Journal of Experimental Botany, Vol. 59, No. 7, pp. 1605–1614, 2008
doi:10.1093/jxb/erm310 Advance Access publication 22 January, 2008
SPECIAL ISSUE REVIEW PAPER
Redox modulation of Rubisco conformation and activity
through its cysteine residues
Joaquı́n Moreno*, Marı́a Jesús Garcı́a-Murria and Julia Marı́n-Navarro
Department Biochemistry and Molecular Biology, Faculty of Biology, University of Valencia, Dr Moliner 50,
Burjassot E-46100, Spain
Received 25 August 2007; Revised 26 October 2007; Accepted 12 November 2007
Abstract
Treatment of purified Rubisco with agents that specifically oxidize cysteine-thiol groups causes catalytic
inactivation and increased proteolytic sensitivity of
the enzyme. It has been suggested that these redox
properties may sustain a mechanism of regulating
Rubisco activity and turnover during senescence or
stress. Current research efforts are addressing the
structural basis of the redox modulation of Rubisco
and the identification of critical cysteines. Redox shifts
result in Rubisco conformational changes as revealed
by the alteration of its proteolytic fragmentation pattern upon oxidation. In particular, the augmented
susceptibility of Rubisco to proteases is due to increased exposure of a small loop (between Ser61 and
Thr68) when oxidized. Progressive oxidation of
Rubisco cysteines using disulphide/thiol mixtures at
different ratios have shown that inactivation occurs
under milder oxidative conditions than proteolytic
sensitization, suggesting the involvement of different
critical cysteines. Site-directed mutagenesis of conserved cysteines in the Chlamydomonas reinhardtii
Rubisco identified Cys449 and Cys459 among those
involved in oxidative inactivation, and Cys172 and
Cys192 as the specific target for arsenite. The physiological importance of Rubisco redox regulation is
supported by the in vivo response of the cysteine
mutants to stress conditions. Substitution of Cys172
caused a pronounced delay in stress-induced Rubisco
degradation, while the replacement of the functionally
redundant Cys449-Cys459 pair resulted in an enhanced catabolism with a faster high-molecular weight
polymerization and translocation to membranes.
These results suggest that several cysteines contribute to a sequence of conformational changes that
trigger the different stages of Rubisco catabolism
under increasing oxidative conditions.
Key words: Chlamydomonas reinhardtii, chloroplast, critical
cysteines, oxidative inactivation, proteolytic enhancement,
redox control, Rubisco catabolism, senescence, site-directed
mutants, thiol-disulphide exchange.
Redox regulation in the chloroplast
Regulation by redox balance is a recurrent feature that
pervades many aspects of plant biology (Buchanan and
Balmer, 2005). It seems particularly fit that redox
regulation mechanisms should exist inside the chloroplast
because redox shifts are strongly correlated with changes
in the physiological state and photosynthetic activity of
the organelle. The shining of light on the thylakoids
produces an electron flow through the components of the
photosynthetic chain up to ferredoxin (Fig. 1). From that
point, reducing power is distributed between NADPH, the
main cofactor transferring reducing equivalents to anabolic pathways, and the chloroplast thioredoxins, signalling molecules which, in turn, activate target proteins
through reduction of critical cysteine residues (Schürmann
and Jacquot, 2000). Besides, the redox state of plastoquinone, one intermediate electron carrier of the photosynthetic chain, is also known to regulate the transcription of
specific chloroplast genes (Pfannschmidt et al., 1999).
From NADPH, reducing power is also transferred to
glutathione (GSH) and ascorbate by means of succesive
redox reactions catalysed by the NADPH–glutathione
* To whom correspondence should be addressed. E-mail: [email protected]
Abbreviations: AP/GR, ascorbate peroxidase/glutathione reductase; CSH, cysteamine; CSSC, cystamine; GSH, glutathione; GSSG, glutathione disulphide;
ROS, reactive oxygen species; Rubisco, ribulose 1,5-bisphosphate carboxylase/oxygenase; SDS-PAGE, polyacrylamide gel electrophoresis with sodium
dodecyl sulphate; SOD, superoxide dismutase.
ª The Author [2008]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved.
For Permissions, please e-mail: [email protected]
1606 Moreno et al.
Fig. 1. Distribution of light-generated reducing power inside the chloroplast. Light drives the electron flow through the photosynthetic chain located
in the thylakoid membrane (depicted here as a double solid line enclosing a series of dots). The last acceptor in this chain of redox reactions is
ferredoxin (Fd) which transfers electrons to thioredoxins (Trx) and NADP+. NADPH is partly consumed at the reductive phases of assimilatory
metabolism, but also regenerates glutathione (GSH) and ascorbate, which constitute the main redox buffers in the chloroplast. Thioredoxins reduce
(and activate) a functionally important set of target proteins in the chloroplast. The redox state of the plastoquinone (PQ) pool has also been reported
to control the expression of some chloroplast genes. Reactive oxygen species (ROS) are generated as a consequence of the escape of electrons from
the photosynthetic redox chain. ROS can be quenched or converted to less harmful species by detoxifying mechanisms, including superoxide
dismutase (SOD), the ascorbate peroxidase/glutathion reductase (AP/GR) system, peroxiredoxin, tocopherols, and carotenoids.
reductase and the glutathione–ascorbate reductase, respectively (Fig. 1). GSH and (specially) ascorbate are
very abundant molecules in the chloroplast stroma, their
concentrations reaching the millimolar range (Law et al.,
1983). Thereby, ascorbate and GSH constitute a pool of
redox buffers, providing inertia that opposes fast redox
fluctuations inside the chloroplast. At a certain point, the
balance of reductive and oxidative activities taking place
inside the chloroplast defines a stromal redox status,
which may be estimated as the ratio of oxidant to
reductant forms of any one of the main redox pairs that
are mutually connected through redox exchange (e.g.
NADP+/NADPH, GSSG/GSH, dehydroascorbate/ascorbate
or oxidized/reduced thioredoxin).
As a side reaction of photosynthetic electron transfer,
and under certain conditions (such as excess light), some
electrons may divert from their path, react with ambient
O2 molecules, and produce incompletely reduced forms of
oxygen, known as ‘reactive oxygen species’ or ROS (Fig.
