Planta (1989)178:463-474
Pl~lclt~ 9 Springer-Verlag1989
A model of photosynthetic C02 assimilation and carbon-isotope discrimination in leaves of certain C 3 - C 4 intermediates Susanne yon Caemmerer Plant Environmental Biology Group, Research School of Biological Sciences, Australian National University, G.P.O. Box 475, Canberra City, A.C.T. 2601, Australia
Abstract. A model of leaf, photosynthesis has been developed for C 3 - C 4 intermediate species found in the genera Panicum, Moricandia, Parthenium and MoIlugo where no functional C4 pathway has been identified. Model assumptions are a functional C3 cycle in both mesophyll and bundle-sheath cells and that glycine formed in the mesophyll, as a consequence of the oxygenase activity of ribulose-l,5bisphosphate carboxylase-oxygenase (Rubisco, EC 4.1.1.39), diffuses to the bundle sheath, where most of the photorespiratory CO2 is released. The model describes the observed gas-exchange characteristics of these C 3 - C 4 intermediates, such as low CO2-compensation points (F) at an 0 2 pressure of 200 mbar, a curvilinear response of F to changing 02 pressures, and typical responses of CO2-assimilation rate to intercellular CO2 pressure. The model predicts that bundle-sheath COz concentration is highest at low mesophyll CO2 pressures and decreases as mesophyll CO2 pressure increases. A partitioning of 5-15% of the total leaf Rubisco into the bundle-sheath cells and a bundlesheath conductance similar to that proposed for C4 species best mimics the gas-exchange results. The model predicts C3;like carbon-isotope discrimination for photosynthesis at atmospheric levels of CO2, but at low CO2 pressures it predicts a higher discrimination than is typically found during Ca photosynthesis at lower CO2 pressures. Key words: C 3 - C 4 intermediates - Carbon isotope discrimination Photosynthesis (model) Abbreviations and symbols: PEP=phosphoenolpyruvate; Ru-
bisco=ribulose-l,5-bisphosphate carboxylase-oxygenase (EC 4.1.1.39); RuBP=ribulose-l,5-bisphosphate; p(COz)=partial pressure of CO2; P(O2)=partial pressure of O2. See also p. 471
Introduction
In recent years, higher plants with anatomical, physiological and metabolic characteristics intermediate between those of C3 and C4 plants have been described extensively (for reviews see Monson et al. 1984; Edwards and Ku 1987). The study of these C a - C 4 intermediates can provide useful insights into both the evolution of C4 photosynthesis, and other aspects of the form and function of photosynthetic carbon metabolism. Plant species with intermediate characteristics are found in the following genera: Moricandia (Brassicaceae-Cruciferae), Panicum and Neurachne (Poaceae), FIaveria and Parthenium (Asteraceae), Mollugo (Aizoiaceae) and Alternanthera (Amaranthaceae). In general, such intermediates have Kranz-like leaf anatomy, C3-1ike carbon-isotope composition (~laC), a reduced oxygen sensitivity of photosynthesis when compared to C3 species, and a CO2-compensation point (F) intermediate between those seen in C3 and C4 plants. For C3 - C4 intermediates of the genera Ftaveria and Neurachne, which have a limited capacity for C4 metabolism (Hattersley and Stone 1986; Monson et al. 1986), carbon metabolism and carbon-isotope discrimination have been modelled by Peisker and Bauwe (1984) and Peisker (1985). However, there is at present no evidence for C~-like metabolism in C3 - C4 intermediate species of the genera Moricandia, Panicum, Parthenium and Alternanthera (Winter et al. 1982; Holaday and Chollett 1983; Rajendrudu et al. 1986; Hunt et al. 1987; Moore et al. 1987). At the present time there is no theoretical model which adequately describes photosynthesis in these intermediates. In the following I present such a quantitative model of leaf photosynthesis which follows upon
464
S. yon Caemmerer: CO2 assimilation and c-isotope discrimination in C 3 --C 4 intermediates
the work of Brown (1980), Monson et al. (1984) and Rawsthorne et al. (1988), where it is assumed that glycine is shuttled from the mesophyll to the bundle-sheath cells for decarboxylation in that compartment. The model described below allows us to evaluate the quantitative significance of such a pathway. Furthermore, within the mathematical framework of the model it can be demonstrated that carbon-isotope discrimination is a useful measure of the presence of compartmentalized glycine decarboxylation in C 3 - C4 intermediate species.
mesophyll
bundle-sheath
(.-~Mm CxlHzO)x IX Frn {glycine}
---~ ) - -
Cx(HzOlx
~--
L
. Ms
IXFm+Fs
Development of the model Figure 1 shows a schematic representation of the proposed carbon fluxes. A complete C3 cycle is assumed to operate in both mesophyll and bundle-sheath ceils. Some or all of the glycine formed in the mesophyll ceils following the oxygenation of ribulose-l,5-bisphosphate (RuBP) is assumed to diffuse to the bundle-sheath cells. There the glycine is decarboxylated, and the photorespiratory COz released supplies the ribulose-l,5-bisphosphate carboxylase-oxygenase (Rubisco) of the bundle sheath with CO 2. For simplicity, it is assumed that the CO2 in the intercellular airspaces is in equilibrium with that of the mesophyll cytoplasm, and that there exists a substantial resistance to COz diffusion at the wall of bundle-sheath cells. The rate of CO2 assimilation (A) is given by
A=V~--(1--c~)F.~--M~--L
(1)
where Vm is the rate of mesophyll Rubisco carboxylation; Fm is half the rate of mesophyll oxygenation, i.e. the rate of photorespiration which would occur if all the resulting glycine were decarboxylated there; and ( 1 - ~) is the fraction that is actually released in the mesophyll. The symbol M,, denotes the rate of mitochondrial CO2 release other than by photorespiration and L is the rate of COz leakage from the bundle-sheath cells to the mesophyll cells or intercellular airspaces. The net rate of assimilation by the bundle-sheath cells (A~) is given by
A~=eFm--L,
(2)
Fig. 1. Simplified scheme of carbon fluxes. A, CO2 assimilation rate; V,,, rate of mesophyll Rubisco carboxylation; ( 1 - c 0 Fro, fraction of photorespiratory COz released in the mesophyll; Mm and M~, rate of mitochondrial respiration in the mesophyll and bundle-sheath not involved in photorespiratory respiration; L, leakage of COz in or out of the bundle-sheath cells; V~ Rubisco carboxylation rate in the bundle-sheath cells; e F m + F ~ rate of photorespiratory COz released in the bundle-sheath cells where e F m derives from mesophyll glycine. Q(H20)x is the fixed carbon
of Rubisco present, by the partial pressures of CO 2 and Oz or by the concentration of its third substrate RuBP. Since the photosynthesis is examined only under conditions of high irradiance, it is assumed in the model that RuBP is present in saturating amounts and that Rubisco is fully activated. The carboxylation rates of mesophyll and bundle-sheath Rubisco are therefore given by the equations (see Farquhar and von Caemmerer 1982) Vm_
(4)
C m Vmm~ Cm + Kr + OJKo)
and
G K~,.~ v~- c~+ Ko(1+ OJKo)
(5)
and
.Fm--L= K--F~-M~.
