Tracking the origins of the Kok effect, 70 years after its discovery

Abstract

The 18th New Phytologist Workshop was dedicated to possible causes of the Kok effect, the typical break in the light response curve of net photosynthesis. Available data obtained since its discovery in 1948 show that the effect is not purely caused by a down-regulation of respiration, contrary to the commonly accepted view. However, estimates of leaf respiratory rates obtained in various ecosystems with techniques including the Kok method appear to be widely consistent across different studies, suggesting that Kok-derived values can be used as a surrogate for actual day respiration values. Gross CO2 assimilation of photosynthetic organs of plants is accompanied by concurrent efflux of CO2 by photorespiration and day respiration (i.e. nonphotorespiratory CO2 evolution in the light). While the rate of photorespiration can be predicted using the internal CO2 mole fraction and equations that describe gas exchange (taking into account the stoichiometry of CO2 liberation with respect to O2 fixation by ribulose-1,5-bisphosphate carboxylase/oxygenase), estimating day respiration is much more challenging because there is no equation that can predict its rate as a function of net photosynthesis, CO2 mole fraction or other environmental parameters. That is, in equations describing gas exchange (or isotopic mass balance), day respiration (Rd) has to be determined separately or simply assumed to model net carbon (C) exchange. At the leaf level, day respiration represents a C loss of c. 5% of gross-fixed CO2 but this proportion is highly variable, depending on species and conditions (see, e.g. Atkin et al., 1997). Estimates of day respiratory CO2 loss rely on specific techniques used to measure Rd: amongst them, the Kok method is certainly the most popular, because it is easy to implement in the laboratory or in the field using classical gas-exchange systems. This method takes advantage of the ‘Kok effect’, a phenomenon first described in the 1940s in unicellular algae (Kok, 1948, 1949). This effect is further described later, and in Fig. 1. The Kok effect is believed to be primarily caused by the inhibition of respiration by light and thus provides a direct way to estimate Rd. At the present date, c. 800 published works have used, or cited, the Kok method, representing c. 40% of articles that involve a measurement of Rd or deal with day respiration. However, some persisting doubt remains about the validity of this method, simply because the Kok effect is inconstant and influenced by environmental conditions (such as O2 mole fraction) in ways that may not be consistent with day respiratory metabolism. Considering the wide range of applications, and the considerable number of articles that have been published, there is an urgent need to clarify the origin of the Kok effect and to evaluate its relevance to measure Rd. This was the objective of the 18th New Phytologist Workshop that took place in July 2016 in Angers (France). The ‘Kok effect’ refers to the change in quantum yield of net photosynthesis (Φ) at low light levels: at very low light levels (typically 0–20 μmol m−2 s−1 of incident photosynthetically active radiation, iPAR), the quantum yield (denoted as Φ1) is larger than that observed at higher light levels (Φ2). In practice, when a light response curve of net photosynthesis is performed, there is a change in the slope and a break point (examples are shown in Fig. 1). In general, Φ1 is c. 0.1 under standard conditions (25°C, 21% O2 and 380 μmol mol−1) while Φ2 is c. 0.06 (Fig. 2). These values are rather similar when net photosynthesis is measured as CO2 fixation or O2 release (but data on the assimilatory quotient at low light, presumably close to 1, are scarce). Thus, the relative change in quantum yield above the break point is about (0.1–0.06)/0.1 = 40% at ambient CO2 (380 μmol mol−1). Note that computing a true value of quantum yield requires a correction for leaf absorbance so as to convert incident radiation into absorbed light. The extrapolated intercept associated with the second portion of the response curve gives an estimate of Rd (illustrated in Fig. 1a), which is typically lower than Rn, the rate of respiration in darkness (night respiration). In other words, in this region of irradiance, the response curve of net assimilation is modelled as A = Φ2·iPAR·α – Rd while at very low light, it is modelled as A = Φ1·iPAR·α – Rn, where α is leaf light absorbance. Problems associated with the Kok method itself should be recognized. First, in practice, carrying out a light response curve at very low light can be difficult due to the small difference between inlet and outlet air in open gas exchange systems (since A is low) and leaks, even very modest, can be an issue. Second, observing the two linear portions of the light response curve (and thus calculating Rd, Φ1 and Φ2) can be rather difficult when the number of data points is limited. A good graphical resolution is also necessary to see the break point (e.g. compare the resolution of Fig. 1a and b). Consequently, there is often some uncertainty in the choice of data points to draw linear regressions. Including or excluding points in the presumed neighbourhood of the break point can change Φ-values and Rd significantly. For example, in Fig. 1(a), excluding and including the third point gives Φ1 values of 0.099 and 0.085, respectively, and gives Rd values of 0.47 and 0.35 μmol m−2 s−1, respectively. A recommendation to solve this problem is to have a sufficient number of measurements: typically, at least three in the 0–10 μmol m−2 s−1 region. The Kok method is used widely to estimate Rd, including in wide-spectrum studies carried out in different species or ecosystems under various conditions. For example, the Kok method has been implemented recently in an unpublished world-wide survey presented at the Workshop by Owen Atkin, Mary Heskel and others, in arctic species (Heskel et al., 2014), in tropical tree canopies (Weerasinghe et al., 2014), in trees in different seasons (Way et al., 2015) and at varying CO2 (Crous et al., 2012; Kroner & Way, 2016), or in different species along a vegetation chronosequence (Atkin et al., 2013). The usefulness of Kok-derived