Triose phosphate utilization and beyond: from photosynthesis to end product synthesis

Photosynthetic triose phosphate utilization limitation is discussed, highlighting misleading points in physiology and focusing on regulation.

Abstract

During photosynthesis, plants fix CO₂ from the atmosphere onto ribulose-bisphosphate, producing 3-phosphoglycerate, which is reduced to triose phosphates (TPs). The TPs are then converted into the end products of photosynthesis. When a plant is photosynthesizing very quickly, it may not be possible to commit photosynthate to end products as fast as it is produced, causing a decrease in available phosphate and limiting the rate of photosynthesis to the rate of triose phosphate utilization (TPU). The occurrence of an observable TPU limitation is highly variable based on species and especially growth conditions, with TPU capacity seemingly regulated to be in slight excess of typical photosynthetic rates the plant might experience. The physiological effects of TPU limitation are discussed with an emphasis on interactions between the Calvin–Benson cycle and the light reactions. Methods for detecting TPU-limited data from gas exchange data are detailed and the impact on modeling of some physiological effects are shown. Special consideration is given to common misconceptions about TPU.

Introduction

Triose phosphate utilization (TPU) is one of the three canonical biochemical limitations of photosynthesis in gas exchange analysis of C₃ plants. It reflects a steady-state condition in which assimilation of carbon is limited by the ability to regenerate phosphate through production of end products of photosynthesis. Phosphate is required by ATP synthase to produce ATP, of which three are needed to fix a single carbon. Although all three ATPs are used for phosphorylation of carbon chains, two are immediately released when the 3-phosphoglyceric acid (PGA) kinase reaction is followed by glyceraldehyde-3-phosphate (GAP) dehydrogenase. Regeneration of ribulose 1,5-bisphosphate (RuBP) releases two phosphates per three fixed carbons, one from fructose-1,6-bisphosphatase (FBPase) and one from sedoheptulose bisphosphatase (SBPase). One phosphate per three carbons remains on the triose phosphates (TPs) GAP and dihydroxyacetone phosphate (DHAP), which are used for synthesis of starch and sucrose. The capacity for end product synthesis relative to carbon fixation can determine the concentration of inorganic phosphate. If the capacity for TPU is high relative to carbon fixation, the concentration of phosphate will be high. A high concentration of phosphate will inhibit starch synthesis and, less so, sucrose synthesis, changing the partitioning of carbon among the end products. A high concentration of phosphate could also make ATP synthesis easier and so interfere with the acidification of the stromal lumen, which is necessary to induce energy-dependent quenching (q_E) in PSII. If TP use is too quick relative to carbon fixation, it may deplete Calvin–Benson cycle intermediates and lead to difficulty in regenerating RuBP. On the other hand, if the capacity for TPU is low relative to carbon fixation, the phosphate concentration decreases, leading to reduced conductivity of protons through thylakoid ATP synthase that ultimately slows photosynthesis (Kanazawa and Kramer, 2002; Takizawa et al., 2008; Kiirats et al., 2009). One minute after becoming TPU limited, the ATP/ADP ratio can fall from 2.3 to 1.2 although after 18 min other regulatory processes can allow it to recover to 1.6 (Sharkey et al., 1986b).

The decline in ATP is a form of feedback limitation and is potentially quite dangerous to the plant. Feedback conditions are known to cause photodamage due to the inability to move energy downstream (Pammenter et al., 1993; Takizawa et al., 2008; Kiirats et al., 2009). To avoid photodamage, instead of maintaining phosphate-restricted feedback, a series of regulatory steps are engaged to slow photosynthetic electron transport and carbon fixation by Rubisco. While the capacity is determined by phosphate balance, the steady-state rate is set by regulatory effects that serve to ameliorate feedback conditions. This includes reduction in the PSII quantum yield (Φ_PSII) (Sharkey et al., 1988; Kiirats et al., 2009) and reduced activation state of Rubisco (Sharkey et al., 1986a; Socias et al., 1993; Viil et al., 2004; Cen and Sage, 2005). In this review, we discuss the effect of end product synthesis on the overall rate and regulation of photosynthesis.

How are triose phosphates used?

The maximal photosynthetic rate under TPU limitation is primarily, but not exclusively, determined by the rate of conversion of TPs into starch and sucrose. The synthesis of sugar alcohols in some plant species (Escobar-Gutiérrez and Gaudillère, 1997; Loescher et al., 2000) has the same effect as sucrose synthesis. The limitation on assimilation is based on the release of phosphate from Calvin–Benson cycle intermediates that leave the cycle, and the most immediate release is from the activity of FBPase in the chloroplast for starch synthesis or in the cytosol for sucrose synthesis. Sucrose synthesis begins with the translocation of TPs through the triose phosphate/phosphate translocator (TPT) (Riesmeier et al., 1993). This removes carbon from the Calvin–Benson cycle and returns phosphate from the cytosol to the chloroplast. Each sucrose molecule requires the combination of two hexose molecules, for a total of four triose phosphates. Net phosphate release from organic phosphates during sucrose synthesis occurs at FBPase (2), UDP-glucose pyrophosphorylase (1), and sucrose-phosphate phosphatase (1). Sucrose synthesis is typically measured at between 25% and 50% of total carbon assimilation (Sharkey et al., 1985; Escobar-Gutiérrez and Gaudillère, 1997; Szecowka et al., 2013; Abadie et al., 2018), with some studies demonstrating up to 75% (Stitt et al., 1983). It is likely that the species and environmental conditions have an effect on partitioning of carbon into sucrose.