(2a)
The symbol c~ denotes the fraction of photorespiratory CO2 release derived from mesophyll glyeine and evolved in the bundle-sheath ceils; V~is the carboxylation rate of the bundle-sheath Rubisco; F~ is the flux of photorespiratory CO 2 release based upon bundle-sheath Rubisco oxygenation, and M~ denotes other mitochondrial respiration in the bundle-sheath cells. Substituting V~-F~-Ms for aFm--L in eq. (1) gives a second equation describing overall CO2-assimilation rate
A=(Vm--Fm-- Mm)+(Vs-- Fs-- Ms),
(3)
For comparison the CO2-assimilation rate of a C3 leaf is given by
where Cm and Om are the CO2 and 02 concentrations in the mesophyll cells and Cs and Os those of the bundle-sheath cells. Kc, Ko are the Michaelis-Menten constants for CO2 and 02 respectively, Vmmaxand Vs,,ax are the maximum activities of Rubisco in the mesophyll and bundle-sheath cells. Following the oxygenation of 1 tool of RuBP, 0.5 tool of CO/ are evolved in the photorespiratory pathway. Farquhar et al. (1980) showed that the ratio of the rate of oxygenation to carboxylation can be expressed as
Vo/V~= 2 F,/C
where C stands for COz concentration, F, is the CO2-compensation point in a C3 plant in the absence of other mitochondrial respiration, and
F,=[0.5 A = V-F-M.
(3a)
The rates of carboxylation and photorespiration. The rate of carboxylation of Rubisco can be limited by the amount and activity
(6)
VornaxKc/(VemaxKo)]0=~/,
O.
(7)
The symbols Vom~xand V~ma, are the maximum activities of the oxygenase and carboxylase, respectively, and O stands for 02 concentration. The rates of photorespiration Fm and F~ are
S. von Caemmerer: CO2 assimilation and c-isotope discrimination in C3--C4 intermediates then given by Fro=0-5 Vom = 7 * Om Vm/Cm
(8)
and
f~=0.5 Vow=7, o~ WCs
(9)
where Vom and Vo~ are the oxygenation rates of the mesophyll and bundle-sheath Rubisco, respectively. The kinetic properties of Rubisco of whole-leaf extracts have been examined for several intermediate species (Yeoh et al. 1980; Holbrook and Chollet 1986). Holbrook and Chollet (1986) found that the Rubisco of Panicum milioides and of Moricandia arvensis had K m (CO2) values and specificity factors (another name for 0.5 7,) in the range of those found for C3 species. The following values, typical for C3 species at 25 ~ C, are used for both the mesophyll and bundle-sheath Rubisco; K~ =310 gbar, Ko=360 mbar (Seemann et al. 1981; Bird et al. 1982; Makino et al. 1985). The ratio Vomax/Vcmax=0.46 was chosen to agree with the experimentally measured F,=41.6 gbar at 210 mbar Oz (Brooks and Farquhar 1985), and gives F, = 39.6 ~tbar at 200 mbar O2. The Km(O2) of Rubisco has frequently been measured, but uncertainty exists in the values of Km(Oz) and the ratio Vomax/Vcmax, Kirschbaum and Farquhar (1984) have used values of Ko= 155 mbar and VomJV~m~ =0.21. This combination also gives a F, =41.6 ~tbar at 210 mbar 02.
7he leakage of C02 and 02 to and from the bundle-sheath cells. A higher CO2 or O2 concentration within either the bundlesheath cells or the mesophyll compartment will result in net diffusion of these gases. The rate of COz leakage, L, from the bundle-sheath cells to the mesophyll cells is given by L=g~(C~-Cm)=L~--L m
(lO)
where g~ is the conductance to leakage of C O 2 a c r o s s the bundle sheath, and L~=g~ C s is the leak rate that would occur from the bundle-sheath cells to the mesophyll cells if C~=0, and Lm = gs Cm is the leak rate that would occur from the mesophyll cells to the bundle-sheath cells if C~=0. Very little is known about the magnitude of the conductance (g~) in C 3 - C4 intermediates. If not otherwise specified a value of gs = 1 m m o l . m - 2 . s -~ has been used in the simulations; in other cases bundle-sheath conductance is treated as a variable. It is known that bundle-sheath chloroplasts in some C4 species have very little photosystem-II activity. The granal nature of bundle-sheath chloroplasts of C3 - C4 intermediates indicates that they are capable of photosystem-II activity (Edwards and Ku 1987). As pointed out by Raven (1977) and Berry and Farquhar (1978), this has implications for the steady-state 02 concentration inside the bundle-sheath cells. Following the development of Berry and Farquhar (1978), it is assumed that net Oz evolution (Eo) in the bundle-sheath cells equals its leakage (Lo) out of the bundle, that is Eo = Lo = go (Os -- Om)"
(11)
The conductance to leakage of Oz across the bundle sheath (go) is related to the conductance to CO2 by way of the ratio of diffusivities and solubilities (Berry and Farquhar 1978) by go=gs(Do~
So2)/(Dco 2 Sco2),
(12)
where Do~ and Dco ~ are the diffusivities of O2 and C O 2 in water, respectively, So~ and Sco~ are the respective Henry constants, and go = 0.047 g~ at 25 ~ C (Farquhar 1983).
465
The net O2 evolution in the bundle-sheath cells depends upon the balance of O2 evolution at photosystem II and the various O2-consuming processes. In particular the balance depends upon how the glycine shuttle is perceived to operate. In the model the solution which generates no extra 02 consumption based upon the glycine shuttle to the bundle-sheath cells has been chosen. Briefly, the integrated carbon reduction and oxidation cycle in the bundle-sheath cells consumes N A D P H at the rate of 2V~+2Vos which results in an 02 evolution rate of V~ + Vow.Oxygen is consumed by bundle-sheath Rubisco, during glycotate oxidation (resulting from glycolate produced in the bundle-sheath), and by mitochondrial respiration at the rate of 1.5 Vo~+ M s. In Moricandia arvensis, enzymes necessary for the synthesis of glycine are present in both bundle-sheath and mesophyll compartments and it is therefore likely that glycine rather than glycolate is the compound transported (Rawsthorne et al. 1989). Thus it is assumed that carbon skeletons are returned to the mesophyll such that 3-phosphoglyceric acid derived from mesophyll glycine is reduced in the mesophyll such that also no extra O2-evolution rate occurs in the bundle-sheath cells. In summary:
Eo=V~+Vo~--I.5Vo~-M~=A s.
(13)
Mathematical solution of the equations. For calculating CO2-assimilation rates, solutions for C~ and O~ are required. Combining eqs. (5), (9), (10), eq. (2) can be rewritten as Vsmax( C s - 7 , Os) = (~Fm _~_M~ -- g, (Cs - Cm).
C~+ Kr
(14)
+O~/Ko)
To obtain an expression for O~, one uses eqs. (2), (10), (11) and (13) to obtain o, = o., + (c~Vm-- gs (C~ -- C .,))/go"
(15)
For convenience the following expression is used
o~ = o , - c~. g2go
(16)
where O , = Om-}- (~Frn-}- g s Cm)/g o.
(17)
Substituting for O s in eq. (14) and solving for C~ a quadratic of the following form is obtained:
a.C~+bCs+d=O,
(18)
so that Cs= ( - b + ~ ) / ( 2
a),
(19)
where
a =(1 -g~ Ko/goKo)
(20)
b = { Vs max (1 -]- 7$ gs/go)}/g~- {(c~Fm + Ms)/g s + Cm}
( 1 -- Ko gjKo go) + K~(1 4- O ,/Ko) and d = -(Vsmax 7, 0 , ) / g s - {(~ VmArMs)/g~+ am} {Kc (1 ~- O./Ko) }. (22) Assuming values for Cm, Om, Vmmax, Vs . . . . M S and Mm, bundlesheath CO2 concentration (C~), O s, A and all other parameters can be calculated. The equation for the compensation point is cubic in nature and F has been calculated with an iterative computer calculation searching for Cm at which A = 0.
