Báo cáo khoa học: Chromophore attachment in phycocyanin Functional amino acids of phycocyanobilin – a-phycocyanin lyase and evidence for chromophore binding doc

Chromophore attachment in phycocyanin
Functional amino acids of phycocyanobilin – a-phycocyanin lyase
and evidence for chromophore binding
Kai-Hong Zhao
1
, Dong Wu
1
, Ling Zhang
1
, Ming Zhou
1
, Stephan Bo
¨
hm
2
, Claudia Bubenzer
2
and Hugo Scheer
2
1 College of Life Science and Technology, Huazhong University of Science and Technology, Hubei, China
2 Department Biologie I – Bereich Botanik, Universita
¨
tMu
¨
nchen, Germany
Phycobiliproteins are a homologous family of light-
harvesting proteins present in cyanobacteria, red algae,
and cryptophytes [1,2]. They absorb light in the
regions where chlorophyll absorbs poorly, and transfer
excitation energy with high quantum efficiency to the
photosynthetic reaction centres. Directed energy trans-

fer is determined by the spectroscopic properties and
relative positions of the various chromophores present,
Keywords
biliproteins; biosynthesis; cyanobacteria;
photosynthesis; post-translational
modification
Correspondence
K H. Zhao, College of Life Science and
Technology, Huazhong University of Science
and Technology, Wuhan 430074, Hubei,
China
Tel. ⁄ Fax: +86 27 8754 1634
E-mail:
H. Scheer, Department Biologie I – Bereich
Botanik, Universita
¨
tMu
¨
nchen,
Menzinger Str. 67, D-80638 Munich,
Germany
Fax: +49 89 17861 271
Tel.: +49 89 17861 295
E-mail:
(Received 10 November 2005, revised 17
January 2006, accepted 20 January 2006)
doi:10.1111/j.1742-4658.2006.05149.x
Covalent attachment of phycocyanobilin (PCB) to the a-subunit of C-phy-
cocyanin, CpcA, is catalysed by the heterodimeric PCB : CpcA lyase,
CpcE ⁄ F [Fairchild CD, Zhao J, Zhou J, Colson SE, Bryant DA & Glazer

AN (1992) Proc Natl Acad Sci USA 89, 7017–7021]. CpcE and CpcF of
the cyanobacterium, Mastigocladus laminosus PCC 7603, form a 1 : 1 com-
plex. Lyase-mutants were constructed to probe functional domains. When
in CpcE (276 residues) the N terminus was truncated beyond the
R33YYAAWWL motif, or the C terminus beyond amino acid 237, the
enzyme became inactive. Activity decreases to 20% when C-terminal trun-
cations went beyond L275, which is a key residue: the K
m
of CpcE(L275D)
and (L276D) increased by 61% and 700%, k
cat
⁄ K
m
decreased 3- and
83-fold, respectively. The enzyme also lost activity when in CpcF (213 resi-
dues) the 20 N-terminal amino acids were truncated; truncation of 53
C-terminal amino acids inhibited complex formation with CpcE, possibly
due to misfolding. According to chemical modifications, one accessible
arginine and one accessible tryptophan are essential for CpcE activity, and
one carboxylate for CpcF. Both subunits bind PCB, as assayed by Ni
2+
affinity chromatography, SDS⁄ PAGE and Zn
2+
-induced fluorescence. The
bound PCB could be transferred to CpcA to yield a-CPC. The PCB transfer
capacity correlates with the activity of the lyase, indicating that PCB bound
to CpcE ⁄ F is an intermediate of the enzymatic reaction. A catalytic mech-
anism is proposed, in which a CpcE ⁄ F complex binds PCB and adjusts via
a salt bridge the conformation of PCB, which is then transferred to CpcA.
Abbreviations

APC, allophycocyanin; CHD, 1,2-cyclohexanedione; CPC, C-phycocyanin; CpcA, a subunit apoprotein of C-phycocyanin; CpcE(x-y) or CpcF(x-y),
truncated CpcE or CpcF, respectively, extending from amino acid ‘‘x’’ to amino acid ‘‘y’’; CpcE, CpcF subunits of the heterodimeric
PCB:CpcA lyase; CpcE ⁄ F, complex of CpcE and CpcF, DEPC, diethylpyrocarbonate; EDAC, 1-ethyl-3-[3-(dimethylamino)propyl]
carbodiimide; IAA, iodoacetic acid; KPB, potassium phosphate buffer; NBS, N-bromosuccinimide; PC, phycocyanin; PCB, phycocyanobilin;
PCMS, p-chloromercuriphenylsulfonic acid; PE, phycoerythrin; PEB, phycoerythrobilin; PEC, phycoerythrocyanin; PecE, PecF, subunits
of PVB:PecA isomerase-lyase; PGO, phenylglyoxal; PLP, pyridoxal-5’-phosphate; PUB, phycourobilin; PVB, phycoviolobilin;
TX100, Triton X-100.
1262 FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS
i.e. linear tetrapyrroles (phycobilins) of which one to
four are covalently attached to the subunits by thio-
ether bonds to conserved cysteines. Phycobiliproteins
from cyanobacteria are heterohexamers (ab)
3
, which
are organized by linker proteins to large antenna com-
plexes, the phycobilisomes. Some of the linkers also
carry phycobilin chromophores [3].
In cyanobacteria, four classes of biliproteins have
been assigned on the basis of their visible absorptions
and sequence homologies: phycocyanin (PC), phyco-
erythrin (PE), allophycocyanin (APC) and phyco-
erythrocyanin (PEC). They contain, alone or in
combination, four different types of isomeric bilin
chromophores: phycourobilin (PUB), phycoerythrobi-
lin (PEB), phycobiliviolin (PVB), and phycocyanobilin
(PCB) [4]. C-phycocyanin (CPC) from Mastigocladus
laminosus PCC 7603 studied in this work, carries three
PCB at cysteines a-C84, b-C82, and b-C155 [5,6].
The last step in phycobilin biosynthesis [7] is the
addition of phycobilin to the apoproteins. In vivo, the

correct attachment of most chromophores is catalysed
by binding site- and chromophore-specific lyases, of
which only a few have hitherto been characterized
[8–10]. Since chromophore addition is autocatalytic in
some biliproteins (phytochromes, ApcE) [11–15] and
proceeds spontaneously, but more slowly and with less
fidelity, also in the phycobiliproteins like CPC [16,17],
a chaperone-like action has been proposed for these
lyases [8,17]. This view has been strengthened by the
observation of low activities of chromophore binding
to all binding sites studied, including a-C84 of PecA
[16–18], by the site-selective effect of Triton X-100
(TX100) on chromophore binding to PecB, and by its
inhibition of side reactions on binding to CpcA [17].
The first and hitherto best studied lyase attaches
PCB to the highly conserved a-C84 of CpcA from Syn-
echococcus sp. PCC7002 [19,20]. In this and several
other cyanobacteria, it is coded by two genes, cpcE
and cpcF, which are located in the cpc operon down-
stream of the structural genes for the two CPC sub-
units. A similar organization has been found for other
biliprotein:a-C84 lyase genes, but other arrangements
including isolated and fused genes have been reported
[8,21,22].
The lyase function of the proteins has been demon-
strated in vitro for CpcE and CpcF from Synechococcus
sp. PCC7002, Anabaena sp. PCC7120, Synechocystis sp.
PCC6803, M. laminosus PCC7603, but has only been
studied in some detail for Synechococcus sp. PCC7002
[13,23,24]. Mutants lacking cpcE and ⁄ or cpcF (or their

