel 6, the apophytochrome that had been previously
incubated with P@R (+BVR sample
from panel a ) is shown after treatment
with oxidized glutathionine and glutathione reductase to remove NADPH
(-Pal?) and after the subsequent addition of P@B(+P@B).Details of incubation conditions are given under Experimental Procedures. Difference minima
and maxima (in nanometers) were determined from mathematically smoothed
plots using H P 8450A software. Because
of a light scattering artifact of the H P
8450A instrument, absorbance values for
the four points from 652 to 658 nm were
deleted and the plotted values interpolated from the measured values at 650
and 660 nm.

Biliverdin Reductase Substrate Specificity

I

0.004

0.006
0.004

0.002

0.002
0.000

0.000

-0,002
-0.002
-0.004

-0.006

-0.004

550

600

650 7 5700 0

(Beale, 1993;Terry et al., 1993).For algae, this would seriously

affect the ability to harvest light for photosynthesis, while for
plants, this would be expected to have a profound influence
on phytochrome-mediated photomorphogenesis. In this regard, phytochrome chromophore-deficient mutants have already been described, notably the hyl, hy2, and hy6 mutants
of Arabidopsis thaliana, which exhibit an etiolated phenotype
in the light (Chory et aL, 1989; Parks et aL, 1989). Recent
studies suggest that thegenetic lesions in these mutantsaffect
an early step in the tetrapyrrole biosynthetic pathway (Parks
and Quail, 1991; Nagatani et al., 1993). For this reason, the
phenotypes of these mutants may not only be due to reduced
phytochrome activity, but may also reflect a consequence of
altered heme and chlorophyll biosynthesis. Expression of
BVR in transgenic plants is expected to only affect the
biosynthesis of phytochrome and thuswill permit assessment
of those phenotypes that aredirectly attributable to a phytochrome deficiency. Since phytochrome is encoded by a small
multigene family (Quail, 1991), expression of BVR under a
strongconstitutivepromoter
should inactivateall phytochrome family members. Individual phytochrome family
members could also be targeted for inactivationthrough
expression of BVR under control of gene-specific phytochrome promoter sequences. If successful, this approach
might provide insight into thebiological roles of the different
phytochrome family members which are at present poorly
understood. This experiment may also address whether multiple phytochromes are expressed in the same cell. Furthermore, tissue- and/or cell-specificexpression of BVR may yield
new information on the involvement of individual plant cells
andtissues in phytochrome-mediated photornorphogenetic
responses in plants. In viewof these considerations, BVR
expression in transgenic plants offers great potential for the
experimental manipulation of photomorphogenetic responses
in plants. Given the recent cDNA cloning of BVR from rat
kidney (Fakhrai and Maines, 1992), these experiments can
now be undertaken.

I
800

550

600

650 750
700

800

wavelength (nm)
REFERENCES
Arciero, D. M., Bryant, D. A., and Glazer, A. N. (1988a)J. Bbl. Chem. 263,
18343-18349
Arciero, D. M.,Dallas, J. L., and Glazer, A.N. (1988b)J. Biol. Chem. 263,
18350-18357
Arciero,D.M., Dallas, J. L., and Glazer, A. N. (198&) J. Bwl Chem. 263,
18358-18363
Awruch, J., Tomaro, M. L.,Frydman, R. B., and Frydman, B. (1984)Biochim.
Biophys. Acta 787,146151
Beale, S. I. (1993)Chem. Rev. 93,785-802
Beale, S . I., and Cornejo, J. (1991)J. Biol. Chem. 266,22333-22340
Chapman, D. J., Cole, W. J., and Siegelman, H.W. (1967)J. Am. Chem. Soc.
89,5976-5977
Chory, J., Peto, C. A,, Ashbaugh, M., Saganich, R., Pratt, L. and Ausubel, F.
(1989)Plant Cell 1,867-880
Cole. W. J.. Chauman. D. J.. and Siecelman.
H. W. (1967)
.
. J. Am. Chem. Soc.
89, 36431364g

Cornejo, J., Beale, S. I., Terry, M. J., and Lagarias, J. C. (1992)J. BWL Chem.
267,14790-14798
Elich, T. D., and Lagarias, J. C. (1989)J. Bbl. Chem. 264,12902-12908
Elich, T. D., McDonagh. A. F., Palma, L. A,, and Lagarias, J. C. (1989)J. Biol.
Chem. 264,183-189
Fakhrai, H., and Maines, M. D. (1992),J.Biol. Chem. 267,4023-4029
Falk, ,H. (1989)The Chemistry of LanearOlzgopyrrolesand Bile Pigments,
Spnnger-Verla NewYork
Fevery, J., Blancfaert, N., Leroy, P., Michiels, R., and Heirwegh, K. P. M.
(1983)Heputolo 3,177-183
Frydman, R.B., yomaro, M. L., Rosenfeld, J., Awruch, J., Sambmtta, L.,
Valaslnas A., and Frydman, B. (1987)Bwchm. Bwphys. Acta 916,500-511
Glazer, A. N. (1989)J. Bwl. Chem. 264,l-4
Gossauer, A., and Hirsch, W. (1974)Liebigs Ann. Chem. 1974,1496-1513
Huang, T.-J., and Maines, M. D. (1990)Mol. Phrnuacol. 37, 25-29
Huang, T.-J., Trakshel,G. M., and Maines, M.D. (1989a)J. Biol. Chem. 264,
7844-7849
Huang, T.-J., Trakshel, G. M., and Maines, M.D. (1989b)Arch. Biochem.
Biophys. 274,617-625
Kutty, R. K., and Maines, M.D. (1981)J. Biol. Chem. 266,3956-3962
Lajmias, J. C., Kelly, J. M., Cyr, K. L., and Smith,W. O., Jr. (1987)Photochem.
hotobwl. 46,5-13
Li, L., and La arias, J. C. (1992)J. Bwl. Chem. 267,19204-19210
Lowry, 0.H.bsebrough, N. J., Farr, A. L., and Randall, R.J. (1951)J. Biol.
Chem. 1931,265-275
Maines, M. D. (1990)MOL P h o r m o l . 38,481-485
McDonagh, A. F.,and Palma, L.A. (1980)Bioehem. J. 189,193-208
Na atani A. Reed J. W. and Chory, J. (1993)Plant Physiol 102,269-277
Pa!ks, B.M.: and b a i l , P.H. (1991)Plant Cell 3,1177-1186
Parks, B. M.,Shanklip, J., Koornneef, M., Kendrick, R. E., and Quail, P. H.
(1989)Plant Mol. Bwl. 12,425-437
Quail, P. H. (1991)Annu. Reu. Genet. 26,389-409
Sin leton, J. W., and Laster, L. (1965)J. E d . Chem. 240,4780-4789
St08 M. S and Gray, C. H. (1977)Sixhem. J. 163.59-101
T e r j , M. 2, and Lagariaa, J. C. (1991)J. Bio2. Chem. 266,22215-22221
Terry, M. J., Wahleitbner, J. A,, and Lagarias, J. C. (1993)Arch. Biochem.
Biophys. 306, 1-15
Wahleithner J. A., Li, L., and Lagarias, J. C. (1991)Proc. Natl. Acad. Sei.
U.S. A. 88, 10387-10391
Weller, J.-P., and Gossauer, A. (1980)Chem. Ber. 113,1603-1611