466
S. von Caemmerer: CO2 assimilation and c-isotope discrimination in C 3 - Ce intermediates
Model predictions Parameters and criteria. Since the modelis designed for testing of ideas rather than empirical fitting, as many parameters as possible have been preassigned from literature values. For example, the kinetic parameters of Rubisco have been chosen from in-vitro measurements (see Model development section) and can be viewed as constants except for the maximum carboxylase activity, which one would expect to vary with leaf age, irradiance and nutrient conditions during growth. A value of 107 g m o l - m - 2 . s -1 was chosen for the total maximum Rubisco activity. Since we know little about mitochondrial respiration and its compartmentation, the values of mitochondrial respirations, M m and M~ were somewhat arbitrarily set at 0.2 g m o l . m - 2 . s -1 The parameter ~ determines what fraction of the photorespiratory CO2 is respired in the bundlesheath cells. Recent evidence by Rawsthorne et al. (1988) has shown that in Moricandia arvensis the glycine decarboxylase was solely located in the bundle-sheath cells. Similar results have subsequently been found for Panicum milioides, Flaveria floridana and F. Iinearis (Hylton et al. 1988). In all results presented here except for Fig. 3 we have assumed that ~ = 1 , i.e. that all photorespiratory CO2 is released in the bundle-sheath cells. Two main variables emerge as important parameters for which values need to be chosen and both parameters, the conductance to leakage between bundle-sheath and mesophyll cells (gs), and the partitioning of Rubisco between mesophyll and bundle-sheath cells affect the leakage rate between these two compartments. The major criteria used to identify C 3 - C4 intermediates experimentally have been the CO2-compensation point (F) and the O2 inhibition of CO2-assimilation. In Fig. 2, CO2-assimilation rate, bundle-sheath CO2 concentration, F, and 02 inhibition are examined with respect to the proportion of Rubisco in the bundle-sheath and mesophyll cells and at two bundle-sheath conductances. Values of g s = l m m o l ' m - 2 ' s - ~ and gs = 1 0 m m o l - m - 2 . s -1 were selected. Apel and Peisker (1978) have estimated a conductance of 3.1 mmol. m - 2. s - 1 from anatomical studies in Zea mays, and a conductance of 1 m m o l . m - 2 , s- 1 is sufficient to simulate the 02 insensitivity of CO2-assimilation rate in a Cr model (Berry and Farquhar 1978). Brown et al. (1983) showed that the Panicum C 3 - C 4 intermediates had a well-developed vascular bundle sheath and that the number of plasmodesmata were similar to that in-
C4 Panicum species and more than that found in C3 Panicum species. The estimate of 1 mmol. m - 2. s- 1 may therefore be appropriate for these intermediates. The CO2-assimilation rate at a mesophyll partial pressure of CO2 of 250 gbar was found to be greater in the intermediate, than in a C3 leaf with the same total amount of Rubisco (Fig. 2a) when up to 15% of the total Rubisco was partitioned to the bundle-sheath cells. This was true at both conductances, although the enhancement was less at the higher conductance. The rate of CO2 assimilation fell below that of a C3 leaf at 250 lxbar if much more than 15% was partitioned into the bundle-sheath cell. There are two main reasons for this. As the fraction of Rubisco in the bundlesheath cells increases, the fraction of Rubisco in the mesophyll cells decreases, hence the oxygenase activity and associated capacity to shuttle glycine also decreases. At some point the capacity of the bundle-sheath Rubisco is too great for the capacity of the shuttle and the bundle-sheath Rubisco is disadvantaged for direct fixation of intercellular CO2 because of the low bundle-sheath conductance to CO2. This is illustrated in Fig. 2b, where the bundle-sheath CO2 pressures were greatest when a small portion of the total amount of Rubisco was in the bundle sheath. It quickly fell below the mesophyll pressure of 250 gbar as the portion of Rubisco in the bundle sheath increased. As a consequence of the kinetics of Rubisco the oxygenase activity is greatest at low CO2 and this can be seen in Fig. 2 a in the enhancement of assimilation rate over that of a C3-1eaf at Cm = 50 gbar COz. The CO2-compensation point decreased initially rapidly from the Ca value of 42 gbar as the capacity to refix respired CO2 increased with an increasing fraction of Rubisco in the bundle-sheath cells. The value of F was dependent upon the values of bundle-sheath conductance chosen (Fig. 2c). The compensation point has often been measured to characterize C 3 - C 4 intermediates and Edwards and K u (1987) have tabulated the results of the literature in their review. They reported for Panicum milioides, F values between 11 and 17 gl/1 at 21% Oz. Brown (1980) measured values as low as 6.9 and 4.4 gl/1. For Moricandia arvensis, values of 28 and 16 gbar have been reported (Edwards and Ku 1987). These values are within the range predicted in Fig. 2c at a Rubisco partitioning factor of 0.1. The model predicts that the glycine shuttle can achieve very low compensation points because of the high rates of oxygenation that occur at low COz pressures. In a C3 leaf the CO2-assimilation rate is greater
S. von Caemmerer: CO2 assimilation and c-isotope discrimination in C 3 - C 4 intermediates
C --
~"
C3
30
-
~...
N"
'E "6
\
467
t,O
30
O
20
L_
Cm=250
"~
20-
\N\
E
-
0
,
I
,
Cl
l
8000 cooo
I
1
c~
0
I
d
b
50
O
1.6
-1.t,
~.~... =
10
E O O
--....~
1,....
