Sugar sensing in C4 source leaves: a gap that needs to be filled

Abstract Plant growth depends on sugar production and export by photosynthesizing source leaves and sugar allocation and import by sink tissues (grains, roots, stems, and young leaves). Photosynthesis and sink demand are tightly coordinated through metabolic (substrate, allosteric) feedback and signalling (sugar, hormones) mechanisms. Sugar signalling integrates sugar production with plant development and environmental cues. In C3 plants (e.g. wheat and rice), it is well documented that sugar accumulation in source leaves, due to source–sink imbalance, negatively feeds back on photosynthesis and plant productivity. However, we have a limited understanding about the molecular mechanisms underlying those feedback regulations, especially in C4 plants (e.g. maize, sorghum, and sugarcane). Recent work with the C4 model plant Setaria viridis suggested that C4 leaves have different sugar sensing thresholds and behaviours relative to C3 counterparts. Addressing this research priority is critical because improving crop yield requires a better understanding of how plants coordinate source activity with sink demand. Here we review the literature, present a model of action for sugar sensing in C4 source leaves, and suggest ways forward.


Introduction
Sugar sensing has emerged as a research topic of interest within the plant science community where mechanisms of action have predominantly been uncovered in the C 3 model dicot Arabidopsis thaliana.Sugar sensing in plants was first proposed more than three decades ago when examining the metabolic regulation of gene expression in mesophyll protoplasts (Sheen, 1990).Despite intensive research on Arabidopsis, there has not been a transition into studying these mechanisms within cereals/monocots.Even though cereals such as wheat (Triticum aestivum), barley (Hordeum vulgare), maize (Zea mays), rice (Oryza sativa), and millets (e.g.Setaria spp.) make up a large proportion of the world's food production, a huge knowledge gap exists on understanding how sugar sensing and signalling occur within these agronomically important crops (Godfray et al., 2010;Lal, 2016;Springmann et al., 2016;Simkin et al., 2019).Some of the most productive plants on the planet, including many grasses and important crop species, utilize the C 4 pathway of photosynthesis (Byrt et al., 2011).Unlike C 3 plants that fix CO 2 directly from the atmosphere in the mesophyll cells using Rubisco, C 4 plants have evolved a biochemical CO 2 -concentrating mechanism in which the enzyme phosphoenolpyruvate carboxylase (PEPC) fixes atmospheric CO 2 in the mesophyll cells to form a 4-carbon acid (C 4 dicarboxylic acid).This C 4 acid or its derivatives diffuses to the bundle sheath cells surrounding the vasculature and is decarboxylated to release CO 2 where Rubisco is localized (Box 1A) (Hatch and Slack, 1966;Furbank, 2016).This two-cell system concentrates CO 2 around Rubisco, reducing the wasteful process of photorespiration and ensuring Rubisco operates at its maximum catalytic capacity.
The cellular specialization necessary for C 4 photosynthesis also has implications for the regulation of carbohydrate synthesis and partitioning, and potentially for sugar signalling.In C 4 plants utilizing the NADP-malic enzyme (NADP-ME) pathway of C 4 photosynthesis (and possibly in all C 4 plants), reduction of 3-phosphoglycerate (3PGA), required to regenerate ribulose 1,5 bisphosphate (RuBP) and also to generate 6-carbon sugar phosphates for starch and sucrose synthesis (Fig. 1), is spatially separated between the mesophyll and bundle sheath chloroplasts (von Caemmerer and Furbank, 2016).Up to two-thirds of the 3PGA diffuses from the Calvin cycle in the bundle sheath to the mesophyll chloroplasts and returns as glyceraldehyde 3-phosphate [G3P; also known as triose phosphates (TP)] (von Caemmerer and Furbank 2016; Furbank and Kelly, 2021).This diffusion requires a large concentration gradient with high levels of 3PGA present in the bundle sheath compartment and high G3P levels present in the mesophyll (Stitt and Heldt, 1985;Arrivault et al., 2017).Since the transport of G3P out of the chloroplast and 3PGA and Pi levels inside the chloroplast are primary regulators of partitioning between sucrose and starch in C 3 plants (McClain and Sharkey, 2019), this specialization has considerable ramifications for regulation of carbon partitioning in C 4 leaves (Furbank and Kelly, 2021).In addition, in many C 4 grasses, sucrose and starch synthesis are spatially separated, whereby sucrose is preferentially synthesized in the mesophyll compartment and starch in the bundle sheath chloroplast (see Box 1) (Lunn andFurbank, 1997, 1999;Furbank and Kelly, 2021).In some cases, this spatial specialization is achieved by transcriptional regulation of expression of key genes in sucrose biosynthesis (Furbank and Kelly, 2021).Global climate warming has drawn attention to C 4 crops as they tend to be more climate resilient compared with their C 3 counterparts.Since little attention has been paid to carbon partitioning and sugar signalling in leaves of C 4 species, this review explores the impact of metabolic and cellular specialization in C 4 leaves on these processes in this important group of plants.