1). Although ROS can also act as intracellular signals
(Apel and Hirt, 2004), these species are very active as
oxidants, recruiting the missing electrons from a variety of
nearby molecules and, thereby, potentially damaging
functional photosynthetic structures. ROS production is
usually minimized under physiological conditions but it is
altogether unavoidable in a changing environment. Moreover, generation of ROS is increased when the components of the photosynthetic electron transport chain are
dismantled (during senescence) (McRae and Thompson,
1983) or damaged (as a result of a variety of environmental stresses). ROS bursts are usually counteracted by (i)
the glutathione–ascorbate redox buffer pool (Noctor and
Foyer, 1998); (ii) ROS-deactivating enzymes (converting
the most damaging species to less reactive forms) such as
superoxide dismutase (SOD) (Alscher et al., 2002),
peroxiredoxin (Dietz et al., 2006), or the ascorbate
peroxidase/glutathione reductase (AP/GR) system (Shigeoka
et al., 2002); and (iii) ROS-quenching (i.e. electron donor)
substances such as tocopherol (Krieger-Liszkay and
Trebst, 2006) and carotenoids (Davison et al., 2002) (Fig.
1). However, under prolonged acute stress or at later
stages of senescence, the protective agents are eventually
overridden, and the chloroplast milieu turns increasingly
oxidative. Lipid peroxidation and protein oxidative modifications are common markers for monitoring the degree
of oxidation achieved inside the chloroplast at a given
stage.
Redox regulation of Rubisco
The CO2 (and O2) -fixing enzyme, ribulose 1,5-bisphosphate
carboxylase/oxygenase (EC 4.1.1.39, Rubisco), controls
the crucial step of partitioning ribulose 1,5-bisphosphate
between the assimilatory photosynthetic metabolism and
its carbon-releasing counterpart, the photorespiratory pathway. Besides, it is a very abundant protein, densely filling
the chloroplast stroma of photosynthetic eukaryotes and
storing a remarkable percentage of the macronutrients
(specially N) allocated inside the chloroplasts (Evans and
Seeman, 1989). Several aspects of Rubisco synthesis and
activity have been proposed to be regulated by redox
poise under physiological conditions. For instance, the
stability of the chloroplastic rbcL mRNA (encoding the
large subunit of Rubisco) has been shown to decrease as
a result of the daily dark-to-light transition in a redoxdependent manner (Salvador and Klein, 1999). This effect
is thought to be mediated by redox-sensitive protein
factors binding to mRNAs. Besides, it has been proposed
Cysteine-mediated redox modulation of Rubisco 1607
that the nascent Rubisco large subunit may itself block its
own translation under oxidative conditions by exposing
a domain that binds to RNA when oxidized (Yosef et al.,
2004). The enzymatic activity of Rubisco may also be
affected by redox balance in some species through the
activity of the thioredoxin-dependent large isoform of
the Rubisco activase, an auxiliary enzyme that releases
tightly-binding natural inhibitors from the catalytic site of
Rubisco (Zhang et al., 2002). The nuclear-encoded small
subunit of Rubisco has also been identified as a direct
target of thioredoxin in a comprehensive screening among
chloroplast proteins (Motohashi et al., 2001). However,
the potential regulatory effects derived from this interaction are yet to be described.
During senescence or under different types of stress,
Rubisco is known to undergo several oxidative modifications, such as thiol oxidation (Garcı́a-Ferris and Moreno,
1994), formation of carbonyl adducts at certain residues
(Eckardt and Pell, 1995; Junqua et al., 2000), acidification
and multimerization mediated by an enzymatic oxidase
system (Ferreira and Davies, 1989; Ferreira and Shaw,
1989) and non-enzymatic fragmentation caused by ROS
(Ishida et al., 1998; Nakano et al., 2006). As a result of
these modifications, an increasing fraction of Rubisco
becomes cross-linked to high molecular weight polymers
(Ferreira and Shaw, 1989, Marı́n-Navarro and Moreno,
2006), and associated to membranes (Mehta et al., 1992;
Garcı́a-Ferris and Moreno, 1994). All these processes are
thought to underlie or facilitate the fast and selective
degradation of the enzyme that takes place habitually
under such conditions (Albuquerque et al., 2001; Ferreira
et al., 2000). However, not all types of modifications are
detected under different stresses, suggesting that Rubisco
catabolism may follow different courses depending on the
nature of the stress and the specific conditions. Indeed, the
co-existence of different catabolic routes for Rubisco
turnover seems to be a reasonable assumption in view of
the divergent experimental evidence on Rubisco degradation found under different conditions (Hörtensteiner and
Feller, 2005). In this review, the focus is exclusively on
cysteine thiol oxidation, and its consequences for Rubisco
performance.
Cysteine-dependent redox properties of Rubisco
Cysteine-thiol oxidation is a widely encountered means of
modifying the activity of functional proteins in redox
changing environments. Besides, it is also frequent as
a marking step for protein turnover (Stadtman, 1990).
Thiol groups may undergo progressive degrees of oxidation, from the mild disulphide (a reversible S–S bond
established with another thiol) to the sulphenic (–SOH),
sulphinic (–SO2H), and sulphonic (–SO3H) acid derivatives. The lower oxidation states of disulphide and
sulphenic acid may easily be reverted again to the
sulphydryl state. Disulphides can be regenerated by disulphide exchange with free thiols, while sulphenic acid may
be reduced by thioredoxins and glutaredoxins (Hancock
et al., 2006). Some sulphinic acid derivatives can also be
reduced back to sulphenic acid through the action of
specific ATP-dependent sulphiredoxins, which are known
to be present in the chloroplast (Rey et al., 2007), while
sulphonic acid is habitually considered an irreversible
final oxidation step.
