-
Chapter 2The Calvin Cycle and Its Regulation
William MartinInstitut fr Genetik, Technische Universitt
Braunschweig,
Spielmannstr. 7, D-38023 Braunschweig, Germany
Renate ScheibePflanzenphysiologie, FB 5 Biologie/Chemie,
Universitt Osnabrck,
D-49069 Osnabrck, Germany
Claus SchnarrenbergerInstitut fr Pflanzenphysiologie und
Mikrobiologie, Freie Universitt Berlin,
Knigin-Luise-Str. 12-16a, D-14195 Berlin, Germany
SummaryI.II.
IntroductionThe Enzymes of the Calvin Cycle
A.B.C.D.E.F.G.H.
Ribulose-1,5-bisphosphate Carboxylase/oxygenasePhosphoglycerate
KinaseGlyceraldehyde-3-phosphate DehydrogenaseTriosephosphate
IsomeraseFructose-1,6-bisphosphate/Sedoheptulose-1,7-bisphosphate
AldolaseFructose-1,6-bisphosphataseSedoheptulose-1,7-bisphosphataseTransketolase
I.J.K.
Ribulose-5-phosphate 3-epimeraseRibose-5-phosphate
IsomerasePhosphoribulokinase
III.IV.
V.VI.
Calvin Cycle Gene Organization, Expression, and Regulation in
EubacteriaCalvin Cycle Expression in Plants
A.B.C.D.E.
Quantification of ActivitiesExpression Studies of Enzyme
Activities and TranscriptionGene Regulation Through High Sugar
Sensing, and Redox StateRegulation in Specific SystemsCalvin Cycle
Enzymes and Expression in Euglena gracilis
Enzyme Interactions and Multienzyme-like ComplexesBiochemical
Regulation in Chloroplasts
The Ferredoxin/Thioredoxin SystemA.B.C.
Target EnzymesPhysiological Consequences
VII.VIII.
Studies of Calvin Cycle Enzymes with Antisense RNAConcluding
Remarks
AcknowledgmentReferences
R. C. Leegood, T. D. Sharkey and S. von Caemmerer (eds),
Photosynthesis: Physiology and Metabolism, pp. 951. 2000 Kluwer
AcademicPublishers. Printed in The Netherlands.
101012121414151516171818191919202021222327283131323334363636
-
10
Summary
The Calvin cycle is the starting point ofcarbon metabolism in
higher plants. It is a typically eubacterial pathway,as comparative
biochemistry of all of its enzymes from prokaryotes and eukaryotes
has revealed. The structuralbasis of Calvin cycle function is
reviewed with an attempt at a balanced consideration of biochemical
andmolecular findings. The structural diversity ofprokaryotic
enzymes is emphasized, since the genes encoding thepathway in
eukaryotes have all been inherited by plants from prokaryotes
through endosymbiosis. Curiously,the enzymes that constitute the
pathway in different organisms are often structurally unrelatedwhat
isconserved in evolution is merely the set of substrate
conversions, not the enzymes that catalyze them. Some ofthe
structural and regulatory properties of the enzymes were present in
the antecedents oforganelles, but otherswere newly acquired at the
eukaryotic level. The expression of Calvin cycle genes is regulated
by a widespectrum of factors, though the molecular details of the
regulation have yet to be unraveled. Findings thatsuggest the
existence of multienzyme-like Calvin cycle complexes are
summarized. The molecular basis ofredox-modulated light regulation
through the thioredoxin system and its importance for flexible
control of thepathway under varying conditions is illustrated.
Expression of Calvin cycle enzymes in response to external
orinternal stimuli is briefly reviewed, as are newer findings from
the expression of antisense constructs of Calvincycle enzymes in
transgenic plants.
I. Introduction
The Calvin cycle is one of four known pathways offixation in
nature, the three other pathways
being the reverse (or reductive) citric acid cycle(Evans et al.,
1966; Beh et al., 1993; Schnheit andSchfer, 1995), the reductive
acetyl-CoA (or Wood-Lungdahl) pathway (Fuchs and Stupperich,
1986;Ragsdale, 1991; Schnheit and Schfer, 1995), andthe recently
discovered 3-hydroxypropionate pathway(Strauss and Fuchs,
1993;Ishii et al., 1996). However,the Calvin cycle is the only
pathway of fixationknown to occur in plants (Fig. 1). It therefore
figuresprominently in plant biochemistry, albeit undervarious
acronyms, among them the reductive pentose
Abbreviations: 1,3BPGA - 1,3 bisphosphoglycerate; 3PGA
3-phosphoglycerate; Aldolase fructose- 1,6-bisphosphate
aldolase;CBB Calvin-Benson-Bassham; cpDNA chloroplast DNA;CTE
C-terminal extension; DHAP dihydroxyacetonephosphate; DTT
dithiothreitol; E4P erythrose 4-phosphate;F1,6BP fructose
1,6-bisphosphate; F6P fructose 6-phosphate;FBPase
fructose-1,6-bisphosphatase; FTR ferredoxin/thioredoxin reductase;
GA3P glyceraldehyde 3-phosphate;GAPDH glyceraldehyde 3-phosphate
dehydrogenase; MDH malate dehydrogenase; PGK phosphoglycerate
kinase;inorganic phosphate; PRK phosphoribulose kinase;
R5Pribose-5-phosphate; RPE ribulose-5-phosphate 3-epimerase;RPI
ribose 5-phosphate isomerase; Ru1,5BP ribulose 1,5-bisphosphate;
Ru5P ribulose-5-phosphate;Rubisco ribulose-1,5-bisphosphate
carboxylase/oxygenase; SBPase sedohep-tulose-l,7-bisphosphatase;
Su1,7BP sedoheptulose 1,7-bisphosphate; Su7P sedoheptulose
7-phosphate; Td thioredoxin; TKL transketolase; TPI
triosephosphateisomerase; Xu5P xylulose 5-phosphate
phosphate pathway (RPPP), the photosyntheticcarbon reduction
(PCR) cycle, the Calvin-Benson-Bassham (CBB) pathway, the
Benson-Calvin cycle,the C3 cycle, and so on. The enzymes of the
Calvincycle have been previously reviewed by Latzko andKelly
(1979), Robinson and Walker (1981) andLeegood (1990). Regulation of
the Calvin cycle hasbeen reviewed by Buchanan (1980), Macdonald
andBuchanan (1990), Geiger and Servaites (1994) and,in
cyanobacteria, by Tabita (1994). Historicaldevelopments surrounding
the elucidation of thepathway have been briefly summarized
elsewhere(Schnarrenberger and Martin, 1997).
In the Calvin cycle, ATP and NADPH from thelight reactions of
the photosynthetic membrane areexpended to reduce to carbohydrate.
From thestandpoint of ATP investment per mole of fixed,the Calvin
cycle is the most costly of the fourfixation pathways known
(Strauss and Fuchs, 1993).The basics of the pathway were clarified
throughtracer studies in eukaryotic algae over 40 years ago(Calvin,
1956). The net reaction can be summarizedas
Mutants defective in fixation in the facul-tatively anaerobic,
chemoautotrophic proteobacteriaRhodobacter sphaeroides (Gibson and
Tabita, 1996)and Ralstonia eutropha (previously
Alcaligeneseutrophus) (Kusian and Bowien, 1997) have been a
William Martin, Renate Scheibe and Claus Schnarrenberger
-
11
powerful tool for understanding the molecular biologyand the
genetic regulation of the pathway in theseorganisms. Molecular
sequences are known for all ofthe enzymes of the pathway from
spinach chloroplasts(Martin and Schnarrenberger, 1997) and from
thegenome sequence of the cyanobacterium Synecho-cystis PCC6803
(Kaneko et al., 1996) and, with a fewexceptions, from Rhodobacter
sphaeroides (Gibsonand Tabita, 1997) and Ralstonia eutropha
(Bommeret al., 1996). Several lines of reasoning support theview
that in order to understand the Calvin cycle ofhigher plant
chloroplasts in a broader context, it isuseful to consider
regulation and structural diversitywithin the pathway among
eubacteria.
First, the pathway did not evolve de novo in plants,but rather
was inherited from eubacteria via theendosymbiotic origins of
organelles. As a conse-quence, manybut not allof the
regulatoryproperties that are observed among the enzymes ofthe
higher plant pathway arose at the prokaryoticlevel and were simply
maintained within the plant
lineage, having been genetically transmitted fromthe
cyanobacterial antecedents of plastids. Theenzymes of the pathway
in chloroplasts are not allacquisitions from cyanobacteria, some
are acquisi-tions from mitochondria (Martin and Schnar-renberger,
1997; Martin and Mller, 1998) that werererouted during evolution to
a new target organelle(Martin and Herrmann, 1998).
Second, in most plastids, at least one enzyme ofthe pathway (one
or both subunits of Rubisco) is stillencoded in chloroplast DNA
(cpDNA), establishinga requirement for coordination of gene
expressionbetween plastids and the nucleus in order to
properlyexpress the pathway. Indeed, the plastids of somealgae even
still possess the cbbR gene (Stoebe et al.,1998) which encodes a
homologue of the trans-criptional regulator of Calvin cycle gene
expressionin eubacteria.
Third, the quantitatively most important mechan-ism governing
the activity of higher plant Calvincycle enzymeslight activation
via the thioredoxin
Chapter 2 Calvin Cycle
-
12
systemis present and active in cyanobacteria. Thisregulatory
mechanism was also inherited by plantsfrom the eubacterial
antecedents of plastids, althoughthe molecularbasis for the
covalent transitions in thetarget enzymes can differ between
cyanobacteria andplants.
Fourth, although the mechanisms of Calvin cyclegene regulation
in eubacteria are probably much lesscomplex than those in
eukaryotes, by no means arethey irrelevant to our understanding of
eukaryoticCalvin cycle gene regulation. On the contrary, due
totheir simplicity and tractability, mechanisms ofCalvincycle gene
regulation are much better understood ineubacteria than in
eukaryotes. And with the recentdiscovery of a cyanobacterial
homologue ofphytochrome (Hughes et al., 1997; Yeh et al., 1997),it
appears that at least some of the basic machineryfor Calvin cycle
gene regulation through light ineukaryotes were simply inherited
from prokaryotesthrough endosymbiosis, although the actual
signaltransduction pathways in prokaryotes and eukaryotesthat lead
to gene regulation through light will, inmany cases, turn out to be
quite different.
Finally, although the series ofsubstrate conversionsthat
constitute the Calvin cycle are strictly conservedacross eubacteria
and eukaryotes, the same degree ofconservation does not apply to
the enzymes thatcatalyze those reactions. In fact, in this chapter
wewill see that the pathway in proteobacteria,cyanobacteria and
higher plant chloroplasts consistsof enzymes that catalyze
identical reactions, but, insome cases, that are altogether
unrelated at the levelof sequence, structure and reaction mechanism
(seeMartin and Schnarrenberger, 1997).
Many of the subsequent chapters in this volumedeal, in one way
or another, with various aspects ofthe Calvin cycle, including
Rubisco itself (Chapters3, (Roy and Andrews) and 4 (von Caemmerer
andQuick)), metabolite transport (Chapter 6, Flgge),C4 metabolism
(Chapters 18 (Furbank et al.) and 19(Leegood)) and
chloroplast-cytosol interactions(Chapters 7 (Aiken et al.) and 8
(Foyer et al.)). In thischapter, we will focus on structural,
functional andregulatory aspects of the enzymes that constitute
thepathway, emphasizing insights provided by molecularapproaches,
but considering classical biochemicalaspects as well.