Sugar transport and sensing in C 4 photosynthesis
The implication of sucrose synthesis occurring preferentially in the mesophyll tissues of many of the important C 4 crops is that sugars may move to the phloem for export via various routes.Sucrose could move from the mesophyll cells to the apoplast tissues and be reimported to the phloem sieve element-companion cell (SE-CC) complex (as is common in C 3 plants) or sucrose could diffuse passively through the abundant plasmodesmatal connections between mesophyll and bundle sheath cells of the leaf (Danila et al., 2016(Danila et al., , 2018) ) and be exported to the phloem either symplastically or apoplastically in the bundle sheath cells.For the latter options, sucrose will need to move into the bundle sheath cells of C 4 leaves down a concentration gradient which must be maintained between the two photosynthetic cell types (Turgeon and Ayre, 2005).The consumption of soluble sugars to produce starch in the bundle sheath cell would probably aid in maintaining this sucrose gradient.However, maintenance of high sugar levels in the mesophyll cells could potentially have a major impact on sugar perception and subsequent signalling pathways.Sucrose export pathways in C 4 leaves have been more extensively elucidated from recent work on the passive membrane-embedded proteins SUGARS WILL EVENTUALLY BE EXPORTED TRANSPORTERS (SWEETs) which export photoassimilates out of the cell (Bezrutczyk et al., 2021;Chen et al., 2022;Hua et al., 2022) and sucrose transporters (SUTs) which actively import sucrose into the SE-CC complex against a gradient from the apoplast (Box 1) (Aoki et al., 2004;Carpaneto et al., 2005).Uniquely, C 4 plants must be able to coordinate carbon fixation within the highly metabolically active bundle sheath cell with temporarily housing hexoses for starch synthesis and providing a conduit for sucrose to be loaded into the phloem.
When sugars accumulate, such as under high light or high CO 2 in some C 3 species, synthesis can exceed the export capacity of these cells, especially when sink demand is low.If the accumulation of sugars is sensed in the photosynthetic cells, it triggers a signalling cascade inducing a down-regulation of photosynthesis by suppressing the expression of key genes encoding proteins in the photosynthetic pathway (for a detailed review of recent literature, see Fichtner et al., 2021).Although there are few data on sugar concentrations of individual photosynthetic cells in C 3 or C 4 species, it is possible that C 4 species accumulate more sugars within these cells due to their increased photosynthetic efficiency.A pertinent question then is whether the coordination of photosynthesis with sink demand in C 4 leaves occurs via mesophyll, bundle sheath, or both cells and how might the sugar sensors involved in this process have been adapted through evolution to accommodate the cellular specialization of C 4 photosynthesis.
C 4 photosynthesis has independently evolved at least 62 times over 60 million years (Sage et al., 2011).During C 4 evolution, pre-existing genes from C 3 species were co-opted shuttled from the bundle sheath to form PEP [6].Sucrose and starch synthesis pathways of C 4 grasses are unique compared with C 3 dicots.3PGA from the bundle sheath can be used in the mesophyll cytosol to form G3P from DHAP catalysed by TPI [7].These substrates form fructose-1,6-bisphosphate (FBP) via fructose-1,6-bisphosphate aldolase (FBA).Hexose-phosphates (Hex-P) are formed from FBP through a series of reactions [8] to eventually result in sucrosephosphate (Suc-P) via sucrose phosphate synthase (SPS) [9].Sucrose phosphate phosphatase (SPP) cleaves Suc-P to form sucrose that can diffuse via plasmodesmata into the bundle sheath [10].FBP in the bundle sheath chloroplast can be used to form Hex-P for starch synthesis via starch synthase (SS) [11] which is temporarily stored in the leaf and depleted through the night.When stored starch is needed, glucosidases break it down to glucose to be exported to the cytosol [12] and used for sucrose synthesis.Once sucrose reaches the bundle sheath it is exported by SUGARS WILL EVENTUALLY BE EXPORTED TRANSPORTERS (SWEET; pink circle), probably by SWEET13, into the apoplast for active uptake into the phloem tissues via sucrose transporter1 (SUT1; blue circle) [13].Not all reactions are represented in this schematic for simplicity.The schematic is adapted from Furbank and Kelly (2021), where full pathways can also be found.
Box 1. Continued for altered functions in C 4 species (Ludwig et al., 2021).For example, of the two paralogues of the phosphoenolpyruvate transporter (PPT), only PPT1 was co-opted to function in the C 4 pathway, but not PPT2 (Lyu et al., 2020).Gene duplication events have also been common in C 4 evolution, including a key sugar transporter SWEET13 (Emms et al., 2016).This study highlighted that the phloem loading strategies between C 3 and C 4 species might differ given the duplication of this transporter.Subsequently, many studies have since focused on this transporter and its role in phloem loading in C 4 grasses such as sugarcane, Setaria, and maize (Bezrutczyk et al., 2018(Bezrutczyk et al., , 2021;;Hu et al., 2018;Chen et al., 2022).A recent analysis of protein sequences of known sugar sensors from the genomes of major groups of C 4 grasses saw no positive selection of codons or duplication events during the evolution of C 4 photosynthesis (Benning et al., 2023).However, further studies on the evolution of sugar sensors and other proteins related to carbohydrate metabolism in C 4 plants need to be expanded on.

Sugar sensing and signalling in C 3 species
Sugar sensing and signalling are usually studied either by artificially increasing sugars within the source leaves of plants through elevated CO 2 or by exogenous sugar feeding.The relevance to climate change of crop responses to elevated CO 2 means that this process has been thoroughly documented in free-air CO 2 enrichment (FACE) experiments designed to better understand how ecological and plant systems respond to increased CO 2 in the atmosphere (Leakey et al., 2009).These FACE experiments usually used elevated CO 2 between 550 ppm and 600 ppm or +200 ppm above ambient levels.In these conditions, some plants increased non-structural carbohydrates such as sugars and starches by between 30% and 40%, resulting in increased biomass (Ainsworth and Long, 2005;de Graaff et al., 2006;Ainsworth, 2008).In the C 3 crops soybean, wheat, and rice, the harvestable yield increased by 12-14% when grown in elevated CO 2 (Long et al., 2006;Ainsworth, 2008).Notably legumes and root crops have shown a greater increase in yield when compared with cereals (Ainsworth and Long, 2021).A comprehensive review of FACE experiments by Leakey et al. (2009) also observed that increased CO 2 failed to meet predicted theoretical yield increases in many crop plants.This perhaps is due to not accounting for metabolic feedback of sugar accumulation within the leaves that causes suppression of photosynthetic activity through a range of transcriptional and post-translational processes (Sheen, 1990;Moore et al., 2003;Oszvald et al., 2018;Paul and Eastmond, 2020;Fichtner et al., 2021).This has been demonstrated in a range of species under elevated CO 2 , with varying magnitudes of effects.In wheat, an increase in sink strength could not overcome photosynthetic acclimation and down-regulation under high CO 2 when photoassimilates accumulated in the leaves (Aranjuelo et al., 2011).Elevated CO 2 over prolonged periods saw similar results in soybean and Arabidopsis, but the effect varied across ecotypes (J.-Y.Li et al., 2008;Zheng et al., 2019).