Rubisco was first described as a ‘cysteine-sensitive’
enzyme on the basis of its inactivation by thiol-directed
reagents such as p-chloromercuribenzoate (Sugiyama et al.,
1968) or the affinity label N-bromoacetylethanolamine
phosphate (Schloss et al., 1978). The latter labelling
allowed the identification of Cys172 and Cys459 as the
targets for the affinity reagent and, thus, as the residues
that were critical for enzymatic activity. Therefore, when
the structure of the enzyme was solved, it came as a
surprise that no cysteine appeared to play a significant
role in the catalytic mechanism nor to be directly implicated in the active site architecture (Andersson et al.,
1989). Hence, inactivation by cysteine reagents appears to
result from indirect structural effects caused by the
covalent bonding of cysteines to relatively bulky moieties.
Indeed, it had already been detected that some of these
modifications led to holoenzyme disassembly (Sugiyama
et al., 1968).
It was later found that the activity and conformation of
Rubisco can be modified by mild oxidative treatments
affecting cysteines (Tenaud and Jacquot, 1987; Peñarrubia
and Moreno, 1990; Garcı́a-Ferris and Moreno, 1993) that
do not disrupt the quaternary structure of the enzyme
(Marı́n-Navarro and Moreno, 2003). To achieve this
transition mimicking different redox ambients in vitro, the
purified Rubisco is incubated in the so-called redox
buffers (Gilbert, 1982). These are (pH-buffered) aqueous
solutions containing a mixture of a small disulphide (such
as cystamine) and its corresponding thiol (cysteamine)
at different ratios; the higher the disulphide/thiol ratio,
the more oxidizing the environment. The use of these
disulphide/thiol buffers under a nitrogen atmosphere (to
avoid other spontaneous oxidation processes) ensures that
only cysteine residues are affected and, furthermore, that
cysteine oxidation progresses only to the disulphide state.
Under those conditions, a typical redox-dependent behaviour is exemplified by the Chlamydomonas reinhardtii
Rubisco in Fig. 2. The exposure to an increasingly oxidative environment triggers a sharp transition to a fully
inactive form of Rubisco with a midpoint corresponding
to a disulphide/thiol ratio of about 1.5 (Fig. 2A). Moreover, Rubisco oxidation also induces structural alterations,
as revealed by a change of the fragmentation pattern
produced by low-specificity proteases (such as subtilisin or
proteinase K) acting on purified Rubisco upon oxidation
1608 Moreno et al.
Fig. 2. Redox-dependent inactivation (A) and modification of the
proteolytic pattern (B) of Rubisco. (A) Redox buffers were prepared by
combining cystamine (CSSC) and cysteamine (CSH) at different ratios
while keeping the monomeric concentration (CSH+2CSSC) constant
and equal to 40 mM, in a Rubisco activation buffer (100 mM TRIS–
HCl, 10 mM NaHCO3, 10 mM MgCl2 pH 8.2). Purified Rubisco (0.13
mg ml1) was incubated with the redox buffers for 2 h at 30 C under
a nitrogen environment and the residual carboxylase activity was
determined. The activity as a percentage of that of the fully reduced
enzyme (CSSC/CSH¼0) is represented. Error bars indicate standard
deviation from triplicates. (B) Purified Rubisco (0.13 mg ml1) either
reduced with 40 mM CSH or oxidized with 20 mM CSSC as above was
subjected to a proteolysis with subtilisin (0.5 lg ml1) in Rubisco
activation buffer. At the indicated times, the reaction was stopped
adding 2 mM phenylmethylsulphonyl fluoride and keeping on ice for 10
min. The samples were subsequently analysed by SDS-PAGE. The 55
kDa Rubisco large subunit (LS) and its degradation products of 53 kDa
(band 1) and 47 kDa (band 2) are indicated with arrows.
site-directed mutagenesis of each of the conserved residues, exchanging it by serine (a conservative substitution replacing a single sulphur atom by oxygen). Perhaps
surprisingly, all single mutants of these highly conserved
residues were recovered as photosynthetically competent
strains (Moreno and Spreitzer, 1999; Marı́n-Navarro and
Moreno, 2006), thereby reinforcing the idea that these
cysteines are not crucial for catalytic activity and may be
conserved for other reasons. The characteristic properties
and phenotypes of some of these Rubisco mutants will be
described below. It should be noted, however, that redox
modulation of Rubisco can result from the additive, cooperative, or even redundant contribution of several
cysteines and, thus, the role of a specific residue may not
be readily appreciable in a single-cysteine mutant. A hint
on the implication of several cysteine residues in the
oxidative transitions of Rubisco comes from the fact that
the midpoint of proteolytic sensitization of the enzyme
does not match that of inactivation, but it is shifted to
more oxidative conditions (corresponding to a cystamine/
cysteamine ratio between 4 and 5) (Fig. 3). This suggests
that oxidation of further cysteines is needed for triggering
the conformational changes that modify the proteolytic
pattern. Moreover, it predicts that, when subjected to increasing oxidative conditions, Rubisco will first become
inactive (with little structural change) and afterwards
experience the conformational alterations that render it
sensitive to proteases. It has been proposed that these
features can be relevant in vivo, specially under
100
100
80
90
60
80
40
20
% Large subunit
RuBP carboxylase activity (%)
(Fig. 2B). These features have been observed with enzymes from species ranging from unicellular algae to
higher plants (Garcı́a-Ferris and Moreno, 1993; Marı́nNavarro and Moreno, 2003), and appear to be universal
among green-type eukaryotic Rubiscos. Moreover, both
transitions (inactivation and proteolytic sensitization)
proved to be fully reversible upon reduction with a free
thiol (Garcı́a-Ferris and Moreno, 1993).
An obvious goal to understand the nature of the redox
modulation of Rubisco has been the identification of the
critical residue (or residues) whose oxidation results in the
observed transitions. The fact that redox modulation of
Rubisco is extended among eukaryotic enzymes points to
phylogenetically conserved residues. However, this consideration still leaves a number of candidates. For
example, in C. reinhardtii, nine (out of a total of 16) Cys
residues are highly conserved among green-type eukaryotic Rubiscos. These are cysteines 84, 172, 192, 247, 284,
427, and 459 of the large subunit, and 41 and 83 of the
small subunit. As a first approach, the scanning of the
functional roles of these cysteines was started through
70
0
0,1
1
10
100
CSSC/CSH
Fig. 3. Inactivation (open circles) and proteolytic sensitization (closed
circles) of Rubisco incubated in redox buffers at different disulphide/
thiol ratios. Purified Rubisco was incubated with redox buffers at
different CSSC/CSH ratios, as detailed in Fig. 2A. The residual
carboxylase activity, as a percentage of that of the fully reduced
enzyme, is shown on the left axis. A proteolytic assay with subtilisin
was performed with the treated Rubisco samples as described in Fig.
