Guard cell photosynthesis and stomatal function


Author for correspondence:
Tracy Lawson
Tel: +44 (0) 1206 873327
Fax:+44 (0) 1206 873416






intercellular CO2 concentration




dihydroxyacetone phosphate


steady-state fluorescences

inline image

quantum efficiency of photosystem II (PSII) photochemistry


Mehler-ascorbate peroxidase


nicotinamide adenine dinucleotide phosphate




phosphoenolpyruvate carboxylase


3-phosphoglyceric acid


ribulose-1,5-bisphosphate carboxylase/oxygenase


ribulose-1,5-bisphosphate carboxylase.

I. Introduction

Stomata are small adjustable pores found in large numbers on the surface of most aerial parts of higher plants and have been documented in the fossil record from as early as the late Silurian, 411 Myr ago (Edwards et al., 1992, 1998). A stoma is formed from two specialized cells in the epidermis (guard cells) which are morphologically distinct from general epidermal cells and are responsible for controlling stomatal aperture (Franks & Farqhuar, 2007). Paired guard cells, in some species together with epidermal subsidiary cells, form the stomatal complex (Fig. 1). Subsidiary cells can play a role in stomatal movements either mechanically or as ion reserves (Raschke & Fellows, 1971). In most plants stomata can be found on both the upper (adaxial) and lower (abaxial) leaf surfaces, such leaves being termed amphistomatous, with the majority of stomata found on the lower surface (Tichà, 1982). In some species (particularly trees) stomata are found only on the lower surface (i.e. the leaf is hypostomatous), whilst some aquatic plants (such as water lilies) have stomata only on the upper surface (i.e. the leaf is epistomatous) (Morison, 2003). As the leaf cuticle is almost impermeable to water and CO2, the central role of stomata is regulation of gas exchange between the inside of the leaf and the external environment (Cowan & Troughton, 1971; Jones, 1992). Through their role in controlling transpiration, stomata also aid in leaf cooling, metabolite fluxes, and long-distance signalling (Brownlee, 2001; Lake et al., 2001; Jia & Zhang, 2008) as well as acting as a barrier to harmful pollutants such as ozone and pathogens (Meidner & Mansfield, 1968; Mansfield & Majernik, 1970).

Figure 1.

Stomatal complex illustrating the pair of guard cells, complete with chloroplasts and subsidiary cells in which calcium oxalate crystals are often found. Calcium oxalate is believed to play a role in regulation of stomatal aperture (see Ruiz & Mansfield, 1994).

Plants require sufficient CO2 to enter the leaf for photosynthesis whilst conserving water to avoid dehydration and metabolic disruption. When fully open, stomatal pores only occupy between 0.5 and 5% of the leaf surface (Hetherington & Woodward, 2003; Morison, 2003); however, almost all the water transpired as well as CO2 absorbed passes through these pores. For this reason, stomatal function has significant implications for global hydrological and carbon cycles. The quest to understand stomatal control of photosynthetic CO2 fixation and plant water relations is becoming increasingly important with changing climatic conditions. Knowledge of stomatal function is critical to determine plant responses to environmental stresses, particularly reduced water availability, and is necessary to identify plants with decreased water use that are capable of high yields in more extreme environments (Morison et al., 2007). On a global scale, drought causes more yield losses than any other single biotic or abiotic factor (Boyer, 1982), resulting in increasing pressure for agronomists and plant breeders to identify crop varieties that are drought tolerant for sustainable production of food and biofuels on drought-susceptible land. Increased knowledge of stomatal function could provide the key to such crop improvements (Jones, 1987; Wang et al., 2007).