Many studies on high CO 2 have also highlighted the importance of the nexus between carbon and nitrogen availability.During photosynthetic acclimation, insufficient nitrogen is being acquired and assimilated at high CO 2 , which in turn causes nitrogen limitation and a lower CO 2 assimilation (P.Li et al., 2008;Taub and Wang, 2008;Zheng et al., 2019).The carbon to nitrogen sensing mechanism allows plants to efficiently modulate carbon and nitrogen transport and metabolism depending on the energy status of the plant.Therefore, when carbon is abundant and internal nitrogen is low, nitrogen assimilation genes can be activated, and conversely be halted when photoassimilate availability is low and internal nitrogen is high (Coruzzi and Zhou, 2001).For many C 3 plants, yield responses can only be maintained if nitrogen is abundantly available (Ainsworth and Long, 2021).
Apart from imposing high CO 2 conditions, sugar feeding on leaves excised from the main plant can also artificially increase sugar within source leaves.During glucose feeding of spinach (Spinacia oleracea) leaves, a marked decrease in key photosynthesis proteins such as Rubisco was observed, subsequently decreasing photosynthetic activity (Krapp et al., 1991).Interestingly, a comparative study on the C 3 dicot tobacco  et al., 2005;Schmid et al., 2005;Waese et al., 2017), Oryza sativa (Os) OsHXK5 (B) (Jain et al., 2007), and Zea mays (Zm) ZmHXK5 (C) (Hoopes et al., 2019) across source and sink tissues.Expression is depicted as the log 2 ratio where red, yellow, and blue represent high, moderate, and low expression, respectively.The gene name and ID are denoted in the top left corner.eFP figures were obtained from http://bar.utoronto.ca/(Winter et al., 2007).
(Nicotiana tabacum) and C 4 dicot Flaveria bidentis demonstrated that when exogenous sucrose was supplemented in the growth medium, it stimulated maximal photosynthetic rates comparable with when plants are grown in glasshouse conditions (Furbank et al., 1997).It must be noted that in the latter study by Furbank et al. (1997) nitrogen was not limited, unlike in Krapp et al. (1991).These highly controlled experiments provided evidence for the link between photosynthesis and sugar sensing and signalling in C 3 species, underlining the complexities of this pathway (Burnett, 2019).
A key sugar sensor that was first identified in Arabidopsis was HEXOKINASE1 (HXK1) (Moore et al., 2003).AtHXK1 phosphorylates hexoses but also senses glucose within the leaves, where an increase in glucose caused a decrease in expression of photosynthesis-related genes such as RbcS which encodes the small subunit of Rubisco.Similar mechanisms have been documented in other C 3 species such as tobacco (N.tabacum), rice, and potato (Solanum tuberosum) (Veramendi et al., 2002;Cho et al., 2009;Kim et al., 2013), where homologues in monocots are usually HXK5 and HXK6.The relationship between sugar sensing and photosynthesis has been most well documented with hexokinase, but there has been evidence for other sugar sensors as well.For example, photosynthesis was altered when manipulating the trehalose synthesis pathway, more specifically by changing the levels of the sensing metabolite trehalose 6-phosphate (Tre6P) which acts as a proxy for sucrose levels (Pellny et al., 2004;Yadav et al., 2014).Photosynthesis-derived glucose has been shown to modulate target of rapamycin (TOR) signalling, changing the transcriptome to regulate plant growth (Xiong et al., 2013).TOR forms part of a larger complex known as TOR complex 1 (TORC1) that also includes the regulatory proteins regulatory-associated protein of TOR (RAPTOR) and Lethal with Sec Thirteen 8 (LST8).Snf1-related protein kinase1 (SnRK1) has been shown to have a role in the starvation response in plants when photosynthesis is limited by up-regulating catabolic processes (Baena-González et al., 2007).SnRK1 comprises catalytic kinase α subunits and regulatory β, γ, and βγ subunits, the latter of which are specific to plants.
Under conditions of sugar build up in source leaves, glucose can act as a potent inhibitor of photosynthetic gene expression in plants.However, sucrose, fructose, and more recently Tre6P have also been implicated as key signalling molecules, but sugar-specific signalling can be difficult to separate from a role in central metabolism (Cho and Yoo, 2011;Ponnu et al., 2011;Li and Sheen, 2016;Baena-González and Lunn, 2020;Yoon et al., 2021).Expression of invertase within different cell compartments such as the apoplast, cytosol, and vacuole of the source leaves has shown that sucrose is not responsible for regulating photosynthesis since it is down-regulated when hexoses accumulate in the photosynthetic cells after sucrose hydrolysis (Sonnewald et al., 1991;Heineke et al., 1992;Büssis et al., 1997;Kingston-Smith et al., 1999).Sugar sensors can form complexes or post-transcriptionally regulate proteins by modifying the overall transcription of genes.Under high glucose, HXK1 can interact with the vacuolar H + -ATPase B1 (VHA-B1) and the 26S proteasome AAA-ATPase subunit, specifically the Regulatory Particle 5b (RPT5B), to form a nuclear complex that regulates transcription (Cho et al., 2006).TORC1 is known as a master regulator, where TOR phosphorylates ETHYLENE-INSENSITIVE 2 (EIN2) or PIN-FORMED 2 (PIN2), usually mediating processes associated with plant growth such as cytokinesis and cell elongation and expansion (Yuan et al., 2020;Fu et al., 2021).Conversely, SnRK1 usually suppresses energy-demanding processes during the starvation response.This complex phosphorylates C group basic leucine zipper 63 (bZIP63) affecting dimerization with S1 group bZIPs, which results in transcriptional changes (Mair et al., 2015;Pedrotti et al., 2018;Han et al., 2020;Muralidhara et al., 2021).Arabidopsis plant extracts from young seedlings showed that Tre6P suppresses SnRK1 activity, but not in extracts from mature leaves (Zhang et al., 2009;Li et al., 2021).The authors noted that the intermediary factor present in young seedlings was not present in mature tissues that would cause Tre6P inhibition of SnRK1.Tre6P and/or the sucrose:Tre6P ratio is critical for controlling carbohydrates during different developmental stages, as well as detecting nitrogen status through changes in sucrose levels (Schluepmann et al., 2003;Wahl et al., 2013;Yadav et al., 2014).Similarly, it is possible that SnRK1 activity corresponds to tissue types and stage of growth in plants where in mature leaves anabolic pathways might be less active when compared with younger tissues.Although sugar sensing is already difficult to study in C 3 species, this is even more complicated in C 4 species (for more detail about the pathways, see Box 2 and references therein).

