Functional studies of the
Synechocystis
phycobilisomes organization
by high performance liquid chromatography on line with a mass
spectrometer
Lello Zolla, Maria Bianchetti and Sara Rinalducci
Dipartimento di Scienze Ambientali, Universita’ degli Studi della Tuscia, Viterbo, Italy
This study was designed to yield data on the supramolecular
organization of the phycobilisome apparatus from
Synechocystis, a nd the possible effects of environmental
stress on this arrangement. Phycobilisomes were dissociated
in a l ow ionic strength solution and a quantitative estimation
of the protein components present in each subcomplex was
obtained using liquid chromatography coupled on-line with
a mass spectrometer equipped with an electrospray ion
source (ESI-MS). An advantage of this approach is that
information can be collected on the initial events, which take
place as this organism adapts to environmental ch anges.
Ultracentrifugation of whole phycobilisomes revealed five
subcomplexes; the lightest c ontained four linker proteins
plus free phycocyanin, the second the core complex, while
the last three bands c ontained the rod complexes. Four
linkers were found in band 1 with higher molecular masses
than those expected from the DNA s equence, indicating that
they also con tain linked c hemical groups. U V-B irradiation
specifically destroyed the b-phycocyanin and one rod linker,
which resulted in the disintegration of the rod complexes.
The two bilins present in b-phycocyanin give a greater
contribution to the UV absorption than the s ingle bilin of the
other b ilinproteins and probably react with atmospheric
oxygen forming toxic radicals. The protein backbone is, in

allophycocyanin are composed of two polypeptide chains,
a and b, of approximately 17 000 and 18 000 [7]. The a and
b polypeptides contain one or two chromophores, respect-
ively. In a low ionic strength aqueous medium, phycobili-
somes dissociate into various components, and the
individual biliproteins, either with or without attached
linkers, are obtained [6]. The relative stability of biliprotein-
linker complexes varies among the biliproteins f rom differ-
ent sources.
Information on the supramolecular organization of
phycobilisomes has come from electro n microscopy
research, which showed that cyanobacteria contain various
structural types of phycobilisomes but the hemidiscoidal
phycobilisomes having a tricylindrical c ore and six rods are
found extensively in most cyanobacteria [1]. Allophycocy-
anin is organized as a trimer near neutral pH, having three a
and three b polypeptides; each of these polypeptides has one
chromophore (bilin). Trimers ( a
3
b
3
) are ringlike a ssemblies
of three monomers (ab) having threefold symmetry. Phy-
cocyanin is found in solution as a complex mixture of (a
3
b
3
)
(a
6

tion. It is generally accepted that the linkers govern the
assembly of the biliproteins into phycobilisomes, and,
despite being colorless, in certain cases they have been
shown to improve the energy migration process [7]. As some
of the linkers mediate assembly of the biliproteins, they
produce changes in the spectra of biliproteins, and t his may
serve to direct energy migration more efficiently through the
phycobilisomes [9,10]. L inkers are found in both c yanobac-
terial and red algae phyc obilisomes and m ay constitute 10–
15% of the total mass [11–13]. Details of how linkers
perform their tasks are still largely unknown.
In this paper we use a recently developed HPLC-ESI-MS
method that allows rapid and efficient separation of
phycobilisomes upon injection of entire subcomplexes [14].
The nondisruptive nature of this method has made it
possible to collect information on the composition of each
subcomplex and on the organization of phycobilisomes
under different environmental conditions. Preliminary
studies on how both environmental factors su ch as UV
radiation and physiological stress such as starvation may
affect this supraorganization are also presented.
MATERIALS AND METHODS
Chemicals
Reagent-grade phosphoric acid, magnesium chloride,
sodium chloride, trifluoroacetic acid, methanol, ethanol,
formamide, as well as HPLC-grade water and acetonitrile,
were obtained from Carlo Erba (Milan, Italy). Sucrose,
Tris, Mes, sodium nitrate, magnesium sulfate, calcium
cloride, citric acid, manganese cloride, cupper sulfate, zinc
sulfate, Hepes, from Sigma, acrylamide and N,N¢-methy-

