Triose phosphate utilization limitation: an unnecessary complexity in terrestrial biosphere model representation of photosynthesis

Abstract

Triose phosphates are the principal product of photosynthesis. They are used within the chloroplast for starch synthesis, or translocated to the cytosol where they are used to fuel sucrose synthesis. Use of triose phosphate releases inorganic phosphate, and is under strict metabolic control that matches the supply of triose phosphate from the Calvin–Benson cycle to the demand for carbon by sinks (Heldt & Piechulla, 2011; McClain & Sharkey, 2019). However, a low rate of triose phosphate utilization (TPU) can deplete the phosphate pool, restrict adenosine triphosphate (ATP) synthesis and reduce the availability of ATP to power the Calvin–Benson cycle, thereby limiting photosynthesis (Sharkey, 1985). Recent work has demonstrated the sensitivity of terrestrial biosphere models (TBMs) to TPU (Lombardozzi et al., 2018) and showed that models predict limitation of photosynthesis by TPU most consistently at high latitudes and future elevated CO2 concentration ([CO2]). However, a global scale analysis provided empirical evidence that TPU limitation rarely limits photosynthesis under present day growth conditions and is unlikely to limit photosynthesis at elevated [CO2], even at the low temperatures typical at high latitudes (Kumarathunge et al., 2019a). Additionally, Walker et al. (2021) revealed an artifact in TBM representation of photosynthesis that exaggerates the limitation of TPU on modeled CO2 assimilation. This artifact arises from quadratic smoothing of the transition among the potential limiting processes governing photosynthesis and is closely associated with TBM representation of TPU limitation. Collectively, these recent advances, highlight the need for an examination of the representation of TPU in TBMs. However, current TBM formulations only consider scenarios whereby carbon is maximally conserved by the photorespiratory cycle, i.e. three quarters of the carbon translocated to the peroxisome as glycollate is returned to the chloroplast as glycerate, and αg = 0. Under this assumption Eqn 4 can be simplified to Eqn 3, as originally proposed by Sharkey (1985). Therefore, in many TBMs, representation of the capacity for TPU is based on a speculative relationship with Vc,max. Note that two models (IBIS and LM3) use a different ratio (Table 1) to describe the TPU rate that was derived from a synthesis of photosynthetic CO2 response (A–Ci) curves (Wullschleger, 1993). There is good evidence for the occurrence of TPU limitation of photosynthesis when leaves are exposed to high irradiance and high [CO2] (Sage et al., 1989; von Caemmerer, 2000; Ellsworth et al., 2015; Busch & Sage, 2017; see blue line Fig. 1a) but also at current [CO2] when plants are measured at a low temperature relative to their growth temperature (Yang et al., 2016; Busch & Sage, 2017), or following an abrupt modification of source–sink balance (Fabre et al., 2019). Recent work suggests that TPU rate is typically poised just above the prevailing rate of photosynthesis (Yang et al., 2016; Fabre et al., 2019), highlighting the importance of understanding the short-term dynamics associated with TPU. However, once Ag becomes limited by TPU, adjustments result in a shift to limitation by RuBP regeneration, or more commonly rubisco (Sharkey, 2019). For example, the work of Busch & Sage (2017) considered plants that were grown at current [CO2] and 25°C (F. A. Busch, pers. comm.). When measured at low temperature they observed TPU limited A. However, they did not investigate the effect of photosynthetic acclimation to lower growth temperature that increases investment in carboxylation capacity and the ratio of Jmax : Vc,max (Kumarathunge et al., 2019b), reducing the likelihood of TPU limitation. When plants are measured in their natural growth environment TPU limitation of photosynthesis is rarely observed (Sage & Sharkey, 1987). Recent work (Kumarathunge et al., 2019a) comprehensively demonstrated that TPU limitation of photosynthesis is a rare phenomenon in natural ecosystems. Kumarathunge et al. (2019a) used a large global dataset of A–Ci curves representing 141 species that ranged from the Arctic tundra to tropical rainforests and demonstrated that TPU did not limit light saturated photosynthesis at current atmospheric [CO2] when plants were measured under natural growth conditions, including plants growing at low temperature. Furthermore, they showed