In starch synthesis, phosphate release occurs at stromal FBPase and ADP-glucose pyrophosphorylase. The flux to starch varies considerably with the growth conditions of the plant; for example, Arabidopsis growing in an 18 h photoperiod committed only 24% of fixed carbon to starch but in a 6 h photoperiod committed 51% (Sulpice et al., 2014). Other studies show that between 30% and 60% of fixed carbon goes to starch (Sharkey et al., 1985; Escobar-Gutiérrez and Gaudillère, 1997; Szecowka et al., 2013; Abadie et al., 2018), but the amount of carbon partitioned to starch can vary greatly among plant species (Huber, 1981). A small amount of phosphate is added to starch in photosynthesizing leaves by glucan-water dikinase and phosphoglucan-water dikinase, but the amount is very low, 0.1–0.9% of glucose moieties (McPherson and Jane, 1999; Ritte et al., 2002; Kötting et al., 2005), and so is not relevant for understanding gas exchange properties of photosynthesis.

There are a number of other routes by which carbon is exported from the Calvin–Benson cycle (Fig. 1). Any carbon metabolism pathway that begins with a phosphorylated Calvin–Benson cycle intermediate and ends with a non-phosphorylated molecule will contribute to TPU. The shikimate pathway to aromatic amino acid synthesis begins with the export of GAP from the chloroplast to make phosphoenolpyruvate (PEP). PEP is reimported into the chloroplast through the phosphoenolpyruvate/phosphate translocator (PPT) and combines with erythrose 4-phosphate (E4P) and ends with chorismate, accounting for 1–2% of fixed carbon (Escobar-Gutiérrez and Gaudillère, 1997; Abadie et al., 2018). Fatty acids and branched chain amino acids are synthesized from acetyl-CoA from pyruvate and account for 1–3% of fixed carbon (Bao et al., 2000). It has been shown that oil biosynthesis can be increased as a carbon sink, and this would contribute to a higher capacity for TPU (Sanjaya et al., 2011). The methyl-erythrtitol 4-phosphate (MEP) pathway begins with GAP and pyruvate to produce isoprenoids consuming up to 3% of fixed carbon (Rasulov et al., 2014). Pyruvate is made from TP exported from the chloroplast and dephosphorylated by pyruvate kinase, freeing phosphate in the cytosol, or by beta elimination of phosphate during the Rubisco reaction (Andrews and Kane, 1991) freeing phosphate in the stroma.

Fig. 1.

A depiction of the major phosphate and carbon exits from the Calvin–Benson cycle. Rates: sucrose, 25–50%; starch, 30–60%; photorespiratory amino acids, 7–15%; shikimate pathway, 1–2%; lipids 1–3%; methylerythritol pathway, 1–3%; PEP carboxylation, 0.5–4%; CO₂ release from photorespiration*, 7–12.5% of fixed carbon lost and does not contribute to TPU capacity. Abbreviations: E4P, erythrose 4-phosphate; F6P, fructose 6-phosphate; GAP, glyceraldehyde 3-phosphate; PEP, phosphoenolpyruvate; PGA, 3-phosphoglyceric acid; SBP, sedoheptulose bisphosphate; TP, triose phosphates; Xu5P, xylulose 5-phosphate.

Amino acid intermediates in the photorespiratory pathway can be exported from the leaf or used in the cytosol as carbon skeletons, for transamination, or for protein construction. It is estimated that an average of 30% to a high of 70% of photorespiratory glycolate carbon is exported from the Calvin–Benson cycle as modeled from gas exchange measurements (Busch et al., 2018). If the ratio of oxygenation to carboxylation (ϕ) is assumed to be 0.25, this represents carbon export from the Calvin–Benson cycle equivalent to 7–15% of fixed carbon. In addition, CO₂ lost from conversion of glycine to serine will allow for increased rates of carboxylation, though it does not increase the maximum assimilation rate. If we assume ϕ is 0.25 and no glycine export, this would represent 12.5% of fixed carbon lost, but under TPU-limited conditions excess carboxylation capacity allows fixation of the same amount of CO₂. This is part of the reason photosynthesis becomes insensitive to CO₂ even though the rate of photorespiration varies with CO₂.

Plants are capable of carboxylating PEP and releasing the phosphate on PEP. The resulting oxaloacetate can be transaminated to aspartate or reduced to malate for use in anapleurotic reactions or storage in the vacuole (sometimes as fumarate). PEP carboxylation contributes to TPU as PEP may come from TPs exported from the chloroplast, and the carboxylation consumes atmospheric carbon which would be measured in gas exchange. Gauthier et al. (2010) found that amino acids made from α-ketoglutarate are quickly labeled by [¹⁵N]ammonium nitrate but not ¹³CO₂ fed to photosynthesizing leaves, indicating that the carbon for these amino acids comes from pre-existing pools and so does not contribute to TPU. Szecowka et al. (2013) showed that no more than 2.6% of label goes through PEP to organic acids or amino acids, including non-carboxylation reactions. Ma et al. (2014), using extensive in silico modeling combined with MS measurements, found that PEP carboxylation represented 0.5–4% of fixed carbon, depending on how much PEP carbon is assumed to be directly from the Calvin–Benson cycle and the overall rate of photosynthesis. Another study found that the rate of PEP carboxylation varied with the rate of photosynthesis, increasing significantly in its proportion at low assimilation from 2% to 25% of fixed carbon (Abadie and Tcherkez, 2019). In Arabidopsis, a significant amount of carbon is stored in the vacuole as fumarate; it is not known how much of this carbon is recent (and therefore contributes to TPU) and how much is pre-existing carbon (Chia et al., 2000; Pracharoenwattana et al., 2010; Zell et al., 2010; Ma et al., 2014). This is also true of sunflower (Abadie et al., 2018).