and disrupted by 15 cycles of 30 s in a Braun Homogenizer.
The cell debris was eliminated by centrifugation as before
and the supernatant was spun at 148 000 g
3
in a TFT 50.38
Kontron centrifuge at 4 °C [16]. The supernatant was
collected and used for HPLC separation without any
further purification.
Sucrose gradient ultracentrifugation
The experimental conditions were as reported by Sinha
et al. [17] with the following modification: the dissociated
complexes was loaded onto a 0–60% sucrose gradient in
0.75
M
phosphate buffer pH 7, and
4
spun at 272 000 g
for 40 h using a Kontron TST 41.14 rotor. The blue
pigmented bands were harvested with a syringe and
analyzed directly.
UV-B and visible light treatment
Culture cells in suspension as well as isolated phycobili-
someswereexposedfor4hat1.8WÆm
)2
at room
temperature to artificial UV-B produced by a transillumi-
nator (Bio-Rad), with its main output at 312 nm. Suspen-
sions were gently agitated by a magnetic stirrer during
irradiation to ensure uniform distribution.
For the visible ligh t treatment, the c ells and phycobili-

packed with 5-lm po rous butyl silica particles [20,21]. This
column was operated at a flow rate of 1 mLÆmin
)1
for
optimum separation efficiency. All solutions were filtered
through a Millipore (Milan, Italy) type FH 0.5-lmmem-
brane filter and degassed by bubbling with helium before
use. Optimization of chromatographic separations was
performed using a Beckman (Fullerton, CA, USA) Gold
System with of Model 126 solvent delivery pumps. Samples
were introduced onto the column by a Model 210 A sample
injection valve with either a 20- or a 50-lL sample l oop.
The V ydac C-4 columns were pre-equilibrated with 20%
(v/v) aqueous acetonitrile solution containing 0.1% (v/v)
trifluoroacetic acid and samples were eluted using a gradient
from 20 to 95% (v/v) acetonitrile in 60 min, at a flow rate of
1mLÆmin
)1
[14].
Electrospray mass spectrometry
The HPLC-ESI-MS experiments were carried out by
splitting the outlet of the HPLC and co upling it with a
Perkin Elmer API 2000 or API 365 triple quadrupole mass
spectrometer equipped with the electrospray ion source [22].
For HPLC-MS analysis, with pneumatically assisted elec-
trospray, a spray voltage o f 5 kV and a sheath gas pressure
of 500 kPa were employed. Protein mass spectra were
Ó FEBS 2002 Study of phycobilisomes organization by HPLC-ESI-MS (Eur. J. Biochem. 269) 1535
recorded by scanning the first quadrupole; the scan range
was 500–1800 average mass units

Synechocyst is PCC 6803, schematically represented in
Fig.1A,intoaC4RP-HPLCsystemresultedinthe
separation of the phycobilisome complex in four main
peaks and many smaller p eaks (Fig. 1B), as observed by
absorbance detection at 214 nm. Simultaneously, the spec-
trum of each eluting peak was recorded by a photodiode
array detector. The c hromatogram recorded at 600 nm (see
left inset of 1B) shows that only peaks 6, 7, 8 and 9 had an
absorption at 600 nm, typically due to the presence of bilin
pigments still connected with the polypeptide backbone [14].
Coupling the HPLC on line with a mass spectrometer using
electrospray as source allowed us t o determine the molecu-
lar masses of the proteins in each HPLC peak. The peaks
eluting within 25 min represented linker p roteins, while the
four main peaks between 28 and 33 min were the phyco-
cyanins and allophycocyanins [15]. In Table 1 the experi-
mental molecular masses determined by the deconvolution
of the ESI-MS spe ctra have been correlated t o the expected
Table 1 . List of the Synechocystis 6803 phycobilisome protein components determined by HPLC-ESI-MS compared with the proteins expected from
DNA sequence and those observed by SDS/PAGE. The molecular masses were determined by the deconvolution of the ESI-MS spectra recorded
during the chromatographic run into a C-4 reverse phase column coupled on-line with by a mass spectrometer equipped with an electrospray ion
source (ESI-MS). The values o f molecular mass deduced from DNA sequence have been c a lculated by u sing the Prot Parameter t oo ls in the
EXPASY
program. The SDS/PAGE molecular masses were deduced by the marker used in the e lectrophoresis reported in the right inset of Fig. 1B.
Type of protein Proteins
Molecular masses
HPLC
peak number
Expected from
DNA sequence