estimates of leaf respiration in the light for ecosystem C budget studies has been extensively discussed (Heskel et al., 2013). Interestingly, the Kok effect has been shown to scale up to the ecosystem, that is, with a break in the response curve of net ecosystem uptake of CO2 to measured irradiance (Bruhn et al., 2011). At the Workshop, it has been recognized that in general, Kok-derived estimates of Rd are lower than Rn by 20–40%, consistent with the well-accepted inhibition of leaf respiratory metabolism by light. Comparisons with Rd values obtained using other techniques (such as the Laisk method, which takes advantage of response curves to CO2 mole fraction) have also been shown to be rather satisfactory despite some variability (see, e.g. Villar et al., 1994). Further data presented during the Workshop also showed a relatively good agreement between Kok-, Laisk- and isotope-derived Rd values in spinach, cocklebur and Magnolia leaves (Barbour et al., 2017). The widely-accepted (historical) origin of the effect is the inhibition of respiratory metabolism by light (linear decrease of Rd with light) and in fact, mechanisms for the down-regulation of respiratory decarboxylation reactions by light have been described (reviewed in Tcherkez et al., 2012). In addition, the pentose phosphate pathway (PPP), which also liberates CO2, has been shown to be inhibited by light, even at very low light levels (Singh et al., 1993; Farr et al., 1994). A metabolic steady-state model has also suggested that at low light, the enhancement of the PPP can potentially explain the Kok effect (Buckley & Adams, 2011). However, a purely respiratory (catabolic) origin of the Kok effect is highly unlikely. In fact, it strongly depends on gaseous conditions whereas Rd is not expected to be very sensitive to CO2 and O2 mole fraction. The Kok effect disappears at low oxygen (Fig. 1b,c; Cornic & Jarvis, 1972; Ishii & Murata, 1978; Sharp et al., 1984), suggesting that photorespiration could be involved. The Kok effect also depends on CO2: the relative difference between Φ1 and Φ2 decreases, but does not disappear, at high CO2 mole fraction (Fig. 2), suggesting again that photorespiration could explain part of the effect. It should nevertheless be noted that the Kok effect disappears at extremely high CO2 (≥ 1%) (Björkman & Demmig, 1987; Evans, 1987) but this observation might not be very conclusive due to side effects of extremely high CO2 on C metabolism (including cellular acidification and inhibition of respiration). Potentially, a photorespiratory origin could be due to: (1) a different photorespiratory metabolism at low light (such as a change in O2/CO2 stoichiometry) thereby making the ‘scaling factor’ (cc − Γ*)/(cc + 2Γ*) erroneous in Eqn 1. Recently, slight changes in photorespiratory stoichiometry have been found at high O2 or low CO2 but significant changes at very low photorespiration rates seem unlikely (Abadie et al., 2016); or (2) changes in cc along a light curve. Usually, the classical correction used to adjust A values to what they would be if intercellular CO2 (ci) were constant (Kirschbaum & Farquhar, 1987) is minimal and does not suppress the Kok effect. Still, the second hypothesis appears very likely, through the influence of internal conductance so that cc/Γ* (rather than ci/Γ*) increases considerably at low light (see the companion article Farquhar & Busch, 2017). It is nevertheless improbable that an effect on cc only can explain the Kok effect in totality. In fact, the effect persists at high CO2 (Fig. 2). Furthermore, it has been originally described in unicellular algae with a carbon concentrating mechanism (CCM) (Kok, 1948) and has also been found in other CCM-containing algae (Peltier & Sarrey, 1988). Also, the break in the light response curve, when it happens to be visible (as in Fig. 1c), would not be easy to explain since there is no clear reason for a discontinuous effect of internal conductance on cc/Γ*. It should also be noted that the effect of gaseous conditions might not be inconsistent with metabolism: under the steady-state hypothesis, the balance of reductive power predicts that PPP activity should depend on CO2 mole fraction (Buckley & Adams, 2011). Unfortunately, there is presently no published data (of either metabolomics or fluxomics) obtained at very low light along a Kok curve. Therefore, fluxes in catabolic pathways responsible for CO2 generation at very low light are not very well known. Recent unpublished data obtained using isotopic (13C) labelling and presented during the Workshop by Gauthier and co-workers have nevertheless suggested that at very low light, decarboxylation by the pyruvate dehydrogenase is up-regulated. Finally, other mechanisms associated with electron transport cannot be excluded. First, at very low light, there is an abrupt decrease in the cyclic electron flux around PSI that disappears under 2% O2, thereby suggesting that γ can change, may be due to the Mehler reaction (Laisk et al., 2005; Kou et al., 2013). Second, both the light partition to PSII (a) and the photochemical yield of PSII (ΦPSII) have been found to increase at low light (Oberhuber et al., 1993; Yin et al., 2014). A summary of possible explanations of the Kok effect is shown in Table 1. It is clear that the origin of this effect is not unique, and it is likely a combination of several processes that lead to an increase in the quantum yield of photochemistry, and cause gaseous (decrease in photorespiration due to the increase in cc/Γ*) and metabolic changes at very low light. In an effort to disentangle the mechanism of the Kok effect, more experiments should be done at very low light to ascertain catabolic pathways involved, examine electron transport parameters and the CO2/O2 assimilatory quotient, and use species where the Kok effect does not occur like C4 plants (Cornic & Jarvis, 1972; Ishii et al., 1979) and perhaps, C3/C4 intermediates. There is little doubt that the rate of day respiration Rd is lower than Rn because it has been shown using several methods (for a review, see Tcherkez & Ribas-Carbó, 2012). However, it seems clear that the Kok effect is not purely respiratory and thus, the values of Rd or Rd/Rn obtained with the Kok method have to be considered as proxies.