that TPU limitation is unlikely to limit photosynthesis until [CO2] is greater than 800 µmol mol−1. TPU limitation has been incorrectly linked to phosphate deficiency (McClain & Sharkey, 2019). Limitation of photosynthesis by TPU is a highly dynamic process influenced by rapid turnover of Calvin–Benson cycle intermediates and is not influenced by whole plant phosphate acquisition. Potential marked variation in leaf level phosphate status resulting from different phosphate nutrition is also unlikely to influence TPU. Acting over a period of hours the vacuole buffers inorganic phosphate concentration in the cell, maintaining cytosolic and plastidic phosphate concentration within a relatively narrow operational range (McClain & Sharkey, 2019). Therefore, plants with a lower total foliar phosphate level are not more susceptible to TPU limitation. As TBMs begin to include representation of the P-cycle (Yang et al., 2014; Goll et al., 2017; Thum et al, 2019; Wang et al., 2020) consideration of TPU limitation of photosynthesis should not be a motivating factor. A recent investigation (Lombardozzi et al., 2018) provided the first assessment of the impact of TPU limitation in a TBM (The Community Land Model, CLMv.4.5). They demonstrated that CO2 assimilation and ecosystem carbon gain were limited by TPU in a future climate, and that by 2100, TPU limitation could reduce global terrestrial carbon gain by 4.5%. When evaluated with a lower TPU rate (0.0835 Vc,max) Lombardozzi et al. (2018) showed a markedly greater TPU limitation of gross primary productivity, particularly at high latitudes, clearly demonstrating the sensitivity of TBM outputs to decreasing the TPU rate (Lombardozzi et al., 2018). Although CLMv.4.5 does include a formulation for photosynthetic acclimation to growth temperature (Lombardozzi et al., 2015), acclimation in CLMv.4.5 is not implemented below 11°C due to the limited temperature range of the underlying study (Kattge & Knorr, 2007), effectively limiting in silico acclimation at high latitudes (Oleson et al., 2013) and exacerbating potential TPU limitation. Furthermore, as noted in Lombardozzi et al. (2018), the prescribed Vc,max at high latitudes was approximately half the rate recently reported for plants growing in the high Arctic (Rogers et al., 2017b) and since TPU is modeled as a fixed fraction of Vc,max (Eqn 6), TPU limitation would be markedly greater when models are parameterized with a lower value of Vc,max. With the exception of FATES, all TBMs that include TPU assume that the temperature response of TPU is identical to Vc,max (Table 1). However, there is evidence that the temperature response of TPU is independent of Vc,max, although the community lacks consensus on whether the TPU rate is more or less sensitive to temperature than Vc,max (Harley et al., 1992; Yang et al., 2016; Kumarathunge et al., 2019a). The extensive analysis conducted by Kumarathunge et al. (2019a) showed that the ratio of the basal rate of TPU to Vc,max at 25°C was 0.09. This estimate is close to the low ratio (0.0835) Lombardozzi et al. (2018) used in their sensitivity analysis, almost half the rate originally assumed by Collatz et al. (1991) and 26% lower than the ratio used by LM3 and IBIS (Eqn 6, Table 1). Note that the ratio used by LM3 and IBIS was derived from just 23 A–Ci curves, from 16 species, grown mostly in controlled environments, and fitted with what are now arguably outdated kinetic constants (Wullschleger, 1993). Importantly, and in contrast to the key TBM assumption of a fixed ratio between TPU and Vc,max, Kumarathunge et al. (2019a) also showed that the ratio between TPU and Vc,max decreases with rising growth temperature dropping from c. 0.2 at 5°C to c. 0.09 at 25°C. The term θ describes the degree of smoothing of the transition between Ac and Aj. When θ = 1 the transition between Ac and Aj is abrupt, and as θ is reduced, progressively greater smoothing is simulated. The resulting term (Acj) represents the smoothed intermediate rate. The transition between Acj and Ap is described by Eqn 8 where β substitutes for θ. This quadratic smoothing is present only in TBMs that implement TPU limitation of photosynthesis, and of those, is present in all but one model (LM3, Table 1). Walker et al. (2021) demonstrated that including quadratic smoothing in model representation of photosynthesis introduces an artifactual fourth limitation that results in a modeled Ag that is always below Ac, Aj and Ap, sometimes markedly so (Fig. 1). Most TBMs that