molecular masses observed in S DS/PAGE (right inset of
Fig. 1B). The molecular mass values obtained are very close
to those expected; the differences observed (ranging from
580 to 607 Da) were probably due to the presence of
residual bilin pigment still bound to the proteins, whose
molecular mass is 587 Da. Proteins with molecular masses
ranging from 32 000 to 37 000 as well as proteins with
molecular masses under 10 000 have been tentatively
attributed to linker p roteins. However, most of the proteins
found by HPLC-ESI-MS correspond to the bands observed
by SDS/PAGE with the exception of the band over
45 kDa, which was previously interpreted as particle
containing both phycocyanin and allophycocyanin [23] or
ferredoxin [24].
When whole phycobilisomes were suspended in low ionic
strength buffer the interactions between the linkers and
different protein components of p hycobilisomes were
destroyed, and various incomplete subcomplexes were
formed [6]. With the aim to get information on the
localization and distribution of the phycobilisome compo-
nents, the mixture of subcomplexes was subjected to sucrose
gradient ultracentrifuge separation. The results are shown in
Fig. 2A: four main blue bands were observed, plus a faint
blue band at the t op. The first band, corresponding to the
smaller complexes, was faintly colored, the second was the
most colored and the more abundant, whereas the other two
bands, observed at a higher sucrose concentration, con-
tained the higher density subcomplexes. Each band was
collected with a syringe and analyzed by RP-HPLC. Besides
being more efficient than SDS/PAGE in discriminating

analysis performed on the corresponding ESI-MS spectra of
the HPLC peaks indicated by an arrow. Obviously sucrose-
gradient separation has resulted in an enrichment of linker
proteins in band 1. However, the heaviest linker ( M
r
of
100 000.5) is not revealed, b ecause it is not recovered under
these separation conditions, as confirmed by SDS/PAGE.
In fact, it h as been reported t o remain enclosed i n the
phycobilisomes unless treated with high salt [25]. Further-
more, in this sucrose band 1 linker p roteins with m olecular
masses under 1 0 000 were not found. They are detected by
ESI-MS into band 5 (data not shown). This result agrees
with the hypothesis that these small linkers remain tightly
bound to the ro d complexes [3].
From this preliminary analysis it may be concluded
that the breakdown of phycobilisomes by low ionic
strength causes the supramolecular complex to be degra-
ded into many subcomplexes, which differ from each
other by the percentage and type of b ilin proteins they
contain.
In order to study the influence of certain environmen tal
changes on t he supramolecular organization of the
Fig. 2. HPLC analysis of phycobilisome complexes. (A) Ultracentri-
fuge tube containing the phycobilisome apparatus from Synechocystis
PCC6803oncefractionatedinitscomponentsandloadedontoa
0–60% sucrose gradient in phosphate buffer. The ind ividual bands
obtained are labeled with number 1–5 st arting from the t op of tube.
(B) Panels B 1–B5 s how t he RP- HPLC p rofi le whe n e ach s ucrose b and
is loaded onto a reversed phase C4 column. B2 reports the identifica-