The aims of this review are to examine some of the potential functions of the guard cell chloroplasts and how these are linked to stomatal behaviour. Contrasting evidence has led to a number of controversies regarding guard cell chloroplast function, in particular concerning the contribution of guard cell photosynthesis to stomatal regulation, and several opposing views are discussed in the following sections. Whilst there are still gaps in our knowledge regarding stomatal regulation, as well as sensory and signalling mechanisms, a wealth of evidence exists on stomatal responses to various environmental stimuli, guard cell osmoregulation and mechanisms of movement. A historical account of the various osmoregulatory pathways and solutes used in stomatal movements is given, which demonstrates the extreme plasticity in guard cell function and illustrates the difficulties faced by stomatal researchers attempting to elucidate specific stomatal mechanisms. The review also briefly examines the use of recent advancements in modern techniques (such as antisense technology and the use of DNA mutation) to address the function of chloroplasts within guard cells. Recent studies using such techniques have already revealed interesting and unexpected results, paving the way for future research, which may allow us to fill gaps in our understanding of the role of the guard cell chloroplasts in stomatal regulation.

1. Stomatal regulation

Despite the controversies mentioned above, decades of research have provided a substantial amount of information regarding stomatal responses to various environmental stimuli. The details of these responses, along with mechanisms of pore opening and closing, are now widely accepted. Stomatal aperture is regulated by both internal physiological and external environmental factors (Farquhar & Sharkey, 1982; Morison, 1987; Mansfield et al., 1990; Hetherington & Woodward, 2003; Buckley, 2005) and can respond in time-scales of seconds to hours (Assmann & Wang, 2003). In general, pore opening through guard cell movements is stimulated by illumination with light in the photosynthetically effective waveband, low CO2 concentrations and high humidity, whilst closure is promoted by darkness, low humidity, high temperature and high CO2 concentrations (see reviews by Assmann, 1993; Willmer & Fricker, 1996; Outlaw, 2003; Vavasseur & Raghavendra, 2005; Shimazaki et al., 2007) as well as plant hormones such as abscisic acid (see review by Weyers & Paterson, 2001). However, there are exceptions, the most obvious being crassulacean acid metabolism (CAM) plants which in general maintain closed stomata during the light period and open stomata in darkness (Osmond, 1978; see Black & Osmond, 2003). Additionally, stomata of Gunnera (Osborne, 1989) and Lemna spp. (Park et al., 1990) are unresponsive to a number of environmental stimuli, whilst there are also examples of several other species that do not respond to numerous plant hormones (e.g. Ridolfi et al., 1996; see Weyers & Paterson, 2001).

It is recognized that stomatal responses to light have at least two components. One component is the photosynthesis-independent, specific blue-light response that saturates at low fluence rates, and is often associated with rapid stomatal opening (see Zeiger et al., 2002) believed to involve the activation of a plasma membrane H+-ATPase (Kinoshita & Shimazaki, 1999; Shimazaki et al., 2007). The other component, a photosynthesis-mediated response (termed the red-light response here and in many other publications), saturates at high fluence rates similar to those that saturate guard and mesophyll cell photosynthesis and is inhibited by 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU, an inhibitor of photosystem II (PSII)), indicating that it is photosynthesis-dependent (e.g. Kuiper, 1964; Sharkey & Raschke, 1981a; Tominaga et al., 2001; Olsen et al., 2002; Zeiger et al., 2002; Messinger et al., 2006) and suggesting that chlorophyll is the receptor (Assmann & Shimazaki, 1999; Zeiger et al., 2002). This photosynthesis-dependent response can be observed under either blue or red light capable of driving photosynthesis (Sharkey & Raschke, 1981a) and is often believed to operate through mesophyll-driven consumption of CO2 reducing the internal CO2 concentration (Ci) (Roelfsema et al., 2002, 2006), to which stomata are known to respond (Mott, 1988). However, there is also evidence for a direct guard cell red-light response, independent of mesophyll photosynthesis, as discussed in Section V.

Stomatal opening is brought about by the accumulation of ions and/or solutes (Imamura, 1943; Fujino, 1967; Outlaw & Manchester, 1979; Outlaw, 1983) in guard cells, which increases the osmotic potential, thus lowering the water potential, causing water uptake from the apoplast (Weyers & Meidner, 1990; Willmer & Fricker, 1996). Increases in guard cell volume and hence turgor pressure widen the stomatal pore (Franks & Farqhuar, 2007; Shimazaki et al., 2007). Closure is brought about by the reverse, a loss or release of solutes, accompanied by a loss of water and consequently turgor pressure. Both pore opening and closure are energy-dependent processes (Willmer & Fricker, 1996).