.z~
-1.2
400C
O .IQ
-10
I
0
01
0.2
0.3
O/~
0.1
I
0.2
I
03
[
0.~
E O G1
0 0.5
Fraction of total Rubisc0 in bundle-sheath. Fig. 2. a COz-assimilation rate (A) versus fraction of total leaf Rubisco partitioned into the bundle-sheath cells at 200 m b a r Oz and at two mesophyll COz-concentrations; Cm (250 and 50 pbar) calculated with two different bundle-sheath conductances: g, = 1 retool, m - 2 s - 1 (solid lines), g~ = 10 mmol. m - 2. s - 1 (dashed lines). Vmax of total Rubisco = 107 pmol. m - 2. s - 1. Also shown are the rate of CO2-assimilation of a modelled C3 leaf at Cm = 250 and 50 gbar with that Rubisco Vm,x(horizontal lines), b Corresponding bundle-sheath CO2 pressures (Cs) versus the fraction of total leaf Rubisco partitioned into the bundle-sheath cells at Cm=250 and 50 gbar and g , = l m m o l - m - Z . s -1 (solid lines), g s = 1 0 m m o l - m - Z . s -1 (dashed lines), e CO2-compensation point (F) versus the fraction of total leaf Rubisco partitioned into the bundle-sheath cells at g s = l m m o l - m - 2 . s - 1 (solid line) and gs= 10 m m o l . m - 2 . s -~ (dashed line), d The ratio of CO2-assimilation rate (A) at 20 mbar 0 2 and 250 gbar CO2 to A at 200 mbar 02 and 250 gbar CO2 at g ~ = l m m o l - m - 2 . s -~ (solid line) and g~=lOmmol.m-2.s -1 (dashed line). In all cases, it is assumed that all photorespiratory CO2 is released in the bundle-sheath cells (~ = 1)
at 20 mbar p(O2) than at 200 mbar p(O2). An increase in the partial pressure of 02 increases the rate of photorespiration and decreases the rate of carboxylation due to increased competition from O2 (eqn. 3a). Figure 2d shows the ratio of CO2-assimilation rate at 20 mbar p(O2) to that at 200 mbar p(O2) for different Rubisco partitioning. The ratio declines from the C3 value of 1.46 to a value close to 1.27, depending on the bundlesheath conductance. The value 1.27 is the ratio of the mesophyll carboxylase rate of Rubisco at the two O2 pressures brought about by the competitive effect of 02 on carboxylation, i.e. at the lower conductances all photorespired CO2 is effectively refixed. Dependence of F on 02 concentration. The compensation point increases linearly with 02 concentration in leaves of C3 species (Forrester et al. 1966; Bj6rkman et al. 1970; Peisker and Apel 1977). This
response can be explained by considering only the kinetics of Rubisco (Laing et al. 1974; Azcon-Bieto et al. 1981). In C 3 - C 4 intermediates such as Panicure milioides or Moricandia arvensis non-linear relationships have been observed (Keck and Ogren 1976; Quebedeant and Chollet 1977; Apel 1980; Brown and Morgan 1980; Holaday et al. 1982; Hattersley et al. 1986; Hunt et al. 1987). Figure 3a shows that the model incorporating a glycine shuttle predicts a curvilinear relationship. This differs from the biphasic dependence of F on 02 concentration predicted by the model of Peisker and Bauwe (1984), which includes the operation of a C4 pathway. At present the available data do not allow us to distinguish between biphasic and curvilinear responses. Figure 3a shows F with changing p(O2) pressure at several different fractions of Rubisco in the bundle-sheath cells. The data points relating F and p(O2) for Panicum milioides have been replotted
468
S. von Caemmerer: COz assimilation and c-isotope discrimination in C 3 - C , intermediates
10C a
" ~ " / / / I ~9/ i,//,/
8c "C u
/E/=0.05
" y=.01
//;/,
6c
40
y=02
2O 0
b
s = lOOmmol.m-2s1
100
3//
80
0/I //
0.1
60 L_ 40 20 0
~
10
~/ / / /
8,
///
"E E1
X3
60
L-
40
CL
/Z0,75 //////~ =1.0
0 200
/*00 600 800 0 m (mbor)
1000 1200
Fig. 3. a CO2-compensation point (s versus mesophyll P(O2) , O m at three different fractions (y) of total leaf Rub&co (V~, x = 107 gmol.m Z.s-~) partitioned into the bundle-sheath cells. The bundle-sheath conductance g~= 1 mmol. m - 2 . s-1, and all photorespiratory CO2 is released in the bundle sheath (c~= 1). Also shown is the predicted F versus Om for a C3 leaf. The data points joined by the dotted line are measurements made on Panicum milioides taken from Hattersley et al. (1986). b CO2-compensation point (F) versus Om at different bundlesheath conductances (g~). The fraction (y) of total Rubisco in the bundle-sheath is 0.1 and c~-1. Also shown is the predicted F versus O,,, for a C3 leaf. e CO2-compensation point (F) versus Om at different fractions of photorespiratory CO2 released in the bundle-sheath cells (c0. The fraction of total Rubisco in the bundle-sheath cells=0.1 and g = l m m o l - m - Z - s -~. Also shown is the predicted F versus Om for a C 3 leaf. The dashed line is common to a, b and e (y=0.1, g~=l m m o l . m - 2 . s -1, ~=1)
from Hattersley et al. (1986). At low p(O2), the compensation point was low, as CO2 released in the bundle-sheath cells is refixed with little leakage out. As the rate of glycine shuttling increases with increasing p(O2) bundle-sheath Rubisco eventually
becomes saturated with bundle-sheath CO2 and the leakage rate rises relative to the rate of CO2 released, and F increases. Therefore an increase in bundle-sheath Rubisco keeps F low until higher p(O2). Bundle-sheath conductance also influences the relationship between F and p(O2), by influencing the leak out of the bundle-sheath cells (Fig. 3 b). Changing the proportion of photorespiratory CO2 that is released in the bundle-sheath cells affects F by the amount of CO 2 lost directly from the mesophyll cells (Fig. 3 c). Clearly a number of different choices of Rubisco partitioning, bundlesheath conductance and 0~could satisfy values ob9served in the literature (Brown 1980; Edwards and Ku 1987; Hunt et al. 1987).
Dependence of C02-assimilation rate on C02 and 0 2 . In Fig. 4 the rate of net CO2 assimilation (eqn. 2) and the bundle-sheat p(CO2) (C~) are plotted against mesophyll p(CO2) (Cm) at 200 mbar and 20 rnbar P(O2). For simplicity the model contains no limitation on RuBP regeneration capacity or electron transport and so no saturation occurs at Cm above 300 gbar (unlike the previous C3-model of Farquhar et al. 1980). It is likely that, at high irradiances and ambient partial pressures of CO2 up to 300 gbar, CO2 assimilation is limited by the amount of Rubisco present, but the model should not be used above 300 gbar. Incorporation of a light and electron-transport limitation will require more knowledge on the partitioning of energy between bundle-sheath and mesophyll cells. The bundle-sheath CO2 concentration was highest at low Cm where the rate of oxygenation is highest and declines with increasing Cm (Fig. 4). This result differs from C4 model predictions for increasing bundle-sheath CO2 concentrations with increasing mesophyll p(CO2) (Berry and Farquhar 1978). Recently, direct measurements of the inorganic carbon pool in the bundle-sheath cells by Furbank and Hatch (1987) have confirmed these C4 predictions. Such measurements may provide a means of testing the present hypothesis, and it may be possible to identify C 3 - C4 intermediates operating with a glycine shuttle. At 20 mbar p(O2) the bundle-sheath p(CO2) was around 50 ~tbar, only marginally above the C3-compensation point, and the contribution of the glycine shuttle was almost negligible. The increase in assimilation rate over that at 200 mbar p(O2) is mainly the result of the increased mesophyll carboxylation rate caused in turn by less competition from 02. The enhancement in CO2-assimilation rate is less than that seen in modelled C3 photosynthesis because of the substantial
S. von Caemmerer: CO, assimilation and c-isotope discrimination in C 3 - C4 intermediates 60
j f A
50 ~
/j
J
~ - ~
4o
E ~
E