II. The Enzymes of the Calvin Cycle
A schematic comparison of Calvin cycle enzymes in
the Ralstonia eutropha (formerlyAlcaligenes eutrophus) (Bowien
et al., 1993) andthose encoded in the genome of the
cyanobacteriumSynechocystis PCC6803 (Kaneko et al., 1996a,1996b)
reveals that pathways in these bacteriacomprise the same sets of
substrate conversions, butin several cases with the help of enzymes
that arenon-homologousor very nearly so (Fig. 2). Suchstructurally
distinct but functionally homologousenzymes are traditionally
designated as class I/classII enzymes, a term that will be used
here. Differencesalso exist between the pathways in
spinachchloroplasts and Synechocystis (e.g. use of class I vs.class
II aldolase, respectively), but as depicted inFig. 2, these
differences are less grave than across thetwo eubacteria compared.
The following sectionsprovide a synopsis of structural and
functionaldiversity for each Calvin cycle enzyme. Regulationof
individual enzymes by covalent modificationthrough the
ferredoxin/thioredoxin system (recentlyreviewed by Jacquot et al.,
1997b) will be consideredlater in this chapter.
A. Ribulose-1,5-bisphosphate Carboxylase/oxygenase
Ribulose-1,5-bisphosphate carboxylase/oxygenase(EC 4.1.1.39,
Rubisco) catalyzes the initialfixation step. The mechanism involves
an activatingcarbamylation reaction between and thegroup of an
active site lysine residue in the largesubunit (Lorimer and
Miziorko, 1980). Car-bamylation is promoted by Rubisco activase
(Portis,1992). For details of Rubisco kinetics, catalyticmechanism
and regulation, see Chapters 3 (Roy andAndrews) and 4 (von
Caemmerer and Quick). Thecrystal structure ofRubisco from spinach
chloroplastsis known at great resolution, it is a stout
cylindricaltetramer of dimers that are glued together byfour small
subunits at each end (Shibata et al., 1996;Andersson, 1996). Two
structurally distinct Rubiscoenzymes are known. Class I (or form I)
Rubisco hasa native of about 560 kDa and consists of eightlarge
subunits (LSU, ~55 kDa each) and eightsmall subunits (SSU, ~ 15 kDa
each) comprisingthe holoenzyme. Assembly of the hetero-hexadecamer
requires the aid of chaperonins in bothchloroplasts and eubacteria
(Goloubinoff et al., 1989;Gatenby and Viitanen, 1994; Gutteridge
and Gatenby,1995).
William Martin, Renate Scheibe and Claus Schnarrenberger
-
13
Class II Rubisco consists only of large subunits of~55 each
holoenzyme) (Gibson and Tabita,
1996; Kusian and Bowien, 1997). The class I andclass II large
subunits share about 30% amino acididentity, indicating that they
share a common ancestor.Rubisco gene diversity is a complicated
matter andhas been discussed in detail elsewhere (Watson andTabita,
1996; Martin and Schnarrenberger, 1997). Atleast two very ancient
gene duplications (or lateraltransfers) have occurred in Rubisco
evolution, onethat gave rise to the class I and class II enzymes
anda second that gave rise to the two distinct families ofclass I
Rubisco found in chlorophytic (green or G-type Rubisco) and
rhodophytic plastids (red or R-type Rubisco), respectively (Martin
and Schnar-renberger, 1997). Cyanobacteria and many
andproteobacteria studied to date possess G-type Rubisco(Watson
andTabita, 1996), whereas the proteobacteriaRalstonia, Rhodobacter
and Xanthoflavus encode R-
type Rubisco in their cbb (Calvin-Benson-Bassham)operons. But
Rhodobacter, like Rhodospirillum,Hydrogenovibrio and several other
eubacteria, alsoencode and express the class II enzyme (Falcone
andTabita, 1991; Stoner and Shively, 1993; Gibson andTabita, 1996).
Notably, the curious Rubisco gene inthe Methanococcus genome
encodes an active, O2-sensitive enzyme (Watson et al., 1999).
Higher plants, like cyanobacteria, possess G-typeclass I
Rubisco, whereby in all species studied, thelarge subunit is
encoded as a single copy in thecpDNA and the small subunit is
encoded in thenucleus, usually as a gene family (see below).
Theprimitive photosynthetic protist Cyanophoraparadoxa is an
exception, however, in that both thelarge and the small subunits of
its G-type Rubiscoare encoded in the cpDNA (Lambert et al.,
1985).Rhodophytes and photosynthetic protists that haveobtained
their plastids from rhodophytes via
Chapter 2 Calvin Cycle
-
14
secondary endosymbiosis (McFadden et al., 1996;Van de Peer et
al., 1996) encode both subunits of R-type Rubisco in their cpDNA,
and also encode ahomologue of cbbR, the transcriptional regulator
ofCalvin cycle operons in proteobacteria. The onlyexamples in which
eukaryotes have been shown topossess class II Rubisco have been
described for thevery diverse group of photosynthetic protists
ofsecondary symbiotic origin known as dinoflagellates(Morse et al.,
1995; Rowan et al, 1996), where, quitesurprisingly, the gene for
the class II Rubisco largesubunit is encoded in the nucleus. The
diversity ofeukaryotic Rubisco genes is, in all likelihood,
simplythe result of sampling from ancient eubacterial genediversity
present in the common ancestor ofendosymbiotic organelles, very
similar to allelesampling in population genetics, but on a
geologicaltime scale (Martin and Schnarrenberger, 1997).
B. Phosphoglycerate Kinase
Phosphoglycerate kinase (EC 2.7.2.3, PGK) catalyzesthe
reversible transfer of the of ATP tothe carboxyl group of
3-phosphoglycerate (3PGA),forming 1,3-bisphosphoglycerate (1,3
BPGA) for thesubsequent reduction step. In all prokaryotic
andeukaryotic sources studied to date, the active enzymeis a
monomer with an of ~44 kDa (Fothergill-Gilmore and Michels, 1993).
The crystal structure ofthe enzyme from several sources is known.
PGK isunusual in that substrate binding induces a
dramaticconformational change: the two wings of thebutterfly
structure are bent upon 3 PGA and ATPbinding by over 30 degrees,
displacing distal regionsof the domains by some 27 (Bernstein et
al., 1997).The chloroplast and cytosolic isoenzymes can beseparated
with conventional techniques, roughly 90%of the PGK activity is
localized in higher plantchloroplasts (Pacold and Anderson, 1975;
Kpke-Secundo et al., 1990; McMorrow and Bradbeer,1990). In
Chlamydomonas reinhardtii, a cytosolicisoenzyme seems to be lacking
(Schnarrenberger etal., 1990; Kitayama and Togasaki, 1995).
ChloroplastPGK from various sources shows biphasic kineticswith
of~400 and of~500 atlow substrate concentrations, with a pH
optimumaround 7.5 (Kpke-Secundo et al., 1990). The enzymehas not
been found to be strongly regulated byallosteric effectors or by
light (Leegood, 1990). The
enzyme has been cloned from several higher plants(Longstaff et
al., 1989; Bertsch et al. 1993) and wasmapped in wheat (Chao et
al., 1989). The higherplant nuclear genes for both the chloroplast
and thecytosolic enzymes were obtained from cyanobacteriathrough
endosymbiotic gene transfer (Brinkmannand Martin, 1996; Martin and
Schnarrenberger,1997).
C. Glyceraldehyde-3-phosphateDehydrogenase
Glyceraldehyde-3-phosphate dehydrogenase (EC1.2.1.13, GAPDH)
catalyzes thereversible reductive step of the Calvin cycle.
Incatalysis, 1,3BPGA forms a highly reactive thioesterbond with the
thiol moiety of the active site cysteineresidue under elimination
of the acylphosphate. The carbonyl group of the covalentlybound
intermediate is reduced to a hemithioacetal byhydride transfer from
NADPH and glyceraldehyde3-phosphate (GA3P) is released from the
enzymethrough cleavage of the hemithioacetal bond (Brndenand
Eklund, 1980). ChloroplastGAPDH has activity with both NAD(H)
andNADP(H). Although the of the enzyme istoo low to be relevant
during (anabolic) fixation,NADPH being strongly preferred by the
enzyme(Cerff, 197 8a), recent studies indicate that theactivity may
play an important role during (catabolic)ATP-synthesis in the dark
(Backhausen et al., 1998).The kinetic properties are complex and
depend uponthe activation state of the enzyme (Cerff,
1978a;Wolosiuk and Buchanan, 1978). In the forwardreaction the
fully active purified enzyme has a
of roughly 30 and a of 40(Baalmann et al., 1995). For the
reaction in intactchloroplasts, a of1 has been estimated(Fridlyand
et al., 1997).
Class I and class II GAPDH enzymes are knownthat share only
1520% sequence identity, both aretetramers of ~150 kDa, consisting
of ~37 kDasubunits. Eukaryotes, eubacteria, and one
halophilicarchaebacterium possess class I GAPDH (Pr etal., 1994;
Brinkmann and Martin, 1996). The tertiarystructure of class I GAPDH
is known for numeroussources (Biesecker et al., 1977; Michels et
al., 1996).Class II GAPDH has been found only in archae-bacteria
(Hensel et al., 1987; Fabry and Hensel,
William Martin, Renate Scheibe and Claus Schnarrenberger
-
15
1988; Zwickl et al., 1990). No crystal structures havebeen
published for the class II enzyme.
Higher plantCalvin cycle GAPDH differs from allother known GAPDH
enzymes in that it is anheterotetramer rather than a homotetramer
(Cerffand Chambers, 1979; Ferri et al., 1990; Scagliarini etal.,
1993).The tetrameric enzyme can reversiblyaggregate to a multimeric
form of 600 to 800 kDa,being less active than the dissociated form
(Cerff,1978a, 1978b; Pupillo and Faggiani, 1979; Wara-Aswapati et
al., 1980; Trost et al., 1993). Thismultimeric form was probably
the form ofthe enzymefirst purified from plants (Yonuschot et al.,
1970).Cytosolic GAPDH (EC 1.2.1.12) and non-phos-phorylating GAPDH
(EC 1.2.1.9), although tetra-mers, do not show this reversible
oligomer formation(Pupillo and Faggiani, 1979). No strong
allostericeffectors are known for higher plant GAPDH, but inthe
cyanobacterium Synechocystis PCC6803 a lowMW fraction has been
described that reduces theactivity of Calvin cycle GAPDH in the
reverse(oxidative) direction (Koksharova et al., 1998).
A novel, plastid-specific GAPDH(GapCp) was recently described
from Pinuschloroplasts that coexists with GAPDH of theCalvin cycle.