2B. The amount of intact large subunit remaining after 20 min of
proteolysis, determined by SDS-PAGE analysis, is indicated on the
right axis as a percentage of that obtained from fully reduced Rubisco in
the same conditions. Error bars indicate standard deviation from
triplicates.
Cysteine-mediated redox modulation of Rubisco 1609
senescence or stress processes, which create oxidative
conditions and lead to an enhanced catabolism of Rubisco
(Garcı́a-Ferris and Moreno, 1993; Moreno et al., 1995). In
the following, current knowledge about each of these
redox regulated transitions (inactivation and proteolytic
sensitization) will be reviewed and the evidence of their
possible physiological role will be examined.
Oxidative inactivation of Rubisco
As shown in Fig. 2A, oxidation of Rubisco with cystamine/cysteamine buffers leads to a complete inactivation
of the enzyme. Under the experimental conditions used in
this assay, the activity loss should result from modification of thiol groups to disulphides, either internal ones
(i.e. those bonding two protein cysteines) or mixed
disulphides with cysteamine. Because the transition to the
inactive form takes place around a disulphide/thiol ratio of
1.5, the critical cysteines should have a standard redox
potential similar to (or slightly more oxidizing than) the
cystamine/cysteamine pair. The spinach Cys172 and
Cys459 residues have been identified as the target of the
affinity labelling that causes also Rubisco inactivation
(Schloss et al., 1978). However, the absence of Cys172 in
the C. reinhardtii C172S mutant Rubisco does not shift,
or otherwise alter, the shape of the transition to the
inactive form in redox buffers (Moreno and Spreitzer,
1999) suggesting that this residue is not involved in
disulphide inactivation. Nevertheless, Cys172 and the
proximal Cys192 are thought to be responsible for
inactivation by arsenite (a reagent for vicinal dithiols)
because the C172S mutant is not sensitive to arsenite
(Moreno and Spreitzer, 1999). This further supports the
notion that disulphide inactivation operates through
a mechanism which is different from those due to
chemical modification by a variety of other reagents.
By contrast, mutation of Cys459 and Cys449 (a
moderately conserved residue located within a disulphide
bonding distance from Cys459) affects redox-dependent
inactivation. These mutations do not prevent a substantial
loss of activity upon oxidation, but the enzyme cannot be
completely inactivated even at very high cystamine/
cysteamine ratios (Marı́n-Navarro and Moreno, 2006).
This indicates that these cysteines contribute (among other
residues) to the changes leading to Rubisco oxidative
inactivation. The fact that the percentage of activity
remaining after oxidation is even more noticeable in the
C449S-C459S double mutant (in which both Cys449 and
459 have been replaced by serines) (Marı́n-Navarro and
Moreno, 2006) implies that the contribution of these
cysteines to the inactivation process may be additive and/
or redundant, but not co-operative (as would be expected
if the effect would be derived from establishing a disulphide bridge between these residues). Indeed, a previous
kinetic analysis had indicated that cysteines involved in
inactivation do not appear to engage in internal disulphide
bonding (Garcı́a-Ferris and Moreno, 1993). The possible
physiological advantages of this fact have been discussed
elsewhere (Moreno et al., 1995). Besides, the case of
Cys449 and Cys459 illustrates the feasibility of affecting
the catalytic performance from a considerable distance
(about 2 nm from the Cys449/Cys459 pair to the nearest
active site). Thus, it seems that cysteine oxidation can
induce subtle conformational changes that may propagate
through the Rubisco structure reaching remote regions.
Structural rearrengements that are known to occur as
a result of cysteine oxidation will be described in the next
section.
Conformational changes induced by oxidation
Oxidation of Rubisco has been shown to weaken its
quaternary structure, as demonstrated by the thermal and
solvent sensitivity of the oxidized enzyme (Marı́n-Navarro
and Moreno, 2003). Nevertheless, the oxidized holoenzyme still remains assembled at physiological temperature
in aqueous solutions. An analysis of the proteolysis of
C. reinhardtii Rubisco with subtilisin revealed that, besides
the clipping of the unstructured N-terminus of the large
subunit (which does not affect the structural stability of
the enzyme and occurs in a redox-independent manner),
the oxidized enzyme becomes more sensitive to proteolysis due to an enhanced processing at a specific loop
(Ser61–Thr68) located between the B a-helix and a neighbouring 310-helix within the N-terminal domain of the
large subunit (Marı́n-Navarro and Moreno, 2003). Thus,
oxidation of Rubisco involves conformational changes
that facilitate the access of the protease to the Ser61–Thr68
loop. Moreover, it has been shown that clipping of this
piece (up to Thr68), which remains transiently attached to
the holoenzyme, favours the disassembly of the quaternary structure and the subsequent unrestricted proteolysis
of the released subunits (Marı́n-Navarro and Moreno,
2003). A quantitative analysis of the kinetics of proteolysis suggests that holoenzyme disassembly follows
the proteolytic processing at the Ser61–Thr68 loop of both
of the large subunits integrating any of the four dimers
that scaffold the holoenzyme core (Marı́n-Navarro and
Moreno, 2003). Hence, the clipped N-terminal pieces
apparently remain attached to the holoenzyme only until
the structure is destabilized by a cut at the Ser61–Thr68
loop in one large subunit whose dimerizing neighbour has
already been processed at the same site. Since the
proteolytic cut at the Ser61–Thr68 loop is obstructed in
the reduced enzyme, this mechanism offers a doublecheck redox-dependent path for full proteolytic access to
the otherwise protease-resistant holoenzyme structure of
Rubisco (Fig. 4).