30 500&~20
(..) 250
10 0
T-----I------T--~
100
200 300 400 Cm, (ubar CO2)
Ases1As 0 500 600
Fig. 4. CO2-assimilation rate (A) bundle-sheath CO2 assimilation rate (As) and bundle-sheath COz pressure (Cs) versus mesophyll CO2 pressure (Cm) at two oxygen pressures: 200 mbar (solid lines), 20 mbar 02 (dashed lines). The fraction of total Rubisco in the bundle-sheath is 0.1 and total Vmax = 107 ~rnol. m - 2. S- 1, bundle-sheath conductance gs = 1 m m o l - m - 2 . s -1, and all photorespiratory C02 is respired in the bundle-sheath cells (~ = 1)
5C 4C 3C E
-5 E
Pl~lclt~ 9 Springer-Verlag1989
A model of photosynthetic C02 assimilation and carbon-isotope discrimination in leaves of certain C 3 - C 4 intermediates Susanne yon Caemmerer Plant Environmental Biology Group, Research School of Biological Sciences, Australian National University, G.P.O. Box 475, Canberra City, A.C.T. 2601, Australia
Abstract. A model of leaf, photosynthesis has been developed for C 3 - C 4 intermediate species found in the genera Panicum, Moricandia, Parthenium and MoIlugo where no functional C4 pathway has been identified. Model assumptions are a functional C3 cycle in both mesophyll and bundle-sheath cells and that glycine formed in the mesophyll, as a consequence of the oxygenase activity of ribulose-l,5bisphosphate carboxylase-oxygenase (Rubisco, EC 4.1.1.39), diffuses to the bundle sheath, where most of the photorespiratory CO2 is released. The model describes the observed gas-exchange characteristics of these C 3 - C 4 intermediates, such as low CO2-compensation points (F) at an 0 2 pressure of 200 mbar, a curvilinear response of F to changing 02 pressures, and typical responses of CO2-assimilation rate to intercellular CO2 pressure. The model predicts that bundle-sheath COz concentration is highest at low mesophyll CO2 pressures and decreases as mesophyll CO2 pressure increases. A partitioning of 5-15% of the total leaf Rubisco into the bundle-sheath cells and a bundlesheath conductance similar to that proposed for C4 species best mimics the gas-exchange results. The model predicts C3;like carbon-isotope discrimination for photosynthesis at atmospheric levels of CO2, but at low CO2 pressures it predicts a higher discrimination than is typically found during Ca photosynthesis at lower CO2 pressures. Key words: C 3 - C 4 intermediates - Carbon isotope discrimination Photosynthesis (model) Abbreviations and symbols: PEP=phosphoenolpyruvate; Ru-
bisco=ribulose-l,5-bisphosphate carboxylase-oxygenase (EC 4.1.1.39); RuBP=ribulose-l,5-bisphosphate; p(COz)=partial pressure of CO2; P(O2)=partial pressure of O2. See also p. 471
Introduction
In recent years, higher plants with anatomical, physiological and metabolic characteristics intermediate between those of C3 and C4 plants have been described extensively (for reviews see Monson et al. 1984; Edwards and Ku 1987). The study of these C a - C 4 intermediates can provide useful insights into both the evolution of C4 photosynthesis, and other aspects of the form and function of photosynthetic carbon metabolism. Plant species with intermediate characteristics are found in the following genera: Moricandia (Brassicaceae-Cruciferae), Panicum and Neurachne (Poaceae), FIaveria and Parthenium (Asteraceae), Mollugo (Aizoiaceae) and Alternanthera (Amaranthaceae). In general, such intermediates have Kranz-like leaf anatomy, C3-1ike carbon-isotope composition (~laC), a reduced oxygen sensitivity of photosynthesis when compared to C3 species, and a CO2-compensation point (F) intermediate between those seen in C3 and C4 plants. For C3 - C4 intermediates of the genera Ftaveria and Neurachne, which have a limited capacity for C4 metabolism (Hattersley and Stone 1986; Monson et al. 1986), carbon metabolism and carbon-isotope discrimination have been modelled by Peisker and Bauwe (1984) and Peisker (1985). However, there is at present no evidence for C~-like metabolism in C3 - C4 intermediate species of the genera Moricandia, Panicum, Parthenium and Alternanthera (Winter et al. 1982; Holaday and Chollett 1983; Rajendrudu et al. 1986; Hunt et al. 1987; Moore et al. 1987). At the present time there is no theoretical model which adequately describes photosynthesis in these intermediates. In the following I present such a quantitative model of leaf photosynthesis which follows upon
464
S. yon Caemmerer: CO2 assimilation and c-isotope discrimination in C 3 --C 4 intermediates
the work of Brown (1980), Monson et al. (1984) and Rawsthorne et al. (1988), where it is assumed that glycine is shuttled from the mesophyll to the bundle-sheath cells for decarboxylation in that compartment. The model described below allows us to evaluate the quantitative significance of such a pathway. Furthermore, within the mathematical framework of the model it can be demonstrated that carbon-isotope discrimination is a useful measure of the presence of compartmentalized glycine decarboxylation in C 3 - C4 intermediate species.
mesophyll
bundle-sheath
(.-~Mm CxlHzO)x IX Frn {glycine}
---~ ) - -
Cx(HzOlx
~--
L
. Ms
IXFm+Fs
Development of the model Figure 1 shows a schematic representation of the proposed carbon fluxes. A complete C3 cycle is assumed to operate in both mesophyll and bundle-sheath ceils. Some or all of the glycine formed in the mesophyll ceils following the oxygenation of ribulose-l,5-bisphosphate (RuBP) is assumed to diffuse to the bundle-sheath cells. There the glycine is decarboxylated, and the photorespiratory COz released supplies the ribulose-l,5-bisphosphate carboxylase-oxygenase (Rubisco) of the bundle sheath with CO 2. For simplicity, it is assumed that the CO2 in the intercellular airspaces is in equilibrium with that of the mesophyll cytoplasm, and that there exists a substantial resistance to COz diffusion at the wall of bundle-sheath cells. The rate of CO2 assimilation (A) is given by
A=V~--(1--c~)F.~--M~--L
(1)
where Vm is the rate of mesophyll Rubisco carboxylation; Fm is half the rate of mesophyll oxygenation, i.e. the rate of photorespiration which would occur if all the resulting glycine were decarboxylated there; and ( 1 - ~) is the fraction that is actually released in the mesophyll. The symbol M,, denotes the rate of mitochondrial CO2 release other than by photorespiration and L is the rate of COz leakage from the bundle-sheath cells to the mesophyll cells or intercellular airspaces. The net rate of assimilation by the bundle-sheath cells (A~) is given by
A~=eFm--L,
(2)
Fig. 1. Simplified scheme of carbon fluxes. A, CO2 assimilation rate; V,,, rate of mesophyll Rubisco carboxylation; ( 1 - c 0 Fro, fraction of photorespiratory COz released in the mesophyll; Mm and M~, rate of mitochondrial respiration in the mesophyll and bundle-sheath not involved in photorespiratory respiration; L, leakage of COz in or out of the bundle-sheath cells; V~ Rubisco carboxylation rate in the bundle-sheath cells; e F m + F ~ rate of photorespiratory COz released in the bundle-sheath cells where e F m derives from mesophyll glycine. Q(H20)x is the fixed carbon
of Rubisco present, by the partial pressures of CO 2 and Oz or by the concentration of its third substrate RuBP. Since the photosynthesis is examined only under conditions of high irradiance, it is assumed in the model that RuBP is present in saturating amounts and that Rubisco is fully activated. The carboxylation rates of mesophyll and bundle-sheath Rubisco are therefore given by the equations (see Farquhar and von Caemmerer 1982) Vm_
(4)
C m Vmm~ Cm + Kr + OJKo)
and
G K~,.~ v~- c~+ Ko(1+ OJKo)
(5)
and
.Fm--L= K--F~-M~.
