Significance of Phosphoenolpyruvate Carboxylase during - NCBI

Plant Physiol. (1989) 89, 1150-1157 0032-0889/89/89/11 50/08/$01 .00/0

Received for publication August 2, 1988 and in revised form October 24, 1988

Significance of Phosphoenolpyruvate Carboxylase during Ammonium Assimilation Carbon Isotope Discrimination in Photosynthesis and Respiration by the N-Limited Green Alga Selenastrum minutum' Robert D. Guy, Greg C. Vanlerberghe, and David H. Turpin* Department of Biology, Queen's University, Kingston, Ontario, Canada, K7L 3N6 ABSTRACT

for aspartic acid synthesis has also been documented by Melzer and O'Leary (17). In spite of this presumed increase in PEPcase activity during N assimilation in the light, total carbon fixation by N-limited Selenastrum minutum declines significantly during NH4' assimilation (6, 7, 27). This is correlated with a decline in the levels of RuBP below the active site density of Rubisco implying RuBP limitation of photosynthetic carbon fixation (6). Thus, N assimilation must cause a major change in the relative significance of these two carboxylating enzymes to total carbon fixation. In this study, we used stable isotope fractionation to determine the partitioning of carbon fixation between these two carboxylating enzymes prior to and during the assimilation of N by the N-limited green alga S. minutum. This approach is based on the differential discrimination against 13C expressed by Rubisco and PEPcase. Stable isotope composition has been employed extensively to indicate photosynthetic mode and the primary carboxylating enzyme in plants, as well as to provide other useful process and tracer information about photosynthetic gas exchange (reviewed by O'Leary [20]). O'Leary et al. (22), Sharkey and Berry (25), and Evans et al. (9) have presented techniques for the short-term measurement of isotope discrimination. In the present study we employ a modification of the O'Leary et al. (22) model, which accounts for simultaneous respiratory CO2 release and refixation within a closed system, to estimate the partitioning of carbon fixation between Rubisco and PEPcase during N assimilation. The results show that N assimilation by N-limited S. minutum causes a major shift in carbon fixation from Rubisco to PEPcase. This underscores the significance of PEPcase in providing anaplerotic substrate for the TCA cycle during N assimilation and provides an example of a situation where carbon fixation by PEPcase may temporarily exceed that of Rubisco in a C3 plant in the light. Apparently, under these conditions the TCA cycle operates primarily as a pathway through to a-KG which is then used for the assimilation of NH4' into amino acids.

The effect of N-assimilation on the partitioning of carbon fixation between phosphoenolpyruvate carboxylase (PEPcase) and ribulose bisphosphate carboxylase/oxygenase (Rubisco) was determined by measuring stable carbon isotope discrimination during photosynthesis by an N-limited green alga, Selenastrum minutum (Naeg.) Collins. This was facilitated by a two process model accounting for simultaneous CO2 fixation and respiratory CO2 release. Discrimination by control cells was consistent with the majority of carbon being fixed by Rubisco. During nitrogen assimilation however, discrimination was greatiy reduced indicating an enhanced flux through PEPcase which accounted for upward of 70% of total carbon fixation. This shift toward anaplerotic metabolism supports a large increase in tricarboxylic acid cycle activity primarily between oxaloacetate and a-ketoglutarate thereby facilitating the provision of carbon skeletons for amino acid synthesis. This provides an example of a unique set of conditions under which anaplerotic carbon fixation by PEPcase exceeds photosynthetic carbon fixation by Rubisco in a C3 organism.