green pigmentation with time, clearly distinguishable from
thenativeblue-greenofthecontrolstrain;theisolated
phycobilisomes gave a much fainter blue pigmentation than
the control. HPLC analysis of entire phycobilisomes before
and after illumination reve aled the d isappearance of the
b-phycocyanin already after 1 h of illumination (Fig. 4). In
contrast, SDS/PAGE analysis of the total mixture of
phycobilisomes exposed to the same UV-B irradiation did
not reveal any significant decrease in the total intensity of
the stained bands, confirming the h igh sensitivity of the
HPLC method (inset A of Fig. 4). Nevertheless, Fig. 4
(inset B) shows the time course of spectroscopic a bsorption
recorded on intact phycobilisome undergone to UV-B
irradiation. It may be observed t hat the absorption at
578 nm is more effected, suggesting involvement o f bilin
chromophores in UV-B damage.
Sucrose-gradient ultracentrifugation of UV-B treated
phycobilisomes (Fig. 5B) showed only two main bands
plus a faint one at the top instead of the five observed in the
control (Fig. 5A), the two heaviest bands having completely
disappeared. HPLC analysis of the second sucrose bands of
UV-B treated phycobilisome revealed that the b-phycocy-
anin peak at 214 nm (indicated with an arrow) was missing
from both of them, while all the other components were
present. Thus it may be inferred that b-phycocyanin is the
main target of UV-B radiation, and that this component is
essential for the formation of all subcomplexes. However,
the disappearance of a peak at 214 nm in the HPLC
chromatogram does not reveal the extent of the damage to
b-phycocyanin. In fact, it may be that a single chromophore

and biliproteins. It was observed that the b-phycocyanin
peak and the peak corresponding to linker 1 (both indicated
by an arrow) were reduced in treated samples (Fig. 6B).
Because in mass spectroscopy the ionic current is strongly
dependent on the protein and not on prosthetic groups, it
could seem reasonable to conclude that both peak reduc-
tions were related to the destruction of t he polypeptide
backbone of the native protein. However, SDS/PAGE of
each HPLC fraction (inset B, line 2) showed that
b-phycocyanin was partially reduced, but not completely
destroyed, contradicting the RIC and UV results. S uch
discrepancies are commonly observed when active oxygen is
involved in photodamage of proteins. T hey arise from the
fact that in hydrophobic proteins radical attacks occur
preferentially in the external hydrophilic region of the
protein, a region rich in amino acids with a higher than
average contribution to the UV optical absorption and
protonation of amino g roups by electrospray ionization,
leading to falsely high e stimates of protein loss by these
methods (L. Zolla, & S. Rinalducci, unpublished results).
This consideration suggested that oxygen radicals might be
involved in the phe nomena described. Thus, in order to
collect more info rmation on the p ossible molecular mech-
anism by which UV -B affecte d the p hycobilisomes appar-
atus, we subjected our sample to UV-B exposure in the
presence or in absence o f oxygen as w ell as in the presen ce of
ascorbate, a well-known scavenger of oxygen radical
species. In anaerobiosis or the presence of ascorbate the
b-phycocyanin was not affected by UV-B irradiation (data
not shown), confirming the possible involvement of oxygen

Synechocystis 6803, together with the determination of their
molecular mass, has s uccessfully been achieved [14] by the
combined use of HPLC coupled on-line with a mass
spectrometer equipped with an electrospray ion source
(ESI-MS). I n t he present paper, we have employed this
method to make a quantitative estimation of components
present in individual subcomplexes obtained by dissociation
in a low ion ic strength solution. Information was collected
on the possible supramolecolar organization and how some
types of environmental stress may interfere with i t. In low
ionic strength conditions phycobilisomes dissociate into
water-soluble s ubcomplexes, which can be separated into
Fig. 6. RP-HPLC-ESI-MS chromatograms comparison from control
and UV-B irradiated phycobilisomes. (A) Reconstructed ionic cu rrent
(RIC) obtained l oading the sucrose gradient band 1 onto a C4 c olumn
coupled on line with a mass spectrometer inter faced with an electro-
spray. Arrows indicate the peaks affected by UV-B irradiation. (B)
RIC of phycobilisomes irradiated with UV-B at 1.8 WÆm
)2
for 4 h.
Inset of (A) and (B) show the SDS/PAGE of the main HPLC peak
once collected, dried and loaded onto 16.5% T 5 .4% C Tris/tricine gel.
Lines 1–4 refer to the PCa,PCb,APCa and APCb, respectively. PC,
phycocyanin; APC, allophycocyanin.
Ó FEBS 2002 Study of phycobilisomes organization by HPLC-ESI-MS (Eur. J. Biochem. 269) 1539
five different bands upon extended ultracentrifugation. The
relative stability of biliprotein-linker complexes is known to
differ among the biliproteins and also depending on the
species. However, our data show that in Synechocystis the
linkers are preferentially removed by l ow ionic strength

