Photosynthesis Research 29: 23-35, 1991. © 1991 Kluwer Academic Publishers. Printed in the Netherlands.

Regular paper

Calculation of absolute photosystem I absorption cross-sections from P700 photo-oxidation kinetics Warren Zipfel & Thomas G. Owens Section o f Plant Biology, Cornell University, Ithaca, N Y 14853-5908, U S A Received 11 February 1991; accepted in revised form 24 May 1991

Key words:

absorption cross-section, P700, antenna size, photosystem I, photosynthetic unit size,

cation effects Abstract

A procedure is described which permits determination of the absolute absorption cross-section of a photosynthetic unit from the kinetics of reaction center photo-oxidation under weak, continuous actinic illumination. The method was first tested on a simple model compound of known absorption cross-section. We then applied the technique to absorption cross-section and functional antenna size measurements in photosystem I (PSI). A kinetic model is presented that can be used to fit P700 photo-oxidation measurements and extract the effective photochemical rate constant. The procedure is shown to properly correct for sample scattering and for the presence of heterogeneous absorbers (pigments not functionally coupled to P700). The relevance of these corrections to comparisons of antenna size using techniques that measure 'relative' absorption cross-sections is discussed. Measurements on pea thylakoids in the presence and absence of 5 mM MgCl 2 show a 45% increase in P S I absorption cross-section in unstacked thylakoids. Analysis of detergent-isolated 'native' P S I preparations (200 chlorophyll a + b/P700) deafly indicate that the preparation contains a broad distribution of antenna sizes. Finally, we confirm that Chlamydomonas reinhardtii strain LM3-A4d contains a P S I core antenna complex which binds only ~60 chlorophyll a/P700, about half the functional size of the wild type complex. Limitations associated with calculation of functional antenna size from cross-section measurements are also discussed. PS - photosystem; PS 1-200 - detergent-isolated photosystem I preparation containing about 200 Chl a + b/P700; Axxx - absorbance at x x x nm; or- absolute absorption cross-section; I a - rate of light absorption; I o - incident actinic light intensity; ~bp - quantum yield of photochemistry; keff effective rate constant for P700 photo-oxidation measured under conditions of limiting actinic intensity; k r - rate constant for P700 + reduction

Abbreviations:

Introduction

Photosynthetic organisms convert light energy into chemical activity with remarkable efficiency; under optimal conditions, a photon absorbed anywhere in the chlorophyll antenna o f the photosynthetic unit leads to photochemical charge separation on the reaction center Chl

with a probability approaching unity. Although this probability for the propitious use of an absorbed photon has been maximized by parameters such as the organization of the pigments and the protein environment, the organism must also balance the overall probability for absorption of a photon against increasing losses via competing processes in a larger antenna. At the

24 molecular level, efficient photon absorption is facilitated by the use of a strongly absorbing antenna pigments with a broad range of absorption characteristics and by the functional coupling of hundreds of these antenna pigments to each reaction center. It is the efficient coupling of many antenna pigments to a single reaction center pigment that increases the effective absorption cross-section of the reaction center, and therefore increases the rate of photochemistry at a given light intensity. The absorption crosssection of a photosystem (reaction center and coupled antenna pigments) is therefore an important parameter in the study of photosynthesis. In general, the absolute absorption crosssection of a molecule is a measure of the target size the molecule presents for absorption of a photon of a particular wavelength. Although an absolute absorption cross-section has units of area per molecule, it is a function of the quantum mechanical properties of the molecule and is not directly related to the actual physical crosssection. For example, a single Chl a molecule has a physical cross-section of 150/~2 but has an in vitro absorption cross-sections of about 2/~2 at its red and blue absorption maxima. An experimentally determined absorption cross-section is the average probability for absorption by a single molecule measured in an isotropic population of molecules at a specific wavelength and in a given chemical environment. The functional absorption cross-section of coupled aggregate of absorbing molecules that efficiently transfer excitation energy to a single photochemical reaction, i.e., a photosynthetic unit, is defined as the product of the quantum yield for the photochemical reaction multiplied by the absolute absorption cross-section of the complex (Mauzerall and Greenbaum 1989). For a case where the quantum yield of the photochemical reaction is near unity, such as PSI (Hiyama 1985, Owens et al. 1987, 1990), the functional cross-section is equivalent to the absolute absorption cross-section of the complex. The absolute absorption cross-section of PS II units has been determined using flash-induced 0 2 evolution (Ley and Mauzerall 1982) while similar measurements of PSI cross-section have utilized respiratory oscillations (Greenbaum et al. 1987),