The synthesis of some major amino acids depends upon the provision of carbon skeletons from the TCA cycle. Use of TCA cycle intermediates in biosynthesis requires anaplerotic reactions to replenish this drain. It has been shown that NH4I assimilation in the dark enhances PEPcase2 activity which serves this anaplerotic function (2, 6, 7, 13, 19). In the light, N assimilation increases TCA cycle carbon flow in support of amino acid biosynthesis (2, 8, 13) and results in an increase in TCA cycle CO2 release (5, 28). Biochemical evidence is consistent with a simultaneous enhancement in PEPcase activity (2, 6, 8, 13). The role ofthis enzyme in providing carbon 'Supported by the Natural Sciences and Engineering Research Council of Canada. 2Abbreviations: PEPcase, PEP carboxylase; D.n,,5, per mil discrimination against 13C by whole cells corrected for the CO2/HC03equilibrium isotope effect assuming CO2 as substrate; DT, per mil discrimination against '3C including the CO2/HC03- equilibrium isotope effect; DIC, dissolved inorganic carbon; KIE, kinetic isotope

MATERIALS AND METHODS Algal Culture and Preparation Selenastrum minutum (Naeg.) Colfins (Chlorophyta) (UTEX 2459) was grown in N-limited chemostats as previ-

effect; a-KG, a-ketoglutarate; OAA, oxaloacetate; PEP, phosphoenolpyruvate; Rubisco, RuBP carboxylase/oxygenase; RuBP, ribulose bisphosphate.

1150

CARBON ISOTOPE FRACTIONATION DURING NH4+ ASSIMILATION

ously described (7). Most experiments were carried out on cells grown on CO2 enriched air (4 mm DIC at pH 8.0) with the exception of one experiment on air grown cells. The 5'3C value of the high CO2 mixture was near -46%o. Carbon isotope compositions are expressed in terms of 6'3C values in per mil (%o) units according to the expression: i"C = (Rsample/Rstadard - 1) x 1000 where R is the ratio '3CO2/'2C02 and the standard is Pee Dee Belemnite (PDB) limestone. For isotope fractionation experiments, algae were separated from the chemostat medium by centrifugation followed by washing two times in DIC-free 25 mM Hepes buffer (pH 7.0). The final pellet was taken up into 50 mL of the DIC-free buffer and injected into a sealed leak-tight cuvette containing a further 300 mL of degassed buffer with 355gg mL-' carbonic anhydrase and 35 ,ggmL-' catalase. The cuvette was similar to one described by Guy et al. (1 1), being equipped with an oxygen electrode (YSI 5331, Yellow Springs OH) and a fixed magnetic stirrer. The base was water-jacketed and all experiments were at 25°C. Experimental Incubations To produce DIC for isotopic analysis of respired C02, cells were incubated in the darkened cuvette for 1 to 3 h prior to sampling. During this time, oxygen was supplied periodically as H202. The effects of N assimilation on the apparent isotopic composition of respired CO2 were monitored for 1 h following addition of NH4Cl to a concentration of 2 mm. To study isotope discrimination during carbon fixation, cells were first incubated in the dark as described above to produce an extracellular DIC substrate pool. This reduced errors that can be introduced by using a source of substrate with a radically different 6'3C value, such as air. Initial DIC concentrations were generally about 800 gM and were not allowed to fall below 400 gM when using high CO2 grown cells. Light was provided by two Kodak 4200 projectors. For purposes of applying the two process isotope discrimination model (see below) an addendum of Na214CO3 was made shortly after turning the lights on. This permitted simultaneous measurement of both gross and net carbon fixation. Sampling for stable isotopes began 2 min into the light period. When studying the effects of N assimilation, the first sample was taken just as NH4Cl (2 mM) was supplied (5 min into the light period). For scintillation counting, 0.6 mL samples were removed by syringe at 2.5 to 5 min intervals and immediately injected into 0.4 mL of kill solution (5% 6 N formic acid in 80% ethanol). Unfixed counts were removed by sparging with a CO2 gas stream for 1 h prior to counting in Aquasol II (Dupont) liquid scintillation cocktail. Calculations of gross carbon fixation accounted for changes in specific activity over the course of the experiment due to respiratory carbon efflux. Preparation of CO2 Samples for Isotope Analysis To sample DIC for isotopic composition, aliquots were withdrawn from the cuvette directly into preevacuated and weighed vacuum bulbs containing a volume of phosphoric acid sufficient to kill the algae. In this manner, up to six