1610 Moreno et al.
The Cys172 residue of the large subunit was the first to
be related to redox conformational changes of Rubisco
because its substitution by serine produces a mutated
enzyme which displays a protease-sensitivity transition
shifted to more oxidative conditions (Moreno and Spreitzer,
1999). Nevertheless, the mutated enzyme eventually
becomes as degradable as the wild type when incubated
at high disulphide/thiol ratios. Thus, the absence of
Cys172 delays the structural changes rendering the
enzyme sensitive to proteolysis until compensating conformational changes are brought about by the oxidation of
other (less reducing) cysteines. The latter residues do apparently fulfil a somewhat redundant role for this transition in the wild-type enzyme. Similarly, the simultaneous
mutation of both Cys449 and Cys459 produces a diminished sensitivity to processing at the Ser61–Thr68 loop,
even at high disulphide–thiol ratios (Marı́n-Navarro and
Moreno, 2006). Therefore, it seems that the oxidation of
Cys449 and Cys459 contribute to the conformational
changes that expose the Ser61–Thr68 loop to proteolytic
Fig. 4. Steps for the proteolytic degradation of Rubisco. (A) Schematic
representation of the N-terminal domain of Rubisco large subunit
highlighting the two regions (site 1 and site 2) that are accessible to
proteolytic attack in vitro. A conformational change between the
reduced (RR) and oxidized (RO) forms can be detected through an
enhanced access of broad specificity proteases to site 2, while
proteolysis at site 1 is not affected by redox conditions. (B) Scheme for
the mechanism of Rubisco degradation in vitro. Rubisco holoenzyme is
composed by four dimers of large subunits (represented by complementary half ovals) and eight small subunits (represented by grey ovals).
The N-terminal end (up to Lys 18) of each large subunit is depicted as
a protruding stretch. The proteolytic cut at site 1, which releases this Nterminal fragment does not compromise the structural stability of the
holoenzyme. Processing at site 2 of the two large subunits integrating
a dimer promotes the disassembly of the holoenzyme and the subsequent unrestricted degradation to small peptides and amino acids of
all the subunits. (C) Rubisco holoenzyme structure highlighting one
dimer of large subunits in blue and green. The N terminal fragment
arising from the cut at site 2, which remains attached to the holoenzyme
core until a double cut occurs at any of the dimers of the same
molecule, is shown in red.
attack. Because no alterations were found in the C449S
and C459S single mutants (Marı́n-Navarro and Moreno,
2006), it might be assumed that Cys449 and Cys459 are
also functionally redundant with regard to this structural
transition.
The role of cysteine-mediated redox modulation
of Rubisco in vivo
While the conditions needed for Rubisco inactivation or
conformational transition may be too oxidizing to be
reached during the daily light/dark redox oscillation, it
has been repeatedly suggested that these oxidative processes may take place in vivo under senescence or stress
scenarios, which are known to trigger a fast catabolism of
Rubisco (Peñarrubia and Moreno, 1990; Mehta et al.,
1992; Garcı́a-Ferris and Moreno, 1993; Moreno and
Spreitzer, 1999; Marı́n-Navarro and Moreno, 2003). There
is indeed ample evidence of the development of strong
oxidative conditions inside the chloroplast during senescence or stress processes, as detected by increase of the
GSSG/GSH ratio (Law et al., 1983), rise of ROS (McRae
and Thompson, 1983), or lipid peroxidation (Mishra and
Singhal, 1992). Besides, during stress processes the
number of titrable cysteine–thiol groups of Rubisco has
been shown to decrease (Garcı́a-Ferris and Moreno, 1994)
and oxidative cross-linking of Rubisco subunits can
be demonstrated comparing reducing and non-reducing
SDS–PAGE electrophoregrams (Garcı́a-Ferris and Moreno,
1994; Marı́n-Navarro and Moreno, 2006). Moreover,
inactivation of Rubisco has been shown to precede its
degradation during senescence of leaves (Wittenbach,
1978; Hall et al., 1978) or isolated chloroplasts (Seftor
and Jensen, 1986), as predicted by Fig. 3. Furthermore,
a Rubisco proteolytic fragment, generated by incubation
with lysates from senescing leaves, has been shown to
have Val69 at its N-terminus (i.e. it has been processed at
the Ser61–Thr68 loop) (Yoshida and Minamikawa, 1996).
It cannot be ignored that other identified fragments produced by endogenous processing have displayed different
N-terminal sequences (Desimone et al., 1996; Kokubun
et al., 2002) or proved to be generated by a different (nonenzymatic) mechanisms (Ishida et al., 1999). As already
stated above, it transpires from the accumulated evidence
that many different paths of Rubisco degradation are
likely to exist, one or several of them prevailing under
specific conditions. Cysteine-dependent oxidative inactivation and proteolytic sensitization may be just one of
these catabolic paths.
Even if consideration of the whole of the above
circumstantial facts is strongly suggestive, the most
compelling evidence for a physiological role of cysteine
oxidation in Rubisco catabolism comes from the analysis
of the phenotype of the C. reinhardtii cysteine mutants
under stress conditions. For instance, Fig. 5 shows the
Cysteine-mediated redox modulation of Rubisco 1611
response of the C172S and C192S mutants to salt stress
compared with the wild-type strain. It can be appreciated
that 10 h of stress with 0.3 M NaCl caused only a slight
decrease of chlorophyll content (Fig. 5A) (indicating little
dismantling of thylakoidal structures) but a noticeable
loss of total protein per cell (Fig. 5C). The amount of
Rubisco in the wild-type strain fell in the same period to
an even lower percentage (Fig. 5B). Thus, the Rubiscoto-total-protein ratio experienced a steady decrease
throughout the treatment (Fig. 5D) showing a preferential
degradation of Rubisco compared with the bulk of other
proteins. However, the C172S mutant, when subjected to
the same stress intensity (as witnessed by the chlorophyll
and total protein loss in Fig. 5), showed a distinctly
slower Rubisco degradation rate (Fig. 5B). Monitoring
of the Rubisco-to-total-protein ratio indicated that the
substitution of Cys172 by serine supressed the preferential degradation of Rubisco in the C172S mutant during
the first 6 h of stress (Fig. 5D). A similar delay of
Rubisco turnover has been reported for the same mutant
under osmotic or oxidative stress (Moreno and Spreitzer,
1999). Interestingly, a C172A mutant (i.e. having Cys172
replaced by alanine) of a cyanobacterial Rubisco has also
been shown to display retarded catabolism under nitrogen
B
100
80
Rubisco (%)
Chlo rophyll (%)
A
deprivation (Marcus et al., 2003). This suggests that the
mechanism for timing of Rubisco degradation that involves
Cys172 may be shared by prokaryotic and eukaryotic
green-type Rubiscos. By contrast, the substitution of the
vicinal Cys192 does not alter the Rubisco degradation rate
in the C. reinhardtii C192S strain (Fig. 5), nor in the
cyanobacterial mutant (Marcus et al., 2003).