Sugar sensing pathway in C 4 species
Sugar sensing studies in C 4 species have so far produced conflicting evidence about the role of sugar sensors and their relationship with photosynthesis.In contrast to C 3 plants, many C 4 species in FACE experiments did not experience an increase in net photosynthesis (Leakey et al., 2009;Ainsworth and Long, 2021).For example, sorghum did not increase photosynthesis rates under high CO 2 but did increase its carbon gain when subjected to drought by improving water relations due to enhanced photosynthetic performance at lower stomatal conductance under high CO 2 compared with ambient levels (Wall et al., 2001).FACE experiments with the C 4 grass Paspadalum dilatatum or the major crop maize also did not show a net increase in photosynthesis (von Caemmerer et al., 2001;Leakey et al., 2004Leakey et al., , 2006)).The paradigm is that in elevated CO 2 , photosynthetic rates increase marginally or not at all in C 4 species since the C 4 photosynthetic mechanism has evolved to saturate Rubisco with CO 2 within the bundle sheath cells at ambient mesophyll CO 2 concentrations (Hatch and Osmond, 1976;von Caemmerer and Furbank, 2003).Many C 4 species are only more productive under high CO 2 when droughted Box 2. Schematic of sugar sensing/signalling pathways What researchers know about the sugar sensing/signalling pathway is derived from studies in Arabidopsis, the C 3 dicot model species.Hexokinase (HXK), Snf1-related kinase 1 (SnRK1), and Target of Rapamycin complex 1 (TORC1) are known proteins/complexes involved in the sugar sensing/signalling pathways along with the sensing metabolite trehalose 6-phosphate (Tre6P) (Li et al., 2021).HXK catalyses the phosphorylation of hexoses, mainly glucose (Glu), to form glucose 6-phosphate (Glu6P), and can also sense glucose independently from its catalytic function, usually HXK1 in dicots or HXK5 and HXK6 in monocots (Moore et al., 2003;Cho et al., 2009).Trehalose phosphate synthase 1 (TPS1) catalyses the reaction between UDP-glucose (UDPGlu) and Glu6P to form trehalose 6-phosphate (Tre6P), a sugar signalling metabolite which can indicate the sucrose (Suc) status (Paul et al., 2020).Trehalose phosphate phosphatase (TPP) cleaves Tre6P to form trehalose (made up of two glucose molecules) which is only detected at very low amounts in plants.It has been established that a Suc-Tre6P nexus exists that can regulate metabolic pathways, but the specific mechanisms behind this remain ambiguous (Zhang et al., 2009;Yadav et al., 2014).It is hypothesized that there is a relationship between Suc and TPS1 activation, but it is unknown how this occurs and if there is also a direct link between Suc and TPP activity.Tre6P is a negative feedback regulator of Suc, modulating its levels in leaves to increase anabolic processes which can occur by suppressing SnRK1 activity (Baena-González and Lunn, 2020).Suppression of SnRK1 activity can also occur through Glu6P.During plant starvation, SnRK1 activity is increased, increasing catabolic (ATP-producing) processes but decreasing anabolic (ATP-consuming) processes.In contrast, TORC1 activity is decreased during starvation since its function usually increases anabolic and decreases catabolic processes (da Silva et al., 2021).Although still unclear, sugar metabolism activates TORC1 alongside certain hormones such as auxin.While not discussed in detail here, SnRK1 comprises catalytic kinase α subunits and regulatory β, γ, and βγ subunits, the latter of which is specific to plants.TORC1 comprises the TOR kinase interacting with Regulatory-Associated Protein of TOR (RAPTOR) and the Lethal with Sec Thirteen 8 (LST8) proteins.The sugar metabolism pathway and these sensors have been implicated in regulating photosynthesis, but the mechanisms behind their role remain ambiguous.Furthermore, information on whether sugar sensing or signalling is ubiquitous or unique between the bundle sheath mesophyll cells of C 4 species has been limited.Sugar sensing and signalling is difficult to study given the complexities of sensing and signalling pathways (see Fichtner et al., 2021 and references therein for more detail), transduction cascades, and hormones involved in a C 3 plant, but is further complicated in a C 4 species where photosynthesis and sucrose and/or starch synthesis is compartmentalized, requiring orchestration between two cell types.
and stomatal occurs to conserve water that evaporates during transpiration (Ainsworth and Long, 2021).As discussed earlier, C 3 species increase photosynthetic rates due to the increased availability of CO 2 within leaf mesophyll cells where Rubisco is located (Grodzinski et al., 1998;Moore et al., 1999;Ainsworth and Rogers, 2007).Since C 4 species already operate at 'CO 2 saturation' under ambient CO 2 concentrations, it is possible that either they can bypass sugar feedback regulation of photosynthesis by not allowing photoassimilates to accumulate through cell type compartmentalization, by exporting it rapidly, or they are less sensitive to sugar increases within these cells.Interestingly, a 20 year FACE experiment has shown a reversal of the C 3 and C 4 trends in grasses under elevated CO 2 in the latter 8 years where C 3 grasses decreased and C 4 grasses increased in total biomass (Reich et al., 2018).This study highlighted that prediction of long-term results using currently known short-term drivers of plant responses could still be unreliable.Sugar sensing and signalling mechanisms may not hold true during prolonged exposure to elevated CO 2 where longterm adaptation to increases in photoassimilates may occur.