1151

samples could be taken over the experimental period (usually 1 h or less). Samples were extracted within the next several hours. The contents of each vacuum bulb were discharged into the stripping vessel of a preparation line and the CO2 extracted by bubbling with He under partial vacuum for 45 min. Water vapour was cryogenically removed and the CO2 was trapped on ten loops passing in and out of two liquid nitrogen baths. After further cryogenic separation from contaminating gases, CO2 yields were determined with a calibrated pressure transducer. This provided an accurate measure of the DIC concentration at the time of sampling. The stability of the acidified algal samples waiting on the bench was verified in two experiments by taking several samples simultaneously and extracting them sequentially. Yields of CO2 and 6'3C values were unaffected. At the end of each experiment, cells remaining in the cuvette were cleaned by centrifugation and washing in deionized water. These were dried for whole cell 6"3C determinations of CO2 prepared by combustion (two replicates each) at 800C according to Macko (15). Isotope analyses were performed on Nucide 6-60 and Micromass 903 ratio mass spectrometers. Precision was better than ±0.1 %o. For DIC samples containing `4C, corrections for 'N) content were based on prior knowledge of the 18Q content of medium water with which the DIC had equilibrated. Other Methods Light response curves for gross carbon fixation were carried out under control and NH4'-pulsed conditions. Cells were separated from the chemostat medium by centrifugation and resuspended in 25 mm Hepes (pH 7) buffer. Procedures were as above except that a water-jacketed glass cuvette was used. Incubations were for 30 min and DIC concentration was determined as in Miller et al. (18). Chl concentration was measured by extraction in methanol (6). Algae used in the respiration experiments were pooled and the starch extracted according to methods outlined in Hassid and Abraham (12) and Abraham and Hassid (1). The starch was combusted in two replications for 6'3C analysis as described above.

Calculations and Two Process Model Although all discrimination factors reported here have been determined by substrate analysis (4), two methods of calculation were used. Where substrate levels are depleted due to an uptake process in the absence of a replenishing source, per mil discrimination factors (D) can be calculated from the "Rayleigh" equation (modified from Kroopnick and Craig

[14]): ln R/Ro (1) In f x1I000, where R is the isotope ratio of the substrate at the time of sampling, Ro is the initial isotope ratio, and f is the fraction of substrate left. A plot of ln R/Ro x 1000 versus -lnfyields a straight line, which can be fitted by regression through the origin, with slope D (16). Where respiration rates are low and D

=

1152

GU]Y ET AL.

if respired carbon is not isotopically too dissimilar from the original substrate, Equation 1 can provide a good estimate of discrimination during photosynthetic carbon fixation. If respiration rates are high, a more sophisticated approach is required. Changes in substrate concentration (C) in time (t) as a function of a negative first order rate process (e.g. carbon fixation, with rate constant k1) and a simultaneous positive zero order rate process (e.g. respiratory carbon release, with rate constant k2) are described by: dC =

-k, C + k2 .(2)

Plant Physiol. Vol. 89, 1989

which is corrected for the C02/HCO3 equilibrium isotope effect assuming CO2 as substrate. The presence of saturating concentrations of DIC in a rapidly stirred carbonic anhydrase containing medium would nullify any contribution of isotope diffusion effects to measured discrimination (23). Discrimination, as calculated in Equations 1 and 4, describes the difference in isotope ratios between substrate (Rs) and instantaneous product (Rp) such that D = (1 - Rp/Rs) x 1000. This differs slightly from some of the physiological literature (e.g. Farquhar and Richards [10]). Discrimination values reported here can be converted to A as used by Farquhar and co-workers by: D/1000