In summary, TPU is primarily starch and sucrose synthesis (~80%). The next most important ‘use’ of TPs may be in removal of glycine or serine from the photorespiratory cycle, potentially reaching 15% but probably usually well below 10%. Many other metabolic pathways account for the remainder, but none of these is likely to exceed 5% of the rate of carbon fixation and so they usually do not have a significant impact on TPU limitation behavior.

TPU and gas exchange

TPU is typically assessed from gas exchange data obtained using infrared gas analyzers to measure rates of CO₂ uptake. Because of the usefulness of fluorescence parameters in analyzing gas exchange data, gas exchange measurements are frequently combined with chlorophyll fluorescence analysis. Measuring the stomatal conductance to gas exchange by transpiration allows the calculation of the partial pressure of CO₂ inside the leaf (C_i) (Sharkey et al., 1982). Diffusion resistance within the mesophyll will further reduce the effective partial pressure of CO₂, resulting in the partial pressure of CO₂ at the site of carboxylation (C_c). TPU-limited photosynthesis is mostly insensitive to CO₂, so resistance to diffusion of CO₂ has little or no effect on TPU-limited photosynthesis.

Plots of carbon assimilation (A) as a function of C_i (or better C_c when mesophyll conductance can be estimated) can be interpreted using Rubisco kinetics to predict what biochemical process is limiting assimilation. At low C_c, assimilation is typically limited by binding affinity of Rubisco for CO₂ (and the inhibition by oxygen), known as the Rubisco limitation (often abbreviated as C limitation). At intermediate C_c or when given insufficient light, assimilation is typically limited by the rate of regeneration of ribulose 1,5-bisphosphate (RuBP), frequently referred to as J limitation. TPU limitation, sometimes called P limitation, only happens when the plant has a greater capacity to fix carbon than it has to remove carbon from the Calvin–Benson cycle in end product synthesis. In many plants, this can be seen at high C_c and saturating light. The requirement for a high photosynthetic rate may be why TPU limitation is so hard to detect in plants with low inherent photosynthetic rates such as Arabidopsis (Yang et al., 2016).

Lack of, or reverse, sensitivity of A to oxygen partial pressure changes and CO₂ partial pressure increases is the primary gas exchange behavior of TPU limitation (Sharkey, 1985a). Insensitivity had been reported for many years (Ludwig and Canvin, 1971; Jolliffe and Tregunna, 1973; von Caemmerer and Farquhar, 1981). Critically, Harris et al. (1983) found insensitivity following feeding with mannose, which sequesters phosphate. Later it was shown that oxygen insensitivity was correlated with CO₂ insensitivity (Sharkey, 1985a). Leegood and Furbank (1986) found that oxygen-insensitive photosynthesis in leaf discs was induced by a combination of low temperature and high CO₂ partial pressure. Feeding of phosphate restored normal oxygen sensitivity and also increased CO₂ assimilation rate, showing that phosphate metabolism was involved in both oxygen sensitivity and the limitation of assimilation. From this and other considerations, Sharkey (1985a) concluded “(a)s the rate of CO₂ assimilation increases, starch and sucrose synthesis must increase as well. If not, triose-P and PGA will build up and phosphate will decline. These changes in pool size will stimulate starch and sucrose synthesis. However, there is a limit to how far the phosphate pool can fall before it begins to limit photophosphorylation. Once this limit is reached, CO₂ will be assimilated at the rate at which starch and sucrose synthesis can metabolize triose-P, regardless of whether oxygenation occurs or not”.

When photosynthesis is limited either by Rubisco or by RuBP regeneration, increasing CO₂ or decreasing O₂ should increase A. When A is Rubisco limited, A will increase because of (i) the affinity of Rubisco for CO₂ and the effects of O₂ on CO₂ affinity and (ii) the reduced CO₂ release in photorespiration. When A is limited by RuBP regeneration, A will increase because of (i) the reduced CO₂ release in photorespiration (as above) and (ii) the diversion of RuBP from oxygenation to carboxylation when photorespiration is suppressed. TPU-limited photosynthesis does not exhibit this increase or exhibits a reduced stimulation when photorespiration is suppressed (Badger et al., 1984; Sharkey, 1985a). The insensitivity of A while TPU is limited happens because the controlling factor is the ability of the leaf to make end products and this is not affected by CO₂, O₂, or the rate of photorespiration. Increasing photorespiration by increasing O₂ or decreasing CO₂ partial pressures will be compensated by increased RuBP regeneration and carboxylation but, because these capacities are in excess in a TPU-limited state, this will not affect A. Use of oxygen or CO₂ insensitivity to determine photosynthetic limitations in A/C_i curves is discussed in greater detail in Busch and Sage (2017).

It is not possible to determine whether C₄ plants suffer TPU limitation. The carbon pump of C₄ metabolism makes it difficult to see the gas exchange behaviors that characterize TPU limitation. C₄ plants at high photosynthetic rates are interpreted to be limited by CO₂-saturated Rubisco activity, and at lower rates by PEP carboxylase activity (Collatz et al., 1992). Even if Rubisco is not saturated with CO₂, oxygen-dependent changes in the rate of photorespiratory CO₂ release change the CO₂ concentration in the bundle sheaths, making the C₄ photosynthesis rate independent of the photorespiration rate (von Caemmerer, 2000). Thus, the CO₂ and O₂ dependence that results from the variation in the ratio of carboxylation to oxygenation is not observed in C₄ photosynthesis and, because this is the gas exchange characteristic that is used to diagnose TPU limitation, it is not possible to tell if C₄ plants have a TPU-limited state.