The delayed degradation of Rubisco in mutants lacking
Cys172 could be a consequence of the shift of the proteolytic sensitivity of the enzyme to more oxidative conditions as observed in vitro (Moreno and Spreitzer, 1999).
However, in the case of Cys449 and Cys459, the single
substitutions produced no apparent phenotype while the
C449S-C459S double mutant displayed an accelerated
catabolism with enhanced oxidative cross-linking and
association to membranes (Marı́n-Navarro and Moreno,
2006). This agrees with the observation of an altered
conformation of the double mutant (but not of the single
mutants) upon oxidation, as detected in vitro by divergent
kinetics of proteolysis using subtilisin as a structural probe
(Marı́n-Navarro and Moreno, 2006). Nevertheless, there is
an apparent paradox in that the elimination of cysteines
(i.e. of potentially oxidizable thiol groups) results in an
enhanced oxidative catabolism. In that instance, it has to
60
40
20
0
100
80
60
40
20
0
2
4
6
8
0
10
0
2
4
D
100
Rubsico/protein
Total protein (%)
C
80
60
40
20
0
6
8
10
time (h)
time (h)
120
100
80
60
40
20
0
2
4
6
time (h)
8
10
0
0
2
4
6
8
10
time (h)
Fig. 5. Evolution of chlorophyll (A), Rubisco (B), total protein (C), and Rubisco to total protein ratio (D) in C172S (closed circles), C192S (closed
squares), and wild-type (open circles) strains of C. reinhardtii subjected to salt stress. C. reinhardtii cultures of the different strains were grown up to
a density of 23106 cells ml1 before the application of stress (0.3 M NaCl). At the indicated times of stress, an aliquot of the culture was taken to
determine the chlorophyll content, and the amount of total protein and Rubisco as described elsewhere (Marı́n-Navarro and Moreno, 2006). In all
cases, the amounts are given as a percentage of the initial value. Error bars represent standard deviation from 9 (C172S), 15 (C192S) or 21 (wild
type) experiments.
1612 Moreno et al.
be admitted that the oxidation of Cys449 and/or Cys459,
and the derived structural changes, may act to shield or
protect other cysteine residues from further oxidation (Fig.
6). Hence, the Cys449–Cys459 pair could serve to adjust
the rate of Rubisco degradation by delaying further stages
of catabolism.
Concluding remarks
Modulation of Rubisco activity and structure by the redox
state of its conserved cysteine residues appears as a sug-
gestive regulatory mechanism for adapting the enzyme to
the changing functional status of the chloroplast. While
the potential for regulation has always been clear from in
vitro experiments, the actual implication of this mechanism as physiologically relevant in vivo has been difficult
to prove. The definitive evidence has come from the
verification that site-directed mutants on conserved cysteine residues display an abnormal Rubisco catabolism
in vivo. Moreover, the fact that the mutants on highly
conserved cysteines are all photosynthetically active and
show an altered phenotype only under stress conditions
Fig. 6. A model for cysteine-mediated redox modulation of Rubisco catabolism. Under sustained stress, photosynthetic organisms experience
increasing oxidative conditions within the chloroplast stroma (indicated by the gradient of grey shade at the top bar). (A) In wild-type Rubisco, each
cysteine has a characteristic redox potential determined by its structural microenviroment and represented in the figure by a grey-shaded cube, with
a lighter shade intensity correlating with a lower redox potential. As an arbitrarily simplified example, only five critical cysteines (numbered from 1 to
5) are shown. When the redox potential in the chloroplast milieu (r) matches that of a specific cysteine, the originally reduced residue (open circle)
becomes oxidized (closed circles). The oxidation of different cysteines results in specific conformational changes that progressively direct Rubisco to
the catabolic pathway, including inactivation, polymerization, mobilization to a membrane fraction, and degradation. The rate at which each
conformational variant is guided to these routes is represented by an increasing number of vertical arrows. A higher versatility of the stress response
is achieved by a higher number of structural variants, and this gradual modification may be tuned by cysteine networking. In the figure, the oxidation
of both Cys 1 and 2 causes a conformational change that converts Cys 3 into a less reducing thiol. Besides, Cys4 and Cys 5 are assumed to be
functionally redundant because the oxidation of Cys 5 alone is supposed to reproduce (among other changes) the conformational distortions caused
by the oxidation of Cys 4. Replacement of cysteines by serines in (B) and (C) is represented by switching from circles to triangles. (B) Double
mutation of Cys 1 and 2, which have a redundant role (these would be exemplified by Cys449 and Cys459 from Rubisco large subunit), causes an
acceleration of the catabolic processes because the critical Cys 3 is now oxidized at a lower redox potential. (C) Mutation of Cys 4 retards catabolism
(this would be the case of Cys172 from Rubisco large subunit) because completion of the conformational transition now has to wait for the oxidation
of the least reducing Cys 5.
Cysteine-mediated redox modulation of Rubisco 1613
further reinforces the idea that these residues are conserved because of their redox properties which are
important for regulating Rubisco degradation. However,
redox modulation of Rubisco through its cysteines appears
as an intricate process where several residues seem to
contribute in a additive, co-operative or redundant manner. The set of critical cysteines seems to act as a network
of redox sensors that drives the progression of Rubisco
through the different stages of catabolism by means of
a sequence of structural changes taking place under
increasingly oxidative conditions (Fig. 6). Each of these
conformational changes is induced by the oxidation of
specific cysteine residues and seems to act as a checkpoint
to ensure an adequate timing for the different steps of
Rubisco catabolism within the general framework of
chloroplast dismantling following acute stress or senescence conditions.