Solution of this differential equation yields: C, = (C0 - k2/ki)e-klt + k2/kl, (3) where C0 is the initial DIC concentration and C, is the DIC concentration after time t. Equation 3 is very similar to one provided by O'Leary et al. (22) where the term k2/kl was replaced with the CO2 compensation point concentration as a means of obtaining information about the respiration rate. Our approach differs in that we have measured the rate of respiratory CO2 efflux more directly as being the difference between gross fixation (14C uptake) and net fixation (change in total DIC concentration). Separate equations in the form of equation (3) can be written for '3C and for 12C. Although uptake of "3C and 12C are actually competing processes, the change in relative abundance of these isotopes during an experiment is so small that this can be ignored. Then `3k2 and 12k2 are the rates of '3CO2 and 12CO2 respiratory efflux and 13k1 and '2k, are the rate constants for 3GCO2 and 12C02 fixation. The ratio '3k2/12k2 will equal the 13CO2/'2C02 ratio of the respired CO2, which we take to be +7.85%o relative to whole cells (see results). For our purposes, '3k, and 12k1 were calculated by iteration until the known values of `3k2, 12k2, "3Co and '2C0 yielded the known values of 13C, and '2C,. The ratio 12k1/'3k1 is then equal to the kinetic isotope effect (KIE) for carbon fixation and is related to per mil discrimination through: D

(xK I1000. -

(KIE)

1 -(D/000)

A

(4)

Discrimination against 3GCO2 thus obtained was then used for correcting gross carbon fixation rates for discrimination against 14CO2 (assumed to be twice that of 3GCO2). This correction in turn had a slight effect upon D. This process was repeated until the actual value of D had been calculated to one decimal place. Because our experiments took place in an aqueous medium, D was not only dependent upon biochemical fractionation but was also affected by the pH dependent C02/HCO3- equilibrium isotope effect in an additive manner (30). Dissolved CO2 differs from HC03- by -9%o. Since most DIC is in the form of HC03- at pH 7.0, reactions fixing CO2 will have an additional apparent discrimination of about 7.5%o. Reactions fixing HC03- will have an apparent discrimination which is lower by about -1.5%o. In addition to DT, which is the measured total discrimination factor, for some experiments presented here we also report a value (D,,,,)

(5)

RESULTS Gas Exchange in the Dark

Table I shows net CO2 and O2 exchange and the resulting RQ values of control and NH4+-pulsed cells 0 to 30 and 30 to 60 min following NH4' addition. The assimilation of ammonium by these cells resulted in a major increase in respiratory CO2 release and 02 uptake. Partial recovery was observed in the 30 to 60 min period. The RQ value decreased dramatically during NH4' assimilation. These trends in the gas exchange data were consistent with results previously reported (28) but batch variability accounted for some differences in absolute rates. Gas Exchange in the Light

The addition of NH4' resulted in a suppression in both net and gross carbon fixation (Table II). As was the case in the dark, NH14 assimilation also caused an enhancement in respiratory CO2 release. Effects of Light and NH4,

on

Gross Carbon Fixation

Control cells exhibited a standard light saturation curve (Fig. 1). Saturation was reached at an incident irradiance of approximately 250 ,E* m2 *s'. In cells which were assimilating NH+, dark carbon fixation was high. Increased irradiance caused a much lower increase in carbon fixation relative to control cells (Fig. 1). Saturation was reached at the same light Table 1. Effects of Ammonium Assimilation on Net Rates of Gas Exchange in the Dark Condition

Control NH4+-pulsed (030 min) NH4+-pulsed

Rate of Net Gas Exchange in Dark Net CO2 Net 02 UPtake evolufon

RQ

mol. mg ' Chl*h-' 116.1 ± 16.4a 98.6 ± 14.0 256.4 ± 20.5 312.8 ± 29.4

1.19 ± 0.06 0.82 ± 0.02

178.3 ± 22.1

0.74 ± 0.04

240.0 ± 19.8

(30-60 min) a Values are reported as mean ± SE (n

=

4).