However, our understanding of the perception of, and precise response of stomata to, different environmental stimuli is not complete. The fact that stomata in isolated epidermal peels respond to various environmental factors suggests that part of the sensory mechanisms is located in the epidermis (Willmer & Fricker, 1996; Frechilla et al., 2002).

It is also well established that there is a strong positive correlation between stomatal conductance and mesophyll photosynthesis (e.g. Wong et al., 1979; Zeiger & Field, 1982), and a close correlation between photosynthetic efficiencies in guard and mesophyll cells has also been observed (Lawson et al., 2002, 2003). The majority of guard cells have chloroplasts, and these would therefore provide an ideal and convenient location for sensory or regulatory mechanisms. Although guard cell chloroplasts are a characteristic feature of most plants, the role of these highly conserved organelles in osmoregulation and their importance in stomatal function largely remain unclear.

2. Guard cell chloroplasts

In most species studied, guard cells contain chloroplasts, which vary in number depending upon the species (Willmer & Fricker, 1996; Lawson et al., 2003; Fig. 2). Most species typically contain 10–15 chloroplasts per guard cell (Humble & Raschke, 1971), compared with 30–70 in a palisade mesophyll cell. However, numbers of chloroplasts per guard cell range from 3–6 in Selaginella (Allaway & Milthorpe, 1976) to up to 100 in Polypodium vulgare (Stevens & Martin, 1978), and guard cells of Paphiopedilum species entirely lack chloroplasts (Nelson & Mayo, 1975; Rutter & Willmer, 1979; D’Amelio & Zeiger, 1988) but still maintain functional stomata (Nelson & Mayo, 1975). Guard cells are formed from epidermal cells, which notably also lack chloroplasts (again there are exception such as Polypodium species; Fig. 2).

Figure 2.

Images of stomata from intact leaves. A reflected light image from Commelina communis (a) and steady-state fluorescence images from Commelina communis (b), Vicia faba (c), Nicotiana tabacum (d), Polypodium vulgare (e). Chlorophyll fluorescence image of epidermal tissue from Polypodium vulgare (f) showing similar photosynthetic efficiency of epidermal and guard cell chloroplasts. Bars, 10 µm.

Guard cell chloroplasts are often smaller, with less granal stacking, and some are less well developed than those in mesophyll cells (Sack, 1987; Shimazaki & Okayama, 1990), although these features vary across plant families (see reviews by Pemadasa, 1981; Willmer & Fricker, 1996). Another noticeable feature of most guard cell chloroplasts is that starch accumulates in the dark and disappears in the light (Willmer & Fricker, 1996), the reverse of the situation in mesophyll cells. However, this may not be the case for all species, as work by Stadler et al. (2003) has revealed that guard cells of Arabidopsis are practically free of starch in the morning and accumulate starch during the day.

Before examining guard cell photosynthesis and its possible role in stomatal behaviour, including the controversial topic of guard cell Calvin cycle activity, it is essential to provide a brief account of the osmoregulatory pathways that occur in guard cells.

II. Osmoregulation in guard cells

Many decades of research have focused on the osmoregulatory mechanisms found in guard cells. To put this review into context, a brief history of stomatal osmoregulation is given in this section; however, this is by no means exhaustive (I refer readers to the following comprehensive reviews: Talbott & Zeiger, 1998; Zeiger et al., 2002; Outlaw, 2003; Roelfsema & Hedrich, 2005; Vavasseur & Raghavendra, 2005).