Acknowledgements
We thank Dr Robert J Spreitzer (University of Nebraska at Lincoln,
USA) for advice and support in obtaining the cysteine site-directed
mutants. Work in our laboratory described here has been supported
by grants BMC2003-03209 and BFU2006-07783 from the Ministerio de Ciencia y Tecnologı́a (MCyT) and QLK3-CT-2002-01945
from the European Commission, and by fellowships from the
MCyT (to MJG-M) and from the Generalitat Valenciana (to JM-N).
References
Albuquerque J, Esquı́vel MG, Teixeira AR, Ferreira RB. 2001.
The catabolism of ribulose bisphosphate carboxylase from higher
plants. A hypothesis. Plant Science 161, 55–65.
Alscher RG, Erturk N, Heath LS. 2002. Role of superoxide
dismutases (SODs) in controlling oxidative stress in plants.
Journal of Experimental Botany 53, 1331–1341.
Andersson I, Knight S, Schneider G, Lindqvist Y, Lindqvist T,
Branden CI, Lorimer GH. 1989. Structure of the active site of
ribulose-bisphosphate carboxylase. Nature 337, 229–234.
Apel K, Hirt H. 2004. Reactive oxygen species: metabolism,
oxidative stress, and signal transduction. Annual Review of Plant
Biology 55, 373–399.
Buchanan BB, Balmer Y. 2005. Redox regulation: a broadening
horizon. Annual Review of Plant Biology 56, 187–220.
Davison PA, Hunter CN, Norton P. 2002. Overexpression of betacarotene hydroxylase enhances stress tolerance in Arabidopsis.
Nature 418, 203–206.
Desimone M, Henke A, Wagner E. 1996. Oxidative stress Induces
partial degradation of the large subunit of ribulose-1,5-bisphosphate
carboxylase/oxygenase in isolated chloroplasts of barley. Plant
Physiology 111, 789–796.
Dietz KJ, Jacob S, Oelze ML, Laxa M, Tognetti V, Nunes de
Miranda SM, Baier M, Finkemeier I. 2006. The function of
peroxiredoxins in plant organelle redox metabolism. Journal of
Experimental Botany 57, 1697–1709.
Eckardt NA, Pell EJ. 1995. Oxidative modification of Rubisco
from potato foliage in response to ozone. Plant Physiology and
Biochemistry 33, 273–282.
Evans JR, Seeman JR. 1989. The allocation of protein nitrogen
in the photosynthetic apparatus: costs, consequences and control.
In: Briggs WR, ed. Photosynthesis. New York: Alan R. Liss Inc.,
183–205.
Ferreira RB, Esquı́vel MG, Teixeira AR. 2000. Catabolism of
ribulose bisphosphate carboxylase from higher plants, Current
Topics in Phytochemistry. 3, 129–165.
Ferreira RB, Davies DD. 1989. Conversion of ribulose-1,
5-bisphosphate carboxylase to an acidic and catalytically inactive
form by extracts of osmotically stressed Lemna minor fronds.
Planta 179, 448–455.
Ferreira RB, Shaw NM. 1989. Effect of osmotic stress on protein
turnover in Lemna minor fronds. Planta 179, 456–465.
Garcı́a-Ferris C, Moreno J. 1993. Redox regulation of enzymatic
activity and proteolytic and susceptibility of ribulose-1,5bisphosphate carboxylase/oxygenase from Euglena gracilis.
Photosynthesis Research 35, 55–66.
Garcı́a-Ferris C, Moreno J. 1994. Oxidative modification and
breakdown of ribulose 1,5-bisphosphate carboxylase/oxygenase
induced in Euglena gracilis by nitrogen starvation. Planta 193,
208–215.
Gilbert HF. 1982. Biological disulfides: The third messenger?
Journal of Biological Chemistry 257, 12086–12091.
Hall NP, Keys AJ, Merret MJ. 1978. Ribulose-1,5-diphosphate
carboxylase protein during flag leaf senescence. Journal of
Experimental Botany 29, 31–37.
Hancock J, Desikan R, Harrison J, Bright J, Hooley R, Neill S.
2006. Doing the unexpected: proteins involved in hydrogen peroxide
perception. Journal of Experimental Botany 57, 1711–1718.
Hörtensteiner S, Feller U. 2005. Nitrogen metabolism and
remobilization during senescence. Journal of Experimental
Botany 53, 927–937.
Ishida H, Shimizu S, Makino A, Mae T. 1998. Light-dependent
fragmentation of the large subunit of ribulose-1,5-bisphosphate
carboxylase/oxygenase in chloroplasts isolated from wheat leaves.
Planta 204, 305–309.
Ishida H, Makino A, Mae T. 1999. Fragmentation of the large
subunit of ribulose-1,5-bisphosphate carboxylase by reactive
oxygen species occurs near Gly-329. Journal of Biological
Chemistry 274, 5222–5226.
Junqua M, Biolley JP, Pie S, Kanoun M, Duran R, Goulas P.
2000. In vivo occurrence of carbonyl residues in Phaseolus
vulgaris proteins as a direct consequence of a chronic ozone
stress. Plant Physiology and Biochemistry 38, 853–861.
Krieger-Liszkay A, Trebst A. 2006. Tocopherol is the scavenger
of singlet oxygen produced by the triplet states of chlorophyll in
the PSII reaction center. Journal of Experimental Botany 57,
1677–1684.
Kokubun N, Ishida H, Makino A, Mae T. 2002. The degradation
of the large subunit of ribulose-1,5-bisphosphate carboxylase/
oxygenase into the 44 kDa fragment in the lysates of chloroplasts incubated in darkness. Plant and Cell Physiology 43,
1390–1395.
Law MY, Charles SA, Halliwell B. 1983. Glutathione and
ascorbic acid in spinach (Spinacia oleracea) chloroplasts. The
effect of hydrogen peroxide and of paraquat. Biochemical Journal
210, 899–903.
Marcus Y, Altman-Gueta H, Finkler A, Gurevitz M. 2003. Dual
role of cysteine 172 in redox regulation of ribulose 1,5bisphosphate carboxylase/oxygenase activity and degradation.