C02/02

1153

CARBON ISOTOPE FRACTIONATION DURING NH4+ ASSIMILATION

Table II. Effects of Ammonium Assimilation on Mean Rates of Net and Gross Carbon Fixation and C02 Evolution over a 60 min Period at Saturating Light t450.uE.m2.sec'1) C02 evolution was calculated as the difference between net and gross carbon fixation. Carbon Fluxes at Saturating Ught Condition Evolution Net fixation Gross fixation Chl DIC ;&mol .mg-' -Ih182.0 ± 10.9a 230.9 ± 13.2 48.9 ± 5.5 Control 17.5 ± 7.9 181.8 ± 33.9 164.3 ± 33.9 NH4+-pulsed (60 min) a Values are reported as mean ± SE (n = 3).

24

0 0 x

1I--

-In(f)

Figure 2. Carbon discrimination by N-limited, high C02 grown S. minutum during photosynthesis in the absence of NH4+. The results are presented as a Rayleigh plot which represents the change in isotopic composition of DIC (as In (R/Ro 1000) as a function of the fraction of consumed substrate [expressed as -In (f)] (see Eq. 1). Data were pooled from five experiments. In those in which 14C was used, gross carbon fixation and C02 evolution averaged 231 and 46.2 Mmol * mg-1 Chi * h-' respectively (not corrected for 14C02 discrimination). Cell densities ranged from 1.8 to 3.1 ug Chil ml-' and the light intensity was saturating to carbon fixation.

350 300

x

- 250

'm

150

o~

100

35.-

50

30-

0

0

200

400

600

800

1000 O

(jzE-M 2s 1) on gross carbon fixation 1. Effects of incident light intensity Figure by N-limited S. minutum during ammonium assimilation and under control conditions for a representative experiment. Culture cuvette maintained at 21.50C and DIC was 4.7 mm. Chi was 1.5 Mg-1 Chi. Irradiance

25-

CO

DT,W = 27.52 7/ DC.iis == 20.02 SE

±0.16

%.

0 20

Ckf15_2 r2 / I-,5 -

.9987

' 10-

ml-1. 5

intensity but the maximum rate of carbon fixation was approximately 50% of the control reflecting the NH44-induced suppression in photosynthetic carbon fixation. Isotopic Discrimination in Control Cells

Figures 2 and 3 report Rayleigh type analysis (Eq. 1) of discrimination against '3C during carbon fixation in the light by cells grown on high and low CO2, respectively. These analyses ignore any effect of respiration which, under control conditions, occurs at low rates relative to photosynthesis (Table II). Low CO2 grown cells had a lower discrimination factor by 9.7%o relative to high CO2 grown cells. Although the coefficients of determination (r2) for the regressions from both data sets were very good, there was noticeably greater scatter about the line for high CO2 grown cells. This increased scatter is probably due in part to higher rates of respiratory CO2 efflux by these cells relative to the low CO2 grown cells. (See legends to Figs. 2 and 3.) Isotopic Composition of Starch and Respired CO2

CO2 respired from N-limited cells in the dark was consistently enriched in 13C relative to whole cells by 7.85%o (SE =

-

0

0.2

0.4

0.6

0.8

1

1.2

-In(f)

Figure 3. Carbon isotope discrimination by N-limited low C02 grown S. minutum during photosynthesis in the absence of NH4+. Results are presented as described in Figure 2. Gross carbon fixation and C02 evolution were 21 1 and 13 Mmol C02 mg-'Chl *h-, respectively (not corrected for 14C02 discrimination). Cell density was 7.65 ,g Chl * ml-' and the calculated mean irradiance was 422 ME. m2 * ± 0.81 %o, n = 4). A nearly identical enrichment was observed in the starch from these same cells (7.36%o). The composition of respired CO2 was unaffected by NH4' assimilation even though the rate of CO2 release increased severalfold (Fig. 4), diluting the exogenous pool of DIC 2 to 3 times.