Reverse sensitivity to CO₂ and O₂ partial pressures

While the TPU limitation offered understanding of insensitivity to increasing O₂ and CO₂ partial pressures, it did not immediately explain reverse sensitivity. It has long been known that oxygen inhibits photorespiration due to competitive binding to Rubisco and photorespiratory CO₂ release (Warburg, 1919; Ludwig and Canvin, 1971; McVetty and Canvin, 1981). It was therefore unexpected to find that reducing oxygen or increasing CO₂ partial pressures could sometimes reduce the rate of CO₂ assimilation. As photorespiration releases CO₂, it is counterintuitive that altering the gas composition to favor carboxylation would result in decreased carbon assimilation. Yet data dating back decades show that once at high CO₂, increasing CO₂ can cause a decrease in net assimilation (Jolliffe and Tregunna, 1973; Canvin, 1978; von Caemmerer and Farquhar, 1981), and increasing O₂ can cause an increase in net assimilation (Viil et al., 1977).

Photorespiration was one key to understanding the reverse oxygen sensitivity under TPU-limiting conditions. Phosphoglycolate is dephosphorylated by phosphoglycolate phosphatase before export through PLGG1 or BASS6 (South et al., 2017). Photorespiratory metabolism of two glycolate molecules leads to re-import of carbon as glycerate, which is phosphorylated to phosphoglyceric acid. The extra phosphate released can be used to make ATP that phosphorylates ribulose 5-phosphate to produce RuBP that will be used to accept a CO₂, balancing the photorespiratory loss of one carbon. However, the two amino acid intermediates in the photorespiratory pathway can be used in the cytosol, resulting in net carbon export from the Calvin–Benson cycle. This carbon is effectively lost from RuBP and not directly from CO₂ fixed from the atmosphere. Photorespiratory carbon that never returns to the chloroplast was parameterized as α, the fraction of glycolate carbon that leaves the photorespiratory cycle as amino acids (Harley and Sharkey, 1991). The α parameter was later refined to α_G and α_S, the fraction of glycolate carbon that leaves as glycine and serine, respectively (Busch et al., 2018). When glycine is exported instead of serine, no CO₂ is released. As these amino acids come from phosphorylated plastidic metabolites, and permanently leave the Calvin–Benson cycle, they contribute to TPU capacity. Adjusting the gas composition to decrease ϕ reduces the export of glycine and serine, and therefore reduces TPU capacity, reducing the maximum photosynthetic rate. This can explain the reverse sensitivity of A to CO₂ and O₂.

Starch synthesis is also affected by oxygen partial pressure and can contribute to severe reverse sensitivity. Beans photosynthesizing quickly then transferred to low oxygen were found to have reduced rates of starch synthesis but a minimal change in the rate of sucrose synthesis. A concurrent reduction in the ratio of glucose-6-phosphate to fructose-6-phosphate indicates inhibition of phosphoglucose isomerase (Dietz, 1985; Vassey and Sharkey, 1989). The precise mechanism of this inhibition is unclear.

Modeling

TPU models have seen some recent changes to account for our enhanced understanding of the possible role of photorespiration in nitrogen metabolism. Original models that account for TPU relied on simple stoichiometry (Sharkey, 1985b):

$${\displaystyle \begin{array}{l}{\displaystyle W_{\mathrm{p}}=\frac{3\times \mathrm{T}\mathrm{P}\mathrm{U}}{1-0.5\varphi }}\end{array}}$$ (1)

where W_p is the rate of carboxylation when limited by phosphate metabolism. Under this model, photosynthetic carboxylation would equal the rate of carbon export from the Calvin–Benson cycle for starch and sucrose synthesis (numerator) adjusted by the amount of carbon released during photorespiration (denominator). Under TPU limitation, A is given by

$${\displaystyle \begin{array}{l}{\displaystyle A=W_{\mathrm{p}}\times (1-0.5\varphi )R_{\mathrm{L}}.}\end{array}}$$ (2)

When Equation 1 is plugged into Equation 2, the (1–0.5ϕ) term cancels out and so A is independent of the rate of photorespiration. This is because Rubisco is not limiting so the amount of CO₂ released during photorespiration can be compensated by increased Rubisco activity.

However, this model did not account for reverse sensitivity of assimilation to oxygen or CO₂ frequently observed. The model also describes all carbon export as TPU, which is not directly true. Any carbon that leaves the Calvin–Benson cycle and is dephosphorylated will contribute to the maximum TPU capacity. While all carbon in the Calvin–Benson cycle derives from TP, some of the end products are made from Calvin–Benson cycle intermediates other than TPs. Despite this, the simple model has some advantages. It requires no estimation of R_L, mesophyll conductance (g_m), or Γ*. These three parameters are currently impossible to measure directly, and there is some debate about our ability to fit them accurately and the constancy of these parameters.

A recent model for TPU incorporates parameters for glycine or serine exit from the photorespiratory cycle. The glycine and serine need not accumulate and could have a range of metabolic fates, as long as the carbon does not re-enter the Calvin–Benson cycle. From Busch et al. (2018):

$${\displaystyle \begin{array}{l}{\displaystyle W_{\mathrm{p}}=\frac{3\times \mathrm{T}\mathrm{P}\mathrm{U}}{1-0.5(1+3{\mathrm{\alpha }}_{\mathrm{G}}+4{\mathrm{\alpha }}_{\mathrm{S}})\varphi }}\end{array}}$$ (3)

The denominator in the equation has three terms to account for carbon that exits photorespiration as glycine (α_G) or serine (α_S). As one carbon out of four is lost as CO₂ in the formation of serine, α_S cannot be greater than 0.75. If α_G and α_S are zero, Equations 1 and 3 are identical. Unlike the simple model of Equation 1, Equation 3 requires knowledge of the relative rate of photorespiration, and therefore relies on fitting for Γ*. There is little signal to differentiate α_S and α_G by gas exchange, which can make fitting these two parameters challenging. For conversion of Equation 3 to assimilation as would be measured by gas exchange, W_p must be adjusted for respiratory carbon loss:

$${\displaystyle \begin{array}{l}{\displaystyle A=W_{\mathrm{p}}\times (1-\frac{\mathrm{\Gamma }{\ast }_{{\mathrm{\alpha }}_{\mathrm{G}}}}{C_{\mathrm{c}}})-R_{\mathrm{L}}}\end{array}}$$ (4)

where Γ*_αG is the Rubisco–C_c compensation point given the reduced rate of photorespiratory CO₂ release due to export of glycine. Γ*_αG/C_c is equivalent to 0.5ϕ if α_G=0.