It shows no detectable activity with
and has a of 62 pM and of344 (Meyer-Gauen et al., 1994;
Meyer-Gauen etal., 1998). GapCp from Pinus is possibly similar
tothe reported from isolated, non-photosynthetic plastids of
developing cauliflowerbuds (Neuhaus et al., 1993). There is also
biochemicalevidence for a similar plastid GAPDHin ripening sweet
pepper fruits where, in cooperationwith an MDH, it appears to
beimportant in the distribution of reducing equivalentsbetween
plastidand cytosol (Backhausen et al., 1998).In some photosynthetic
tissues, for example in pineseedlings, GapCp (an enzyme)appears to
coexist with Calvin cycle GAPDH
(Meyer-Gauen et al., 1994;Schnarrenberger, unpublished). In some
non-photosynthetic tissues, GapCp may functionallyreplace the
enzyme. By analogy, in somephotosynthetic protists, an GAPDHenzyme
has been recruited from anancestral enzyme (Liaud et al., 1997;
Fagan et al.,1998). The nuclear gene for higher plant Calvincycle
GAPDH was obtained by plants fromcyanobacteria, the cytosolic
enzyme appears to havebeen obtained from the mitochondrial
symbiont
Triosephosphate isomerase (EC 5.3.1.1, TPI)catalyzes the rapid
and reversible ketose-aldoseisomerization of dihydroxyacetone
phosphate(DHAP) and GA3P. The native enzyme in eubacteriaand
eukaryotes is a homodimer of ~27 kDa subunits(Fothergill-Gilmore
and Michels, 1993), in hyper-thermophilic archaebacteria TPI is a
homotetramerof 25 kDa subunits (Kohlhoff et al., 1996). TheCalvin
cycle enzyme of higher plant chloroplasts is ahomodimer of ~27 kDa
subunits (Kurzok andFeierabend, 1984; Henze et al., 1994; Schmidt
et al.,1995). For both the chloroplast and cytosolic
enzymesseparated from leaves is ~2 mM andis ~700 (Kurzok and
Feierabend, 1984). Thecrystal structure of the enzyme from many
sources isknown (Velanker et al., 1997).
As for PGK, class I/class II forms ofTPI have notbeen described.
Calvin cycle TPI of higher plantchloroplasts arose through a
duplication of the pre-existing eukaryotic nuclear gene for
cytosolic TPI,accompanied by the acquisition of a transit
peptide(Henze et al., 1994; Schmidt et al., 1995). But sincethe
prexisting nuclear gene was itself acquired viaendosymbiotic gene
transfer from ancestors ofmitochondria (Keeling and Doolittle,
1997), theCalvin cycle of higher plant chloroplasts functionswith
TPI enzyme of mitochondrial origin that wasrerouted to the plastid
during evolution (Martin andSchnarrenberger, 1997).
genome (Martin et al., 1993; Henze et al., 1995). TheGapA and
GapB subunits of the enzyme arosethrough gene duplication during
chlorophyteevolution (Meyer-Gauen et al., 1994). The B subunitis
implicated in regulatory properties of the enzyme(Scagliarini et
al., 1998) and possesses a CTE ofroughly 30 amino acids relative to
the A subunit thatis involved in thioredoxin-dependent regulation
(seeSection VI).
D. Triosephosphate Isomerase
E. Fructose-1,6-bisphosphate/Sedoheptulose-1,7-bisphosphate
Aldolase
Chapter 2 Calvin Cycle
-
16
Fructose-1,6-bisphosphate aldolase (EC 4.12.1.13,aldolase)
catalyzes the reversible aldol condensationof dihydroxyacetone
phosphate and either GA3P orerythrose-4-phosphate to yield
fructose-1,6-bisphosphate or
sedoheptulose-l,7-bisphosphate,respectively. Both activities are
part of the Calvincycle. Two very distinct types of aldolase
enzymesoccur in nature that differ in their catalytic
mechanism(Rutner, 1964; Marsh and Lebherz, 1992). Class Ialdolase
enzymes form a Schiff-base with thesubstrate during catalysis via
condensation of theamino group of an active-center lysine residue
withthe carbonyl group of the substrate. Class II aldolaseenzymes
require divalent cations such as
as cofactors which stabilize the carbanionintermediate formed
during the reaction. The dualspecificity for F1,6BP and Su1,7BP
formation byaldolase applies to the chloroplast enzyme and to
thecytosolic enzyme, both of the class I type in higherplants
(Brooks and Criddle, 1966; Moorehead andPlaxton, 1990) and of the
class II type in Cyanophoraparadoxa (Flechner et al., 1999). Class
I aldolasesare homotetramers whereas class II aldolases
arehomodimers. The subunit size of both classes ofaldolase enzymes
is ~40 kDa, but class I and class IIaldolase monomers share no
detectable sequencesimilarity. This, in addition to the different
catalyticmechanisms and unrelated crystal structures forclass I
(Blom and Sygusch, 1997) and class II (Cooperet al., 1996)
aldolase, clearly indicates that these twoclasses of aldolase
enzymes are the result ofevolutionary functional convergence. For
separatedspinach chloroplast and cytosolic class I aldolase,
is 20 and 1 respectively, whereasis 6 and 4 respectively.
The
corresponding values for chloroplast and cytosolicclass II
aldolase from Cyanophora paradoxa are
1 mM and 660 respectively, whereasis 200 and 230
respectively
(Flechner et al., 1999).Class I and class II aldolases have a
very complex
phylogenetic distribution across prokaryotes andeukaryotes
(Henze et al., 1998). Most eubacteria,including all cyanobacteria
studied to date, typicallypossess class II aldolase (Rutter, 1964;
Antia, 1967),although class I aldolase is known in eubacteria(Witke
and Gtz, 1993). Halophilic archaebacteriacan possess either class I
or class II aldolase (Dharand Altekar, 1986). Interestingly, the
Methanococcusgenome does not encode a recognizable
homologueofeither class I or class II aldolase (Bult et al.,
1996),
although methanogens are known to possess aldolaseactivity (Yu
et al., 1994; Schnheit and Schfer,1995), raising the possibility
that a class III aldolasewill eventually be found. A possible
candidate forsuch a new class of aldolase has been described froma
halophilic archaebacterium (Krishnan and Altekar,1991) that
possesses a (mechanistically) class 1aldolase consisting of 27 kDa
(rather than 40 kDa)subunits with novel properties. Among
highereukaryotes, fungi typically possess class II aldolasewhereas
metazoa and higher plants possess class Ialdolase (Schnarrenberger
et al., 1990; Marsh andLebherz, 1992; Tsutsumi et al., 1994).
Euglenagracilis is exceptional among eukaryotes in that itpossesses
both class I aldolase (in the chloroplast)and class II aldolase (in
the cytosol) (Pelzer-Reith etal., 1994b). In addition to class I
and class II aldolase,ancient eubacterial gene duplications are
known inclass II aldolase evolution that have given rise
toaldolase-related enzymes specialized for substratesother than
sugar phosphates (Plaumann et al., 1997).The Calvin cycle of both
proteobacteria andcyanobacteria operates with class II aldolase,
whilealdolase of higher plant chloroplasts is a class Ienzyme, and
the paucity of sequences for class Ialdolase from prokaryotes makes
it currentlyimpossible to tell whence the class I gene for
thechloroplast enzyme arose (Plaumann et al., 1997).To add to this
conundrum of diversity, the Calvincycle in cyanelles of Cyanophora
paradoxa operateswith a class II aldolase (Gross et al., 1994).
Thus,Calvin cycle aldolase of plastids has arisen at leasttwice in
evolution, and the data for class I aldolase ofEuglenas
chloroplasts suggest that a third inde-pendent origin ofCalvin
cycle aldolase in plastids islikely (Plaumann et al., 1997).
F. Fructose-1,6-bisphosphatase
Fructose-1,6-bisphosphatase (EC 3.1.3.11, FBPase)catalyzes the
cleavage of the phosphoester bond onC1 to yield
fructose-6-bisphosphate (F6P). In mostproteobacteria and
cyanobacteria, the FBPase andSBPase reactions of the Calvin cycle
are catalyzedby a single enzyme (F/SBPase) with dual specificityfor
both substrates (Gerbling et al., 1986; Gibson andTabita, 1988; Yoo
and Bowien, 1995; Paoli et al.,1995). Xanthobacter flavus is an
exception (seebelow). F/SBPase from cyanobacteria (Gerbling et
William Martin, Renate Scheibe and Claus Schnarrenberger
-
17
al., 1985) is a tetramer of ~40 kDa subunits, as isFBPase from
spinach (Marcus and Harrsch, 1990).The crystal structure of spinach
chloroplast FBPasehas been determined (Villeret et al., 1995).
FBPase catalyzes a highly exergonic reaction thatis virtually
irreversible under physiological condi-tions, and it is one ofthe
key targets for regulation ofthe Calvin cycle. Activity of the
enzyme isundetectable in the dark (oxidized state), but increasesto
maximum activities within a few minutes ofillumination due to thiol
reduction via the thioredoxinsystem (Buchanan, 1980). FBPase is
specificallyactivated by thioredoxin (hence the
designation(Buchanan, 1980; Lopez-Jaramillo et al., 1997).
Themechanism of chloroplast FBPase regulation wasrevealed by
altered kinetics observed in the presenceof pH, and thiols
(Zimmermann et al., 1976).Activation of chloroplast FBPase by
reduced thiolsaffects a dramatic increase of substrate affinity
of>20-fold (Charles and Halliwell, 1980),of the fully activated
enzyme is 6 as compared to130 for the oxidized enzyme (Cadet and
Meunier,1988b). The sensitivity of the chloroplast enzyme tolow
concentrations ofmercuric ions has been studiedin several species
(Ashton, 1998a).
Both chloroplast and cytosolic FBPase of higherplants are highly
regulated (see Section VI), but byquite different mechanisms
(Latzko et al., 1974;Zimmermann et al. 1976). The cytosolic enzyme
is acontrol point for regulating flux through gluconeo-genesis.
Like its homologues from the cytosol ofnon-photosynthetic
eukaryotes, it is subject to strongallosteric inhibition by AMP and
regulation throughF2,6BP (Stitt, 1990a, 1990b), whereas
thioredoxinhas no effect. The chloroplast enzyme on the otherhand,
is insensitive to both AMP and F2,6BP.Curiously, these distinct
regulatory properties seemto have evolved specifically in the plant
lineage. Thisis because higher chloroplast FBPase arose throughgene
duplication of the prexisting nuclear gene forcytosolic FBPase,
that itself appears to have beenacquired from mitochondria (Martin
et al., 1996a;Schnarrenberger and Martin, 1997), indicating thatas
in the case of TPIthe higher plant Calvin cyclefunctions with an
FBPase enzyme of mitochondrialorigin. The archaebacteria
Methanococcus mari-paludis and Haloarcula vallismortis possess
highFBPase activity (Altekar and Rangaswamy, 1992; Yuet al., 1994)
but the enzyme has not been purifiedfrom any archaeon and the
Methanococcus genomedoes not encode a recognizable gene for
FBPase
A highly specific sedoheptulose-1,7-bisphosphatase(EC 3.1.3.37,
SPB) is not known from prokaryotes,although Xanthobacter flavus
differentially expressestwo distinct F/SBPase isoenzymes that both
acceptF1,6BP and Su1,7BP as substrates. The isoenzymeexpressed
during autotrophic growth (CbbF) hasnearly equal activities with
F1,6BP and Su1,7BP(3:1, respectively) as substrates, with a of3 the
other isoenzyme possesses much loweractivity with Su1,7BP (van den
Bergh et al., 1995).Similarly, Synechococcus PCC 7942 possesses
twoimmunologically distinct tetrameric FBPase isoen-zymes, one
ofwhich is specific for F1,6BP, the otherof which efficiently
cleaves both F1,6BP and Su1 ,7BP(Tamoi et al., 1996). The
cyanobacterial F/SBPaseisoenzymes thus have slightly more similar
propertiesto those found in higher plants, where chloroplastFBPase
and SBPase in plants are separate enzymesencoded by distinct but
distantly related nucleargenes (Raines et al., 1988,1992; Martin et
al., 1996a).In contrast to FBPase and bacterial F/SBPase, whichare
tetramers, SBPase from higherplant chloroplastsis a dimer of ~35
kDa subunits (Nishizawa andBuchanan, 1981; Cadet et al., 1987).