Isotope Discrimination during NH44 Assimilation and the Application of the Two Process Model

Changes in the 6'3C of exogenous DIC were monitored during photosynthesis at various light intensities. Representative examples of a control and an NH4+-pulsed experiment at saturating light are presented in Figure 5. Isotope data are superimposed upon model-generated scenarios for the ex-

1154 1lant Physiol. Vol. 89, 1989 GUY ET AL.

pected changes in 6'3C of DIC over time for a range of hypothetical discrimination factors and the measured rates of carbon fixation and respiratory CO2 release. Unlike the situation in the dark the isotopic composition of DIC during photosynthesis did not remain constant (Fig. 5). This was true under both control and NH4' pulsed conditions. However, the divergence from the 6'3C value of the source (i.e. respired) carbon was much more dramatic in the absence of NH4+ than in its presence. Approximately 40 mn following NH4' addition, isotopic discrimination had become more pronounced but remained less than controls. Figure 6 shows the complete data set. Discrimination by controls was constant and not significantly affected by light intensity. Total discrimination was 34.7 ± 1.5%o (SE, n = 8) at an irradiance -50 -55 I--

-60 -0

0

-65

'0

-70 -75 1000

I-,

800

C

1-

0 a

600

c

C) Q

c

400

0 C. C. 0

200 0

Time (min) Figure 4. Effects of ammonium assimilation on the isotopic composition and net accumulation of respired C02 by N-limited S. minutum in the dark. Results are reported for two separate experiments with cell densities of 1.3 ,ug Chi * ml' (0) and 2.5 jsg Chi *mlm' (0), respectively. The isotopic composition of whole cells is also reported.

of greater than 450 ,uE m2.s' and 32.0 11.6%o (SE, n =7) at irradiance between 80 and 100 uE. m2s'. The mean total control discrimination of 33.4%o is reported in Figure 6 and was 3.8%o lower that obtained through Rayleigh analysis (Fig. 2). This value was close to the DT expected if Rubisco were the only carboxylating enzyme. Figure 6 also reports discrimination factors determined over the indicated time intervals during NH4' assimilation. These values were based upon the appropriate respiratory rates and the initial rates of gross carbon fixation specific to each interval. It was apparent that NH{' assimilation caused a decrease in DTwhich more closely matched the discrimination expected from PEPcase. This was most pronounced at low light intensity. Over the duration of the experiments (approximately 60 min) some recovery was observed. DISCUSSION Partitioning of Carbon Fixation btwen Rubisco and PEPcase When S. minutum is grown under N-limited conditions, its maximum capacity for NH4' assimilation increases dramatically (28, 29). It is therefore not surprising that the resupply of NH4' to such cells results in major changes in cell metabolism. Previous work has shown that carbon fixation in the light declines dramatically during NH4 assimilation by Nlimited S. minutum (6-8, 27). This suppression does not result from an uncoupling of photosynthetic electron transport but is due to other mechanisms which result in RuBP limitation of Rubisco (6, 7). Figure 1 illustrates both the NH4' induced suppression of gross carbon fixation at saturating light and the enhancement of carbon fixation by PEPcase which occurs in the dark (2, 8, 13). The question then arises as to the relative contributions of PEPcase and Rubisco to cellular carbon fixation during NH4' assimilation in the light. One approach to solving this problem employs stable isotope fractionation techniques. It is well known that Rubisco and PEPcase differ ignificantly in their discrimination against 13C. Rubisco fixes CO2 with a discrimination factor of 29.4%o (11, 24), whereas PEPcase fixes HC03- and discriminates against H13C03 by

NH+- -Pulsed

Control

Figure 5. Model generated scenarios for the change in 613C of DIC over time for a range of discrimination values plotted with actual experimental resuits for representative control and NH4+ pulsed experiments. The measured initial rate of gross carbon fixation (302 Mmol C02-mg Chl-' *h-' for the control and 182 Mmol C02.mg Chl-'- h-1 during NH4+ assimilation) and mean rate of respiratory C02 release were used in the model predictions. In the control experiment, cell density was 1.8 Mg Chl-mlm' and irradiance was 533 ME . m-2.*s1. In the NH4+ pulsed experiment cell density was 2.8 Mg Chl.ml-' and irradiance was 493 ME.