Conflicting evidence in elevated CO 2 conditions for C 3 and C 4 plants arises due to different species, CO 2 levels, temperature, water/nutrient availability, light, and experimental conditions (e.g.pot size, duration of measurements), contributing to varying conclusions.A meta-analysis of biomass and photosynthetic rates found that it generally increased in C 3 and C 4 species but usually at a higher percentage for C 3 plants.In Amaranthus edulis, a C 4 dicot species, it was shown that elevated CO 2 did not affect photosynthetic rates or carbohydrate accumulation (Blechschmidt-Schneider et al., 1989).However, after cold treatment which inhibits sucrose translocation, rates of photosynthesis rapidly declined.This suggests that in this C 4 species at least, feedback regulation of photosynthesis still occurs but the threshold at which this occurs might be different to C 3 species.
In elevated CO 2 conditions, sugarcane (Saccharum spp.), a major C 4 grass crop, was found to accumulate 29% more sugars in the leaf than under ambient conditions, with an up-regulation of photosynthesis and its associated genes between the 13th and 22nd week of a 50 week experiment (De Souza et al., 2008).Total biomass also increased by 40% compared with ambient conditions and, unlike observations in other C 4 grasses, the sugarcane exhibited these trends under well-watered and fertilized conditions.In the C 4 model grass species Setaria viridis, the down-regulation of photosynthesis was not observed under high light (1000 µmol m −2 s −1 ) even though there was high accumulation of sugars compared with medium light (500 µmol m −2 s −1 ).Transcriptional changes related to photosynthesis and sugar sensor genes were stronger under low light (50 µmol m −2 s −1 ) than high light (Henry et al., 2020).These findings appear to contradict previous results established in C 3 species where an increase in sugars downregulates photosynthesis gene expression (Moore et al., 2003;Cho et al., 2009).This trend was also observed in isolated maize mesophyll cells when applying sucrose and glucose exogenously (Sheen, 1990).Sucrose concentrations >300 mM were shown to significantly decrease the promoter activity of key photosynthesis genes, pyruvate, phosphate dikinase (PPDK), PEPC, and RbcS in maize protoplasts when compared with normal levels of sucrose (30 mM) (Gerhardt et al., 1987).The author postulated that photosynthetic regulation by sugars may only occur above certain physiological limits.Therefore, it is possible that although no down-regulation of photosynthesis was detected in some studies on C 4 species despite the accumulation of sugars, the threshold to cause this repression of genes had not been reached.
Sugarcane is a major C 4 grass crop and widely cultivated across the world due to its ability to store substantial amounts of sugars within its stems.During sugarcane phloem loading perturbation, it was found that photosynthesis rates and photosynthetic enzyme activity decreased across 5 d (McCormick et al., 2008b, c).Similar findings have been uncovered when increasing sugars within the leaf by other methods such as exogenous application of sucrose and in elevated CO 2 , but with the latter decreasing assimilation only after a long period of exposure (De Souza et al., 2008;Lobo et al., 2015).Sugar accumulation within the leaf is an early symptom (prior to yellowing) of Yellow Canopy Syndrome (YCS) in sugarcane and can be used as a system to study effects of feedback regulation on photosynthesis (Marquardt et al., 2019(Marquardt et al., , 2021)).Sugarcane suffering from YCS displays an accumulation of sugars in the leaf that precedes yellowing and a decrease in photosynthesis.RbcS and Rubisco activase (RCA) as well as PEPC transcripts and protein were not down-regulated in early-stage sugar accumulation.This was similarly observed with RbcL transcript except that the protein which it encodes was up-regulated early in sugar accumulation, perhaps suggesting some post-transcriptional modifications of RbcL.Early-stage accumulation of sugars could be attributed to the apoplast rather than the cytosol of photosynthetic cells as the infection of pathogens can often lead to an increase in apoplastic sugars.This response differs for different pathogens where for viruses they replicate within the cells unlike fungal and bacterial infections that usually access sugars via the apoplast (Liu et al., 2022).The apoplastic and cytosolic sugars were not measured under YCS infection and therefore changes cannot be attributed to an increase of intracellular sugars specifically and perhaps this accounts for the absence of transcriptional and proteomic changes related to Rubisco and PEPC (Marquardt et al., 2021).Many transcripts and proteins were only down-regulated during late-stage sugar accumulation, perhaps suggesting that a certain level of sugars might need to be reached within the cytosol before a down-regulation of photosynthetic genes occurs.