homologues) produce significantly reduced amounts of
CPC [20,25,26]. Homologous lyases of the ‘E ⁄ F-type’
are involved in the attachment of PEB and PCB to C84
of the a-subunits of CPE and PEC, respectively; the
latter reaction involves a concomitant isomerization of
PCB to PVB [9,18]. In vitro, CpcE and CpcF produced
in Escherichia coli jointly catalysed the correct attach-
ment of PCB to CpcA-C84, while CpcE or CpcF alone
were ineffective. CpcE and CpcF can also transfer the
bilin homo- and heterologously from a chromophoryl-
ated to a nonchromophorylated CpcA: this reaction
was reversible and specific for the a-84 site [23]. CpcE
and CpcF from Nostoc sp. PCC7120 can transfer PCB
from chromophorylated CpcA to PecA, and even to
apo-phytochrome AphA [13]. Addition of the CpcE ⁄ F
complex to a-CPC alters its absorption and dramatic-
ally reduces the fluorescence yield, no such changes are
seen with b-CPC [27].
Enzymes catalysing the biosynthesis of PCB (haem
oxygenase and biliverdin reductase) were introduced
together with the lyase (CpeE ⁄ F) and CpcA into
E. coli to generate a-CPC [24].
There is currently no structure known for any of the
biliprotein lyases. CpcE and CpcF from different spe-
cies show up to 60% identities, while they are only
20–40% homologous with other enzymes that are
known or suggested as lyase for phycobilin addition,
such as PecE and PecF [9], CpeY and CpeZ [21]. The
current study was initiated by the finding of conserved
motifs in alignments among lyases from different

species [28]. We report truncations, site-directed muta-
tions and chemical modifications which were guided by
such sequence comparisons, and propose a model of
action that involves transient covalent binding of the
chromophore to the lyase.
Results
Expression and purification of wild-type and
mutant enzymes
Full-length and truncated CpcE and CpcF proteins
were expressed with N-terminal His- and S-tags using
the pET-30a vector. The proteins were generally well
soluble (unlike those from Anabaena sp. PCC 7120
[13]), and the yield of extracted protein was > 70%.
An exception was CpcF(1–160) which tended to
precipitate (see below). The tags interfere with neither
the functions of the lyase (CpcE ⁄ F), nor with the
reactivity of the apoprotein (CpcA), but they facilitate
protein purification and improve their solubilities
[13,17,18]. After expression in E. coli, full-length
CpcE, CpcF, and their mutants were purified and
quantified by the Bradford method, and then their
concentrations adjusted to 50 lm for further
experiments.
K H. Zhao et al. Chromophore binding by a-phycocyanin lyase
FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS 1263
Mutation of CpcE and CpcF
The enzyme activities of the mutated subunits are com-
pared with those of the wild-type subunits in Table 1,
they were quantified by the fluorescence emission of
chromophorylated CpcA at 640 nm [23]. In these tests,

a mutated subunit was always combined with the full-
length complementary one. As mutations may affect
the interactions among subunits, all enzyme activities
of CpcE and CpcF were measured as before [28] in
three ways: with the nonpurified proteins (supernatants
of the disrupted E. coli), with the subunits purified via
Ni
2+
affinity chromatography, and with the purified
subunits which were first denatured together with the
full-length complementary one in 8 m urea and then
slowly corenatured by dialysis against urea-free buffer
(20 mm KPB, 0.5 m NaCl, pH 7.2). The individual
subunits, CpcE or CpcF, did not show any enzyme
activities, in agreement with previous studies using
Synechococcus sp. PCC 7002 [23]. The full-length lyase
showed highest activity when purified CpcE (276 resi-
dues) and CpcF (213 residues) were renatured jointly
in a 1 : 1 mixture from 8 m urea.
For CpcE(1–272) and CpcE(L275D), the superna-
tants showed higher activities than the purified
proteins, which may be due to the deletion and site-
directed mutation that affect the interaction among
the subunits, or some unknown factors in E. coli.
CpcE(1–274) corenatured with CpcF showed some-
what higher activity than the supernatants and purified
forms, indicating that the function lost by deletion of
the two amino acids could be improved by CpcF.
The mutants were constructed according to a
sequence comparison of PCB:CpcA lyases. The N-ter-

minal motif shown in Fig. 1A is highly conserved in
CpcE and PecE, therefore the truncated CpcE(42–276)
was constructed to delete the motif. Similarly, the
truncation CpcE(1–272) was generated to remove the
highly conserved C-terminal motif (i.e. DSLL, in
Fig. 1B). In CpcE, deletion of 41 amino acids at the N
terminus [CpcE(42–276)] and 39 amino acids at C ter-
minus (CpcE(1–237)) abolished the enzymatic activity.
If judged from the Ni
2+
-column binding assay (see
Experimental procedures) using bound His-tagged
CpcF as a bait, CpcE(1–237) has lost the ability to
form a complex with CpcF (data not shown), indica-
ting that the amino acids in the region 238–276 are
involved in the interaction of the two subunits.
CpcE(1–272) had only 17–28% activity of the wild-
type. This activity is retained if two more amino acids
were removed in CpcE(1–272), but only if this subunit
was not corenatured with CpcF. Possibly, the deletion
interferes with the refolding of CpcE. Replacement of
the conserved leucine-275 with the polar aspartate in
CpcE(L275D) resulted only in a moderate decrease of
activity. Kinetic studies (Table 2) showed that the K
m
Table 1. Comparisons of relative enzymatic activities of CpcE ⁄ F and their mutants, of covalent binding of PCB to CpcE and CpcF, and of
their capacity for transferring PCB. All test were done under standard reconstitution conditions, with the omissions of specific components
given in the headings and footnotes.
Lyase subunits
Relative lyase activity [%]

a
Subunit bound
PCB
b
[%]
Yield of a-CPC [l
M]
from PCB transfer
c,d
PCB transfer
[%]
c
Non-purified Purified Co- renatured
CpcE 0 0 0 9.2 0 (–) 0
CpcF 0 0 0 6.8 0 (–) 0
CpcE(42–276) + CpcF 0 0 0 7.2 0 (0) 0
CpcE(1–237) + CpcF 0 0 0 13.8 0 (0) 0
CpcF(21–213) + CpcE 0 0 0 7.2 0 (0) 0
CpcF(1–160)
e
+ CpcE 0 0 18 – – – –
CpcF(10–213) + CpcE 0 26 21 7.4 0.044 (0.038) 20
CpcE(1–272) + CpcF 29 21 0 10.8 0.072 (0.049) 33
CpcE(1–274) + CpcF 17 22 28 9.6 0.077 (0.061) 35
CpcE(L275D) + CpcF 81 65 65 9.0 0.14 (0.12) 64
CpcE(L276D) + CpcF 22 27 38 10.6 0.061 (0.061) 28
CpcF(I9K) + CpcE 100 100 100 6.6 0.099 (0.094) 45
(CpcE + CpcF) 100 100 120 11.0 0.16 (0.16) 73
a
Purified CpcE ⁄ F ¼ 100%.

b
No CpcA added, otherwise identical conditions as for reconstitutions. Yields are given with respect to the con-
centrations of CpcE or CpcF (5 l
M), the concentration of PCB was 10 l M.
c
No PCB added, otherwise identical conditions as for reconstitu-
tions.
d
Values in parentheses are controls with added extra CpcE (5 lM) and CpcF (5lM) to test for free PCB; compared with the
fluorescence brightness of the band on the Zn
2+
SDS ⁄ PAGE, the lyase and mutants in these tests had 0.22 lM PCB bound.
e
CpcF(1–160)
has very low solubility, so PCB binding to the lyase could not be evaluated.
Chromophore binding by a-phycocyanin lyase K H. Zhao et al.
1264 FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS
value increased by 61% and that k
cat
⁄ K
m
decreased
by 32%. The same replacement at the neighbouring
position L276, had much more drastic effects. The
activity of CpcE(L276D) decreased fourfold, the K
m
increased nearly eightfold, and k
cat
⁄ K
m

decreased by
almost two orders of magnitude. Obviously, L276 is a
critical residue, likely to be involved in the substrate
affinity.
The truncated protein CpcF(20–213) was generated
upon removal of the first ATG serving as initiation
site, because there is a second ATG 60 bases down-
stream. This product was inactive, in spite of a large
degree of heterogeneity in the N-terminal region
among the different lyases. Therefore, the truncation
mutant CpcF(10–213) was generated in order to
investigate this region more closely. CpcF(10–213) only
lost activity in the supernatant form. When purified
and corenatured, the mutant showed 26% and 21%
enzyme activity, respectively, indicating that amino
acids 1–9 in CpcF are only moderately relevant for the
activity. In this region, only I9 shows high homology
with other lyases. However, the activity of the mutant
CpcF(I9K) did not show any changes, which was veri-
fied by kinetic studies (Table 2). The C-terminally
truncated CpcF(1–160) was mostly deposited in inclu-
sion bodies (Fig. 2B), and the soluble fraction of this
mutant has lost activity. It was partly regained, how-
ever, when it was corenatured with CpcE, indicating
that the 53 amino acids in the C-terminal region of
CpcF are important for CpcF folding.
PCB binding to CpcE and CpcF
Work with the isomerizing lyase, PecE ⁄ F, had indica-
ted that the chromophore is bound transiently to the
lyase [28]. Such binding was investigated now in more