Also in contrastto many bacterial F/SBPase enzymes, higher
plantchloroplast FBPase and SBPase show a highspecificity for their
respective substrates, wherebychloroplast SBPase is highly, but not
completelyspecific for Su1,7BP (Zimmermann et al., 1976;Breazeale
et al., 1978; Cadet and Meunier, 1988b).Chloroplast SBPase, as
FBPase, is redox-modulatedby thioredoxin (Breazeale et al., 1978;
Nishizawaand Buchanan, 1981). The reduced (activated)enzyme has a
of 50 and a of380 (Cadet and Meunier, 1988b). But since the
of reduced (activated) chloroplast FBPaseis 6 the F1,6BP
activity of SPB is probably oflittle or no physiological relevance.
Similar reasoningapplies to the SBPase activity of chloroplast
FBPase(Ashton, 1998b). SBPase from wheat has beenexpressed in E.
coli (Dunford et al., 1998). F2,6BP,a potent allosteric regulator
ofcytosolic FBPase, hasno allosteric effect on chloroplast SBPase,
but can
(Bult et al., 1996), raising the possibility thatstructurally
unrelated class I and class II FBPaseenzymes may exist.
G. Sedoheptulose-1,7-bisphosphatase
Chapter 2 Calvin Cycle
-
18
act as a competitive inhibitor (Cadet and Meunier,1988b). The
evolutionary relationship betweenchloroplast SBPase and eukaryotic
FBPase andeubacterial F/SBPase enzymes is unclear (Martin etal.,
1996a), but it appears that the specialization ofSBPase from a
bifunctional F/SBPase ancestoroccurred at the prokaryotic
level.
H. Transketolase
Transketolase (EC 2.2.1.1, TKL) catalyzes thereversible,
thiamine diphosphate-dependent transferof a two carbon ketol group
from either fructose-6-phosphate or sedoheptulose-7-phosphate
(Su7P) toglyceraldehyde-3-phosphate to yield xylulose-5-phosphate
(Xu5P) and either erythrose-4-phosphateor ribose-5-phosphate (R5P),
respectively. TKL fromvarious sources is a homodimer of 74 kDa
subunits(Feierabend and Gringel, 1983). The crystal structureof the
yeast enzyme is known (Nikkola et al., 1994;Nilsson et al., 1997).
The catalytic mechanisminvolves nucleophilic attack ofthe substrate
carbonylgroup via the C2 carbanion of thiamine diphosphate(ThDP):
the rate-limiting C2 deprotonation steprequires interaction of N1'
in the ThDP pyrimidinering with (Kern et al., 1997).
Beyond the studies of Murphy and Walker (1982),who purified the
enzyme 400-fold, and Feierabendand Gringel (1983), who found only a
singlechloroplast species, little attention has been given tothe
biochemistry of this Calvin cycle enzyme.Substrate affinities for
the plant enzyme have beenreported as 100-130 for Xu5P, E4P and
R5P(Murphy and Walker, 1982), for human erythrocytesthe values 20
30 and 2mM were found(Himmo et al., 1989). For the purifiedenzyme
from spinach chloroplasts 77and 330 were found (Teige et al.,
1998).TKL has been cloned from Craterostigma (Bernacciaet al.,
1995), spinach (Flechner et al., 1996) andpotato (Teige et al.,
1996). The enzyme from spinachchloroplasts has been expressed in
highly activeform in E. coli (Flechner et al., 1996). Spinach
leavesappear to possess only a single TKL enzyme,
localizedexclusively in the chloroplast (Feierabend andGringel,
1983; Schnarrenberger et al., 1995). TKLshows structural similarity
to several other enzymes
involved in ThDP-dependent C2 metabolism:pyruvate decarboxylase
and the E1 subunit ofpyruvate dehydrogenase (Robinson and Chun,
1993).The nuclear gene for the Calvin cycle enzyme ofhigher plants
was acquired from cyanobacteria(Martin and Schnarrenberger,
1997).
I. Ribulose-5-phosphate 3-epimerase
Ribulose-5-phosphate 3-epimerase (EC 5.1.3.1, RPE)catalyzes the
reversible interconversion of ribulose-5-phosphate and
xylulose-5-phosphate. RPE is ahomodimer of~23 kDa subunits in
animals (Karmaliet al., 1983), Ralstonia (Kusian et al., 1992)
andspinach (Nowitzki et al., 1995). The spinach enzymehas been
purified to homogeneity and N-terminallysequenced (Teige et al.,
1998). The purified spinachenzyme migrates as an octamer, the
wasdetermined as 250 (Teige et al., 1998). RPEfrom the red alga
Galdieria sulphuraria has aof ~830 (J. Girnus, W. Gross and C.
Schnar-renberger, unpublished). Spinach leaves appear topossess
only a single RPE enzyme, localized inchloroplasts (Schnarrenberger
et al., 1995). RPE hasbeen cloned from sorghum and spinach
(Nowitzki etal, 1995) and potato (Teige et al., 1995), the
enzymefrom spinach chloroplasts has been expressed inactive form in
E. coli (Nowitzki et al., 1995). Morerecently, the enzyme from
spinach chloroplasts wascloned again, and was expressed in E. coli
again(Chen et al., 1998). Neither the mechanism of catalysisnor the
tertiary structure have been reported fromany source. Class I /
class II RPE enzymes have notbeen described, but three very
distantly rpe- relatedgenes exist in the E. coli genome, indicating
thepresence of relatively ancient eubacterial genefamilies
(Nowitzki et al., 1995). The nuclear gene forhigher plant Calvin
cycle RPE was acquired fromcyanobacteria (Martin and
Schnarrenberger, 1997).
J. Ribose-5-phosphate Isomerase
Ribose-5-phosphate isomerase (EC 5.3.1.6, RPI)catalyzes the
reversible isomerization of ribose-5-phosphate and
ribulose-5-phosphate. RPI has notbeen identified in the cbb operons
of photosyntheticproteobacteria (Gibson and Tabita, 1996). No
crystal
William Martin, Renate Scheibe and Claus Schnarrenberger
-
19
structures have been reported for this enzyme. Rutner(1970)
purified RPI from spinach 2800-fold. Only asingle enzyme was found,
a homodimer of 23 kDasubunits, as later shown for Arabidopsis
(Babad-zhanova and Bakaeva, 1987), that had a of460 Chloroplast RPI
from spinach (Martin etal., 1996b) has been cloned, it has sequence
similarityto RpiA from E. coli (Hove-Jensen and Maigaard,1993). But
E. coli also possesses a gene for a secondfunctional RPI enzyme,
RpiB, that is a homodimerof 16 kDa subunits. It shows no sequence
similarityto RpiA, but very high similarity to
galactose-6-phosphate isomerases (Srensen and Hove-Jensen,1996).
Thus for RPI, class I (e.g. spinach RPI orRpiA of E. coli) and
class II (RpiB of E. coli)enzymes should be distinguished. No
cytosolicisoenzyme of RPI was found in spinach
leaves(Schnarrenberger et al., 1995). Calvin cycle (class I)RPI
from spinach has identifiable homologuesencoded in the
Synechocystis and Methanococcusgenomes, but due to paucity of
reference sequences,the evolutionary origin of the plant nuclear
gene instill unclear.
K. Phosphoribulokinase
Phosphoribulokinase (EC 2.7.1.19, PRK) transfersthe of ATP to
the C1 hydroxyl group ofribulose-5-phosphate, regenerating the
primaryacceptor. Class I and class II PRK enzymes areknown (Tabita,
1994; Brandes et al., 1996a; Martinand Schnarrenberger, 1997).
Class I PRK is encodedin proteobacterial cbb operons. It is an
octamer of~30 kDa subunits with allosteric inhibition throughAMP
and allosteric activation through NADH(Runquist et al., 1995).
Crystal structure data hasbeen reported for class I PRK (Roberts et
al., 1995;DHT Harrison et al., 1998). Class II PRK is found
incyanobacteria and higher plants. It is a dimer of ~44kDa subunits
in chloroplasts (Porter et al., 1986;Clasper et al., 1994). The
enzyme can associate totetramers in Synechocystis (Wadano et al.,
1995).The 300-foldpurified enzyme from the chromophyticprotist
Heterosigma carterae is a tetramer of 53 kDasubunits with a of 208
and of226 (Hariharan et al., 1998). Crystal structureshave not been
reported for the class II enzyme. Thecatalytic properties of class
I and class II PRK differmarkedly (Tabita, 1988). The nuclear gene
for the
higher plant Calvin cycle enzyme was acquired fromcyanobacteria
(Martin and Schnarrenberger, 1997).
PRK catalyzes a highly exergonic reaction and isstrongly
regulated by the thioredoxin system(Buchanan, 1980). The oxidized
(dark) enzymepossesses only about 2% of the activity of the
fullyactive (reduced) form (Surek et al., 1985). Kineticvalues of
of 60 and of 110were reported for the spinach enzyme expressed
inthe yeast Pichia pastoris (Brandes et al., 1996a),similar to
values determined for the purified nativeactivated wheat enzyme
(Surek et al., 1985). Incontrast to GAPDH, FBPase, and SBPase,
thio-redoxin activation of PRK does not involve loweringof values,
but affects the (Porter et al., 1986).
III. Calvin Cycle Gene Organization,Expression, and Regulation
in Eubacteria
Several excellent reviews on this topic haveappeared recently
(Tabita, 1994; Gibson and Tabita,1996; Bommer et al., 1996; Gibson
andTabita, 1997;Kusian and Bowien, 1997; Shively et al.,
1998).Mutant strains of Rhodospirillum rubrum (Falconeand Tabita,
1993), Rhodobacter sphaeroides (Gibsonet al., 1991), and Ralstonia
eutropha (formerlyAlcaligeneseutrophus)
(Bowienetal.,1993)defectivefor autotrophic growth continue to
uncover newgenes involved in Calvin cycle function andregulation.
The structure and regulation of Calvincycle operons and gene
clusters has been investigatedin several eubacteria. Among
eubacteria, the mostcomplete picture of gene organization exists
for the
Ralstonia eutropha and thecyanobacterium Synechocystis PCC6803.
Thestructural organization ofCalvin cycle genes in theseorganisms
could not possibly differ more.
Ralstonia eutropha possesses the largest cbboperon characterized
to date (Bowien et al., 1993;Bommer, 1996). With two exceptions
(ribose-5-phosphate isomerase and triosephosphate isomerase)it
encodes the entire pathway, and is transcribed asone polycistronic
mRNA under the regulation ofCbbR (Windhvel and Bowien, 1991), a
member ofthe LysR family of transcriptional regulators
(Tabita,1994). The opposite extreme is realized in Syne-chocystis,
where no two genes for Calvin cycleenzymes occur as neighbors in
the genome (Kanekoet al., 1996a, 1996b). The Synechocystis genes
arenot separatedbyjustafewhundredorafewthousand
Chapter 2 Calvin Cycle
-
20
bases, they are strewn around the 3.6 Mb genomewith no
recognizable pattern whatsoever. Even therbcL/rbcS operon is
disrupted, the genes for the twosubunits being separated by an ORF
of still unknownfunction, rbcX. Synechocystis possesses two
geneshomologous to cbbR, but neither the function of theirproducts
are known, nor whether Calvin cycle genesin Synechocystis form a
regulon. Comparatively littleis known about regulation, coordinated
or otherwise,of cyanobacterial Calvin cycle genes (Beuf et
al.,1994; Li and Tabita, 1994; Gibson and Tabita, 1996;Xu and
Tabita, 1996).