Current modeling software is available with varying numbers of parameters to fit. Sharkey (2016) presented an Excel tool which allows picking of points from A/C_i curves, with options to fit R_L, g_m, α_G, and α_S. Bellasio et al. (2016) provide a highly detailed Excel tool that uses combined gas exchange and fluorescence to fit R_L, g_m, J_max, V_cmax, Γ*, and Rubisco specificity for CO₂ versus oxygen (S_c/o), but not α; much of the basis of this fitting is also discussed by Yin et al. (2009). Dubois et al. (2007) provide an SAS program which allows fitting of R_L, g_m, J_max, V_cmax, Γ*, S_c/o, and α. Moualeu-Ngangue et al. (2017) propose to improve the Dubois fitting by reducing the number of assumptions made, though they do not fit α. Gu et al. (2010) provide a website for fully automated leaf data analysis called LeafWeb which does not require selecting limitations point-wise or specific software. It should be noted that no current model attempts to incorporate other carbon sinks, and TPU is treated as a single variable.

Temperature sensitivity

Photosynthesis under TPU limitation is highly temperature sensitive. Though the other photosynthetic limitations demonstrate temperature sensitivity (Cen and Sage, 2005; Sage and Kubien, 2007; Sharkey and Bernacchi, 2012; Busch and Sage, 2017), TPU-limiting conditions are the most temperature sensitive (Sharkey and Bernacchi, 2012; Yang et al., 2016) perhaps because of the strong temperature sensitivity of sucrose-phosphate synthase (Stitt and Grosse, 1988; Leegood and Edwards, 1996) or altered sensitivity of cytosolic FBPase to 2,6-fructose bisphosphate (Stitt and Grosse, 1988). Other enzymes implicated in TPU limitation are also temperature sensitive, such as nitrate reductase (Leegood and Edwards, 1996; Busch et al., 2018). Because of the different ways by which temperature affects the three limitations, the conditions in which they appear change with temperature. At temperatures lower than growth conditions, the plant is significantly more likely to become TPU limited (Stitt, 1986; Sage and Sharkey, 1987; Labate and Leegood, 1988). Labate and Leegood (1988) demonstrated a temperature-sensitive increase in photosynthesis from phosphate feeding. Leaf discs floated on a solution containing phosphate at 25 °C saw a marginal reduction in assimilation. However, discs fed phosphate at 10 °C experienced significant photosynthetic gains, indicating that reduced temperatures result in greater limitation of photosynthesis by TPU (Fig. 2).

Fig. 2.

Rate of CO₂ assimilation of barley versus C_i at 10 °C (top) and 25 °C (bottom) with and without the addition of phosphate. A temperature-dependent increase in photosynthetic assimilation is observed upon addition of phosphate. Adapted by permission from Springer: Springer Planta. Limitation of photosynthesis by changes in temperature, Labate, CA, Leegood, RC, Copyright 1988.

Acclimation of TPU

The capacity for TPU is not immutable. Plants grown under poor conditions are highly adaptive, and those grown under low temperature tend to have greatly elevated TPU capacity (Guy et al., 1992; Holaday et al., 1992; Sage and Kubien, 2007). This acclimation largely comes from increased expression of sucrose biosynthesis enzymes (Guy et al., 1992; Holaday et al., 1992; Strand et al., 1999; Hurry et al., 2000), and it has been proposed that this acclimation is signaled by low phosphate levels (Hurry et al., 2000). This increased capacity offsets the decreased activity of starch synthase and sucrose-phosphate synthase at low temperature and makes it less likely that the plant will be TPU limited (Cornic and Louason, 1980; Sage and Sharkey, 1987). Plants transferred to an elevated CO₂ environment developed increased phosphate regeneration capacity, demonstrating acclimation (Sharkey et al., 1988; Sage et al., 1989).

Plants experiencing water stress reduce their TPU capacity, possibly reflecting the reduced internal CO₂ partial pressure that results from stomatal closure (von Caemmerer and Farquhar, 1984; Vassey and Sharkey, 1989; Cornic et al., 1992). Transgenic plants overexpressing alternative oxidase cope better with water stress (Dahal et al., 2014, 2015) and experience reduced negative effects on assimilation from TPU capacity. The reduced occurrence of TPU limitation in plants overexpressing the alternative oxidase was correlated with higher amounts of chloroplast ATP synthase, which might allow ATP synthesis at lower phosphate concentration. This adaptability shows that TPU will influence the metabolic investments of the plant; it will enhance the ability to handle high TP production, but only when it is required for the current output of photosynthesis.