detail with CpcE ⁄ F. Wild-type and mutant proteins of
CpcE, CpcF and their 1 : 1 complexes were incubated
with PCB under reconstitution conditions, but omit-
ting the acceptor, CpcA. They were then re-purified
using a Ni
2+
affinity column, where unbound PCB
was removed in the 1 m NaCl wash step, and dialysed
against KPB (pH 7.2). The absorption spectra of
these fractions under native (Fig. 3A), and denaturing
conditions (Fig. 3B) showed that PCB could be bound
by CpcE, CpcF, the CpcE ⁄ F complex, and also by
their mutants (data not shown). Obviously, binding is
strong enough to retain the chromophore under the
Fig. 1. (A, B) Comparison of conserved N- and C-terminal domains
in CpcE and PecE from different organisms, and (C) of N-terminal
amino acids in CpcF and PecF.
CLUSTAL W (1.8) multiple sequence
alignment method was used. The number in front of the sequence
gives the accession code of the protein sequence in the Swiss-port
database. (A, B) CpcE Query: M. laminosus PCC7603 (acc. no.
AF506031, protein id AAM69288.2, note that the sequence has
been updated on 12.7.2005); compared to CpcE from Anabaena sp.
PCC7120 (PO7125); Fremyella diplosiphon PCC7601 (P07126);
Pseudanabaena sp. PCC7409 (Q52448); Synechococcus sp.
PCC7002 (P31967); Synechocystis sp. PCC6803 (P73638); Syn-
echococcus elongatus (P50037), and to PecE of Anabaena sp.
PCC7120 (P35791); and M. laminosus PCC7603 (P29729). (C)
Query: CpcF from M. laminosus PCC7603 (accession number.
AF506031, protein i.d. AAM69289.2, note that the sequence has

been updated on 12.7.2005) compared to CpcF from Anabaena sp.
PCC7120 (P29985; Synechococcus sp. PCC7002 (P31968); Syn-
echocystis sp. PCC6803 (P72652); Synechococcus sp. PCC7942
(Q44116); Pseudanabaena sp. PCC7409 (Q52449); Synechococcus
elongatus (P50038) and to PecF of M. laminosus PCC7603
(P29730).
Table 2. Kinetic analyses for the wild-type lyases and site-directed
mutants.
Enzyme K
M
[lM]
v
max
[pM
S)1
]
k
cat
[s
)1
]
k
cat ⁄
K
M
[s
)1
ÆlM
)1
]

Wild-type 0.38 ± 0.06 6.3 ± 0.5 1.3 · 10
)3
3.4 · 10
)3
CpcF(I9K) 0.41 ± 0.08 6.5 ± 0.6 1.3 · 10
)3
3.2 · 10
)3
CpcE(L275D) 0.61 ± 0.05 6.8 ± 1.2 1.4 · 10
)3
2.3 · 10
)3
CpcE(L276D) 2.94 ± 0.64 0.61 ± 0.04 1.2 · 10
)4
4.1 · 10
)5
K H. Zhao et al. Chromophore binding by a-phycocyanin lyase
FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS 1265
high ionic strength conditions (1 m NaCl) used during
chromatography. Covalent binding was supported by
the following observations: (a) denaturation with acidic
urea (8 m, pH 2.0) caused a loss of the distinct peak
at  650 nm on top of a broad background band
> 600 nm. The 650-nm peak, on top of a broad
absorption, was recovered in 40–70% yield when the
urea was dialysed out again, with the losses probably
due to irreversible oxidation or denaturation. Band
narrowing and absorption increase are characteristic
for chromophores bound to specific sites in native bili-
proteins, which are reversibly lost upon uncoupling of

the chromophore by denaturation of the protein [4,8].
The reversible loss of the distinct peak in chromo-
phore-treated CpcE ⁄ F is reminiscent of such changes;
(b) the small absorption decrease upon denaturation,
the relatively broad background in Fig. 3 extending
well beyond 700 nm, and the absence of fluorescence
(see below) indicate, however, a less tight coupling
between protein and chromophore, and a conforma-
tional heterogeneity of the latter. This binding
situation is rather different from the well-defined one
typical for phycocyanin and phytochromes [6,29,30];
(c) PCB remained bound to the protein during
SDS ⁄ PAGE, as shown by Zn
2+
-staining [31], even
though the amount is small when compared to the
fluorescence intensity of a CPC control (Fig. 3D,
Table 1).
Transfer of enzyme-bound chromophore to CpcA
PCB bound to CpcE or CpcF has very low fluores-
cence (Fig. 4C). This opened a way to test for the
transfer of enzyme-bound chromophore to the final
acceptor, CpcA, because the product, a-CPC, is
strongly fluorescent. CpcA was incubated, under stand-
ard reconstitution conditions, but in the absence of
free PCB, with an aliquot of the samples shown in
Fig. 3A, which induced the fluorescence typical for
a-CPC (Fig. 4C). Obviously, bound PCB could be
transferred from the lyase to CpcA to give the correct
product, a-CPC. As shown in Table 1, this capacity

roughly parallels the enzymatic activities of the lyase
and its mutants, indicating that the capacity of the
lyase to transiently bind and subsequently transfer
PCB is part of its enzymatic activity.
This is supported by another observation. When
PCB, CpcA, CpcE and CpcF were added simulta-
neously in the reconstitution system, there is generally
a by-product obtained with maximum absorption at
640 nm and fluorescence at 660 nm, which arises from
a spontaneous, nonenzymatic reaction [9,16,17]. Its
formation depends on the amount of PCB added, and
is particularly pronounced at higher concentrations
(Fig. 4A). If, however, PCB (0.05–1 lm, i.e. substo-
ichiometric amounts with respect to the lyase) was first
incubated with CpcE (0.8 lm) and CpcF (0.8 lm) for
1 h, and then CpcA (5 lm) was added, no such non-
natural PCB-CpcA adduct was detectable even at high
PCB concentrations (Fig. 4B). Obviously, the nonenzy-
matic reaction was inhibited when PCB was preincu-
bated with the lyase. This nonenzymatic reaction was
restored when CpcE and CpcF mutants were used that
lost the transfer ability. It is therefore concluded that
binding of PCB by the lyase during the preincubation
period inhibits the side reaction.
Chemical modifications of amino acids
Arginine modification by 1,2-cyclohexanedione (CHD)
and phenylglyoxal (PGO) [32] resulted in inactivation
of CpcE but not of CpcF (supplementary material
Fig. S1A and B). The semilogarithmic plots of remain-
ing activity against reaction time are linear, indicating

that the inactivations followed pseudo-first-order kinet-
ics. Second-order rate constants of 0.1 ± 0.02 and
0.7 ± 0.09 mm
)1
Æmin
)1
were obtained from the linear
plots of the first-order rate constants of inactivation
against modifier concentrations, for the reactions with
CHD and PGO, respectively. The numbers of modified
residues were obtained from plots of log(1 ⁄ t
0.5
) against
log[PGO] or log[CHD]. They resulted in straight lines
A
B
Fig. 2. SDS ⁄ PAGE of Ni
2+
affinity column purified mutant proteins.
Lane assignments: (A) M, protein marker; 1, CpcE(42–276); 2,
CpcE(L275D); 3, CpcE(1–274); 4, CpcE(1–237); 5, CpcE(1–272); 6,
CpcE(L276D); (B) M, protein marker; 1, CpcF(10–213); 2, CpcF(21–
213); 3, CpcF(I9K); 4, CpcF(1–160); 5, CpcF(1–160) purified from
inclusion bodies. The last mutant was poorly soluble, when corena-
tured with CpcE, it showed a little activity (see text). The different
mobilities irrespective of the similar sizes of CpcE(42–276) and
CpcE(1–237) were reproducible.
Chromophore binding by a-phycocyanin lyase K H. Zhao et al.
1266 FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS
with slopes 1.09 and 0.89, respectively; it is therefore

concluded that one accessible arginine residue is
required for the catalytic activity of CpcE. The modifi-
cations may affect the lyase activity by interfering with
the interactions of CpcE and CpcF. This was tested by
binding His-tagged CpcF to the Ni
2+
affinity column,
800750700650
0
100
200
300
400
A
1 M PCB
0.5
M PCB
Fluorescence intensity
Wavelength [nm]
B
800750700650
Wavelength [nm]
100
200
300
400
0
Fluorescence intensity
C
800750