H 2 production (Boichenko and Litvin 1986) or electron paramagnetic resonance (Weaver and Weaver 1969) to quantify photochemical reactions. The average number of pigment molecules that compose the photosynthetic unit of a reaction center (the functional antenna size) can be calculated, assuming that the cross-section for the photosynthetic unit is simply a summation of the individual cross-sections of its pigment molecules (Mauzerall and Greenbaum 1989). Relative absorption cross-sections for PSI and PS II have also been estimated from a variety of kinetic measurements (see Mauzerall and Greenbaum 1989 for a review) including the rate of light-induced absorbance changes at 700 and 320nm, respectively, measured under limiting actinic illumination. Relative cross-sections, in the form of an effective rate constant for photochemistry, have been used in many studies including light-adaptation responses in Chlamydomonas reinhardtii (Neale and Melis 1986), calculation of the number of Chls per photosystem (Melis and Anderson 1983), comparison. of PS I antenna sizes between phosphorylated and non-phosphorylated spinach thylakoids (Haworth and Melis 1983) and between thylakoids suspended in the presence and absence of Mg 2÷ (Melis 1982). Telfer et al. (1984) also investigated the effect of phosphorylation and Mg 2÷ concentration on relative PSI antenna size by measuring the flashinduced absorption increase at 820 nm. Although this technique differs from that used by Haworth and Melis (1983) and Melis and Anderson (1983), (flash excitation versus continuous low intensity actinic light), both techniques provide a relative measurement of PSI absorption crosssection. However, results from the two sets of experiments are not in agreement. The flashinduced AAs20 measurements indicate an increase in relative PSI cross-section following thylakoid protein phosphorylation or cation deletion (Telfer et al. 1984) while Melis (1982) and Haworth and Melis (1983) report no change in the relative antenna size for either treatment. In this report we demonstrate how absolute PS I absorption cross-sections can be calculated from the effective rate constant for P700 photooxidation (kef~) measured under continuous, low intensity actinic illumination, and provide a

25 kinetic model for extracting k~ff. Unlike the flash yield technique of Mauzerall and coworkers (Greenbaum et al. 1987), which measures the effect of PS I turnover on the subsequent rate of dark respiration (and thus requires whole cells), our procedure is more generally applicable to sub-cellular preparations. This technique rigorously accounts for the effects of non-functional, heterogeneous absorption (pigments whose absorption does not contribute to P S I photochemistry, e.g., PS II pigments) and scattering on the absolute PS I cross-section. We demonstrate the errors that arise from the effects of heterogeneous absorption when using photochemical rate constants directly as a measure of relative absorption cross-section. The limitations of the technique for determining a single crosssection as opposed to a distribution of crosssections are evaluated. Finally, we apply the technique to an evaluation of functional P S I antenna size in isolated P S I preparations, to the effects of divalent cations on P S I antenna size in thylakoids, and to a mutant of C. reinhardtii which was previously shown by time-resolved techniques to have an unusual PS I core antenna size (Owens et al. 1987, 1989).

Materials and methods

Measurement o f the rate constant for P700 photo-oxidation

Photobleaching kinetics were determined on a laboratory-constructed single beam spectrophotometer. The 697 nm measuring beam, which had no detectable actinic effects, was modulated at 100kHz using an photo-acoustic modulator (Morvue Electronic Systems, Newberg, Oregon) between crossed polarizers. The absorbance signal was processed by a lock-in amplifier (time constant 3ms) interfaced to a PC computer. Absorbance values were digitized at 5 ms intervals. Actinic illumination (1.1/~mol photons m -2 s -1) was provided by a quartz-halogen lamp projected through a 436 nm filter (bandwidth = 10 nm). Samples were placed in a cuvette with a pathlength of 10 mm along the measuring axis and 5 mm along the actinic axis. Traces were averaged on the same sample 3 to 5 times for

PS 1-200 preparations, and 8 to 15 times for thylakoid samples. Photobleaching was usually completely reversible in the dark; care was taken to average only those traces which reached the same extent of bleach in order to avoid the creation of artificial multiple exponential kinetics. The averaged data was digitally filtered using standard finite impulse response techniques (Press et al. 1988) to create a lowpass filter with a corner frequency of 20 Hz (10% of the sample frequency). The digitally filtered data returned the same values of the kinetic fitting parameters as the unfiltered data, however deviation from a single exponential fit is much more readily realized (by inspection of residual plots) with filtered data. Effective rate constants for P700 photo-oxidation (kcff) and P700 ÷ reduction (kr) were determined by fitting the kinetic data to the integrated form of the first-order rate equation: keff (1 [P700+](t) = [P700]0 keff + k r

--

e -(keff+kr)t) (1)

(parameterized as A ( 1 - e Bt) using standard non-linear least squares fitting routines (Press et al. 1988). Absorbance values used in the kinetic measurements were normalized between values of 0 (P700 fully reduced) and 1 (P700 fully oxidized); the latter was determined using a 50X higher actinic intensity such that kef f >> k r. In all kinetic measurements at normal actinic intensity, the steady state value of [P700 +] was always

.... ''

"''""-'"': """-"''~''::'"

.......