Perturbations between the source and sink by shading every leaf except one in sugarcane increased sink demand and suppressed the expression of genes encoding HXK sensor homologues, similar to results reported for Setaria (McCormick et al., 2008a;Henry et al., 2020).Disturbance of the source-sink balance between the only unshaded leaf in sugarcane and the subtending internode resulted in an increase in photosynthetic rate and its related genes.The authors suggested that this supported the notion that the sugar sensing HXK can regulate photosynthesis gene expression since decreased HXK sensing would result in less photosynthetic repression.The trehalose metabolism pathway has also been implicated in the sugar sensing mechanisms in sugarcane.Sugar accumulation in the leaves caused genes encoding trehalose 6-phosphate phosphatase (TPP) and trehalose 6-phosphate synthase (TPS) to be up-and downregulated, respectively (McCormick et al., 2008b).Previous evidence in the C 3 dicot tobacco showed that, when Escherichia coli homologues of TPP and TPS were overexpressed, they increased and decreased photosynthesis, respectively (Pellny et al., 2004).Overexpressing the same TPP and TPS homologues from E. coli in sugarcane increased and decreased sucrose levels, respectively (Gabriel et al., 2021).Therefore, it is possible that similar mechanisms of sugar sensing via Tre6P signalling might be involved in regulating expression of photosynthesis genes in sugarcane and other C 4 grasses.The importance of the Tre6P signalling pathway in C 4 plants has recently been summarized by Rojas et al. (2023) and underlined the need to move knowledge beyond the model C 3 dicots Arabidopsis and tobacco given their different physiology and anatomy as well as source-sink demands.

A better understanding of C 4 plant sugar sensing and signalling
In C 4 plants, studying the sugar sensing pathways is complicated because photosynthesis and carbohydrate synthesis occur in both the mesophyll and bundle sheath cells instead of just the one cell type.Expression of genes encoding sugar sensor components could have evolved and been co-opted during C 4 evolution to predominate in regulating sugars in the C 4 source tissues.HXK is well known for its sugar feedback regulation of photosynthesis in the C 3 dicot Arabidopsis, C 3 monocot rice, and, to a lesser extent, in maize, a C 4 monocot (Sheen, 1990;Moore et al., 2003;Cho et al., 2009).Electronic fluorescent pictographs (eFPs) depicting gene atlas expression from published RNA-seq data showed that the sensing HXK does not have preferential expression in source tissues over sink tissues of any of the three species depicted (Fig. 1).Preliminary analyses with available gene expression data and atlases across various tissues showed that there was no particular gene encoding sugar sensor components that was consistently expressed in the source tissues in either C 4 or C 3 species (Benning et al., 2023).A gradient of sugar sensor gene expression was also not consistently observed for grasses as is observed for many sugar and starch metabolism genes during the transition from sink to source tissue along a developing leaf (Chen et al., 2022).Protein sequence comparisons of C 4 and C 3 grasses across a phylogenetic spread showed that for many major sugar sensor components, the amino acid residue identity was often >90% (Table 1) (Benning et al., 2023).This suggests that sugar sensors between C 4 and C 3 species are functionally very similar if not the same.Details on the mechanisms underpinning sugar sensing in C 4 species remain elusive and at times conflicting.For example, in Setaria, a down-regulation of photosynthesis under high light is observed even when exposed for only 4 h at 900 μmol m −2 s −1 (Anderson et al., 2021), but an up-regulation of photosynthesis under high light was observed when exposed for 4 d at 1000 μmol m −2 s −1 (Henry et al., 2020).This highlights the importance of the difference between adaptation to treatments and an immediate response.Early work on maize indicated that this species was very tolerant to high foliar carbohydrate levels even when grown under continuous illumination, precluding diurnal starch breakdown (John et al., 1973).Much of the evidence presented in this review suggests that sugar accumulation in C 4 species can still cause a down-regulation in photosynthesis under certain conditions.Given the lack of cell-specific measurements of sugar levels in C 4 leaves, it is difficult to ascertain whether treatments such as phloem loading perturbation and exogenous sugar feeding reflect physiologically relevant cellular sugar levels.In the case of sugarcane, the process of 'ripening', where sucrose accumulates in the stem following a reduction in growth (usually due to lower temperatures but also by plant growth regulators) but with continual provision of photosynthate from leaves, relies on photosynthesis proceeding even though sink strength has been reduced (Glasziou et al., 1965).This would suggest a lack of sensitivity to sink feedback via sugar signalling in this C 4 grass.
One of the major underlying questions is whether the accumulation and feedback regulation threshold is much higher in C 4 leaves than in their C 3 counterparts, whether dicot or monocot.Although difficult to achieve, examining the amount of sugars specifically in the mesophyll and bundle sheath cells of C 4 species would also provide clarity on sugar sensing mechanisms.In the C 3 grass barley, it has been estimated that the cytosol of the mesophyll cells have ~232 mM sucrose, while the parenchymatous bundle sheath can reach up to 100 mM under cold stress; however, this may in fact be much higher in C 4 species and can change drastically under different conditions (Winter et al., 1992;Koroleva et al., 1998).
Currently the specific roles that HXK, TORC1, SnRK1, and Tre6P play in regulating C 4 photosynthesis remain unclear.Mechanical separation of mesophyll and bundle sheath cells is often used to determine their individual transcriptomes.Analyses of transcript abundance, protein abundance, and activity for photosynthetic and sugar/starch metabolism have been established between the bundle sheath and mesophyll cells (Lunn andFurbank, 1997, 1999;Lunn, 2007;Furbank and Kelly, 2021).In the C 4 grasses Z. mays, Sorghum bicolor, Panicum hallii, Setaria viridis, and Setaria italica, the genes encoding the Rubisco small subunit (RbcS) and NADP-ME and their expression predominates in the bundle sheath whereas PEPC is found in the mesophyll (see Box 1 for pathway) (Fig. 2A).This highlights the known compartmentalization of photosynthesis in C 4 plants.While it was mentioned that sucrose and starch synthesis is often compartmentalized in C 4 grasses, it can vary from species to species (Lunn and Furbank, 1999).Starch synthesis has been more consistently found to occur in the bundle sheath cells, evidenced by transcriptomic, proteomic, and enzyme activity information (Furbank and Kelly, 2021).It is shown in Fig. 2B that the key starch synthesis gene, starch synthase, is more abundant in the bundle sheath for the each of the five C 4 grasses.Genes encoding sucrose phosphate synthase (SPS) and sucrose phosphate phosphatase (SPP) within the sucrose biosynthesis pathway were more inconsistent in their expression between cells.Many SPS and SPP genes were equally abundant in both cell types except for ZmSPS, SiSPS, and SvSPP which seemed to be markedly more abundant in the mesophyll and SbSPS and PhSPP which were more abundant in the bundle sheath.All C 4 grasses Note that for those species where there was no publicly available genome, a de novo assembly was performed which did not map small regions at the N-and C-terminus of the protein.Numbers represent percentage similarity.
showed a strong preferential expression of SWEET13 genes in the bundle sheath cells.SWEET13 has been implicated to an important export step for phloem loading in C 4 grasses in the apoplastic pathway (Box 1) (Bezrutczyk et al., 2018;Chen et al., 2022;Hua et al., 2022).