In 1908, Lloyd (Lloyd, 1908) observed that stomata contained more starch when closed in the dark than when open during the day, which led to the starch–sugar hypothesis. This theory relies on the interconversion of starch to sugars, which results in osmotic changes, leading to alterations in guard cell turgor, and became the most widely accepted theory regarding the osmoregulatory mechanism for many decades (Meidner & Mansfield, 1968). In the 1960s Fischer and co-workers highlighted the importance of potassium (K) uptake in stomatal opening (Fischer, 1968; Fischer & Hsiao, 1968) (although a paper on this by Imamura had already been published; Imamura, 1943), and the starch–sugar hypothesis was effectively replaced with the K+-malate2− theory. This work demonstrated that K+ uptake (Yamashita, 1952; Fischer, 1968; see reviews by Raschke, 1975, 1979) in guard cells was correlated with stomatal opening, with malate2− and/or chloride (Cl) (Allaway, 1973; Schnabl, 1977; Schnabl & Raschke, 1980; Outlaw, 1983; Willmer & Fricker, 1996; Asai et al., 2000) acting as the counterion(s). The general view is that malate2− is the major counterion balancing K+ uptake, although species such as Allium cepa (in which starch is absent) exclusively use Cl (Schnabl & Zeiger, 1977; Schnabl & Raschke, 1980). Uptake of K+, driven by a H+ gradient activated by proton ATPase (Zeiger, 1983; Shimazaki & Kondo, 1987), was shown to be correlated with malate accumulation (Allaway, 1973) and stomatal opening (see review by Outlaw, 1983). A role for sucrose in guard cell osmoregulation was nearly forgotten until several studies suggested that K+ and its counterions could not provide all the osmoticum required to support stomatal apertures in Commelina communis (MacRobbie & Lettau, 1980a,b), and led to the suggestion that soluble sugars account for additional osmoticum to support opening (MacRobbie, 1987; Talbott & Zeiger, 1993). Evidence exists that sugars as well as K+-malate2− can act as osmotica for guard cell osmoregulation (Outlaw & Manchester, 1979; Outlaw, 1983; Reddy & Rama Das, 1986; Talbott & Zeiger, 1993, 1998). For example, Commelina benghalensis accumulates sugars (60% of required osmoticum) and malate2− when treated with fusicoccin (Reddy et al., 1983), a fungal toxin that activates the plasma membrane H+-ATPase (Johansson et al., 1993).

It is easy to see how these changes in osmoregulatory theories may have been problematic for researchers determining the role of guard cell chloroplasts (see Fig. 3). For example, if sucrose is not considered to be an important solute in water movements, a role for guard cell photosynthetic carbon reduction is virtually redundant.

Figure 3.

Schematic diagram showing possible osmoregulatory pathways in guard cells for solute accumulation. Blue lines represent pathways that are believed to be stimulated mostly by blue illumination, and red lines indicate pathways relating to red light- or photosynthesis-dependent pathways. These pathways may not be mutually exclusive. The diagram is not to scale. (Redrawn from information provided by Talbott & Zeiger, 1998; Vavasseur & Raghavendra, 2005; Shimazaki et al., 2007).