include a formulation for TPU also include quadratic smoothing (Eqns 7 and 8, Table 1), and therefore exaggerate the impact of TPU limitation on A (Fig 1c). In the original formulation θ and β were set to 0.98 and 0.95, respectively (Collatz et al., 1991). However, some models use values for θ and β which are considerably lower (Table 1), further exacerbating the impact of TPU limitation (Friend, 1995; Walker et al., 2021). In combination, inclusion of TPU limitation and quadratic smoothing results in a reduction in A at all [CO2] values and most markedly so at high [CO2] (Fig. 1c; Walker et al., 2021). Projections from models, and consideration of observations, suggest that photosynthesis could become TPU limited at high [CO2] (Busch & Sage, 2017; Lombardozzi et al., 2018). However, these assessments do not account for reductions in Vc,max that result from acclimation of photosynthesis to rising [CO2] that have been well documented in Free-Air CO2 Enrichment studies (Leakey et al., 2009). As a result of photosynthetic acclimation, light saturated photosynthesis will continue to be limited predominantly by Ac in plants grown at elevated [CO2] (Rogers & Humphries, 2000; Ainsworth & Rogers, 2007). However, with the exception of optimality approaches (e.g. Smith et al., 2019; Stocker et al., 2020), formulations for the representation of photosynthetic acclimation to elevated [CO2] are absent from TBMs. TPU limitation of A is an important phenomenon to understand but it has received little attention compared to the processes of carboxylation and electron transport. There is a need for further research to better understand the short-term dynamics of TPU, its temperature response, the role TPU might play in signaling source–sink balance (McClain & Sharkey, 2019; Sharkey, 2019), and circumstances where glycollate might not be maximally conserved by the photorespiratory cycle (Ellsworth et al., 2015; Busch et al., 2018). It also remains important for physiologists to continue to consider TPU when analyzing A–Ci curves to avoid potential underestimation of Jmax (Busch & Sage, 2017; Sharkey, 2019). However, when plants are grown in their natural environment the best current evidence shows that TPU does not limit A below a [CO2] of 800 µmol mol−1 (Kumarathunge et al., 2019a) and photosynthetic acclimation of plants to rising [CO2] suggests that TPU limitation of A is also unlikely at higher [CO2] (Leakey et al., 2009) questioning the value of including representation of TPU in TBMs. Current TBM formulations of TPU are founded on uncertain assumptions, most models do not account for the independent temperature sensitivity of TPU, or capture the temperature dependence of the ratio between the basal rate of TPU and Vc,max. The inclusion of TPU in combination with quadratic smoothing, results in an artificial reduction in CO2 assimilation (Walker et al., 2021) that increases the limitation of A by TPU (Fig. 1) and changes the fundamental mechanistic description of how plants respond to rising [CO2]. Collectively these issues suggest that inclusion of TPU limitation of A in TBMs has introduced additional parameter uncertainty which has, or could, result in compensatory tuning of Vc,max. Given that Vc,max is linked to several other important processes in many TBMs through simple multipliers, (e.g. respiration), more mechanistic approaches (e.g. nitrogen allocation), and has a strong influence on diverse model outputs (Rogers, 2014; Ricciuto et al., 2018; Walker et al., 2021), the resulting tuning will have a pervasive influence on the veracity of model projections. We advocate for the removal of current formulations of TPU from TBMs. The authors received support from the Next-Generation Ecosystem Experiments – NGEE Arctic (AR, SPS) and NGEE Tropics (AR, SPS, APW) projects that are supported by the Office of Biological and Environmental Research in the Department of Energy, Office of Science, and through the US Department of Energy contract number DE-SC0012704 to Brookhaven National Laboratory. Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the Department of Energy under contract DE-AC05-1008 00OR22725. DLL was supported by the National Institute of Food and Agriculture (NIFA)/US Department of Agriculture (USDA) grant 2015-67003-23485. The National Center for Atmospheric Research is a major facility sponsored by the National Science Foundation under Cooperative Agreement 1852977. AR wrote the manuscript with valuable contributions from DPK, DLL, BEM, SPS and APW.