Analysis of transcript abundances of genes encoding sugar sensor components shows that there is some preferential expression of certain genes, but results are inconsistent between species (Benning et al., 2023).Notably, TPS1 was strongly expressed in the bundle sheath compared with mesophyll cells Fig. 2. Heatmap expression of major genes encoding photosynthesis, sugar metabolism, and sensors in the bundle sheath and mesophyll cells.Gene expression from Zea mays (Zm) (Denton et al., 2017), Sorghum bicolor (Sb) (Döring et al., 2016), Panicum hallii (Ph) (Washburn et al., 2021), Setaria viridis (Sv) (John et al., 2014), and Setaria italica (Si) (Washburn et al., 2021) bundle sheath and mesophyll cells.Expression of genes encoding Rubsico small subunit (RbcS), NADP-malic enzyme (NADP-ME), and phosphoenolpyruvate carboxylase (PEPC) from the C 4 photosynthetic pathway (A).Expression of genes encoding starch synthase (SS), sucrose phosphate synthase (SPS), sucrose phosphate phosphatase (SPP), and Sugars Will Eventually be Exported Transporters (SWEET) in the sucrose/starch synthesis and sucrose transport pathways (B).Expression of genes encoding components of sugar sensors hexokinase (HXK), Target of Rapamycin (TOR), Snf1-related kinase 1 (SnRK1) α1 subunit, and trehalose phosphate synthase 1 (TPS1) (C).Where there is more than one isoform present, the most abundantly expressed gene is depicted.The scale bar to the right of each heatmap depicts the log 2 bundle sheath/mesophyll (BS/M) ratio where values >1 denote higher expression in bundle sheath, values <1 denote higher expression in mesophyll, and 0 denotes equal abundance in both photosynthetic cells.
for all C 4 grasses examined (Fig. 2C).TPS1 is part of the Tre6P signalling pathway (Box 1) where this metabolite important for controlling diurnal starch degradation by signalling sucrose availability (Lunn et al., 2006;dos Anjos et al., 2018;Ishihara et al., 2022).Therefore, in C 4 species, it could be more functionally relevant if TPS was predominantly located in the bundle sheath to modulate Tre6P levels since this is the primary site of starch synthesis for many C 4 grasses.When Tre6P levels were manipulated in maize, it was shown that this metabolite was important for regulation of photosynthesis, probably due to an increase in sink strength since SWEET expression also increased in reproductive organs (Oszvald et al., 2018).While there has been evidence for Tre6P modulation of SWEET expression, the specific mechanisms behind this are yet to be resolved.Tre6P signalling-controlled expression of SWEET13 in bundle sheath cells of C 4 species could provide a regulatory response to sugar status at the first step of phloem loading.
Unlike the sensors such as kinases, which are limited in action to the cells where they are expressed, the metabolite Tre6P can potentially pass through the abundant plasmodesmatal connections between bundle sheath and mesophyll cells of C 4 leaves, signalling photoassimlate abundance (Danila et al., 2016).Tre6P could be important for sensing and controlling sucrose abundance in the bundle sheath, perhaps coordinating the flux of sucrose from the mesophyll, its synthesis after starch degradation from the chloroplast, and its export to the phloem for loading.Under high light in Setaria, as sucrose increases, levels of Tre6P also increase while total SnRK1 activity does not change when compared with low or medium light over 4 d (Henry et al., 2020).In vitro Tre6P inhibition assays on SnRK1 activity show that activity was reduced by half, suggesting that this type of regulation via Tre6P and SnRK1 also occurs in C 4 species.A comprehensive review of the role of Tre6P in crops by Paul et al. (2018) noted that Tre6P levels can be vastly different between Arabidopsis and crops such as wheat and maize.For example, Tre6P can reach up to 10 nmol g -1 FW in seedlings (Nunes et al., 2013) and is typically lower in rosettes (Lunn et al., 2006), whereas up to 119 nmol g -1 and 50 nmol g -1 FW of Tre6P has been detected in the wheat grain endosperm and maize kernel, respectively (Martínez-Barajas et al., 2011;Nuccio et al., 2015;Bledsoe et al., 2017).Furthermore, the Tre6P:sucrose nexus appears to be weaker in these crops, suggesting that Tre6P may be different from the established mechanisms in Arabidopsis, possibly due to breeding for high yields (Martínez-Barajas et al., 2011;Bledsoe et al., 2017;Paul et al., 2018).The Tre6P levels in the bundle sheath and mesophyll cells of C 4 source leaves is unknown as is whether Tre6P is important for signalling related specifically to C 4 photosynthesis.
SnRK1α1 and TOR subunits show varied expression between the bundle sheath and mesophyll and between C 4 species (Fig. 2) (see Benning et al., 2023 for more detail).These master regulators are likely to be necessary in both bundle sheath and mesophyll cells of C 4 species, possibly modulating photosynthetic activity in both cells, photoassimilate metabolism, and its subsequent transport.HXK sensor transcripts, however, were more abundantly expressed in the bundle sheath, except for SiHXK5, SvHXK6, and SiHXK6.The transcript abundance in the bundle sheath was more preferential for TPS1 than for HXK, suggesting that in C 4 grasses HXK is often present at similar levels in both the mesophyll and bundle sheath cells.This is possibly not surprising given that the known photosynthesis targets of HXK are housed in both the bundle sheath and mesophyll cells (Henry et al., 2020).