(2a)
The symbol c~ denotes the fraction of photorespiratory CO2 release derived from mesophyll glyeine and evolved in the bundle-sheath ceils; V~is the carboxylation rate of the bundle-sheath Rubisco; F~ is the flux of photorespiratory CO 2 release based upon bundle-sheath Rubisco oxygenation, and M~ denotes other mitochondrial respiration in the bundle-sheath cells. Substituting V~-F~-Ms for aFm--L in eq. (1) gives a second equation describing overall CO2-assimilation rate
A=(Vm--Fm-- Mm)+(Vs-- Fs-- Ms),
(3)
For comparison the CO2-assimilation rate of a C3 leaf is given by
where Cm and Om are the CO2 and 02 concentrations in the mesophyll cells and Cs and Os those of the bundle-sheath cells. Kc, Ko are the Michaelis-Menten constants for CO2 and 02 respectively, Vmmaxand Vs,,ax are the maximum activities of Rubisco in the mesophyll and bundle-sheath cells. Following the oxygenation of 1 tool of RuBP, 0.5 tool of CO/ are evolved in the photorespiratory pathway. Farquhar et al. (1980) showed that the ratio of the rate of oxygenation to carboxylation can be expressed as
Vo/V~= 2 F,/C
where C stands for COz concentration, F, is the CO2-compensation point in a C3 plant in the absence of other mitochondrial respiration, and
F,=[0.5 A = V-F-M.
(3a)
The rates of carboxylation and photorespiration. The rate of carboxylation of Rubisco can be limited by the amount and activity
(6)
VornaxKc/(VemaxKo)]0=~/,
O.
(7)
The symbols Vom~xand V~ma, are the maximum activities of the oxygenase and carboxylase, respectively, and O stands for 02 concentration. The rates of photorespiration Fm and F~ are
S. von Caemmerer: CO2 assimilation and c-isotope discrimination in C3--C4 intermediates then given by Fro=0-5 Vom = 7 * Om Vm/Cm
(8)
and
f~=0.5 Vow=7, o~ WCs
(9)
where Vom and Vo~ are the oxygenation rates of the mesophyll and bundle-sheath Rubisco, respectively. The kinetic properties of Rubisco of whole-leaf extracts have been examined for several intermediate species (Yeoh et al. 1980; Holbrook and Chollet 1986). Holbrook and Chollet (1986) found that the Rubisco of Panicum milioides and of Moricandia arvensis had K m (CO2) values and specificity factors (another name for 0.5 7,) in the range of those found for C3 species. The following values, typical for C3 species at 25 ~ C, are used for both the mesophyll and bundle-sheath Rubisco; K~ =310 gbar, Ko=360 mbar (Seemann et al. 1981; Bird et al. 1982; Makino et al. 1985). The ratio Vomax/Vcmax=0.46 was chosen to agree with the experimentally measured F,=41.6 gbar at 210 mbar Oz (Brooks and Farquhar 1985), and gives F, = 39.6 ~tbar at 200 mbar O2. The Km(O2) of Rubisco has frequently been measured, but uncertainty exists in the values of Km(Oz) and the ratio Vomax/Vcmax, Kirschbaum and Farquhar (1984) have used values of Ko= 155 mbar and VomJV~m~ =0.21. This combination also gives a F, =41.6 ~tbar at 210 mbar 02.
7he leakage of C02 and 02 to and from the bundle-sheath cells. A higher CO2 or O2 concentration within either the bundlesheath cells or the mesophyll compartment will result in net diffusion of these gases. The rate of COz leakage, L, from the bundle-sheath cells to the mesophyll cells is given by L=g~(C~-Cm)=L~--L m
(lO)
where g~ is the conductance to leakage of C O 2 a c r o s s the bundle sheath, and L~=g~ C s is the leak rate that would occur from the bundle-sheath cells to the mesophyll cells if C~=0, and Lm = gs Cm is the leak rate that would occur from the mesophyll cells to the bundle-sheath cells if C~=0. Very little is known about the magnitude of the conductance (g~) in C 3 - C4 intermediates. If not otherwise specified a value of gs = 1 m m o l . m - 2 . s -~ has been used in the simulations; in other cases bundle-sheath conductance is treated as a variable. It is known that bundle-sheath chloroplasts in some C4 species have very little photosystem-II activity. The granal nature of bundle-sheath chloroplasts of C3 - C4 intermediates indicates that they are capable of photosystem-II activity (Edwards and Ku 1987). As pointed out by Raven (1977) and Berry and Farquhar (1978), this has implications for the steady-state 02 concentration inside the bundle-sheath cells. Following the development of Berry and Farquhar (1978), it is assumed that net Oz evolution (Eo) in the bundle-sheath cells equals its leakage (Lo) out of the bundle, that is Eo = Lo = go (Os -- Om)"
(11)
The conductance to leakage of Oz across the bundle sheath (go) is related to the conductance to CO2 by way of the ratio of diffusivities and solubilities (Berry and Farquhar 1978) by go=gs(Do~
So2)/(Dco 2 Sco2),
(12)
where Do~ and Dco ~ are the diffusivities of O2 and C O 2 in water, respectively, So~ and Sco~ are the respective Henry constants, and go = 0.047 g~ at 25 ~ C (Farquhar 1983).
465
The net O2 evolution in the bundle-sheath cells depends upon the balance of O2 evolution at photosystem II and the various O2-consuming processes. In particular the balance depends upon how the glycine shuttle is perceived to operate. In the model the solution which generates no extra 02 consumption based upon the glycine shuttle to the bundle-sheath cells has been chosen. Briefly, the integrated carbon reduction and oxidation cycle in the bundle-sheath cells consumes N A D P H at the rate of 2V~+2Vos which results in an 02 evolution rate of V~ + Vow.Oxygen is consumed by bundle-sheath Rubisco, during glycotate oxidation (resulting from glycolate produced in the bundle-sheath), and by mitochondrial respiration at the rate of 1.5 Vo~+ M s. In Moricandia arvensis, enzymes necessary for the synthesis of glycine are present in both bundle-sheath and mesophyll compartments and it is therefore likely that glycine rather than glycolate is the compound transported (Rawsthorne et al. 1989). Thus it is assumed that carbon skeletons are returned to the mesophyll such that 3-phosphoglyceric acid derived from mesophyll glycine is reduced in the mesophyll such that also no extra O2-evolution rate occurs in the bundle-sheath cells. In summary:
Eo=V~+Vo~--I.5Vo~-M~=A s.
(13)
Mathematical solution of the equations. For calculating CO2-assimilation rates, solutions for C~ and O~ are required. Combining eqs. (5), (9), (10), eq. (2) can be rewritten as Vsmax( C s - 7 , Os) = (~Fm _~_M~ -- g, (Cs - Cm).
C~+ Kr
(14)
+O~/Ko)
To obtain an expression for O~, one uses eqs. (2), (10), (11) and (13) to obtain o, = o., + (c~Vm-- gs (C~ -- C .,))/go"
(15)
For convenience the following expression is used
o~ = o , - c~. g2go
(16)
where O , = Om-}- (~Frn-}- g s Cm)/g o.
(17)
Substituting for O s in eq. (14) and solving for C~ a quadratic of the following form is obtained:
a.C~+bCs+d=O,
(18)
so that Cs= ( - b + ~ ) / ( 2
a),
(19)
where
a =(1 -g~ Ko/goKo)
(20)
b = { Vs max (1 -]- 7$ gs/go)}/g~- {(c~Fm + Ms)/g s + Cm}
( 1 -- Ko gjKo go) + K~(1 4- O ,/Ko) and d = -(Vsmax 7, 0 , ) / g s - {(~ VmArMs)/g~+ am} {Kc (1 ~- O./Ko) }. (22) Assuming values for Cm, Om, Vmmax, Vs . . . . M S and Mm, bundlesheath CO2 concentration (C~), O s, A and all other parameters can be calculated. The equation for the compensation point is cubic in nature and F has been calculated with an iterative computer calculation searching for Cm at which A = 0.