are present in equivalent amounts. This is in agreement with
the current model where the functional units of all
phycobilisomes are disk shaped trimers of closely associated
(ab) phycobiliproteins [1–4]. An unexpected exception is
observed for the subcomplex contained i n the su crose band
2 where the amount of the b-phycocyanin component is
significantly reduced. The allophycocyanins are mainly
found in the second band, which p robably corresponds to
the core, where the allophycocyanins are prevalently located
[1]. However, the second band probably contains only
remains of dissociated cores, because the intact core
complex has an expected molecular m ass close to that of a
phycobilisome rod, and so should migrate down the
gradients. The lack of intact cores is probably due to
absence i n our preparation of the L
CM
protein r equired for
the core assembly [25]. Data reported previously confirm
that different rod subcomplexes are present around the core,
which are held together by linker proteins to form the
Synechocyst is phycobilisome supramolecular organization.
In agreement, X-ray crystallography has shown that the
hexamers (a
6
b
6
) are disk shaped, f ormed by face-to-face
assembly of trimers. Rods are formed by face-to-face
assembly of these disks [26].
Treatment of phycobilisomes with UV-B destroys this

attribute the specific damage to variations in aromatic
amino-acid content, because the amino-acid composition of
allophycocyanins and phycocyanins is significantly con-
served, it seems reasonable to attribute their greater
sensitivity to the external allocation in the supramolecular
organization. In a recent p aper isolated a-andb-phyco-
cyanins were irradiated for various lengths of time and
Fig. 7. Effect of nutrient deprivation on phycobilisome protein compo-
sition. (A) RP-HPLC chromatogram of entire phycobilisomes used as
control. (B) The RP-HPLC pattern of phycobilisome isolated from
cells grown in nutrient starvation. The arrows indicate the main peaks
affected. The growing conditions are reported in Materials and
methods section. Insets show the ultracentrifugation tu bes obtained
upon loading phycobilisomes from control and starved cyanobacteria
onto a 0–60% sucrose gradient.
1540 L. Zolla et al. (Eur. J. Biochem. 269) Ó FEBS 2002
analyzed by HPLC on a reversed-phase column; both
phycobiliproteins s howed similar photo destruction quan-
tum yields [29]. On the other hand, irradiation at 280 or
640 nm caused the same extent o f damage, indicating that
both tryptophan and bilin absorption are involved in the
phenomenon observed. In contrast, our experiments
showed that whether entire phycobilisomes or whole cells
were irradiated with UV-B, only the b-phycocyanin com-
ponent was s ignificantly a ffected. This specific damage was
observed a fter the first hour of irradiation by both optical
absorption and RIC decrease. However, although the
optical decreases observed could be correlated to the
damage of aromatic amino acids, this could not explain
the RIC decrease, b ecause protonation of positive a mino

with the previous observation that upon light adaptation
some phycocyanin (ab)
3
units were released and conse-
quently the length of phycobilisome rods resulted reduced
[27]. In any case, cyanobacteria seem to be better adapted
than higher plants to endure h igh intensity visible light [1].
Finally, we have here presented an example of a response
to an environmental change that does not interact directly
with the chromophore, but is supposed to involve the
biosynthetic apparatus of t he cells. Synechocystis grown
under sulfur- and nitrogen-deficient conditions contained
less b-phycocyanin; this only had a marked effect on band 5,
the heaviest one, consistent with th e idea that this band
contains mainly b-phycocyanins. Probably during nitrogen
starvation b-phycocyanin represents the main source of
nitrogen for cyanobacteria [33]. Moreover, the linker with
molecular mass 34 316 is also lost, raising the possibility
that this rod linker, which is different to that destroyed by
UV-B, plays a role in binding the heaviest subcomplex
present in band 5. However, it is worth emphasizing that all
this evidence is obtained by the HPLC method during the
first days of starvation, allowing us to monitor the first steps
of phycobilisome reorganization.
In conclusion, the data reported here demonstrate that
the use of analytical methods with greater resolving powers
can reveal the initial events in the process of damage, which
are well observed by HPLC after 3 h of UV-B irradiation or
3 d ays of starvation, but not by SDS/PAGE analysis.
Moreover, by the HPLC-ESI-MS method, a minimum