In Xanthobacter flavus, cbb genes are distributedacross at least
two operons, the gap-pgk cluster is notcontiguous with the cbb
operon, but it is part of a cbbregulon under CbbR control (Meijer
et al., 1996). Asecond cbb operon is present on a large plasmid
inRalstonia that is nearly identical to the chromosomaloperon
(Bowien et al., 1993). The cbb operons studiedfrom Ralstonia
Rhodobactersphaeroides, Rhodobacter capsulatus, Rhodo-spirillum
rubrum, Xanthobacter flavus, and Nitro-bacter vulgaris (all show
very littleconservation of gene order across species, other thanthe
fact that cbbR is usually transcribed, as inRalstonia, on the
opposite strand from a divergentpromoter (Gibson and Tabita, 1996).
Given thedispersed nature of the Synechocystis genes, it
isconceivable that either these operons were assembledin
independent lineages from an ancestrally dispersedstate, or that
fragmentation of an ancestral operonhas occurred in Synechocystis,
accompanied byrearrangements in proteobacteria. Visible
rearrange-ment of cbb genes in suggests thatconsiderable structural
reorganization of cbb operonshas occurred in these genomes during
evolution.There have been reports of Calvin cycle
specificactivities in halophilic archaebacteria (Rawal et al.,1988;
Altekar and Rajagopalan, 1990; Rajagopalanand Altekar, 1994) but
the enzymes have not beencharacterized in detail.
A general picture of higher level control of signaltransduction
and gene regulation for the Calvin cycleand its integration into
the general metabolism ofphotosynthetic eubacteria is beginning to
emerge.The presumably top level of hierarchy involves theRegA/RegB
(PrrA/PrrB) two-component sensor-kinase system (Joshi and Tabita,
1996). This systemappears to integrate the control, expression
andfeedback of regulons for photosystem biosynthesis(Sganga and
Bauer, 1992; Eraso and Caplan, 1994;
Mosley et al., 1994; Allen et al, 1995), nitrogenmetabolism
(nif-system), and fixation (Qianand Tabita, 1996; Joshi and Tabita,
1996). Theregulatory cascade is apparently influenced by theredox
state of the cells, the level of oxygen, and thepresence of various
carbon and nitrogen sources(Joshi and Tabita, 1996). Since
precisely these factors(redox state, oxygen, carbon and nitrogen)
are knownto have dramatic and, in some cases, interdependenteffects
on plant metabolism and nuclear geneexpression (Turpin and Weger,
1990; Chen et al.,1993; Escoubas et al, 1995; Kozaki and
Takeba,1996; Wingsle and Karpinski, 1996; Karpinski et al.,1997),
these findings from bacterial systems mayhave a degree of model
character for understandingrelated phenomena in eukaryotic systems,
where themolecular basis of regulation is less
thoroughlyunderstood. Although it seems unlikely at first sightthat
these same prokaryotic molecular componentswill be found to be
involved in plant signaltransduction, the principles of regulatory
responseimplemented by eukaryotic signaling/regulationmachinery may
ultimately prove to be very similar. Aremarkable study recently
provided strong evidencethat some components for rapidly
transducing redoxsignals in higher plants perceive the redox state
ofthe plastoquinone pool directly in the thylakoidmembrane
(Pfannschmidt et al., 1999). Whether ornot nuclear encoded
bacterial two-componentsystems, which are still encoded in some
chloroplastgenomes (Stoebe et al., 1998; Martin et al., 1998),might
be involved in such processes, is an attractivequestion.
Regulation of the Calvin cycle enzyme expressionhas often been
monitored using Rubisco as a markerfor the pathway. PRK, TKL RPI
and RPE may fulfillthe same purpose, because they appear to be
localizedin chloroplasts exclusively as well (Schnarrenbergeret
al., 1995). The other enzymes may fulfill functionsin other
pathways. For example PGK, GAPDH, TPI,aldolase and FBPase are
involved in the gluconeo-genetic and partially in the glycolytic
reactionsequence, and the oxidative pentose phosphatepathway relies
on many enzymes of the Calvin cycle(Schnarrenberger et al., 1995).
Expression studies of
IV. Calvin Cycle Expression in Plants
A. Quantification of Activities
William Martin, Renate Scheibe and Claus Schnarrenberger
-
21
Calvin cycle enzymes that possess cytosolichomologues require
not only measurement of totalactivities in crude extracts but also
quantification ofthe amount of activities attributable to
chloroplastand cytosolic isoenzymes. This is particularlynecessary
for the gluconeogenetic enzyme activitiesPGK, GAPDH, TIM, aldolase
and FBPase isoen-zymes, which may be separated by
ion-exchangechromatography but not by gel filtration. In
specialcases, it is also possible to distinguish cytosol-
andchloroplast-specific activities by virtue of theirdifferent
substrate specificity for NADH and NADPH,as in the case of GAPDH,
or by their differentialresponse to pH, sulfhydryl reagents, and as
forFBPase. But compartmentation of plant carbohydratemetabolism is
not an evolutionarily conservedproperty across species
(Schnarrenberger et al., 1990),and within a given plant it varies
across developmentalstages and tissues. Worse yet, across
species,completely different enzymes are involved that mustbe
assayed by different means, for example class Iand class II Calvin
cycle aldolase (Gross et al., 1994;Plaumann et al., 1997; Flechner
et al., 1999). Thus,Rubisco is a valuable marker for regulation of
Calvincycle expression, but other enzymes may show verydifferent
regulation patterns, and sweeping general-izations to the effect
that, beyond the Calvin cycle,plantshave this, that or the other
pathway of sugarphosphate metabolism in this or that compartmentare
not possible.
The chloroplast activities of PGK, GAPDH,aldolase and FBPase in
green leaves usually accountfor about 90% of the total activity
(Heber et al., 1963;Latzko et al., 1974; Krger and
Schnarrenberger,1983; Schnarrenberger and Krger, 1986; Lebherzet
al, 1984; Kpke-Secundo et al., 1990; McMorrowand Bradbeer, 1990),
chloroplast TPI accounts foronly about 50% of the total activity
(Kurzok andFeierabend, 1984). For other isoenzyme activities
ofstarch metabolism and the oxidative pentosephosphate pathway in
green leaves, the cytosolicisoenzyme appears to account for most of
theprevalent activity (Schnarrenberger, 1987). It appearsthat the
activities of the regenerative part of theCalvin cycle (RPI, RPE,
TKL) may not requireisoenzyme separation in most cases, since they
areprobably located exclusively in the chloroplasts(Schnarrenberger
et al., 1995), except in somespecialized tissues and species (e.g.
TKL inCraterostigma: Bernacchia et al., 1995). For thesethree
enzymes no chloroplast/cytosol isoenzymes
can be separated in spinach leaves (Schnarrenbergeret al.,
1995). It is well known that various Calvincycle enzymes vary
considerably in their maximalactivities among higher plants,
various algae andeubacteria (Smillie, 1963; Heber et al., 1967;
Latzkoand Gibbs, 1968; Latzko and Gibbs, 1969; Kelly andLatzko,
1979), as do activities in other pathways.
B. Expression Studies of Enzyme Activitiesand Transcription
The literature on expression of Calvin cycle enzymesis vast. One
of the most widely studied aspects is theincrease of enzyme
activity and mRNA levels inresponse to light. The influence of
light on expressionof genes involved in photosynthesis has been
reviewed(Chory et al., 1996; Kloppstech, 1997).
Phytochromes(Schopfer, 1977; Pratt, 1995), blue light, and
UV-receptors play important roles in this regulation,
thatultimately reaches genes for many Calvin cycleenzymes. It is
also well-known that the glycolytic,cytosolic isoenzymes of several
Calvin cycle activitiesare, as a rule, not responsive to light and
are inducedunder anaerobic conditions (Sachs, 1994; Kennedyet al.,
1992).
Complete cDNAs have been characterized forseveral Calvin cycle
enzymes from several sources,and all of the Calvin cycle enzymes
from spinachchloroplasts have been cloned (Flechner et al.,
1996;Martin et al., 1996a). Rubisco gene expression hasbeen studied
in by far the greatest detail of all of theCalvin cycle enzymes.
Transcription factors involvedspecifically in rbcS gene expression,
e.g. GT-1 (Lamand Chua, 1990; Sarokin and Chua, 1992) and
GT-2(Gilmartin et al., 1992) have been characterized. Ingeneral,
expression of Calvin cycle genes in plants,particularly in
etiolated seedlings, is stimulated bylight. This can occur through
elevated transcription,or, as recent studies of the Cen gene in
Chlamy-domonas mutants have shown, post-transcriptionallyat the
level of mRNA stability (Hahn et al., 1996).Various cis elements
have been described for Calvincycle genes from different sources,
including theWF-1 element upstream ofthe genes for SBPase andFBPase
ofwheat (Miles et al., 1993), the Gap andAEboxes upstream of the
Arabidopsis GapA and GapBgenes (Conley et al., 1994; Kwon et al.,
1994; Park etal., 1996), and an octameric motif in the first
intronof the maize GapA1 gene (Donath et al., 1995;Khler et al.,
1996). The FBPase promoter alsocontains a DNA binding site for the
GT-1 factor
Chapter 2 Calvin Cycle
-
22
which mediates light activation ofexpression throughphytochrome
in promoters of oat and rice (Lloyd etal., 1991b). In addition to
the small subunit ofRubisco, for which numerous gene structures
areknown (Wolter et al., 1988; DeRocher et al, 1993;Fritz et al.,
1993), several higher plant Calvin cyclegene structures have been
characterized. Theseinclude GapA from maize (Quigley et al,
1988),Arabidopsis (Shih et al., 1992), and several othersources
(Kersanach et al., 1994), SBPase and FBPasefrom wheat (Raines et
al., 1988, 1992), aldolasefrom rice (Tsutsumi et al., 1994) and
Chlamydomonas(Pelzer-Reith et al., 1995), and PRK from wheat(Lloyd
et al., 199la).
In studies of the transcript levels of all Calvincycle enzymes
in various tissues during spinachdevelopment (Henze, 1997), it was
observed thatmost mRNAs were present in all green leaf tissue
inroughly the same relative quantities, with theexception of rbcS
mRNA, that was present at roughly10-fold higher steady-state
levels. In etiolatedcotyledons, all mRNA levels were reduced at
least10- to 20-fold relative to green leaves. Uponillumination,
mRNAs for rbcS, aldolase, SBPaseand PRK increased within 2 h after
illumination,followed by the other mRNAs. After 24 h
ofillumination, mRNA levels were indistinguishablefrom those in
green leaf tissue. It is still too early totell whether genes of
the Calvin cycle in higherplants are regulated as a unit, or
whether their activityis simply modulated as part of a general
greeningresponse of the gene regulatory machinery to lightand redox
state. In the red alga Galdieria sulphuraria,that can grow
heterotrophically or autotrophically(Gross et al., 1999),
isoenzymes of many Calvincycle activities (aldolase, RPE, PGK,
FBPase, andGAPDH) are specifically induced during thetransition
from heterotrophic to autotrophic growth(J. Girnus, C.
Schnarrenberger, W. Gross, unpub-lished)
There are far too many reports involving expressionstudies of
Calvin cycle genes and enzyme activitiesto permit thorough review.
In Table 1 we have tried toprovide access to some of that
literature, includingmany studies that cannot be found by
computer-searching (and omitting many studies that can).Rubisco is
well known for its inducibility byphytochrome. The effects of
phytochrome and otherlight receptors on the remaining Calvin cycle
enzymeshave been studied in less detail and in many of theearly
studies, the quantification of chloroplast vs.
cytosol enzyme activities was not considered. Table1 is by no
means complete, but we hope that readersfind parts of it
useful.