466
S. von Caemmerer: CO2 assimilation and c-isotope discrimination in C 3 - Ce intermediates
Model predictions Parameters and criteria. Since the modelis designed for testing of ideas rather than empirical fitting, as many parameters as possible have been preassigned from literature values. For example, the kinetic parameters of Rubisco have been chosen from in-vitro measurements (see Model development section) and can be viewed as constants except for the maximum carboxylase activity, which one would expect to vary with leaf age, irradiance and nutrient conditions during growth. A value of 107 g m o l - m - 2 . s -1 was chosen for the total maximum Rubisco activity. Since we know little about mitochondrial respiration and its compartmentation, the values of mitochondrial respirations, M m and M~ were somewhat arbitrarily set at 0.2 g m o l . m - 2 . s -1 The parameter ~ determines what fraction of the photorespiratory CO2 is respired in the bundlesheath cells. Recent evidence by Rawsthorne et al. (1988) has shown that in Moricandia arvensis the glycine decarboxylase was solely located in the bundle-sheath cells. Similar results have subsequently been found for Panicum milioides, Flaveria floridana and F. Iinearis (Hylton et al. 1988). In all results presented here except for Fig. 3 we have assumed that ~ = 1 , i.e. that all photorespiratory CO2 is released in the bundle-sheath cells. Two main variables emerge as important parameters for which values need to be chosen and both parameters, the conductance to leakage between bundle-sheath and mesophyll cells (gs), and the partitioning of Rubisco between mesophyll and bundle-sheath cells affect the leakage rate between these two compartments. The major criteria used to identify C 3 - C4 intermediates experimentally have been the CO2-compensation point (F) and the O2 inhibition of CO2-assimilation. In Fig. 2, CO2-assimilation rate, bundle-sheath CO2 concentration, F, and 02 inhibition are examined with respect to the proportion of Rubisco in the bundle-sheath and mesophyll cells and at two bundle-sheath conductances. Values of g s = l m m o l ' m - 2 ' s - ~ and gs = 1 0 m m o l - m - 2 . s -1 were selected. Apel and Peisker (1978) have estimated a conductance of 3.1 mmol. m - 2. s - 1 from anatomical studies in Zea mays, and a conductance of 1 m m o l . m - 2 , s- 1 is sufficient to simulate the 02 insensitivity of CO2-assimilation rate in a Cr model (Berry and Farquhar 1978). Brown et al. (1983) showed that the Panicum C 3 - C 4 intermediates had a well-developed vascular bundle sheath and that the number of plasmodesmata were similar to that in-
C4 Panicum species and more than that found in C3 Panicum species. The estimate of 1 mmol. m - 2. s- 1 may therefore be appropriate for these intermediates. The CO2-assimilation rate at a mesophyll partial pressure of CO2 of 250 gbar was found to be greater in the intermediate, than in a C3 leaf with the same total amount of Rubisco (Fig. 2a) when up to 15% of the total Rubisco was partitioned to the bundle-sheath cells. This was true at both conductances, although the enhancement was less at the higher conductance. The rate of CO2 assimilation fell below that of a C3 leaf at 250 lxbar if much more than 15% was partitioned into the bundle-sheath cell. There are two main reasons for this. As the fraction of Rubisco in the bundlesheath cells increases, the fraction of Rubisco in the mesophyll cells decreases, hence the oxygenase activity and associated capacity to shuttle glycine also decreases. At some point the capacity of the bundle-sheath Rubisco is too great for the capacity of the shuttle and the bundle-sheath Rubisco is disadvantaged for direct fixation of intercellular CO2 because of the low bundle-sheath conductance to CO2. This is illustrated in Fig. 2b, where the bundle-sheath CO2 pressures were greatest when a small portion of the total amount of Rubisco was in the bundle sheath. It quickly fell below the mesophyll pressure of 250 gbar as the portion of Rubisco in the bundle sheath increased. As a consequence of the kinetics of Rubisco the oxygenase activity is greatest at low CO2 and this can be seen in Fig. 2 a in the enhancement of assimilation rate over that of a C3-1eaf at Cm = 50 gbar COz. The CO2-compensation point decreased initially rapidly from the Ca value of 42 gbar as the capacity to refix respired CO2 increased with an increasing fraction of Rubisco in the bundle-sheath cells. The value of F was dependent upon the values of bundle-sheath conductance chosen (Fig. 2c). The compensation point has often been measured to characterize C 3 - C 4 intermediates and Edwards and K u (1987) have tabulated the results of the literature in their review. They reported for Panicum milioides, F values between 11 and 17 gl/1 at 21% Oz. Brown (1980) measured values as low as 6.9 and 4.4 gl/1. For Moricandia arvensis, values of 28 and 16 gbar have been reported (Edwards and Ku 1987). These values are within the range predicted in Fig. 2c at a Rubisco partitioning factor of 0.1. The model predicts that the glycine shuttle can achieve very low compensation points because of the high rates of oxygenation that occur at low COz pressures. In a C3 leaf the CO2-assimilation rate is greater
S. von Caemmerer: CO2 assimilation and c-isotope discrimination in C 3 - C 4 intermediates
C --
~"
C3
30
-
~...
N"
'E "6
\
467
t,O
30
O
20
L_
Cm=250
"~
20-
\N\
E
-
0
,
I
,
Cl
l
8000 cooo
I
1
c~
0
I
d
b
50
O
1.6
-1.t,
~.~... =
10
E O O
--....~
1,....