1. Multi-osmoregulatory pathways in guard cells

To resolve the differences reported in the literature among results obtained using different experimental procedures and different species, Talbott & Zeiger (1996, 1998) outlined three distinct pathways (see Fig. 3) involved in guard cell osmoregulation that incorporate K+, Cl, malate2− and sucrose, also involving guard cell chloroplasts. They suggested that the importance of these pathways may change depending upon time of day, species, and growth and experimental conditions (Talbott & Zeiger, 1996, 1998). The first pathway describes the uptake of K+ and Cl from the apoplast and/or the synthesis of malate2− from carbon skeletons derived from starch (Outlaw & Lowry, 1977; Outlaw & Manchester, 1979), and is believed to be involved in the early morning opening response and under blue light. In the second pathway, which is insensitive to DCMU (Poffenroth et al., 1992), sucrose is supplied from the breakdown of starch (Outlaw, 1982) and is also thought to play a role in blue-light responses (Tallman & Zeiger, 1988; Poffenroth et al., 1992; Talbott & Zeiger, 1993). In the third, DCMU-sensitive pathway, sucrose is supplied as a product of guard cell photosynthetic carbon reduction (Talbott & Zeiger, 1998). In summary, K+ accumulation is used primarily for rapid opening in the morning, whereas turgor maintenance in the afternoon primarily uses sucrose (Talbott & Zeiger, 1993, 1996, 1998). Talbott & Zeiger (1993) also demonstrated that the major solutes change depending upon lighting regimes and the duration of opening. In Vicia faba peels, the initial (30-min) stomatal opening in response to blue-light illumination resulted in a 173% increase in the concentration of malate, which then decreased, and the concentration of sucrose (from starch breakdown) rose continuously, reaching 215% after 2 h. Under red light, there was little increase in organic acid or maltose concentrations, but the sucrose concentration increased to 208% (Talbott & Zeiger, 1993), with no evidence of starch breakdown (Tallman & Zeiger, 1988; Poffenroth et al., 1992; Talbott & Zeiger, 1993). As this was observed in epidermal peels, sucrose must be supplied from guard cell photosynthetic carbon reduction. Such observations support the hypothesis of multi-osmoregulatory pathways that are regulated by measurement (Talbott & Zeiger, 1998) and growth conditions such as humidity (Talbott et al., 2003) and CO2 concentration (Talbott et al., 1996, 1998; Frechilla et al., 2002, 2004; see also Zeiger et al., 2002). The majority of this work was carried out using V. faba and it is possible that different mechanisms are used by different species or groups of plants; for example, Arabidopsis lacks starch in the morning (Stadler et al., 2003).

A lack of evidence for significant carbon reduction in the guard cells (Outlaw, 1989; Tarczynski et al., 1989; Reckmann et al., 1990; Gautier et al., 1991) led Outlaw and co-workers to propose an alternative source of sucrose (Lu et al., 1995, 1997; Ritte et al., 1999; Outlaw & De Vleighere-He, 2001; Outlaw, 2003; Kang et al., 2007a). Based on the work of Hite et al. (1993), who suggested that guard cells act as carbon sinks, taking up sucrose via plasma membrane transporters (e.g. Stadler et al., 2003), Outlaw and colleagues suggested that apoplastic sucrose recently fixed in the mesophyll cells was a source for guard cell symplastic sucrose and acted as an osmoticum for stomatal opening or replacing guard cell carbon stores (Lu et al., 1997; Ewert et al., 2000; Outlaw & De Vleighere-He, 2001). However, this mechanism appears to be dependent on the amount of sucrose in the apoplast, with concentrations lower than 4 mM unable to support stomatal opening (Ritte et al., 1999). The guard cell apoplastic sucrose can also exert an osmotic effect, which can drive stomatal closure, acting as a possible signal between mesophyll assimilation rate and transpiration (Kang et al., 2007a). It was also postulated that sucrose concentrations near the guard cell regulate gene expression, as has been shown in many other tissues (e.g. Baiser et al., 2004). However, the majority of these studies were conducted on V. faba, which is an apoplastic phloem loader, and different mechanisms may regulate stomatal movements in symplastic loaders (Kang et al., 2007b).

The above research emphasizes the importance of environmental growth and experimental conditions, as well as experimental incubation periods, and provides possible arguments for the involvement of K+, malate2− and sucrose in stomatal function. Additionally, it provides a feasible explanation as to why so many conflicting results are reported in the literature (Zeiger et al., 2002).

III. Role of guard cell chloroplasts in stomatal function

By the early 1990s a consensus was reached that in fact little was known about guard cell metabolism (Mansfield et al., 1990) and, despite over a decade of research, in 2003 Ritte & Raschke (2003) reiterated that there was a lack of information about the physiology and role of the guard cell chloroplasts.