C. Gene Regulation Through High SugarSensing, and Redox
State
A sugar-sensing system has been discussed in plantsthat may be
able to significantly influence geneexpression (Sheen, 1990, 1994;
Koch, 1996; vanOosten and Besford, 1996; Jang and Sheen,
1997;Chapter 10, Graham and Martin). Sugars like glucose,fructose
and sucrose cause strong repression of genesfor photosynthetic
functions, resulting in e.g.reduction of photosynthetic pigments
and Calvincycle enzymes while other sugars are largelyineffective.
Glucose feeding can reduce the steady-state mRNA levels of several
Calvin cycle genes inwheat, including FBPase, SBPase, PGK and
rbcS(Jones et al., 1996). Among the Calvin cycle enzymes,activity
and protein of Rubisco decline steadily withinseveral days. This
was demonstrated in cell suspensioncultures of Chenopodium rubrum
and in intacttobacco and potato leaves which were cold-girdledfor
12 h to reduce assimilate export (Krapp et al.,1993), detached
spinach leaves fed with glucosethrough the petioles (Krapp et al.,
1991), mesophyllcells of tobacco (Criqui et al., 1992) and by
comparinggreen and bleached leaves of transgenic tobaccoplants
expressing a yeast-derived invertase in theapoplast (Stitt et al.,
1990). In some ofthese systemsit was shown that glucose treatment
resulted in therepression of these and other
photosynthesis-relatedgenes. In cell suspension cultures of
Chenopodiumrubrum, rbcS mRNA levels are reduced within severalhours
and run-on experiments with isolated nucleiindicated that also the
synthesis of rbcS is reduced,as is incorporation into the
Rubiscoprotein, indicating an inhibition of de-novo synthesis(Krapp
et al. 1993). In glucose-fed tobacco protoplastsand leaf discs,
rbcS transcript levels are reducedwithin hours upon glucose
addition, but rbcLtranscript levels are reduced at a much slower
rate(Criqui et al., 1992). Thus, regulation takes placeprimarily at
a transcriptional level. Information onglucose repression of other
Calvin cycle enzymes islimited to FBPase and GAPDH, both of which,
likeRubisco, show similarly declining activities in thepresence of
glucose (Stitt et al., 1990; Krapp et al.1991, 1993).
Elevated (about 1000 ppm) can also influence
William Martin, Renate Scheibe and Claus Schnarrenberger
-
23
gene expression (reviewed by von Oosten andBesford, 1996),
including those for some Calvincycle enzymes. For example, whole
tomato plantsgrown at elevated relative to ambientgrown plants, for
10 days showed reduced Rubiscoactivity in the second half of this
period, probablyaccounting for the long-term decline in
photosyntheticefficiency under high (Yelle et al., 1989).
rbcStranscript levels are greatly reduced in tomato plantswithin 4
days, while rbcL transcript levels declineless pronounced. This
effect was enhanced in detachedleaves, indicating repression by
elevated internalglucose levels (van Oosten et al., 1994). On
thecontrary, low levels of resulted in an overex-pression of rbcS
(Krapp et al., 1993). Besides Rubisco,also the activities ofPGK and
GAPDH were reducedunder elevated condition, though only in
fullydeveloped leaves (Besford, 1990). However, Krappet al. (1991)
observed less inhibition at highthan under ambient at saturating
irradiation andeven less under low irradiation. It should be
notedthat any inhibition seen under these conditions mightalso be
attributable to nitrogen limitation that canbecome apparent at
increased growth rates (Kozakiand Takeba, 1996).
Van Oosten and Besford (1995) showed thattranscript levels of
plastid-encoded rbcL and othergenes involved in photosynthesis
(psbA, psaAB) werereduced in mature leaves by elevated Expressionof
nuclear genes associated with the Calvin cyclesuch as Rubisco
activase are also reduced by elevated
(van Oosten et al., 1994). Nuclear encoded rbcStranscript leaves
were reduced in tomato plantsexposed to high as were
plastid-encoded rbcLtranscript levels, though less markedly, and
theseeffects could be simulated by sugar feeding (vanOosten and
Besford, 1994). In terms of Rubiscocontent, tomato plants responded
to high in amanner similar to plants grown with low nitrogensupply
(van Oosten et al., 1995). In bird-cherry treesgrown under
conditions where nutrients were notlimiting, Rubisco activity
decreased in response tohigh (Wilkins et al., 1994). Clearly, there
is aninterdependence between availability, nitrogen,redox state,
sugar levels and light levels that influencegene expression. An
Arabidopsis mutant defective ina gene that might be involved in
integrating ortransducing sugar-related signals was
recentlydescribed (van Oosten et al., 1997). Furthermore, theredox
state of the thylakoid membrane itself has beenrecently shown to
regulate the transcription of plastid
genes involved in maintaining redox balance(Pfannschmidt et al.,
1999), a process that certainlyentails the Calvin cycle as the
primary means forregenerating The question of which andhow many
signaling pathways are involved inmaintaining redox balance in a
manner that affectsthe Calvin cycle is still open.
D. Regulation in Specific Systems
A system involving preferential breakdown of 70Schloroplast
ribosomes on Calvin cycle enzymes wasused in early studies, because
it permitted the site ofenzyme synthesis to be determined long
before thecoding capacity of chloroplast genomes had beendetermined
(Feierabend and Schrader-Reichhardt,1967). Nuclear-encodedenzymes
are still synthesizedon 80S ribosomes and imported to chloroplasts
underpermissive low (22 C) temperatures or non-permissive high (32
C) temperatures. Among theenzymes of sugar phosphate metabolism
assayed,GAPDH, PGK, TPI, TKL, FBPase, RPI, PRK andaldolase were
recovered in chloroplasts at non-permissive conditions (Feierabend
and Brassel, 1977;Feierabend, 1979, 1986; Feierabend and
Gringel,1983; Kurzok and Feierabend, 1983,1986; Otto andFeierabend,
1989), however, Rubisco was absent(Feierabend, 1979), transcripts
of Rubisco wererepressed as well (Winter and Feierabend, 1990).
Another well studied system of Calvin cycleexpression is green
and white leaf tissue of thealbostrians mutant of barley. This
mutant shows avariegated pattern of white and green striped
leaveswith non-Mendelian inheritance (Hagemann andScholz, 1962).
Rubisco, GAPDH, aldolase, andFBPase were strongly reduced in white
leaf tissue(Brner et al., 1976; Bradbeer and Brner, 1978;Boldt et
al., 1992). In contrast, the cytosoliccounterparts of the Calvin
cycle enzymes, theenzymes of starch metabolism and the key
enzymesof the oxidative pentose phosphate pathway werevirtually
unchanged (Boldt et al., 1992). Transcriptsof Rubisco were totally
repressed in white tissue butwere enhanced in green tissue through
phytochrome(Hess et al., 1991). The transcripts of chloroplastPRK,
GAPDH, PGK, aldolase, and FBPase wererepressed in white tissue,
while those of cytosolicGAPDH and PGK were slightly enhanced (Hess
etal., 1993, 1994; Boldt et al., 1994). The phenomenonis
interpreted as the action of a plastid derived factoror signal
which represses many (but not all) nuclear-
Chapter 2 Calvin Cycle
-
24 William Martin, Renate Scheibe and Claus Schnarrenberger
-
25Chapter 2 Calvin Cycle
-
26
encoded plastid proteins in the nucleus (Boldt et al.,1990; Hess
et al., 1994; Hess et al., 1997). Thisfactor/signal also represses
most genes of theglycolate pathway in plastids and
peroxisomes,indicating functional unity of repression
ofphotosynthetic functions (Boldt et al., 1997). Hahnet al. (1996)
recently described a nuclear gene inChlamydomonas reinhardtii that
posttranscriptionally
affects mRNA levels for several chloroplast proteins,including
RPE.
Expression ofCalvin cycle genes has been studiedin the
facultative CAM plant Mesembryanthemumcrystallinum (iceplant).
GAPDH mRNA accumulatesin response to salt stress (Vernon and
Bohnert, 1992;Vernon et al., 1993), whereby PRK expression
isreduced by salt stress (Michalowski et al., 1992).
William Martin, Renate Scheibe and Claus Schnarrenberger
-
27
Transcript levels for cytosolic GAPDH increaseduring the
transition from C3 to CAM metabolism(Ostrem et al., 1990). In C4
plants, only the bundlesheath cells contain a complete set of
Calvin cycleenzymes (see also Chapter 18, Furbank et al.).
Enzymes of the Calvin cycle can also be found inplastids of
non-green tissues. An extreme example isthe endosperm tissue of
developing and germinatingcastor bean which never greens (Plaxton,
1996).Many of the plastidic isoenzymes of the Calvin cycleare
present in this tissue and show an expressionpattern following the
typical fat-to-sugar conversionwith maximum activities 5 days after
germination(Nishimura and Beevers, 1981). The question iswhether
these enzymes really function in fixationor whether they are due to
a leaky expression.Alternatively, they could function in the
oxidativepentose phosphate cycle which can provide NADPHfor nitrate
reduction. This has been implied by workof Emes and Fowler (1978)
on TKL and transaldolase.A cytosolic class I aldolase with
specificity for bothF1,6BP and Su1,7BP was found in carrot
storageroots but the plastidic homolog was missing(Moorhead and
Plaxton, 1990). A general consider-ation ofmetabolism in
chromoplasts was summarizedby Camara et al (1995).
E. Calvin Cycle Enzymes and Expression inEuglena gracilis
Euglena gracilis can grow autotrophically andheterotrophically
on various substrates (Kitaoka etal., 1989; Brandt and Wilhelm,
1990). A first screenfor cytosolic and chloroplast enzyme
activities ofsugar phosphate metabolism was presented by
Smillie(1963). Enzyme levels related to photosynthesis
anddegradative reactions like glycolysis seem to beantagonistically
regulated under autotrophic andheterotrophic growth conditions,
respectively. Mostenzyme activities ofthe Calvin cycle increase
duringgreening though to various degrees and decreaseafter transfer
to heterotrophic conditions (Latzkoand Gibbs, 1969; Kitaoka et al.,
1989).
Enzyme activities in Euglena gracilis may also beregulated by
light. This phenomenon was dissectedinto a blue- and a red-light
reaction by the use ofmutants (Schmidt and Lyman, 1974): In
wild-typecells blue light increased the activities of Rubisco,PRK,
chloroplast GAPDH, and chloroplast aldolasetwice as effectively as
red light. Mutant Y9ZNa1Lhad no chlorophyll and showed a blue-light
but no
red-light effect. Mutant Y11P22DL had smallamounts ofchlorophyll
and showed the same activityin red and blue light. Mutant W14ZNa1L
had nochloroplasts and no Rubisco but the same activity ofGAPDH,
PRK and aldolase as dark-grown wild-type cells, but this activity
was not increased by light.In other mutants (W3BLU and W8BHL) with
noplastid DNA small amounts of cytosolic class Ialdolase were
recorded (Karlan and Russell, 1976).Chloroplast development in
Euglena gracilis issensitive to glucose repression. When
heterotrophiccells are transferred to autotrophic
conditions,chloroplast development starts from proplastids.
Thepresence of glucose inhibits greening and synthesisof Rubisco
(Reinbothe et al., 1991a). Duringdedifferentiation from
chloroplasts to proplastids,on the other hand, Rubisco synthesis
ceasedimmediately upon transfer to the dark in the presenceof
glucose (Reinbothe et al., 1991b).
Chloroplast and cytosolic GAPDH are both presentin Euglena
gracilis cells, have been purified andshow no immunochemical
cross-reaction (Grissonand Kahn, 1974; Theiss-Seuberling, 1984).