,

0.06

I¢"-%-

b 7.,m~. ~.;.r.-. . . . . . .

-0

06 [

"""

].0

O

o p.,. [2_

"

i

I

I

I

0,8

0.8 "El (1) N

0.4 E L. o z

0.2

0.0 0

I

I

I

1

2

3

4

Time (s) Fig. 2. PT00 photo-oxidation kinetics from a detergentisolated PS 1-200 preparation. (a) D a t a was normalized relative to the m a x i m u m a m o u n t of photo-oxidizable P700 (actual AA697 = 0.012) and fit to the kinetics model described in the text. Regression line shown in (a) is from a fit to a single exponential term. Calculated rate constant was 1.64 s -1. (b) Residuals resulting from a fit to a single exponential term. (c) Residuals from a fit to three exponential terms; weighted m e a n rate constant was 2.10 s -1.

is composed of two or more sub-populations of different antenna sizes. However, the actual distribution cannot be determined. At the signal-tonoise ratio (~250) of the data, fits with two or more exponential components all provide suitable models for the PS 1-200 particles, with little difference in the distribution of residuals. The weighted mean rate constant converges to a value of 2.10s -1 for fits with four or more exponential components; this value was subsequently used to calculate a weighted mean absolute cross-section of 668/~2/molecule for the sample (Table 1). The distribution of antenna sizes is asymmetric about the mode as indicated by the non-equality of the arithmetic mean (1.96s -1) and the weighted mean (2.10s-1); there are more large antenna size particles (greater than the mode) than small (less than the mode).

29 Table 1. Absolute absorption cross-section of PSI measured at 436 nm in various preparations. Functional antenna sizes (Chl a + b/P700) were determined using Eq. (A.13) and probably underestimate the actual antenna sizes due to assumptions concerning the origin of absorption at the actinic wavelength (see text for details). Data for absolute cross-sections are reported as mean --- standard deviation (n), where n is the number of replicates

Sample

PS 1-200 PS 1-200 pea thylakoids ( + Mg 2+ ) pea thylakoids ( - M g 2÷) C. reinhardtii (wt) C. reinhardtii (A4d)

Absorption cross-section

Chl a + b/PTO0

(/~2/m°iecule)

(total)

(functional)

668 -+ 31 (11) 676 -+ 29 (4) 858 -+ 44 (4) 1292 -+ 38 (4) 966 -+ 40 (4) 436 -+ 34 (4)

201 253 711 689 687 115

191 193 220 340 247 91

One possibility, especially relevant with detergent isolated preparations, would be the formation of in vitro aggregates of PS 1-200 complexes in which excited states could freely migrate between antenna domains of each reaction center. Although the effects of aggregation on the measured rate constant for photobleaching would depend on the kinetics of excited state movement between complexes, such an aggregate would exhibit a larger rate constant then a single complex and could account for the appearance of the largest rate constants in the PS 1-200 kinetics fits. These data and simulations demonstrate that for reasonable signal-to-noise ratios, it is possible to distinguish between a single functional antenna size and a moderate distribution (factor of 2 or greater range) of antenna sizes. Under no physiological circumstances is it possible to describe the actual distribution of antenna sizes. However, this limitation does not preclude the ability to distinguish small differences (10-20%) in mean antenna size between samples, as will be demonstrated in the following sections. Verification of the absorption cross-section calculations