Plant species also differ in their carbohydrate storage molecules, where some species use starch predominantly or sucrose in the vacuole which can be trafficked by SUTs, tonoplastlocalized transporters (TSTs), or SWEETs depending on the gradient between the vacuole and cytosol.Setaria has been shown to contain some levels of fructans within their leaves (Chen et al., 2022) which is more common in C 3 temperate grasses such as barley and wheat (Pollock, 1986;Pollock and Cairns, 1991).Depending on the primary storage strategy of each plant, this could also result in differing sugar sensing pathways.Activity of sugar sensors and their signalling has also not been determined thus far for each photosynthetic cell type of C 4 species.Although it must be noted that many of these sensors have more than one function or signalling role, isolating those functions related specifically to sugar sensing can be difficult in a single photosynthetic cell system of a C 3 leaf, and even more complicated in a two-cell system of a C 4 leaf.
AtHXK1 in Arabidopsis was one of the first sugar sensors whose sugar sensing function was separated from its catalytic function using glucose insensitive (gin) mutants where HXK1 was knocked out and, as a result, could overcome high exogenous glucose application (Moore et al., 2003).This seminal study established a direct link between photosynthesis regulation and a sugar sensor within a plant by using genetic manipulation of the gene encoding the protein of interest.This experiment has not been replicated in C 4 plants to determine if similar phenotypes are observed.
The sensor and signal that coordinates sugar synthesis and export is still unknown in C 3 and C 4 plants.Earlier work on enzymes of sugar synthesis noted that they are relatively insensitive to product inhibition and so it is unlikely sucrose itself acts as a feedback signal (Koch et al., 1996;Koch, 2004).As discussed, there is some level of feedback regulation on photosynthesis during sugar accumulation in C 4 plants, but the mechanisms are also unknown.Given that C 4 plants have substantially higher photosynthetic rates than most C 3 plants (Byrt et al., 2011), one would predict that rates of phloem export would need to be proportionally enhanced to fuel plant growth.Many C 4 grasses follow the apoplastic pathway of phloem loading where there is active transport of sugar into the SE-CC complex via SUTs and passive export via SWEETs from phloem parenchyma and photosynthetic cells (Robinson-Beers et al., 1990;Lohaus et al., 1995;Slewinski et al., 2010;Chen et al., 2022).The efficiencies of this pathway compared with symplastic loading may prevent the build-up of photosassimilate at the point of export from the leaf, or C 4 plants may be more tolerant to sugar accumulation.Currently, it is which of these possibilities is correct. 14CO 2 -labelled gas exchange showed that while C 4 plants showed a higher rate of photoassimilate export, the concurrent rate of export as a percentage of photosynthetic rate did not differ from C 3 species, although this measurement does not accommodate remobilization of starch to sucrose and export at night (Grodzinski et al., 1998).Therefore, it is possible that sugar transporters and diurnal export strategies have adapted to export needs of the efficient photosynthetic cells in a C 4 system (Emms et al., 2016).

Conclusions
Sugar sensing and signalling in relation to the regulation of photosynthetic flux in the source leaf is not well known, especially in C 4 plants, many of which are major cereal crops.Since improving photosynthetic performance has become a major frontier for increasing crop yields, it is pivotal to better understand the coordination of source and sink so that gains in photosynthetic capacity can be realized in increased yield (Reynolds et al., 2012;Lawlor and Paul, 2014;Paul et al., 2020).Current knowledge on sugar feedback regulation of photosynthesis has been established mainly in the C 3 dicot Arabidopsis, but work in this model species should not be presumed to be directly translatable to a C 4 monocot crop.The biochemical and anatomical specialization in C 4 leaves required to support the two-cell photosynthetic mechanism has major ramifications for the regulation of carbon partitioning and sugar signalling, but almost nothing is known about these regulatory mechanisms in this important group of plants.Conflicting results of sugar feedback regulation that occurs in C 4 plants means that the sugar sensing and signalling mechanisms remain ambiguous in these species.Sugar feedback probably does occur, but perhaps the mechanisms regulating it are different from those in C 3 species given the anatomical complexities of C 4 grasses and/or that the threshold which sugars levels must reach to elicit a response is higher.

Box 1 .
Compartmentalization of photosynthesis and carbohydrate synthesis in C 4 grass source leaves In the more common NADP-malic enzyme (NADP-ME) subtype of C 4 plants, phosphoenolpyruvate carboxylase (PEPC) catalyses the first carbon fixation step with phosphoenolpyruvate (PEP) to form oxaloacetate (OAA) [1].OAA enters the chloroplast via a plastidic transporter (white circle) where malate dehydrogenase (MDH) forms the C 4 acid malate (MAL) [2].MAL moves out of the chloroplast and diffuses across to the bundle sheath via the abundant plasmodesmata connections between the two cell types.MAL moves into the bundle sheath chloroplast where NADP-ME decarboxylates to release CO 2 which enters the Calvin cycle to be refixed by Rubisco while also forming pyruvate (PYR) [3].3-Phosphoglycerate (3PGA) is formed in the Calvin cycle and can move back into the mesophyll chloroplast to form glyceraldehyde 3-phosphate (G3P) through reactions first catalysed by phosphoglycerate kinase (PGK) and then glyceraldehyde 3-phosphate dehydrogenase (GADPH) [4].Triose phosphate isomerase (TPI) catalyses the reaction with G3P to form dihydroxyacetone phosphate (DHAP) [5] which can re-enter the Calvin cycle.Pyruvate phosphate dikinase (PPDK) catalyses the reaction with PYR