Chloro-plast GAPDH is activated by dithiothreitol and/orthioredoxin
(Theiss-Seuberling, 1981). Duringchloroplast development of Euglena
gracilis in thelight NADP-GAPDH increases in activity
(Hoven-kamp-Obbema and Stegwee, 1974). Both enzymesare encoded by
nuclear genes and possess severalunusual sequence attributes (Henze
et al., 1995).
The aldolase isoenzymes of Euglena gracilisbelong to the class I
and class II type and arecompartmented in the chloroplasts and in
the cytosol,respectively (Rutter, 1964; Mo et al., 1973;
Pelzer-Reith et al., 1994b; Plaumann et al., 1997).
Underautotrophic conditions, the chloroplast enzyme ismore active
than the cytosolic enzyme (Mo et al.,1973; Karlan and Russel,
1976). This pattern isreversed during growth under heterotrophic
condi-tions (Mo et al., 1973). While chloroplast andcytosolic
aldolase of higher plants differ little in theirbiochemical
parameters (Anderson and Pacold, 1972;Buckowiecki and Anderson,
1974; Krger andSchnarrenberger, 1983; Lebherz et al., 1984),
thecytosolic class II aldolase of Euglena has a muchhigher value
and a broader pH optimumthan class I aldolase (Pelzer-Reith et al.,
1994b).Both aldolases of Euglena gracilis show
endogenousrhythmicity in the light and in the dark (Pelzer-Reithet
al., 1994a; Malik, 1997). The expression oftranscripts and the
enzymes appears to be regulated
Chapter 2 Calvin Cycle
-
28
posttranscriptionally (Malik, 1997). Chloroplast andcytosolic
isoenzymes of TPI were separated fromEuglena gracilis (Mo et al.,
1973). The chloroplasttype A enzyme is high in autotrophic and low
inheterotrophic cells while the cytosolic type Bisomerase
predominates under heterotrophic growthconditions. An antagonistic
regulation underautotrophic and heterotrophic growth conditions
isalso implied for the chloroplast and cytosolic FBPasewhen
measured at pH 8.5 and 6.9, respectively (Latzkoand Gibbs,
1969).
Structure and expression of nuclear genes forchloroplast
proteins in Euglena are unusual in severalrespects. Many encode
polyprotein precursors ofchloroplast proteins (Houln and Schantz,
1988,1993). An example of such a polyprotein is rbcS inEuglena. The
nuclear gene is transcribed as an mRNAencoding eight nearly
identical concatenate smallsubunits that are translated as a 140
kDa cytosolicpolyprotein, all eight subunits are imported
intochloroplasts with the aid of a single transit peptide,and then
proteolytically processed from thepolyprotein into individual
subunits for Rubiscoassembly (Chan at al., 1990). This unusual
polyproteinorganization appears to be restricted to some
nuclear-encoded genes in protists of secondary symbioticorigin,
i.e. protists that acquired their plastids byengulfing
photosynthetic eukaryotes rather thanprokaryotes. [This notion was
first suggested forEuglena (Gibbs, 1978) and subsequently
demon-strated to be the case for several photosyntheticprotists
(Maier, 1992; McFadden et al., 1994;Melkonian, 1996; Van de Peeret
al., 1996; McFaddenet al, 1997)]. Such plastids are surrounded by
threeor more membranes instead of two, and precursorimport is
therefore more complex, involving ER-processing of a signal peptide
prior to chloroplastuptake in the case of Euglena (Kishore et al.,
1993;Sulli and Schwartzbach, 1996). It is likely that anumber of
Euglenas nuclear genes for chloroplastproteins stem from the
secondary symbiont and weretherefore transferred twice in
evolution: once fromcyanobacteria to the chlorophyte nucleus, and
oncemore from the chlorophyte nucleus to the nucleus ofthe
Trypanosoma-like host (Henze et al., 1995;Plaumann et al., 1997).
Euglenas nuclear genes forCalvin cycle GAPDH (Henze et al., 1995),
aldolase(Plaumann et al., 1997), TKL and TPI (W.
Martin,unpublished) are not encoded as polyproteins,indicating that
this unusual organization is restrictedto certain transcripts.
Euglenas nuclear genes for
rbcS and cytosolic GAPDH contain a novel class ofhighly
structured introns that have not been describedfrom any other
eukaryotes (Henze et al., 1995; Tessieret al., 1995). Also, spliced
leader sequences arefound at the 5' end of many of Euglenas
nucleartranscribed mRNAs (Tessier et al., 1991). Suchspliced
leaders have been implicated in the RNA-processing
ofpolycistronically transcribed eukaryoticoperons found in the
euglenozoan lineage and inCaenorhabditis (Hirsch, 1994).
In other photosynthetic protists, little is known atthe
molecular level about Calvin cycle enzymes andgene structure, but
this can be expected to change inthe future since these organisms
are turning up quitea number of surprising findings. For example
thedinoflagellates Gonyaulax (Morse et al., 1995) andSymbiodinium
(Rowan et al., 1996) use class IIRubisco in their Calvin cycle.
Moreover, those class IIRubisco genes are nuclear encodedand
inSymbiodinium as a polyprotein, as in the case ofEuglenas rbcS.
Other photosynthetic protists seemto lack chloroplast- and
cytosol-specific isoenzymesof sugar phosphate metabolism.
Chlamydomonasreinhardtii is an extreme example, since this alga
hasno cytosolic isoenzymes of sugar phosphatemetabolism for at
least eight enzyme activities, amongthem aldolase (Schnarrenberger
et al., 1994),suggesting that the general compartmentation
ofcarbohydrate metabolism may be surprisinglyvariable across
protists.
V. Enzyme Interactions and Multienzyme-like Complexes
In 1970, Rutner noted that ...there are now
severalwell-documented cases of multi-enzyme complexes(e.g. fatty
acid synthase, pyruvic dehydrogenase [...]),there is a tendency to
implicate them in othersequential biochemical reactions (Rutner,
1970) anddelineated some straightforward
mass-activitystoichiometric difficulties encountered when
suchcomplexes are considered in the context ofthe Calvincycle.
Since that time, there have been many reportsthat some enzymes of
the Calvin cycle may formmultienzyme-like complexes, findings that
have oftenbeen discussed in the context of metabolic channelingof
intermediates. In enzymological studies prior to1980, these
complexes were rarely observed. Themajority of reports deal with
complexes isolatedfrom pea and spinach chloroplasts and from
green
William Martin, Renate Scheibe and Claus Schnarrenberger
-
29
algae, similar associations between Calvin cycleenzymes have not
been observed in any cyanobacteriaor photosynthetic proteobacteria.
The reports differsubstantially with respect to the number and
natureof protein-protein interactions observed. There arestill many
open questions in this area, and there iscurrently no consensus
concerning the nature,function or significance of such complexes.
Variouscomplexes have been isolated by ultracentrifugationin
sucrose gradients, by exclusion chromatographyin the presence of
stabilizing compounds such asglycerol, and by ion-exchange
chromatography. Themultienzyme-like complexes should be
distinguishedfrom multimeric forms of individual enzymes
whichthemselves can aggregate in purified form, forexample GAPDH
(Baalmann et al., 1994), FBPase(Grotjohann, 1997) or RPE (Teige et
al., 1998).
Several reports concern complexes consisting oftwo or three
enzymes. A complex containing Rubisco,RPI (90 kDa) and PRK (54 kDa)
was reported frompea leaves (Sainis and Harris, 1986) that
catalyzedR5P-dependent fixation in the presence ofATP,and contained
about 45% of RPI and PRK activitiesin the complexed form. In a
similar complex fromspinach, 75% of PRK and 7% of RPI were found
toassociate and copurify with Rubisco (Sainis et al.,1989). The
ratio ofPRK to Rubisco was estimated tobe 1:1 to 1:3. In another
report, a complex of PRKwith GAPDH was isolated from
Scenedesmusobliquus (Nicholson et al., 1987). The stoichiometrywas
estimated to be with a (too low)molecular mass of560 kDa. Also, PRK
and GAPDHwere found to form a complex coexisting with thefree
enzyme forms in spinach (Clasper et al., 1991).A novel, 12 kDa
chloroplast protein (CP12) hasrecently been described from higher
plants that shareshigh sequence similarity with the CTE of the
GapBsubunit (Pohlmeyer et al., 1996). CP12 interacts withGAPDH in
affinity chromatography and with PRKin the yeast two-hybrid system.
Under oxidizingconditions, CP12 interacts with both proteins toform
~600 kDa complexes and has been suggested tobe involved in
regulation (Wedel et al., 1997; Wedeland Soll, 1998).
A complex was isolated from Chlamydomonasreinhardtii consisting
of with anaverage molecular mass of 460 kDa (Avilan et al.,1997;
Lebreton et al., 1997), equal to the sum ofindividual masses of the
free enzymes. Thedissociation of the complex can be achieved
byreducing agents like DTT, NAD(P)H, reduced
ferredoxin or reduced thioredoxin, accompanied byan increase
particularly in PRK activity. PRK isinactive in the oxidized form
(see Section VI) andgained some activity during complex formation
withGAPDH in a manner similar but not identical tochaperonin action
(Lebreton et al., 1997). However,this increase corresponds to only
a few percent oftheactivity of the reduced form present in the
light. Thesubsequent dissociation of the complex by reducingagents
causes a conformation change in PRK, another20-fold increase in PRK
activity with a 4-folddecrease in and a 2-fold decrease inDuring
complex dissociation, GAPDH showed arelative increase in favor of
overactivity (Avilan et al., 1997). On the other hand, thecomplex
could form spontaneously, upon addition of
or oxidized glutathione (Avilan et al., 1997;Lebreton et al.,
1997). The association betweenGAPDH and PRK in Chlamydomonas
involves twoenzymes that catalyze non-consecutive steps in
thepathway and the complex is present under darkconditions where
there should be no Calvin cycleactivity. Thus, the complex is
unlikely to be involvedin channeling in the classical sense
(Gontero et al.,1994; Ricard et al., 1994).
There have been reports of larger Calvin cyclemultienzyme-like
complexes involving severaladditional enzymes. The first such
larger complexcontained PRK, Rubisco, PGK, and GAPDH andwas
reported by Mller (1972), who recognized thatthe fragile complex is
dissociated by NADPH orATPand that the enzymes involved are
activated duringdissociation. The complex had a molecular mass
of400 kDa, less than the value of 700 to 800 kDaexpected. Sasajima
and Yoneda (1974) found thatRPI, TKL and RPE copurify. More recent
reportshave detected complexes with an in the range of500 to 1000
kDa (Gontero et al., 1988; Gontero et al.,1993; Rault et al., 1993;
Sainis and Srinivasan, 1993;Sss et al., 1993, 1995). The enzymes
involved inthese complexes and their stoichiometry differ
inindividual laboratories, the function is generallyinterpreted as
metabolic channeling.
Gontero et al. (1988) found a complex consistingof Rubisco, RPI,
PRK, PGK and GAPDH. Thecomplex was fairly stable and homogeneous
duringultracentrifugation. DTT increased the activity ofthe
individual enzymes. The complex catalyzedfixation with R5P, ATP,
NADPH and Themolecular mass ofthe complex was estimated as 520kDa
and the ratio of the individual enzymes inferred
Chapter 2 Calvin Cycle
-
30
to be 2PRK:2GapA:2GapB:2RbcS:4RbcL, inaddition to some RPI and
PGK (Rault et al., 1993).Because the molecular mass of the
individualenzymes is anticipated to be much larger than that ofthe
complex, it was suggested that Rubisco mightexist in an form (Rault
et al., 1993), differingfrom that of the crystallized enzyme
(Shibita et al.,1996). If the complex is subjected to
SDS-PAGE,several protein bands are observed corresponding tothe
bands of the individual enzymes.