700
650
Wavelength [nm]
0
5
10
15
Fluorescence intensity
After
Before

Fig. 4. (A) Fluorescence analysis of PCB transfer from CpcE ⁄ F to apo-CpcAA: PCB, CpcE (0.8 lM), CpcF (0.8 lM) and CpcA (5 lM) added
together at the same time, and then incubated for 1.5 h. (B) PCB was first incubated with CpcE (0.8 l
M) and CpcF (0.8 lM) for 1 h, then
CpcA (5 l
M) was added, and the sample incubated for another 1.5 h; PCB concentrations in (A) and (B): 0.05 lM (—); 0.1 lM (— —); 0.3 lM
(– –); 0. 5 lM (– ÆÆ -); 0.8 lM (– Æ– Æ); 1 lM (ÆÆÆÆ). (C) Fluorescence emission of PCB bound to CpcE ⁄ F (– –), and after adding CpcA to the system
(—). All reactions were carried out under standard reconstitution conditions, see Experimental procedures for more details.
C
B
A
D
Fig. 3. Binding of PCB to wild-type CpcE, CpcF and CpcE ⁄ F. (A) Absorption spectra after incubation of the proteins indicated by the labels
(all 10 l
M) with PCB (10 lM) under reconstitution conditions (37 °C, 1 h), subsequent purification by Ni
2+
affinity column to remove unbound
PCB, and by dialysis against KPB (pH 7.2), 12 h, in the dark. (B) Absorption spectra of the same solutions after addition of 8
M acidic urea
(pH 2). (C) Absorption spectra after subsequent renaturation from 8

M acidic urea (pH 2.0) by dialysis against KPB buffer (pH 7.2). See
Experimental procedures for details. (D) SDS ⁄ PAGE of proteins with bound PCB chromophore. Lane assignments: M, protein marker; 1,
CPC; 2, Histag-CpcE; 3, Histag-CpcF treated with PCB (Coomassie blue stain); 4, CPC; 5, Histag-CpcE; 6, Histag-CpcF treated with PCB
(Zn
2+
-induced fluorescence).
K H. Zhao et al. Chromophore binding by a-phycocyanin lyase
FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS 1267
and analysing for a retention of untagged CpcE by
SDS ⁄ PAGE (Fig. 5) and by activity assays. According
to this criterion, the inactivating arginine modification
did not interfere with complex formation between the
two subunits, but reduced the enzymatic activity
(Fig. S1).
Lysine residues were modified with pyridoxal-5’-
phosphate (PLP) [33]. Treatment of CpcE and CpcF
with an excess of the reagent for 30 min resulted only
in minor activity changes (96–98% versus control),
indicating that none of the accessible lysines in CpcE
and CpcF are required for the catalytic activity.
Carboxyl groups were modified with 1-ethyl-3-[3-
(dimethylamino)propyl] carbodiimide (EDAC) [34],
resulting in inactivation of CpcF, but not of CpcE
(Fig. S2). The semilogarithmic plots of remaining
activity against reaction time are linear, indicating
that the inactivations followed pseudo-first-order
kinetics. A second-order rate constants of 0.6 ±
0.06 mm
)1
Æmin

)1
was obtained from the linear plots
of the first-order rate constants of inactivation against
modifier concentrations. The numbers of modified res-
idues were obtained from plots of log(1 ⁄ t
0.5
) against
log[EDAC], they resulted in straight lines with slopes
0.65 for the reactions with EDAC. It is concluded
that one accessible carboxyl group is required for the
catalytic activity of CpcF. This modification also did
not affect the complex formation of CpcE and CpcF
(Fig. 5).
Tryptophan residues were modified by N-bromosuc-
cinimide (NBS) [35], it only affected the activity of
CpcE. There was a gradual decrease of activity, which
was analysed for the number of critical residues, i, by
the statistical method of Tsou [36]. The data can be
fitted to a straight line with i ¼ 1 (Fig. S3), suggesting
that a single accessible tryptophan residue is crit-
ical for the activity of CpcE. This modification of
CpcE did not affect complex formation with CpcE
(Fig. 5).
Histidine residues of CpcE were modified by DEPC
[37], there is no histidine in CpcF of M. laminosus.In
this case, untagged CpcE was used that was purified
via ion-exchange column [28]. DEPC had no effect on
the activity of CpcE, even through all three histidine
residues in CpcE were modified, as determined from
the absorption increase at 242 nm [28].

Cysteine residues were modified by p-chloromercuri-
phenylsulfonic acid (PCMS) and iodoacetic acid (IAA)
[38]. Reactions were carried out in KPB buffer con-
taining 6 m urea in two ways: Either CpcE and CpcF
were modified separately, or the CpcE ⁄ F complex was
modified. In case of individual treatments, the treated
subunit was combined with the complementary before
they were corenatured. PCMS had no effect on CpcE
and reduced the activity of CpcF only slightly to 78%.
When CpcE and CpcF were modified together with
PCMS, the activity still was 82%. When IAA was
used, as a more specific thiol reagent, to separately
modify CpcE and CpcF, there were again only moder-
ate losses of activity to 80% and 67%, respectively.
However, when they were treated together, the activity
was reduced to 16%. It is likely from these data, that
accessible cysteine residues in both CpcE and CpcF
play a role in the reaction, but are not essential. There
is no evidence that complex formation of CpcE and F
involves an intersubunit disulfide bridge. In the Ni
2+
-
affinity assay, the untagged subunit can always be
removed by extensive washing, and at low concentra-
tions the complex also dissociates during gel filtration
(Superdex 200).
Neither this nor any of the other modifications dis-
cussed below produced significant changes in the far-
UV CD-spectra (data not shown); according to this
criterion the secondary structure (mainly a-helix) was

retained after the modifications.
Discussion
CpcE and CpcF, the two subunits of PCB:CpcA lyase,
are involved in the PCB attachment to a-CPC at C84
[23]. There are eight different pairs of CpcE ⁄ Fof
known sequence, they show large regions of high
homology [39–41]. The enzymology has been studied
of CpcE ⁄ F from Synechococcus sp. PCC7002 [23,27],
but the amino acids that play a role in PCB attach-
ment are not yet clear. Also the bifunctional lyase,
PecE ⁄ F, is homologous with CpcE ⁄ F, and some char-
acteristic motifs were identified that distinguish in par-
ticular the F-subunits of the former and the latter [28].
CpcF lacks, for example, the four contiguous histidines
Fig. 5. Effect of chemical modifier on the formation of CpcE or
CpcF complex. SDS-containing gels of the fractions eluted with
500 m
M imidazole from the Ni
2+
-chrelating columns (see Experi-
mental procedures for details). Lane assignments: M, protein
marker; 1, CpcE (modified by CHD) with His-tag-CpcF; 2, CpcE
(modified by PGO) with His-tag-CpcF; 3, CpcE (modified by NBS)
with His-tag-CpcF; 4, CpcE with His-tag-CpcF (modified by EDAC);
5, CpcE with His-tag-CpcF (modified by IAA).
Chromophore binding by a-phycocyanin lyase K H. Zhao et al.
1268 FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS
of PecF, which caused a moderately strong binding to
the Ni
2+