The adaptability of TPU is important for fulfilling the role of stromal phosphate in balancing starch synthesis and ATP synthesis (Fig. 3). Starch synthesis is highly sensitive to phosphate due to inhibition of ADP-glucose pyrophosphorylase (Preiss, 1982), and ATP synthase is kinetically (Takizawa et al., 2007) and thermodynamically sensitive to phosphate. This relationship can help explain the very low partitioning of carbon into starch at a low photosynthetic rate (Escobar-Gutiérrez and Gaudillère, 1997), which is exacerbated by reduced levels of PGA which would otherwise stimulate starch production (Heldt et al., 1977). If sucrose synthesis is in excess, the balance of starch versus sucrose synthesis during the day could become unfavorable for growth, and the extra phosphate could even collapse the Calvin–Benson cycle by driving export of too much TP out of the chloroplast. This has been reported in isolated chloroplasts (Leegood and Walker, 1983) but not in intact leaves. High phosphate outside of chloroplasts has also been shown to result in starch breakdown in the light (Stitt and Heldt, 1981). The highest rate of photosynthesis will be achieved with a fine balance of phosphate usage and phosphate release. In an environment where expected photosynthetic rates are lower, the plant will benefit from reduced TPU capacity. This allows phosphate to fall, correcting several issues with starch and sucrose metabolism and reducing the risk of overconsumption of TPs. When expected photosynthetic rates are higher, the plant will benefit from increased TPU capacity allowing better recycling of phosphate and improved ATP synthase throughput, and alleviating the potential for photodamage due to feedback conditions.

Fig. 3.

As the photosynthetic rate increases, the gap between the phosphate concentration required by the ATP synthase and the phosphate concentration to inhibit starch synthesis narrows. The shapes of the responses are represented by straight lines only for simplicity. When TPU limits the photosynthetic rate, any increase in phosphate required for higher ATP synthase activity would inhibit starch synthesis, restricting phosphate release.

Effects on the light reactions

Elevating CO₂ partial pressure when photosynthesis is limited by TPU will cause a decrease in Φ_PSII. Rubisco binds CO₂ and O₂ competitively, meaning that an increase in CO₂ partial pressure reduces the rate of use of RuBP for oxygenation. This does not lead to an increase in assimilation when TPU is controlling. Rather, it reduces the rate of carboxylation because less carbon is lost through photorespiration, resulting in reduced total Rubisco activity. Both carboxylation and oxygenation require ATP and NADPH, which come from electron transport. Therefore, increasing CO₂ partial pressures over TPU-limited leaves results in an overall reduction in electron transport requirements (Stitt, 1986; Sharkey et al., 1988; Stitt and Grosse, 1988). Regulatory processes lead to reduced Φ_PSII, a phenomenon which can be useful in discriminating TPU limitation using combined gas exchange and fluorescence data (Fig. 4).

Fig. 4.

The decline in electron transport rate is diagnostic of TPU limitation. Combined gas exchange and fluorescence data in A/C_i curves of Nicotiana benthamiana at varying light intensity and 35 °C. At low CO₂, plants are limited by Rubisco activity (C limitation, red), characterized by a sharp upwards slope of both A and ETR with increasing CO₂. When light is insufficient, plants will be limited by the rate of RuBP regeneration (J limitation, green), characterized by a flat slope of the ETR with increasing CO₂. Only when the plant has ample CO₂ and electron transport will TPU limitation (P, yellow) be seen, characterized by a decline in ETR with increasing CO₂. ETR is calculated from fluorescence-derived Φ_PSII. Light intensity (µmol m⁻² s⁻¹): , 250; , 400; , 550; , 750; , 1000; 1500.

There are effects on the kinetics of the light reactions that happen concurrently with reduction of electron transport rate. Proton conductivity across the thylakoid membrane goes down under TPU limitation (Takizawa et al., 2008; Kiirats et al., 2009; Yang et al., 2016). It is proposed that this kinetic change occurs because of a reduced pool of available phosphate in the stroma, which reduces the rate of ATP synthase. The K_m of chloroplast ATP synthase for phosphate has been measured at 0.2–1 mM (Selman-Reimer et al., 1981; Grotjohann and Gräber, 2002). Stromal phosphate concentration during feedback conditions is estimated to be between 0 mM and 1.7 mM depending on how much phosphate is assumed to be free (Sharkey and Vanderveer, 1989), so it is reasonable to suggest that the phosphate concentration may drop below the K_m of ATP synthase. Co-occurring with a decrease in ATP synthase conductivity is an increase in proton-motive force (PMF). The energy needed to make ATP will depend on the concentration of phosphate.

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{\Delta }{G}_{\mathrm{A}\mathrm{T}\mathrm{P}}=\mathrm{\Delta }{G}_{\mathrm{A}\mathrm{T}\mathrm{P}}^{\prime 0}+R\times T\times \mathrm{l}\mathrm{n}\frac{\left[\mathrm{A}\mathrm{T}\mathrm{P}\right]}{\left[\mathrm{A}\mathrm{D}\mathrm{P}\right]\times \left[{\mathrm{P}}_{\mathrm{i}}\right]}}\end{array}}$$ (5)

As the effective [P_i] declines, ∆G_ATP will increase, requiring a greater PMF for ATP synthesis. Increased PMF leads to controls on electron transport through q_E, reducing energy arrival at P680 or reduction in the rate of electron flow at the cytochrome b₆f complex, leading to reduced electron transport rates (ETRs) (Kramer and Crofts, 1996; Owens, 1996). While phosphate seems to play a role in linking the light reactions and the Calvin–Benson cycle, it is less clear what other molecular mechanisms may be important. It is likely that we do not yet know some important regulatory components that control ETR when TPU limits the rate of photosynthesis.