-451 D=40' -50C-)

@30

0

20

-551-1

C.)

I--- 0

-60-

m2 .1

-oa- X

0

.

. 5

10

15

Time

20

25

(min)

30

35

40

Time

(min)

CARBON ISOTOPE FRACTIONATION DURING NH4+ ASSIMILATION 4U 35

.-

-

Rubisco

--*-control discrimination-.

30 -

0

25 20 -

r-

.E

15 -

L..

.e

-E-m2s

r-

O

J~t ----0

T 1

0

-T

r

X

20

30

Time after

40

NH+-Pulsing

=_ 80-100 180-200 _ >450

TI

50

l PEP Case

-T-----

60

7

0

(min)

Figure 6. Changes in the total discrimination factors for carbon fixation, over time, at the indicated irradiance prior to (control) and during NH4+ assimilation by N-limited S. minutum. The total discrimination expected if carbon fixation occurred by either Rubisco (36.9%o) or PEPcase (0.56%o) alone is indicated on the right-hand side.

2.0%o (21). Due to the equilibrium isotope effect of 9%o between CO2 and HCO3- and because most ofthe DIC, which is the substrate analyzed, is in the form of HCO3- at pH 7.0, fixation of CO2 by Rubisco should result in a measured discrimination (ie. DT) of 36.9%o. For similar reasons, fixation of HC03- by PEPcase should result in a measured discrimination of only 0.5%o. These discrimination factors can be used as end members for the estimation of the partitioning of carbon fixation between Rubisco and PEPcase. In the absence of a significant effect of respiration in the light, as in control experiments, Equation 1 can be used to calculate discrimination from measurements of the enrichment of 13C-DIC in the aqueous medium of the cell suspension (Fig. 2). After correcting for the C02/HC03- equilibrium isotope effect, the discrimination exhibited by whole cells was 29.7%o, reflecting the predominance of carbon fixation by Rubisco. If diffusion were at all limiting to photosynthesis (23), this large discrimination would not have been possible. Beardall et al. (3), have suggested that N-limitation may induce a CO2 concentrating mechanism resulting in less discrimination. This is apparently not the case with S. minutum. However, when this alga was grown under conditions of low CO2 we observed a decrease in discrimination to 20%o, presumably reflecting the induction of a CO2 concentrating mechanism (25). To minimize confounding influences, further experiments were therefore carried out with high CO2 grown cells. Although the Rayleigh equation approach to calculating D appears adequate for control conditions, the resulting high respiration rates during N assimilation necessitates a more sophisticated approach. Knowing the rate of gross and net carbon fixation and therefore respiration, a two process model based on Equation 3 can be used to determine discrimination by carbon fixation where the effects of respiration have been taken into account. A requirement of this model is that the isotopic composition of respired CO2 be known. Measurements of the 6'3C of CO2 respired in the dark were approximately 8%o heavier than total cell carbon (Fig. 4). This may

1155

be partially due to the high levels of lipid in N-limited cells and the relative depletion of lipid in "3C (26). The isotopic composition of purified starch was nearly identical to that of the respired carbon from the same cells. This supports our belief that starch was the primary substrate of respiration and is consistent with our observation of NH4W-stimulated starch breakdown in the light (6). Figure 5 presents a family of model-generated curves (based on Eq. 3) representing different scenarios for change in 6'3C over time for a range of possible values for DT. Superimposed on these curves are the actual measured b'3C values. Under control conditions it is clear that discrimination remains fairly constant and is between 30 and 40%o, confirming our earlier conclusions that Rubisco was responsible for the majority of carbon fixed and that there was little or no CO2 concentrating mechanism. In sharp contrast, changes in the b'3C of the DIC upon NH4+-pulsing are much less pronounced (