affinity column and interferes with mutual
binding assays using one His-tagged partner. Complex
formation of CpcE with CpcF, and PCB binding to
them, could therefore be analysed with more confid-
ence than for PecE ⁄ F, using Ni
2+
affinity chromato-
graphy. Also, the amounts of PCB bound by CpcE,
CpcF and their mutants were larger than with PecE,
PecF and their mutants, thus facilitating the quantita-
tive analyses of PCB bound by CpcE, CpcF.
Several interesting N- and C-terminal motifs were
noted when comparing the sequences PecE and PecF
with those of CpcE and CpcF (Fig. 1). For CpcE,
both the N and C termini have conserved regions.
When the motif R33YYAAWWL near the N terminus
was deleted (CpcE(41–273)), the enzyme lost its activ-
ity completely. A 39-amino acid C-terminal truncation
in CpcE also rendered the protein inactive; it also
nearly lost the ability to form a complex (data not
shown), indicating that this region is involved in the
complex formation. When the C-terminal motif
D273SLL was removed in CpcE(1–272), there was still
some activity left, but the mutant lost activity when
CpcE(1–272) was denatured and corenatured with
CpcF, indicating irreversible unfolding. If the two resi-
dues D273 and S274 were maintained, the enzyme still
had 28% relative activity: site-directed mutations of
the two leucines (L275D, L276D) reduced the enzyme
activity only moderately to 65% and 27%, respect-

ively; these mutations also reduced the substrate affin-
ity (Table 2). These regions were also important for
PecE ⁄ F lyase-isomerase activity; truncations rendered
the enzyme inactive but did not affect the stability of
the proteins [28].
The C terminus of CpcF shows only little homology
for as much as 50 amino acids. A truncation by 53
amino acids reduced the solubility of the protein, pos-
sibly due to misfolding, and most of the protein was
deposited in E. coli as inclusion bodies. Only 18% rel-
ative activity was recovered by solubilization with urea
(8 m) and corenaturation with CpcE, indicating that
interaction with CpcE aided the re-folding. The N ter-
minus of CpcF showed more homology: a 10 amino
acid truncation reduced the activity to 26%, and it
was lost completely when 20 amino acids were trun-
cated. Among the 10 N-terminal amino acids, I9 is
highly conserved among different CpcF, but its muta-
tion to lysine resulted in no marked change of the
enzyme activity (Table 2).
There are four cysteines in CpcE and three in CpcF,
of which only C99 of CpcE is highly conserved. In
reconstitution experiments in vitro, reducing reagents
such as mercaptoethanol or dithiothreitol were not
required for enzyme activity, indicating that no disul-
fides are present that interfere with the enzymatic activ-
ity. While the thiol group modification using PCMS
proved ineffective, a more complex picture was obtained
from modification with IAA. When only one of the sub-
units was modified, the enzyme activity was retained,

but modification of both subunits, CpcE and CpcF, in
6 m urea reduced the activity to 16%. These modifica-
tions were done in 6 m urea to reach otherwise inaccess-
ible cysteins, and the protein subsequently renatured.
They did not interfere noticeably with PCB binding,
IAA therefore modifies residues that are otherwise
involved in the catalytic activity (e.g. PCB transfer).
Both CpcE and CpcF bind PCB, as evidenced by
absorption spectroscopy and chromatographic separ-
ation from unbound chromophore. This binding is
only moderately strong and reversible, as judged from
the low amount of chromophore found on the SDS ⁄
PAGE purified proteins (Fig. 3D). Covalent chromo-
phore binding, albeit even weaker than with CpcE ⁄ F,
has also been reported for PecE ⁄ F. In the latter case,
PCB bound to the enzyme was neither transferred to
PecA to form the PCB adduct, nor transferred and
concomitantly isomerized to PVB. By contrast, chro-
mophore transfer from CpcE ⁄ F to CpcA could now be
demonstrated. Furthermore, in mutants this transfer
correlated with their lyase activities (Table 1). In com-
bination, these results are evidence for a transient
chromophore binding to the enzyme as part of the cat-
alytic reaction. A chaperone function of a-84 lyases
had been suggested before as (at least part of) the en-
zymatic activity of a-84 lyases [17,42]. The absorption
spectral changes of PCB upon binding to CpcE ⁄ F are
indicative of a conformational change, but at the same
time indicate that the chromophore conformation is
less restricted not yet extended as it is in a-CPC. The

lack of an intense fluorescence further suggests that
the chromophore retains flexibility upon binding [29],
which further supports a comparably weak binding, in
a conformation that is intermediate between the cyclic
one of the free chromophore, and its extended, rigid
conformation in the a-CPC binding site.
There are 19 arginines, 13 lysines, three histidines
and two tryptophans in CpcE, and 13 lysines, 10 argi-
nines, one tryptophan and no histidine in CpcF.
According to chemical modification of these residues,
only one accessible arginine and one tryptophan is
involved in CpcE function, and one carboxyl group
CpcF. In the bifunctional lyase, PecE ⁄ F, a consider-
ably larger number of critical amino acid residues have
been identified by the same methods. An additional
histidine is required in the PecE subunit, and one tryp-
tophan, one cysteine and one histidine in PecF [28]. Of
K H. Zhao et al. Chromophore binding by a-phycocyanin lyase
FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS 1269
the latter, C121 and H122 are located in a region that
has been related to the lyase function. The only critical
residue that is missing in the isomerizing lyase, is the
essential carboxyl group in CpcF. This may be related
to the chromophore transfer capacity of CpcE ⁄ F,
which is lacking in PecE ⁄ F. Because the optimal pH
for CpcE⁄ F is 7.5–8.0, the carboxyl group in CpcF is
expected to be present as a carboxylate anion. Since
the native PCB chromophore is protonated [43–45], a
possible scenario is the formation of a salt-bridge
between the carboxylate anion in CpcF and the proto-

nated PCB, but this working hypothesis remains to be
tested. An alternative function for the carboxylate, i.e.
an intermolecular salt bridge with the essential Arg in
CpcE, is not supported in view of the fact that the
modification of the carboxyl of CpcF did not inhibit
the CpcE ⁄ F complex formation (Fig. 5).
In summary, the two types of homologous lyases
show some common features (heterodimeric complex,
catalysis of attachment of phycobilin at C84 of a-sub-
unit), but they differ in the manipulation of the chromo-
phore not only by the isomerase capacity of PecE ⁄ F
that is lacking in CpcE ⁄ F, but also in the chromophore
transfer capacity that is present in CpcE ⁄ F, but absent
in PecE⁄ F. Several residues have been identified that
relate to the different functions. However, a general cat-
alytic model is still lacking. Investigations of the new
class of distantly related lyases recently identified are
expected to further clarify the molecular basis of the
variability and specificities of this class of enzymes.
Experimental procedures
Materials and reagents
1,2-cyclohexanedione, PGO, NBS, PCMS and IAA were
from Sigma (Beijing, P.R.C.); diethylpyrocarbonate
(DEPC), PLP and EDAC were from Fluka (Beijing,
P.R.C.). All other biochemicals and separation materials
were of the highest purity available and obtained from the
sources described previously [9,18]. Recombinant proteins
were purified as before [13].
Full-length proteins
Cloning and expression followed the standard procedures

of Sambrook et al. [46]. The integral genes cpcA, cpcE and
cpcF were PCR-amplified from M. laminosus PCC7603.
They were cloned first into pBluescript SK(+) (Stratagene,
Shanghai, P.R.C.), and then subcloned into pET-30a (Nov-
agen, Munich, Germany). Proteins without His- and S-tags
were obtained by expressing pGEMEX containing the
desired DNA [9].
Deletion and site-directed mutants
Truncated and site-directed mutants were prepared by
the PCR method, using the following primers.
P1: 5¢-TGTCCCGGGGCATTGGTCATGACAGAAGCA-3¢,
upstream; P2: 5¢-GGGCTCGAGCGGCAATTAAAGTGG
GAAT-3¢, downstream; P3: 5¢-ATACCCGGGATACTCCT
GACCATGACTGC-3¢, upstream; P4: 5¢-ACCCTCGAGT
TATCTTGAGAGTGGAACAAA-3¢, downstream; P5:
5¢-ATGCCCGGGGGTAAGTTTCGCGTTCG-3¢,upstream;
P6: 5¢-GGGCTCGAGTTACATCAAATTCATGACTCG-3¢,
downstream; P7: 5¢-CCCCTCGAGCTTGCTACAATTAT
GAATCCA-3¢, downstream; P8: 5¢-ACCCTCGAGTTATT
TTCTACCTTGGCCAGC-3¢, downstream; P9: 5¢-TGTCC
CGGGCAAATGACAGCAGCTGTA-3¢, upstream; P10:
5¢-AAACCCGGGCGCAGTGTAGCTGAAG-3¢, upstream;
P11: 5¢-C CCCTCGAGCCCTTAAATTGGTT GTTGTA-3¢,
downstream; P12: 5¢-ATACCCGGGATGACTGCCACTA
CTCAACAATTAA
AACGT-3¢, upstream; P13: 5¢-GGGCT
CGAGCGGCGCTTACAA
ATCTGAATC-3¢, downstream,
P14: 5¢-AGCCTCGAGCGCCTA
GTCAAGTGAATCCAT