Journal of Bacteriology 185, 1509–1517.
Marı́n-Navarro J, Moreno J. 2003. Modification of the proteolytic fragmentation pattern upon oxidation of cysteines from
ribulose 1,5-bisphosphate carboxylase/oxygenase. Biochemistry
42, 14930–14938.
Marin-Navarro J, Moreno J. 2006. Cysteines 449 and 459
modulate the reduction–oxidation conformational changes of
1614 Moreno et al.
ribulose 1,5-bisphosphate carboxylase/oxygenase and the translocation of the enzyme to membranes during stress. Plant, Cell
and Environment 29, 898–908.
McRae DG, Thompson JE. 1983. Senescence-dependent changes
in superoxide anion production by illuminated chloroplasts from
bean leaves. Planta 158, 185–193.
Mehta RA, Fawcett TW, Porath D, Mattoo AK. 1992. Oxidative
stress causes rapid membrane translocation and in vivo degradation of ribulose-1,5-bisphosphate carboxylase/oxygenase. Journal
of Biological Chemistry 267, 2810–2816.
Mishra RK, Singhal GS. 1992. Function of photosynthetic
apparatus of intact wheat leaves under high light and heat stress
and Its relationship with peroxidation of thylakoid lipids. Plant
Physiology 98, 1–6.
Moreno J, Peñarrubia L, Garcia-Ferris C. 1995. The mechanism
of redox regulation of ribulose-1,5-bisphosphate carboxylase/
oxygenase turnover. A hypothesis. Plant Physiology and Biochemistry 33, 121–127.
Moreno J, Spreitzer RJ. 1999. C172S substitution in the
chloroplast-encoded large subunit affects stability and stressinduced turnover of ribulose-1, 5-bisphosphate carboxylase/oxygenase’. Journal of Biological Chemistry 274, 26789–26793.
Motohashi K, Kondoh A, Stumpp MT, Hisabori T. 2001.
Comprehensive survey of proteins targeted by chloroplast
thioredoxin. Proceedings of the National Academy of Sciences,
USA 98, 11224–11229.
Nakano R, Ishida H, Makino A, Mae T. 2006. In vivo
fragmentation of the large subunit of ribulose-1,5-bisphosphate
carboxylase by reactive oxygen species in an intact leaf of
cucumber under chilling-light conditions. Plant and Cell Physiology 47, 270–276.
Noctor G, Foyer CH. 1998. Ascorbate and glutathione: keeping
active oxygen under control. Annual Review of Plant Physiology
and Plant Molecular Biology 49, 249–279.
Peñarrubia L, Moreno J. 1990. Increased susceptibility of
ribulose-1,5-bisphosphate carboxylase/oxygenase to proteolytic
degradation caused by oxidative treatments. Archives of Biochemistry and Biophysics 281, 319–323.
Pfannschmidt T, Nilsson A, Allen JF. 1999. Photosynthetic
control of chloroplast gene expression. Nature 397, 635–628.
Rey P, Bécuwe N, Barrault MB, Rumeau D, Havaux M,
Biteau B, Toledano MB. 2007. The Arabidopsis thaliana
sulfiredoxin is a plastidic cysteine–sulfinic acid reductase involved in the photooxidative stress response. The Plant Journal
49, 505–514.
Salvador ML, Klein U. 1999. The redox state regulates RNA
degradation in the chloroplast of Chlamydomonas reinhardtii.
Plant Physiology 121, 1367–1374.
Schloss JV, Stringer CD, Hartman FC. 1978. Identification of
essential lysyl and cysteinyl residues in spinach ribulosebisphosphate carboxylase/oxygenase modified by the affinity
label N-bromoacetylethanolamine phosphate. Journal of Biological Chemistry 253, 5707–5711.
Schürmann P, Jacquot JP. 2000. Plant thioredoxin systems
revisited. Annual Review of Plant Physiology and Plant Molecular Biology 51, 371–400.
Seftor RE, Jensen RG. 1986. Causes for the disappearance of
photosynthetic CO2 fixation with isolated spinach chloroplasts.
Plant Physiology 81, 81–85.
Shigeoka S, Ishikawa T, Tamoi M, Miyagawa Y, Takeda T,
Yabuta Y, Yoshimura K. 2002. Regulation and function of
ascorbate peroxidase isoenzymes. Journal of Experimental Botany 53, 1305–1319.
Stadtman ER. 1990. Covalent modification reactions are marking
steps in protein turnover. Journal of Biological Chemistry 29,
6323–6331.
Sugiyama T, Nakayama N, Ogawa M, Akazawa T, Oda T. 1968.
Structure and function of chloroplast proteins. II. Effect of pchloromercuribenzoate treatment of the ribulose 1,5-diphosphate
carboxylase activity of spinach leaf fraction I protein. Archives of
Biochemistry and Biophysics 125, 98–106.
Tenaud M, Jacquot JP. 1987. In vitro thiol-dependent redox
regulation of purified ribulose-1,5-bisphosphate carboxylase/oxygenase. Journal of Plant Physiology 130, 315–326.
Wittenbach V. 1978. Breakdown of ribulose bisphosphate carboxylase and change in proteolytic activity during dark-induced
senescence of wheat seedlings. Plant Physiology 62, 604–608.
Yosef I, Irihimovitch V, Knopf JA, Cohen I, Orr-Dahan I,
Nahum E, Keasar C, Shapira M. 2004. RNA binding activity
of the ribulose-1,5-bisphosphate carboxylase/oxygenase large
subunit from Chlamydomonas reinhardtii. Journal of Biological
Chemistry 279, 10148–10156.
Yoshida T, Minamikawa T. 1996. Successive amino-terminal
proteolysis of the large subunit of ribulose 1,5-biphosphate
carboxylase/oxygenase by vacuolar enzymes from French bean
leaves. European Journal of Biochemistry 238, 317–324.
Zhang N, Kallis RP, Ewy RG, Portis Jr AR. 2002. Light
modulation of Rubisco in Arabidopsis requires a capacity for
redox regulation of the larger Rubisco activase isoform. Proceedings of the National Academy of Sciences, USA 99, 3330–3334.