TPU and sink strength

TPU limitation is a form of very short-term sink/source disequilibrium, separate from long-term sinks such as fruit or root growth, though the two could be related. TPU is concerned with the ability to dephosphorylate and remove carbon quickly from the Calvin–Benson cycle. The half-life of Calvin–Benson cycle intermediates tends to be very short, with many <1 s; some larger pools such as glucose 6-phosphate and UDP-glucose have a half-life of <1 min (Stitt et al., 1980; Arrivault et al., 2009). Pool lifetimes this short mean that TPU limitation can build up and diminish very rapidly. Over a longer time frame, a greater sink can be important in freeing up short-term sinks. It has been reported that defruited wheat experiences significant down-regulation of photosynthesis (King et al., 1967), though not all plants experience this effect (Farquhar and von Caemmerer, 1982). Build up of sucrose in source leaves could result in reduced TPU capacity due to reduced sucrose-phosphate synthase activity, as shown in some experiments (Huber, 1981; Paul and Foyer, 2001), or increased invertase activity (Mengin et al., 2017). In some experiments using conditions consistent with TPU limitation, starch builds up and causes a decline in photosynthetic rate (Sasek et al., 1985; Peet et al., 1986; Ramonell et al., 2001). The source of this decline is still to be conclusively determined. A long-term sink which can absorb carbon will allow the plant to recover (Sasek et al., 1985; Arp, 1991).

TPU and plant nutrition

TPU limitation is often incorrectly interpreted as a nutritional deficiency. It is true that plants transferred to media without any phosphate experience significant reduction in photosynthetic capacity (Brooks, 1986; Foyer and Spencer, 1986). However, less dramatic differences in phosphate nutrition result in relatively small changes in photosynthetic rate. This is due to the vacuole buffering the phosphate concentration in the rest of the cell on a time scale of hours (Rebeille et al., 1983; Woodrow et al., 1984). Under increased or decreased phosphate nutrition, large changes in vacuolar phosphate concentration are seen, but only relatively small changes are seen in plastidic phosphate concentration (Rebeille et al., 1983; Foyer and Spencer, 1986). Plants grown with different phosphate nutrition are therefore not significantly more or less likely to experience TPU limitation. Most phosphate in photosynthesizing cells will be used in nucleic acids and phospholipids (Dissanayaka et al., 2018), and growth is more sensitive to phosphate nutrition than is photosynthetic rate (Mo et al., 2018). Ellsworth et al. (2015) showed that Australian plants growing in the wild with varying phosphate availability were adapted to their environment, and TPU limitation was more likely at high phosphate nutrition. Furthermore, TPU limitation can only be seen when the plant is photosynthesizing very quickly, which usually cannot occur if the plant is nutritionally deprived. Plants with reduced nitrogen were not capable of photosynthesizing quickly enough to reach TPU limitation (Sage et al., 1990).

Oscillations

Oscillations in carbon assimilation rate are a common side effect of TPU limitation (Ogawa, 1982; Sivak and Walker, 1986, 1987). They are typically seen after a perturbation in the environment of a plant that results in high photosynthetic rates, such as sharp increases in illumination or CO₂. Oscillations then continue without further input for a variable amount of time. Oscillations include tandem changes in carbon assimilation and fluorescence parameters, indicating simultaneous changes in both the light reactions and the Calvin–Benson cycle (Ogawa, 1982; Walker et al., 1983; Peterson et al., 1988; Stitt and Grosse, 1988). The amplitude of oscillations can increase with conditions that further exacerbate TPU limitation, such as low temperature or low O₂ (Peterson et al., 1988; Stitt and Grosse, 1988). Oscillations showed a significant impact on organic phosphates and their relevant ratios, notably large initial spikes in PGA, and reduction in RuBP and ATP pools (Sharkey et al., 1986b; Sage et al., 1988; Stitt and Grosse, 1988; Laisk et al., 1991).

A few models have been produced to explain oscillations. The most significant theory is that there is a delay in activation of sucrose synthesis after a photosynthetic increase that causes oscillations (Laisk and Walker, 1986). The delay may also originate from cytosolic FBPase inhibition by fructose-2,6-bisphosphate (Stitt et al., 1984; Laisk and Eichelmann, 1989; Laisk et al., 1989) or post-translational regulation (Huber and Huber, 1996). An additional interpretation of these oscillations has been proposed originating from the light reactions, with damping caused by a slow leak of protons across the thylakoid membrane (Kocks and Ross, 1995).

Environmental impact

The changing climate, resulting in large measure from increasing CO₂, has the potential to affect the frequency and severity of TPU limitations to photosynthesis. Since this syndrome occurs when carbon fixation and light capture have a greater capacity than end product synthesis, increasing CO₂ should increase the occurrence of TPU limitation. However, because TPU is stimulated by increasing temperature, there could be a reduction in the occurrence of TPU limitation in the future. It is hard to predict which effect will dominate, and whether TPU limitation will be observed more or less frequently based on climate change predictions. However, beyond the short-term effects of temperature and CO₂, it is important to consider how the plant responds when it is TPU limited. Generally, plants growing in elevated CO₂ show less propensity for TPU limitation because they have reduced capacity for other processes in photosynthesis (Sage et al., 1989). This suggests that plants cannot or do not make full use of the greater potential for photosynthesis. We hypothesize that understanding TPU will help in predicting acclimation responses of plants to increasing atmospheric CO₂. How plants might acclimate could depend on such things as stochasticity of their environment and the typical day/night change in temperature. If night (and dawn) temperature rises more than day temperature, this could affect optimal TPU capacity.

It is often found that TPU limitation occurs whenever photosynthesis is stimulated to be ~20% higher than was occurring in the plant under natural conditions (Yang et al., 2016). Increasing CO₂, decreasing oxygen, or lowering the temperature usually allows TPU limitation to be observed. In a large study of published A/C_i curves, Wullschleger (1993) found 23 cases (out of 109) where investigators reported TPU limitations. It is likely that the phenomenon is observed but not recognized much more often. For example, a curve presented in Wullschleger et al. (Fig. 1B, taken from Ireland et al., 1988), shows evidence of TPU limitation but this was not one of the 26 instances of TPU limitation cited. It is common for the TPU limitation to be ignored even when it is evident in data.