CA-3¢, downstream.
All upstream primers have a SmaI site (CCCGGG) and
the downstream primers have a XhoI site (CTCGAG),
which ensure correct ligation of the fragments to pBlue-
script.
P1 and P2 were used to generate the full-length cpcE, P3
and P4 for full-length cpcF, P5 and P2 for cpcE(42–276),
P1 and P6 for cpcE(1–272), P1 and P7 for cpcE(1–274), P1
and P8 for cpcE(1–237), P9 and P4 for cpcF(21–213), P10
and P4 for cpcF(10–213), P3 and P11 for cpcF(1–160),
P12 and P4 for cpcF(I9K), P1 and P13 for cpcE(L275D),
P1 and P14 for cpcE(L276D), In P12, P13 and P14, the
site-directed mutations are underlined. All mutations were
verified by sequencing.
Expressions
The pET-based plasmids were used to transform E. coli
BL21 (DE3). Cells were grown at 37 °C in Luria–Bertani
medium containing kanamycin (30 lgÆml
)1
). When the cell
density reached OD
600
¼ 0.5–0.7, isopropyl-thio-b-dgal-
actopyranoside (1 mm) was added. 5 h after induction, cells
were collected by centrifugation, washed twice with double-
distilled water, and stored at )20 °C until use. CpcA,
CpcE, CpcF of M. laminosus and all mutants were pre-
pared using the methods described previously [18,47].
SDS/PAGE
SDS ⁄ PAGE was performed with the buffer system of Lae-

mmli [48]. The gels were stained with zinc acetate for bilin
chromophores [31] and with Coomassie brilliant blue R for
the protein. The UV-induced fluorescence of protein-bound
Chromophore binding by a-phycocyanin lyase K H. Zhao et al.
1270 FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS
bilins was recorded digitally with a camera. The amounts
of bilins bound by lyases and their mutants were quantita-
tively evaluated by comparing their scanned fluorescence
intensity to that of a standard, i.e. CPC, on the same
SDS ⁄ PAGE, using photoshop 6.0 (Adobe, San Jose, CA,
USA).
Spectroscopy
Enzyme reactions and amino acid modifications were
followed by UV-visible spectrophotometry (Perkin-Elmer
model Lamda 25) and fluorimetry (Perkin-Elmer LS55).
The formation of chromophorylated-a-CPC was detected
by the emission at 640 nm. Far-uv CD spectra were recor-
ded at 25 °C with a Dichrograph VI (ISA, Jobin Yvon,
Munich, Germany) using 1 mm cuvettes, five spectra were
averaged and the data smoothed by 5-point averaging.
PCB and protein concentration determinations
PCB was prepared as described before [18]. PCB concentra-
tions were determined spectroscopically in methanol ⁄ 2%
HCl using e
690
¼ 37 900 m
-1
Æcm
-1
[18]. Protein concentra-

tions were determined according to [49] using the protein
assay reagent (Bio-Rad, Munich, Germany) according to
the manufacturer’s instructions with BSA as standard. Con-
centrations of overexpressed proteins in crude extract were
determined by first assaying the total protein content by the
Bradford method, and then the relative amount (%) by
SDS ⁄ PAGE.
Phycobiliproteins
CPC and a-CPC from M. laminosus were prepared as
before [17].
Lyase activity assay
Chromophore reconstitution with CpcA was assayed as des-
cribed before [27]. Either full-length CpcE was comple-
mented with mutants of CpcF, or full-length CpcF was
complemented with mutants of CpcE, using the following
standard reaction conditions: potassium phosphate buffer
(KPB, 15–20 mm, pH 7.2) containing NaCl (150–200 mm),
MgCl
2
(5 mm), CpcE and CpcF or their mutants (5 lm
each), and His6-CpcA (5 lm). PCB (final concentration,
5 lm unless stated otherwise) was added as a concentrated
dimethylsulfoxide solution; the final concentration of
dimethylsulfoxide was 1% (v ⁄ v). The mixture was incubated
at 37 °C for 1 h in the dark. Products were quantified by
fluorescence emission at 640 nm [27]. The lyase reactions
were carried out with three different preparations of each
His-tagged CpcE, CpcF and their mutants: (1) nonpurified
proteins, i.e. the supernatants of the sonicated cells after
centrifugation; (2) proteins purified by Ni

2+
chelating affin-
ity chromatography as before [18]; (3) corenatured proteins,
which were obtained by the following procedure: purified
CpcE (or its mutants) and purified CpcF (or its mutants)
were denatured separately with urea (8 m) at room tempera-
ture. They were then mixed in equimolar amounts (5 lm)
and renatured slowly by repeated dialysis against KPB
(20 mm, pH 7.2) containing NaCl (0.5 m)at4°C for 4 h.
For kinetic tests, only purified proteins were used. Either
full-length CpcE was complemented with mutants of CpcF,
or full-length CpcF was complemented with mutants of
CpcE. The purified subunits (5 lm), CpcA (5 lm) and differ-
ent concentrations of chromophore substrate, PCB, were
mixed in the reconstitution system (see above) and incubated
at 20 °C. At regular time intervals, the reaction was termin-
ated by rapidly cooling the samples on ice to 0 °C, then the
product was quantified by the fluorescence emission at
640 nm. The fluorescence was calibrated with a solution of
a-CPC of known concentration. K
m
, v
max
and k
cat
were
calculated from Lineweaver–Burk plots, using origin v6
(Origin Laboratory Corporation, Northampton, MA, USA).
PCB binding to CpcE and CpcF
A mixture of CpcE and CpcF (1 : 1), individual subunits,

or their mutants was incubated in the reconstitution system,
as described above, but using twice the standard concentra-
tion of PCB (10 lm) and omitting CpcA. The products
were purified by Ni
2+
chelating chromatography, and ana-
lysed by absorption spectroscopy (300–800 nm), and by
SDS ⁄ PAGE using Zn
2+
staining [31] and Coomassie brilli-
ant blue staining.
To check if bound PCB could be transferred to CpcA, the
lyase CpcE ⁄ F with bound PCB was first purified by Ni
2+
affinity chromatography to remove free PCB, and then dia-
lysed against KPB (20 mm, pH 7.2) containing NaCl (0.5 m)
at 4 °C for 12 h in the dark. The sample that has been freed
of PCB was divided into three parts. The first part of the
sample was denatured with 8 m acidic urea (pH 2.0), and its
absorption spectrum recorded. Then the sample was rena-
tured against KPB (20 mm, pH 7.2) containing NaCl
(0.5 m), and the absorption recorded again. The second part
of the sample was mixed with CpcA (one or both lyase sub-
units added when needed), and Mg
2+
(5 mm), and incubated
at 37 °C for 1.5 h. Then, fluorescence emission at 640 nm
was measured as described above. The third part was ana-
lysed by SDS⁄ PAGE using Zn
2+

and Coomassie brilliant
blue staining as described above.
Complex formation of CpcE and CpcF
Purified His-tagged CpcE or its mutants were corenatured
(see Lyase activity assay) with untagged CpcF. The mix-
tures were then loaded on a Ni
2+
column, washed three
K H. Zhao et al. Chromophore binding by a-phycocyanin lyase
FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS 1271
times with five column volumes of KPB (20 mm, pH 7.2)
containing NaCl (0.5 m), once with the same buffer con-
taining in addition imidazole (50 mm), and finally with the
same buffer containing 500 mm imidazole. The eluate from
the last wash was analysed by (a) SDS ⁄ PAGE, and (b) for
enzymatic activity as described above, after dialysis against
KPB buffer (pH 7.2).
The1 : 1 complex of His-tagged CpcF and untagged
CpcE could be purified by this method. The purified
CpcE ⁄ F complex after Ni
2+
affinity chromatography was
further checked by size exclusion chromatography with
Amersham Pharmacia FPLC system and a Superdex 200
column at 4 °C. The column was equilibrated in 0.1 m Tris
buffer (pH ¼ 7.2) containing 1 mm EDTA. Fractions were
collected and detected with SDS ⁄ PAGE.
Chemical modifications of amino acids
Chemical modifications of arginine, lysine, carboxyl groups,
tryptophan and histidine were performed as described