A complex isolated from spinach containedRubisco, PRK, GAPDH,
SBPase, ferredoxin-NADPreductase (FNR) and chaperonin 60 (Sss et
al.,1993) and fixed from R5P. The complex wasstable at low salt
conditions (250 mMKCl). Ammonium sulfate (1 M) or pH 4.5
completelydissociated the complex. All enzymes ofthe complexwere
found almost exclusively attached to the outersurface of thylakoid
membranes during goldimmunolabeling except Rubisco, which showed
alsostromal localization (Sss et al., 1993a, 1993b; Adleret al.,
1993). Similar association with thylakoids hadpreviously been
reported for Rubisco (Grisson andKahn, 1974; McNeil and Walker,
1981), PRK (Fischerand Latzko, 1979) and GAPDH (Grisson and
Kahn,1974). A newly described ~600 kDa complex fromtobacco
contained Rubisco, PRK, RPI and carbonicanhydrase (Jebanathirajah
and Coleman, 1998).
In other reports, complexes consisting of PGK-GAPDH (Malhotra et
al., 1987; Macioszek andAnderson, 1987; Macioszek et al., 1990),
GAPDH-TPI, aldolase-TPI, GAPDH-aldolase (Anderson etal., 1995), and
PRI and PRK (Anderson, 1987;Skrukrud et al., 1991) have been found.
CytosolicPGK and GAPDH were also found to form abienzyme complex
(Weber and Berhard, 1982;Malhorta et al., 1987). Complexes of
aldolase-GAPDH, PGK-GAPDH and aldolase-TPI fromchloroplasts were
isolated and characterized in pea(Anderson et al., 1995). It has
been suggested thatinteraction among GAPDH, TKL and aldolase
aroundSBPase may lead to a direct transfer of GA3P amongthese
enzymes (Marques et al., 1987). Finally, PRIand PRK were shown to
have kinetics in a complexstate that differed from those
anticipated for substratesused by non-complexed enzymes (Anderson,
1987).The theoretical kinetics ofCalvin cycle multienzymecomplexes
have been modeled (Gontero et al., 1994;Ricard et al., 1994). In
multienzyme complexes, thekinetics become increasingly complicated
because
of the combined action of several enzymes.Note that the sum of
molecular weights of the
native enzymes from spinach chloroplasts shown inFig. 1 is about
1500 kDa, Rubisco alone contributinga third of that. Since Rubisco
is far more abundantthan any of the other Calvin cycle enzymes in
plastids,it is clear that not all active Calvin cycle enzymes
canexist in a complexed state (Rutner, 1970).
Summing up these findings on Calvin cyclemultienzyme complexes,
it appears clear thatinteractions between various enzymes do exist,
butthere is no consensus on which or how many enzymesinteract and
whether the same enzymes interact indifferent species. The
metabolic relevance of thesecomplexes is still unclear. Pressing
problemsconcerning these associations have yet to be solved.
First, the isolated enzyme complexes are usuallydescribed to
dissociate in the presence of reducingthiols (light), but the key
regulatory factor of overallCalvin cycle activity is light-mediated
(redox)activation through reduced thiols (see below).
Thisdiscrepancy is difficult to reconcile with metabolicchanneling,
since the kinetic data indicate higherenzyme activities for
associated enzymes via transferof substrates in a consecutive
reaction sequence, butflux through the pathway in the dark is
basically nil(associated state) due to severe down-regulation
ofFBPase, PRK and SBPase (and moderate down-regulation of GAPDH) in
the dark (i.e. in absence ofreduced thioredoxins). But if the
overall activities ofthe complexed enzymes are higher, as many
suchstudies indicate, we are left with the question of
thephysiological relevance of improved kinetics for(dark-)
associated enzyme complexes, since the activeforms are dissociated
in the light (see Section VI).
Second, if associations between enzymes are ascritical to Calvin
cycle function as the interpretationsof many multienzyme studies
would suggest, thenproblems in understanding the pathway ensue
whenthe results from antisense inhibition of Calvin cycleenzymes
are considered (see section VII). This isbecause antisense studies
have shown that most (butnot all) Calvin cycle activities must be
reduced onthe order of five- to ten-fold to affect a
significantreduction in assimilation rate under normal
growthconditions. If the brunt of assimilation occurs incomplexes,
limiting any one component should beexpected to have a more drastic
effect. Further workis needed to clarify the general significance
of theseenzyme association phenomena.
William Martin, Renate Scheibe and Claus Schnarrenberger
-
31
VI. Biochemical Regulation in Chloroplasts
Light governs not only Calvin cycle gene expressionin higher
plants, it is also the foremost determinantof enzyme activity, and
hence flux through thepathway (Buchanan, 1980; Wolosiuk et al.,
1993).Light regulation of the Calvin cycle is achieved bymodulation
of enzyme activity of four enzymesthrough the
ferredoxin/thioredoxin system: FBPase,SBPase, PRK and GAPDH
(Buchanan, 1991). Thefirst three enzymes are obvious targets for
regulation,since they catalyze reactions that are irreversibleunder
physiological conditions, their regulatoryprinciple is the same:
reduced activity as a result ofoxidation of regulatory cysteines by
in the dark,full activity through reduction of regulatory
cysteinesby reduced thioredoxin in the light. Activity
iscontinuously adjusted in the light through continuedreoxidation
and reduction (thioredoxin). Thesteady-state between these
interconvertible enzymeforms is individually influenced by
metabolites(Scheibe, 1990). The fourth reaction, catalyzed byGAPDH,
is reversible. Its activity is dependent uponits state of
aggregation, redox modulation being theprerequisite enabling this
metabolite-inducedinterconversion under physiological
conditions.
A. The Ferredoxin/Thioredoxin System
Ferredoxin (Fd) reduced by photosynthetic electronflow provides
electrons for reduction via Fd/NADP reductase, for nitrite
reduction via nitritereductase, for sulfite reduction via sulfite
reductase,for the reductive generation of glutamate
fromoxoglutarate via glutamate synthase (GOGAT), andfor the
reduction of the thioredoxin (Td) viaferredoxin-thioredoxin
reductase (FTR) (reviewedby Buchanan, 1980; Woodrow and Berry,
1988;Scheibe, 1990; Buchanan, 1991; Knaff and Hirasawa,1991;
Wolosiuk et al., 1993; Jacquot et al., 1997b).FTR is composed of
two different subunits, subunitA is rather variable between
organisms, subunit B ismore highly conserved and contains an Fe-S
clusterin addition to conserved cysteines involved in
redoxtransfer, but it is not a flavoprotein (Tsugita et al.,1991;
Falkenstein et al., 1994). Thioredoxins aresmall heat-stable
proteins that occur in all organismsand in many compartments. In
the chloroplast variousisoforms occur that differ in their primary
structuresand specificities: Tdm, and (reviewed by
Eklund et al., 1991). Invitro,Tdm primarily
activatesNADP-dependent malate dehydrogenase (NADP-MDH), and
inactivates chloroplast glucose-6-phosphate dehydrogenase (G6PDH),
whilepreferentially activates chloroplast FBPase, SBPase,PRK, GAPDH
in addition to the chloroplast couplingfactor CF1 (reviewed by
Buchanan, 1991). Whetherthis pattern of specificities also holds in
stroma,where the protein concentration is very high, remainsto be
established.
Traditionally, light/darkmodulationofchloroplastenzymes was
considered as an all-or-nothing on/off-switch, but more recently it
has become apparent thatit is also a means to fine-tune enzyme
activities in thelight (Scheibe, 1990, 1991,1995). This is
becausepresent at high concentrations in the chloroplast(Steiger et
al., 1977) continuously reoxidizes thecysteines
generatedbythioredoxin-mediatedelectronflow to the target enzymes.
Light-modulated enzymesthus exist as two interconvertible enzyme
forms thatare subject to covalent modification (reduction
andreoxidation of cystine/cysteine residues), comparableto those
enzymes that are subject to
proteinphosphorylation/dephosphorylation (Scheibe, 1990).In both
cases, energy is consumed to drive the cyclebetween the two forms,
but in the light, energy in theform of reducing equivalents is
abundant and posesno significant drain on the photosynthetic
membrane.
Target enzymes as well as the chloroplastthioredoxins are
characterized by the very negativemidpoint redox potentials of
their regulatory cysteines(Faske et al., 1995). For NADP-MDH,
FBPase andPRK these are around 380 mV, similar to that of
thenonphysiological reagent dithiothreitol (DTT), andeven more
negative than the value of 350 mV forTdm and (Gilbert, 1984). These
redox potentialsare all more negative than those of NADP(H)
(320mV), and of glutathione (-260 mV), indicating thatthese protein
thiols cannot be reduced by cellularreductants other than reduced
ferredoxin. In somecases, mixed disulfides can be formed with
lowmolecular weight thiols such as glutathione(Ocheretina and
Scheibe, 1994). That certainchloroplast proteins occur in an
oxidized form is arather special attribute, since usually only
extracellularproteins tend to exhibit this property (Fahey et
al.,1977). For chloroplast enzymes it is this specificproperty
which is the basis for a very flexibleregulatory system.
Chapter 2 Calvin Cycle
-
32
B. Target Enzymes
The light/dark-modulated chloroplast enzymes arecharacterized by
their unusually negative redoxpotentials. As a result of this, they
are only in thereduced state when electrons of very negative
redoxpotential from ferredoxin are available in the light;otherwise
they relax to their oxidized state. Theredox potentials themselves,
however, are subject tochange by specific metabolites, mostly the
substrateor the product of the respective enzyme reaction.These
metabolites are also known to act as effectorsofthe reductive
and/or the oxidative part ofthe redoxcycle. At equilibrium (with
the redox buffer, in vitro)or at steady-state (in vivo) changes in
the relativeeffector concentrations result in a more or
lesspronounced shift of the ratio between oxidized andreduced
enzyme form (in a concentration-dependentmanner) (Faske et al.,
1995). Redox-modulatedchloroplast enzymes generally exhibit high
simi-larities with their non-redox-modulated homologuesfrom other
sources, but also tend to possess unique,cysteine-bearing sequence
motifs that are responsiblefor their regulatory properties
(Scheibe, 1990).
Chloroplast FBPase is the classical target for lightregulation
via thioredoxin 1980). Thereis a strong dependence of FBPase
activity upon theF1,6BP concentration, i.e. FBPase cannot easily
beactivated by DTT (or in the light) in the absence ofF1,6BP. The
resulting regulatory pattern is a strictfeedforward mechanism
ofFBPase activation due toincreasing F1,6BP levels (Scheibe, 1991).
Severalstudies have investigated the mechanism of activationusing
FBPase overexpressed in E. coli (Jacquot et al.,1995; Hermoso et
al., 1996; Jacquot et al., 1997a;Lopez-Jaramillo et al., 1997;
Sahrawy et al., 1997).Chloroplast FBPase possesses a
conspicuousinsertion of 1215 amino acids in the central regionof
the primary structure with two conserved cysteineresidues separated
by five amino acids andpreceded by a third conserved cysteine
furtherN-terminal (Marcus et al., 1988; Raines et al., 1988).In
vitro mutagenesis of and results inenzymatically active FBPase
enzymes that can nolonger be regulated by thioredoxin, indicating
thatthese may be specific targets of thioredoxin regulation(Jacquot
et al., 1995). But in a more recent study,was also found to be
responsible for redox dependence(Jacquot et al., 1997a). Replacing
these three cysteineswith serine residues in rapeseed FBPase also
resultedin enzymes that were active in a manner largely
independent of