Since the components of photosynthesis must all work in concert and in strict stoichiometry, it is not surprising that there might be a relationship between V_cmax and TPU capacity. This has been invoked in global models of photosynthesis, although many models do not include TPU. Lombardozzi et al. (2018) used several estimates of the ratio of V_cmax and TPU capacity, and concluded that current global models may overestimate how much CO₂ will be fixed by plants in the future because TPU limitations, or adjustments to avoid TPU limitation, will reduce photosynthetic capacity. It is important to realize that even though plants growing in elevated CO₂ do not show TPU limitation, TPU may still be setting an upper bound and that plants adjust other capacities to keep below the upper bound of TPU because TPU can cause damage.

Conclusions

TPU is a metabolic condition that incorporates numerous signals to reflect the state of photosynthesis across the whole cell. Most metabolites in the chloroplast are phosphorylated, and so phosphate can reflect the metabolic state of the chloroplast. Phosphate is linked throughout the cytosol, where sucrose synthesis takes place, and thus phosphate represents the photosynthetic state across all chloroplasts. Phosphate concentrations are carefully regulated, and TPU limitation is very unlikely to be found at ambient conditions. A low phosphate level naturally signals to the other processes that photosynthesis is very fast, kinetically controls the ATP synthase, and leads to downstream effects on photosynthesis by accumulation of PMF and engaging q_E. The reduction in phosphate signals the plant to build up starch by relieving phosphate inhibition of ADP-glucose pyrophosphorylase (Preiss, 1982). Plants which are photosynthesizing slowly can reduce their TPU capacity, which will lower their phosphate regeneration, helping to produce starch and prevent cycle collapse from overexport of TPs; conversely, increasing TPU capacity in plants which are photosynthesizing quickly will raise their phosphate regeneration and help produce ATP. In this way, TPU sets the span on expected photosynthesis. We believe that the gas exchange behavior in TPU conditions reflects several important regulatory features. Yet, the role of TPU as regulation is relatively unexplored. Experimental determination of the molecular mechanisms that underpin this system, and ecological studies to examine the broader effects of TPU are exciting future directions in this field.

A number of misconceptions cloud the field in regards to TPU. Even the term ‘TPU’ can now be seen not to be wholly accurate. It largely describes phosphate metabolism, but not all effects on carbon metabolism related to phosphate can be accurately described as TPU. At steady state, there are other sources of phosphate release that contribute to the assimilation cap. Amino acid release from photorespiration, methylerythritol 4-phosphate and shikimate pathways, and other carbon sinks for Calvin–Benson cycle intermediates will all contribute to the maximal assimilation rate when photosynthesis is TPU limited. An alternative view is that all Calvin–Benson cycle exports are downstream of TP, and thus constitute a form of TPU. The specific terminology and nuance are less important than the total understanding, which is that TPU limitation is the result of insufficient capacity for carbon export from the Calvin–Benson cycle. Other carbon metabolism pathways in the chloroplast that do not immediately originate in the Calvin–Benson cycle, while important for the overall physiology of the plant, will not be discernible in gas exchange measurements.

Maintaining TPU limitation is unhealthy for the plant due to risk of oxidative stress from photosystem oxidation (Pammenter et al., 1993). Electron transport regulation as assessed by chlorophyll fluorescence quenching analysis and deactivation of Rubisco lead to an overall slowing of photosynthesis lower than TPU, eventually reaching a steady state with the assimilation rate based on the rate of TPU. Excess assimilation when already low on phosphate would further deprive ATP synthase of phosphate it needs. Contrary to what one might expect given the term ‘TPU limitation’, TPs do not necessarily need to build up, though phosphate levels should be low (Sharkey and Vanderveer, 1989). This is why plants can be drained of phosphate via mannose or deoxyglucose feeding and be TPU limited (Herold and Lewis, 1977; Herold, 1980; Sivak and Walker, 1986). It is the relationship between the need for phosphate for ATP synthase and the phosphate sensitivity of starch and sucrose synthesis that results in TPU (Herold, 1980).

Glossary

Abbreviations

A

net assimilation of carbon, or, when negative, net respiration of carbon

C_c

partial pressure of CO₂ at the site of carboxylation

C_i

partial pressure of CO₂ inside the leaf

DHAP

dihydroxyacetone phosphate

E4P

erythrose 4-phosphate

ETR

photosynthetic electron transport rate

FBPase

fructose-1,6-bisphosphatase

GAP

glyceraldehyde 3-phosphate

g_m

mesophyll conductance to CO₂

J_max

maximum rate of electron transport under infinite light

MEP

methylerythritol 4-phosphate

PEP

phosphoenolpyruvate

PGA

phosphoglyceric acid

PMF

proton-motive force

PPT

phosphoenolpyruvate/phosphate translocator

q_E

energy-dependent quenching

R

rate of RuBP consumption

R_L

rate of respiration in the light

RuBP

ribulose 1,5-bisphosphate

S_c/o

specificity of Rubisco for CO₂ versus oxygen

TP

triose phosphate

TPT

triose phosphate/phosphate translocator

TPU

triose phosphate utilization

V_cmax

maximum velocity of carboxylation

W

rate of carboxylation

α

the fraction of glycolate carbon which leaves the Calvin–Benson cycle as amino acids

α_G

the fraction of glycolate carbon which leaves the Calvin–Benson cycle as glycine

α_S

the fraction of glycolate carbon which leaves the Calvin–Benson cycle as serine

Γ*

The Rubisco CO₂ compensation point ignoring the effect of R_L

ϕ

the ratio of oxygenations to carboxylations

Φ_PSII

quantum yield of PSII