cladus laminosus. Photochem. Photobiol. 62, 847–854.
5. Glazer, A. (1989) Minireview: directional energy transfer in
phycobilisomes. J. Biol. Chem. 264, 1–4.
6. Glazer, A.N. (1988) P hycobiliproteins. Methods Enzymol. 167,
291–303.
7. Sidler, W. (1997) Phycobilisomes and phycobiliprotein structure.
In The Molecular Biology of Cyanobacteria (Bryant, A., ed.).
Kluwer, Dordrecht.
11
8. We stermann, M., Reuter, W., Schimek, C. & Wehrmeyer, W.
(1993) Presence of both hemidiscoidal and hellipsoidal phycobili-
somes i n a Phormidium species (cyanobacteria). Z. Naturforsch.
C48, 28–34.
9. Yu, M.H., Glazer, A.N. & Williams, R.C. (1981) Cyanobacterial
phycobilisomes. Phycocyanin assembly in the rod substructure
of Anabaena variabilis phycobilisomes. J. Biol. Chem. 256,
13130–13136.
10. Wendler, J., John, W., Scheer, H. & Holzwarth, A.R. (1986)
Energy transfer in trimeric C-phycocyanin studied by picosecond
fluorescence kinetics. Photochem. Ph otobiol. 44, 79–85.
11. T andeau de Marsac, N. & Cohen-Bazire, G. (1977) Molecular
composition of cyano bacterial phyc obilisomes. Proc. Natl Acad.
Sci. USA 74, 1635–1639.
Ó FEBS 2002 Study of phycobilisomes organization by HPLC-ESI-MS (Eur. J. Biochem. 269) 1541
12. Yamanaka, G., Glazer, A.N. & Williams, R.C. (1978) Cyano-
bacterial phycobilisomes. Characterization of the phycobilisomes
of Shynechoccocus sp 6301. J. Biol. Chem. 253, 8303–8310.
13. K oller, K.P., Wehrmeyer, W. & Morschel, E. (1978) Biliprotein
assembly in its disc-shaped phycobilisomes of Rhodella violacea.
On the molecular composition of energy-transferring complexes

Manera, F. & Corradini, D. (1999) Isolation and characterization
of chloroplast photosystem II antenna of spinach by reversed-
phase liquid chromatography. Photosynth. Res. 61, 281–290.
22. Corradini, D., Huber, C.G., Timperio, A.M. & Zolla, L. (2000)
Resolution and identification of the protein components of the
photosystem II antenna s ystem of h igher plants by r eversed-phase
liquid chromatography with electrospray-m ass spectrometric
detection. J. Chromatogr. A. 886, 111–121.
23. Yamanaka, G., Lundell, D.J. & Glazer, A.N. (1982) Molecular
architecture of a light-harvesting antenna. Isolation and
characterization of phycobilisome subassembly particles. J. Biol.
Chem. 257, 4077–4086.
24.vanThor,J.J.,Gruters,O.W.M.,Matthijs,H.C.P.&
Hellingwerf, K.J. (1999) Localization and function of ferredoxin:
NADP
+
reductase bound to the phycobilisomes of Synechocystis.
EMBO J. 18, 4128–4136.
25. Re uter, W., Westermann, M., Brass, S., Ernst, A., Boger, P. &
Wehrmeyer, W. (1994) Structure, composition, and assembly of
paracrystalline phycobiliproteins in Synechocystis sp. strain BO
8402 and of phycobilisomes in the derivative strain BO 9201.
J. Bacteriol. 176, 896–904.
26. Elmorjani, K., Thomas, J.C. & Sebban, P. (1986) Phycobilisomes
of wild type and pigment mutants of the cyanob acterium
Synechocystis PCC 680 3. Arch. Microbiol. 146, 186–191.
27. Ritz, M., Thomas, J C., Spilar, A. & Etienne, A L. (2000)
Kinetics of photoacclimatation in response to a shift to high light
of the red alga Rhodella violacea adapted to low irradiance. Plant
Physiol. 123, 1415–1425.