before [28]. The number of essential residues of a certain
amino acid was determined by the kinetic method of [50] or
by the statistical method of [36], as carried out for PEC
lyase [28].
Cysteine
Purified CpcE and CpcF (5 lm) in KPB buffer (20 mm,
pH 7.2) were modified by PCMS or IAA [38,51]. For modifi-
cation with PCMS, a stock solution (50 mm in double-
distilled water) was added in 2–5-fold molar excess over the
sulfhydryl groups in the protein. For modification with IAA,
the purified enzyme was first transferred to KPB (20 mm,
pH 7.2) containing urea (6 m), incubated with a fivefold
excess of IAA under argon at room temperature in the dark.
At designated time intervals, the modification reaction was
stopped by addition of mercaptoethanol (0.1 m). After dialy-
sis against KPB (20 mm, pH 7.2, three times for 4 h each),
the enzyme activity was assayed as above.
Acknowledgements
HS and KHZ are grateful to Volkswagen Stiftung for
a Partnership grant (I ⁄ 77900). The laboratory of KHZ
is supported by National Natural Science Foundation
of China (30270326, 30540070) and the Program for
New Century Excellent Talents in University (NCET-
04–0717, P.R. China), that of HS is supported by
Deutsche Forschungsgemeinschaft (SFB 533, TP A1).
References
1 Grossman AR, Schaefer MR, Chiang GG & Collier JL
(1993) The phycobilisome, a light-harvesting complex
responsive to environmental conditions. Microbiol Rev
57, 725–749.

2 Glazer AN (1994) Adaptive variations in phycobilisome
structure. Adv Mol Cell Biol 10, 119–149.
3 Sidler WA (1994) Phycobilisome and phycobiliprotein
structures. In The Molecular Biology of Cyanobacteria
(Bryant, DA, ed.), pp. 139–216. Kluwer, Dordrecht.
4 Glazer AN (1985) Light harvesting by phycobilisomes.
Ann Rev Biophys Biophys Chem 14, 47–77.
5 Frank G, Sidler W, Widmer H & Zuber H (1978) The
complete amino-acid sequence of both subunits of
C-phycocyanin from the cyanobacterium Mastigocladus
laminosus. Hoppe-Seyler’s Z Physiol Chem 359, 1491–
1507.
6 Schirmer T, Bode W, Huber R, Sidler W & Zuber H
(1985) X-Ray crystallographic structure of the light-har-
vesting biliprotein C-phycocyanin from the thermophilic
cyanobacterium Mastigocladus laminosus and its resem-
blance to globin structures. J Mol Biol 184, 257–277.
7 Beale SI (1994) Biosynthesis of cyanobacterial tetrapyr-
role pigments: Hemes, chlorophylls and phycobilins. In
The Molecular Biology of Cyanobacteria. (Bryant, DA,
ed.), pp. 519–558. Kluwer, Dordrecht.
8 Schluchter WM & Glazer AN (1999) Biosynthesis of
Phycobiliproteins in Cyanobacteria. In The Phototrophic
Prokaryotes (Peschek, GA, Lo
¨
ffelhardt, W & Schmet-
terer, G, eds), pp. 83–95. Kluwer ⁄ Plenum Press, New
York.
9 Zhao K-H, Deng M-G, Zheng M, Zhou M, Parbel A,
Storf M, Meyer M, Strohmann B & Scheer H (2000)

Novel activity of a phycobiliprotein lyase: both the
attachment of phycocyanobilin and the isomerization to
phycoviolobilin are catalyzed by PecE and PecF. FEBS
Lett 469, 9–13.
10 Shen G, Saunee NA, Gallo E, Begovic Z, Schluchter
WM & Bryant DA (2004) Identification of novel
phycobiliprotein lyases in cyanobacteria. In PS 2004
light-Harvesting Systems Workshop, book of abstracts.
p. 14.
11 Wu S-H & Lagarias JC (2000) Defining the bilin lyase
domain: lessons from the extended phytochrome super-
family. Biochemistry 39, 13487–13495.
12 Hughes J, Lamparter T, Mittmann F & Hartmann E
(1997) A prokaryotic phytochrome. Nature 386, 663.
13 Zhao K-H, Ran Y, Li M, Sun Y-N, Zhou M, Storf M,
Kupka M, Bo
¨
hm S, Bubenzer C & Scheer H (2004)
Photochromic biliproteins from the cyanobacterium
Anabaena sp. PCC 7120: lyase activities, chromophore
exchange and photochromism in phytochrome and phy-
coerythrocyanin. Biochemistry 43, 11576–11588.
14 Zhao KH, Su P, Bo
¨
hm S, Song B, Zhou M, Bubenzer
C & Scheer H (2005) Reconstitution of phycobilisome
core-membrane linker, L
cm
, by autocatalytic chromo-
phore binding to ApcE. Biochim Biophys Acta 1706,

81–87.
Chromophore binding by a-phycocyanin lyase K H. Zhao et al.
1272 FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS
15 Robben U, Lindner I, Ga
¨
rtner W & Schaffner K (2001)
Analyse der Topologie der Chromophor-Bindungstasche
von Phytochrom durch Variation des Chlromophor-
Substitutionsmuster. Angew Chemie 113, 1080–1082.
16 Arciero DM, Bryant DA & Glazer AN (1988) In vitro
attachment of bilins to apophycocyanin. 1. Specific
covalent adduct formation at cysteinyl residues involved
in phycocyanobilin binding in C-phycocyanin. J Biol
Chem 263, 18343–18349.
17 Zhao KH, Zhu JP, Song B, Zhou M, Storf M, Bo
¨
hm S,
Bubenzer C & Scheer H (2004) Non-enzymatic chromo-
phore attachment in biliproteins: Conformational con-
trol by the detergent Triton X-100. Biochim Biophys
Acta 1657, 131–145.
18 Storf M, Parbel A, Meyer M, Strohmann B, Scheer H,
Deng M, Zheng M, Zhou M & Zhao K (2001) Chromo-
phore attachment to biliproteins: Specificity of PecE ⁄
PecF, a lyase ⁄ isomerase for the photoactive 3
1
-Cys-a84-
phycoviolobilin chromophore of phycoerythrocyanin.
Biochemistry 40, 12444–12456.
19 Swanson RV, Zhou J, Leary JA, Williams T, de Lorimier

R, Bryant DA & Glazer AN (1992) Characterization of
phycocyanin produced by cpcE and cpcF mutants and
identification of an intergenic suppressor of the defect in
bilin attachment. J Biol Chem 267, 16146–16154.
20 Zhou J, Gasparich GE, Stirewalt VL, de Lorimier R &
Bryant DA (1992) The cpcE and cpcF genes of Syne-
chococcus sp. PCC 7002. Construction and phenotypic
characterization of interposon mutants. J Biol Chem
267, 16138–16145.
21 Kahn K, Mazel D, Houmard J, Tandeau de Marsac N
& Schaefer MR (1997) A role for cpeYZ in cyanobac-
terial phycoerythrin biosynthesis. J Bacteriol 179, 998–
1006.
22 Six C, Thomas J-C, Thion L, Lemoine Y, Zal F &
Partensky F (2005) Characterization of two novel
phycoerythrin-associated linker proteins in the marine
cyanobacterium Synechococcus sp. WH8102. J Bacteriol
187, 1685–1694.
23 Fairchild CD, Zhao J, Zhou J, Colson SE, Bryant DA
& Glazer AN (1992) Phycocyanin a-subunit phycocy-
anobilin lyase. Proc Natl Acad Sci USA 89, 7017–7021.
24 Tooley AJ, Cai YA & Glazer AN (2001) Biosynthesis of
a fluorescent cyanobacterial C-phycocyanin holo-alpha.
subunit in a heterologous host. Proc Natl Acad Sci USA
98, 10560–10565.
25 Bhalerao RP & Gustafsson P (1994) Factors Influencing
the Phycobilisome Rod Composition of the Cyanobac-
terium Synechococcus sp. PCC7942 – effects of reduced
phycocyanin content, lack of rod-linkers, and over-
expression of the rod-terminating linker. Physiol Plant