.z~
-1.2
400C
O .IQ
-10
I
0
01
0.2
0.3
O/~
0.1
I
0.2
I
03
[
0.~
E O G1
0 0.5
Fraction of total Rubisc0 in bundle-sheath. Fig. 2. a COz-assimilation rate (A) versus fraction of total leaf Rubisco partitioned into the bundle-sheath cells at 200 m b a r Oz and at two mesophyll COz-concentrations; Cm (250 and 50 pbar) calculated with two different bundle-sheath conductances: g, = 1 retool, m - 2 s - 1 (solid lines), g~ = 10 mmol. m - 2. s - 1 (dashed lines). Vmax of total Rubisco = 107 pmol. m - 2. s - 1. Also shown are the rate of CO2-assimilation of a modelled C3 leaf at Cm = 250 and 50 gbar with that Rubisco Vm,x(horizontal lines), b Corresponding bundle-sheath CO2 pressures (Cs) versus the fraction of total leaf Rubisco partitioned into the bundle-sheath cells at Cm=250 and 50 gbar and g , = l m m o l - m - Z . s -1 (solid lines), g s = 1 0 m m o l - m - Z . s -1 (dashed lines), e CO2-compensation point (F) versus the fraction of total leaf Rubisco partitioned into the bundle-sheath cells at g s = l m m o l - m - 2 . s - 1 (solid line) and gs= 10 m m o l . m - 2 . s -~ (dashed line), d The ratio of CO2-assimilation rate (A) at 20 mbar 0 2 and 250 gbar CO2 to A at 200 mbar 02 and 250 gbar CO2 at g ~ = l m m o l - m - 2 . s -~ (solid line) and g~=lOmmol.m-2.s -1 (dashed line). In all cases, it is assumed that all photorespiratory CO2 is released in the bundle-sheath cells (~ = 1)
at 20 mbar p(O2) than at 200 mbar p(O2). An increase in the partial pressure of 02 increases the rate of photorespiration and decreases the rate of carboxylation due to increased competition from O2 (eqn. 3a). Figure 2d shows the ratio of CO2-assimilation rate at 20 mbar p(O2) to that at 200 mbar p(O2) for different Rubisco partitioning. The ratio declines from the C3 value of 1.46 to a value close to 1.27, depending on the bundlesheath conductance. The value 1.27 is the ratio of the mesophyll carboxylase rate of Rubisco at the two O2 pressures brought about by the competitive effect of 02 on carboxylation, i.e. at the lower conductances all photorespired CO2 is effectively refixed. Dependence of F on 02 concentration. The compensation point increases linearly with 02 concentration in leaves of C3 species (Forrester et al. 1966; Bj6rkman et al. 1970; Peisker and Apel 1977). This
response can be explained by considering only the kinetics of Rubisco (Laing et al. 1974; Azcon-Bieto et al. 1981). In C 3 - C 4 intermediates such as Panicure milioides or Moricandia arvensis non-linear relationships have been observed (Keck and Ogren 1976; Quebedeant and Chollet 1977; Apel 1980; Brown and Morgan 1980; Holaday et al. 1982; Hattersley et al. 1986; Hunt et al. 1987). Figure 3a shows that the model incorporating a glycine shuttle predicts a curvilinear relationship. This differs from the biphasic dependence of F on 02 concentration predicted by the model of Peisker and Bauwe (1984), which includes the operation of a C4 pathway. At present the available data do not allow us to distinguish between biphasic and curvilinear responses. Figure 3a shows F with changing p(O2) pressure at several different fractions of Rubisco in the bundle-sheath cells. The data points relating F and p(O2) for Panicum milioides have been replotted
468
S. von Caemmerer: COz assimilation and c-isotope discrimination in C 3 - C , intermediates
10C a
" ~ " / / / I ~9/ i,//,/
8c "C u
/E/=0.05
" y=.01
//;/,
6c
40
y=02
2O 0
b
s = lOOmmol.m-2s1
100
3//
80
0/I //
0.1
60 L_ 40 20 0
~
10
~/ / / /
8,
///
"E E1
X3
60
L-
40
CL
/Z0,75 //////~ =1.0
0 200
/*00 600 800 0 m (mbor)
1000 1200
Fig. 3. a CO2-compensation point (s versus mesophyll P(O2) , O m at three different fractions (y) of total leaf Rub&co (V~, x = 107 gmol.m Z.s-~) partitioned into the bundle-sheath cells. The bundle-sheath conductance g~= 1 mmol. m - 2 . s-1, and all photorespiratory CO2 is released in the bundle sheath (c~= 1). Also shown is the predicted F versus Om for a C3 leaf. The data points joined by the dotted line are measurements made on Panicum milioides taken from Hattersley et al. (1986). b CO2-compensation point (F) versus Om at different bundlesheath conductances (g~). The fraction (y) of total Rubisco in the bundle-sheath is 0.1 and c~-1. Also shown is the predicted F versus O,,, for a C3 leaf. e CO2-compensation point (F) versus Om at different fractions of photorespiratory CO2 released in the bundle-sheath cells (c0. The fraction of total Rubisco in the bundle-sheath cells=0.1 and g = l m m o l - m - Z - s -~. Also shown is the predicted F versus Om for a C 3 leaf. The dashed line is common to a, b and e (y=0.1, g~=l m m o l . m - 2 . s -1, ~=1)
from Hattersley et al. (1986). At low p(O2), the compensation point was low, as CO2 released in the bundle-sheath cells is refixed with little leakage out. As the rate of glycine shuttling increases with increasing p(O2) bundle-sheath Rubisco eventually
becomes saturated with bundle-sheath CO2 and the leakage rate rises relative to the rate of CO2 released, and F increases. Therefore an increase in bundle-sheath Rubisco keeps F low until higher p(O2). Bundle-sheath conductance also influences the relationship between F and p(O2), by influencing the leak out of the bundle-sheath cells (Fig. 3 b). Changing the proportion of photorespiratory CO2 that is released in the bundle-sheath cells affects F by the amount of CO 2 lost directly from the mesophyll cells (Fig. 3 c). Clearly a number of different choices of Rubisco partitioning, bundlesheath conductance and 0~could satisfy values ob9served in the literature (Brown 1980; Edwards and Ku 1987; Hunt et al. 1987).
Dependence of C02-assimilation rate on C02 and 0 2 . In Fig. 4 the rate of net CO2 assimilation (eqn. 2) and the bundle-sheat p(CO2) (C~) are plotted against mesophyll p(CO2) (Cm) at 200 mbar and 20 rnbar P(O2). For simplicity the model contains no limitation on RuBP regeneration capacity or electron transport and so no saturation occurs at Cm above 300 gbar (unlike the previous C3-model of Farquhar et al. 1980). It is likely that, at high irradiances and ambient partial pressures of CO2 up to 300 gbar, CO2 assimilation is limited by the amount of Rubisco present, but the model should not be used above 300 gbar. Incorporation of a light and electron-transport limitation will require more knowledge on the partitioning of energy between bundle-sheath and mesophyll cells. The bundle-sheath CO2 concentration was highest at low Cm where the rate of oxygenation is highest and declines with increasing Cm (Fig. 4). This result differs from C4 model predictions for increasing bundle-sheath CO2 concentrations with increasing mesophyll p(CO2) (Berry and Farquhar 1978). Recently, direct measurements of the inorganic carbon pool in the bundle-sheath cells by Furbank and Hatch (1987) have confirmed these C4 predictions. Such measurements may provide a means of testing the present hypothesis, and it may be possible to identify C 3 - C4 intermediates operating with a glycine shuttle. At 20 mbar p(O2) the bundle-sheath p(CO2) was around 50 ~tbar, only marginally above the C3-compensation point, and the contribution of the glycine shuttle was almost negligible. The increase in assimilation rate over that at 200 mbar p(O2) is mainly the result of the increased mesophyll carboxylation rate caused in turn by less competition from 02. The enhancement in CO2-assimilation rate is less than that seen in modelled C3 photosynthesis because of the substantial
S. von Caemmerer: CO, assimilation and c-isotope discrimination in C 3 - C4 intermediates 60
j f A
50 ~
/j
J
~ - ~
4o
E ~
E
30 500&~20
(..) 250
10 0
T-----I------T--~
100
200 300 400 Cm, (ubar CO2)
Ases1As 0 500 600
Fig. 4. CO2-assimilation rate (A) bundle-sheath CO2 assimilation rate (As) and bundle-sheath COz pressure (Cs) versus mesophyll CO2 pressure (Cm) at two oxygen pressures: 200 mbar (solid lines), 20 mbar 02 (dashed lines). The fraction of total Rubisco in the bundle-sheath is 0.1 and total Vmax = 107 ~rnol. m - 2. S- 1, bundle-sheath conductance gs = 1 m m o l - m - 2 . s -1, and all photorespiratory C02 is respired in the bundle-sheath cells (~ = 1)
5C 4C 3C E
-5 E