Equation (4) shows that kef f is uniquely determined by the absorption cross-section of the photoactive complex, the incident light intensity, and the total scattering-corrected absorption of the sample, all measured at the actinic wavelength. The validity of this relationship was tested in its simplest form (monochromatic actinic light and homogenous sample), using a pure solution of Reinecke' salt. The photo-aquation

reaction of Reinecke's salt: KCr(NHa)z(SCN)4 + H20 + h v ~ KCr(NH3)2(SCN)3(H20 ) + SCN is first order and is easily measured by quantifying the amount of photo-released SCN at various times during the reaction. Using monochromatic illumination from a helium-neon laser (I o = 0.00246/xmol c m - Z s - 1 ) a n d 2.09/zmol Reinecke's salt in the reaction cuvette initially, a photo-aquation rate of 2.79 × 10 -5/zmol s -1 was determined. The quantum efficiency and absorption cross-section at 633 nm was interpolated from the data provided in Wegner and Adamson (1966) and have values of 0.277 and 0.0204cm 2 ~mole -1, respectively. The calculated value for the absolute absorption crosssection at 633 nm is 0.0200 cm 2/~mole -1, in excellent agreement with the published data. Accurate measurement of photochemical rate constants in photosynthetic samples is complicated by the presence of heterogeneous absorbers (e.g., PS II antenna pigments in a PSI measurement) and by Scattering from biological membranes. One important assumption used in deriving Eq. (4) is that the transmittance of the sample decreases exponentially with increasing pathlength (Eq. (A.4)). This was investigated for thylakoid preparations containing identical Chl concentrations (25/zgm1-1) suspended in the presence or absence of 5 m M M g 2+. In both cases, the absorption at 436nm was a linear function of pathlength (1-10 mm) and extrapolated through the origin (data not shown). This confirms that any attenuation due to off-axis scattering within the sample has been removed

30 (Mauzerall and Greenbaum 1989). Without scattering correction, the absorbance of the + Mg 2÷ sample was 16% greater than the Mg 2÷ depleted sample; after scattering correction, the absorbance remained slightly larger (4%) in the + Mg 2÷ samples. This represents an increase in the true absorption of the + Mg 2÷ sample due to a longer average pathlength in the more highly scattering solution. The validity of Eq. (4), which expresses k as a function of I o and the scattering-corrected sample absorbance can be tested experimentally by the addition of neutral absorber and/or scatterer to a sample with a known absorption crosssection. Figure 3 is a plot of k~ff versus scattering-corrected sample absorbance for a PS 1200 preparation (open symbols) and isolated pea thylakoids (filled symbols). For each sample type, the data points at lowest absorption represent the controls; higher scattering-corrected sample absorptions were produced by successive addition of neutral absorber (FD&C yellow #5, absorption maximum at 436 nm, no absorption at 697 nm; circles) and neutral scatterer (dry milk powder; triangles). The solid lines represent the

,3.5 5.0 2.5 2.0

I

1.5 ,c

1.0 0.5 0.0 0.0

I

I

I

I

1

I

0.5

1.0

1.5

2.0

2.5

,3.0

5.5

Absorption

Fig. 3. P700 photo-oxidation rate constant measured as a

function of the scattering-corrected sample absorbance at 436 nm. Open circles, PS 1-200 with additions of neutral absorber (436nm maximum absorbance, 697 nm no absorbanee); filled circle~, pea thylakoids with additions of neutral absorber; open triangles, PS 1-200 samples with added neutral scatterer plotted at their scattering-corrected 436 nm absorbance; open squares, PS 1-200 samples with added neutral scatterer plotted at their apparent 436 nm absorbance. Curves shown correspond to the variation in rate constant with sample absorbance predicted by Eq. (4).

function predicted by Eq. (4), holding the control values of ~bp, I o and o-1 constant and varying only A. The uncorrected (for scattering) absorbances of the PS 1-200 samples with added scatterer (open squares) are also shown. The data show that when the sample absorbance increases because of non-functional pigments, keff decreases due to additional shading of the functional pigments by the non-functional ones. In contrast, when the (uncorrected) sample absorbance increases due to non-absorbing scatterers, keff actually increases slightly. Again, scattering results in a higher rate of photon absorption because of the longer mean pathlength of the scattered photons. This produces a slightly faster rate of photochemistry than would be expected for a given cross-section (Fig. 3); k~ff for the PS 1-200 sample with added scatterer (open triangles) are approximately 10% higher than the same sample without scatterer (open circles directly below). The agreement between the experimental data and the predicted function indicates that absolute PS I absorption cross-sections can be accurately determined in the presence of PS II or other interfering pigments and in dilute, scattering solution. Figure 3 also demonstrates the difficulties involved when using k~ff directly as an indication of relative antenna size. In Fig. 3, the absorption cross-section is the same in all of PS 1-200 data (open symbols); the only reason k~ff varies is because of overall absorption differences between samples. The same is true for the thylakoid data (filled symbols). If the absorption properties of the samples being compared are not corrected for, misleading interpretations can arise. The steepness of the function at normal 'working' absorbances (~