90, 187–197.
26 Tandeau de Marsac N, Mazel D, Damerval T, Gugli-
elmi G, Capuano V & Houmard J (1988) Photo-
regulation of gene expression in the filamentous
cyanobacterium Calothrix sp. PCC 7601: light-harvest-
ing complexes and cell differentiation. Photosynth Res
18, 99–132.
27 Fairchild CD & Glazer AN (1994) Oligomeric structure,
enzyme kinetics, and substrate specificity of the phyco-
cyanin alpha subunit phycocyanobilin lyase. J Biol
Chem 269, 8686–8694.
28 Zhao KH, Wu D, Zhou M, Zhang L, Bo
¨
hm S, Buben-
zer C & Scheer H (2005) Amino acid residues associated
with enzymatic activity of the isomerizing phycoviolobi-
lin-lyase PecE⁄ F. Biochemistry 44, 8126–8137.
29 Scheer H (1982) Phycobiliproteins: Molecular aspects of
photosynthetic antenna systems. In Light Reaction Path
of Photosynthesis (Fong, FK, ed.), pp. 7–45. Springer,
Berlin.
30 Braslavsky SE, Holzwarth AR & Schaffner K (1983)
Solution conformations, photophysics, and photochem-
istry of bili-pigments – bilirubin and biliverdin dimethyl-
esters and related linear tetrapyrroles. Angew Chem Int
Ed Engl 22, 656–674.
31 Berkelman T & Lagarias JC (1986) Visualization of
bilin-linked peptides and proteins in polyacrylamide
gels. Anal Biochem 156, 194–201.
32 Suckau D, Mak M & Przybylski M (1992) Protein sur-

face topology-probing by selective chemical modification
and mass spectrometric peptide mapping. Proc Natl
Acad Sci USA 89, 5630–5634.
33 Bazaes S, Goldie H, Cardemil E & Jabalquinto AM
(1995) Identification of reactive lysines in phosphoenol-
pyruvate carboxykinases from Escherichia coli and
Saccharomyces cerevisiae. FEBS Lett 360, 207–210.
34 Hoare DG & Koshland JDE (1967) A method for the
quantitative modification and estimation of carboxylic
acid group in proteins. J Biol Chem 242, 2447–2453.
35 Spande TF & Witkop B (1967) Determination of the
tryptophan content of proteins with N-bromosuccini-
mide. Methods Enzymol 11, 498–506.
36 Tsou CL (1962) Relation between modification of func-
tional groups of proteins and their biological activity. 1.
A graphical method for the determination of the num-
ber and type of essential groups. Sci Sin 11, 1535–1558.
37 Miles EW (1977) Modification of histidyl residues in
proteins by diethylpyrocarbonate. Methods Enzymol 47,
431–442.
38 Lundblad RL (1991) Chemical Reagents for Protein
Modification, 2nd edn. CRC Press, Boca Raton.
39 Kaneko T, Sato S, Kotani H, Tanaka A, Asamizu E,
Nakamura Y, Miyajima N, Hirosawa M, Sugiura M,
Sasamoto S et al. (1996) Sequence analysis of the gen-
ome of the unicellular cyanobacterium Synechocystis sp.
strain PCC6803. II. Sequence determination of the
entire genome and assignment of potential protein-
coding regions. DNA Res 3, 109–136.
40 Kaneko T, Nakamura Y, Wolk CP, Kuritz T, Sasamoto

S, Watanabe A, Iriguchi M, Ishikawa A, Kawashima
K H. Zhao et al. Chromophore binding by a-phycocyanin lyase
FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS 1273
K, Kimura T et al. (2001) Complete genomic
sequence of the filamentous nitrogen-fixing cyanobacte-
rium Anabaena sp. strain PCC 7120. DNA Res 8,
205–213.
41 Nakamura Y, Kaneko T, Sato S, Ikeuchi M, Katoh H,
Sasamoto S, Watanabe A, Iriguchi M, Kawashima K,
Kimura T et al. (2002) Complete genome structure of
the thermophilic cyanobacterium Thermosynechococcus
elongatus BP-1. DNA Res 9, 123–130.
42 Schluchter WM & Bryant DA (2002) Analysis and
reconstitution of phycobiliproteins: methods for the
characterization of bilin attachment reactions. In Heme,
Chlorophyll, and Bilins (Smith, AG & Witty, M, eds),
pp. 311–334. Humana Press, Totowa, NJ, USA.
43 Adir N, Vainer R & Lerner N (2002) Refined structure
of c-phycocyanin from the cyanobacterium Synechococ-
cus vulcanus at 1.6 A: insights into the role of solvent
molecules in thermal stability and co-factor structure.
Biochim Biophys Acta 1556, 168–174.
44 Kneip C, Parbel A, Fo
¨
rstendorf H, Scheer H, Siebert F
& Hildebrandt P (1998) FT-NIR-resonance raman spec-
troscopic study of the a-subunit of phycoerythrocyanin
and phycocyanin from the cyanobacterium Mastigocla-
dus laminosus. J Raman Spectr 29, 939–944.
45 Beuhler RJ, Pierce RC & Friedman L (1976) Reversible

addition 3-a
¨
thyliden mit R-SH. J Biol Chem 251, 2405–
2411.
46 Sambrook J, Fritsch EF & Maniatis T (1989) Molecular
Cloning: a Laboratory Manual, 2nd edn. Cold Spring.
Harbour Laboratory Press, New York.
47 Zhao K-H, Wu D, Wang L, Zhou M, Storf M,
Bubenzer C, Strohmann B & Scheer H (2002) Charac-
terization of phycoviolobilin phycoerythrocyanin-
a84-cystein-lyase-(isomerizing) from Mastigocladus lami-
nosus. Eur J Biochem 269, 4542–4550.
48 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriopharge T4.
Nature 227, 680–685.
49 Bradford MM (1976) A rapid and sensitive method for
the quantitation of microgram quantities of protein util-
izing the principle of protein-dye binding. Anal Biochem
72, 248–254.
50 Levy HM, Leber PD & Ryan EM (1963) Inactivation
of myosin by 2,4-dinitrophenol and protection by ade-
nosine triphosphatase and other phosphate compounds.
J Biol Chem 238, 3654–3659.
51 Horowitz P & Criscimagna NL (1982) The specificity of
active-site alkylation by iodoacetic acid in the enzyme
thiosulfate sulfurtransferase. Biochim Biophys Acta 702,
173–177.
Supplementary material
The following supplementary material is available
online:

Fig. S1. Kinetics of inactivation of CpcE by CHD (A)
and PGO (B). The insets show plots of log(1 ⁄ t
0.5
)of
inactivation against log(modifier concentration). In (A),
final CHD concentrations were (n)5mm;(d)10mm;
(
)15mm;( )30mm;(r)45mm; in (B) final PGO
concentrations were (n) 2.5 mm;(d)5mm;(
) 7.5 mm;
( )10mm;(r)15mm.
Fig. S2. Kinetics of inactivation of CpcF by EDAC.
The inset shows a plot of log(1 ⁄ t
0.5
) of inactivation
against log(modifier concentration). Final EDAC con-
centrations were (n) 2.5 mm;(d)5mm;(
) 7.5 mm;
(
)10mm;(r)15 mm.
Fig. S3. Relationship between residual enzyme activity
and the number of tryptophan residues modified by
NBS in CpcE. Squares (n) are the experimental values,
the lines are obtained from the equation of Tsou
35
with i ¼ 1 (—); i ¼ 2(ÆÆÆÆ); i ¼ 3 (– –).
This material is available as part of the online atricle
from
Chromophore binding by a-phycocyanin lyase K H. Zhao et al.
1274 FEBS Journal 273 (2006) 1262–1274 ª 2006 The Authors Journal compilation ª 2006 FEBS