There are four primary ways in which guard cell chloroplasts could contribute to stomatal function (see Outlaw, 1983; Tominaga et al., 2001), with experimental evidence supporting all of these functions:

1. Guard cell electron transport

Guard cells have a pigment composition similar to that of the mesophyll, along with functional photosystem I (PSI) and PSII (Zeiger et al., 1980; Outlaw et al., 1981; Shimazaki et al., 1982; Hipkins et al., 1983). Several researchers have provided evidence for linear electron transport, oxygen evolution and photophosphorylation (Hipkins et al., 1983; Shimazaki & Zeiger, 1985; Willmer & Fricker, 1996; Tsionsky et al., 1997) which can be modulated by CO2 concentration (Melis & Zeiger, 1982) and blue light (Mawson & Zeiger, 1991; Srivastava & Zeiger, 1992). However, studies on fluorescence transients have suggested that induction profiles in guard cells can differ from those in mesophyll cells (Zeiger et al., 1980; Shimazaki et al., 1982; Mawson & Zeiger, 1991; Srivastava & Zeiger, 1992). Changes in the fine structure observed in fluorescence transients are believed to reflect changes in Calvin cycle activity throughout the day or with different wavelengths of light (Mawson & Zeiger, 1991; Srivastava et al., 1998; Zeiger et al., 2002). Additionally, the light-harvesting chlorophyll protein is found in the phosphorylated state in the dark and is dephosphorylated by red light, the opposite situation to that in mesophyll (Kinoshita et al., 1993). This reversal has been suggested to be related to high rates of cyclic electron flow observed in guard cell protoplasts of V. faba, supported by high PSI activity compared with the mesophyll (Lurie, 1977), which could enhance ATP production driven by increased development of the thylakoid proton gradient. However, Shimazaki & Zeiger (1985) did not observe any unusually high PSI activity in guard cells of V. faba but showed linear electron flow to be c. 80% that of the mesophyll. This is consistent with the values of PSII operating efficiency reported later by Lawson et al. (2002, 2003) using high-resolution chlorophyll a (chla) fluorescence.

In the absence of any CO2 fixation, such electron transport rates could provide sufficient ATP to drive ion exchange during stomatal opening (Shimazaki & Zeiger, 1985; Fig. 3), depending on the light wavelength (Schwartz & Zeiger, 1984). Red light induced stomatal opening was shown to be DCMU sensitive and potassium cyanide (KCN) (respiratory poison) resistant whilst under blue light the reverse was observed (Schwartz & Zeiger, 1984). Patch clamp techniques established that a chloroplast modulated red-light response stimulated a proton pump at the plasma membrane in guard cells, suggesting that guard cell photosynthesis may regulate stomatal aperture, through the provision of energy (ATP) and photosynthetic signalling products, such as NADPH (Serrano et al., 1988; see also Wu & Assmann, 1993). However, later studies could not confirm these findings (Roelfsema et al., 2001; Taylor & Assmann, 2001). Experiments conducted under red light with and without the inhibitors oligomycin (an inhibitor of oxidative phosphorylation) and DCMU (an inhibitor of PSII) demonstrated that guard cells supplied ATP to the cytosol under red light, which was utilized by the plasma membrane H+-ATPase for H+ pumping and stomatal opening (Tominaga et al., 2001). An alternative theory for the utilization of photosynthetic electron transport products suggested that ATP and redox power provided from electron transport are used for the reduction of oxaloacetate (OAA) and 3-phosphoglycerate (3-PGA) (from guard cell CO2 fixation or imported from the cytosol; Fig. 3) and exported to the cytosol via a 3-PGA-triose phosphate shuttle (Shimazaki et al., 1989; Ritte & Raschke, 2003).

Sugar as well as K+ accumulation during red light-induced stomatal opening has been reported (Talbott & Zeiger, 1998; Olsen et al., 2002), with sugar production possible either from starch breakdown (Talbott & Zeiger, 1998) or from the utilization of end products of electron transport in photosynthetic carbon reduction within the guard cells themselves (Fig. 3; Talbott & Zeiger, 1998; Olsen et al., 2002). Blue light-induced stomatal opening (see next section) is generally believed not to be dependent on products of guard cell electron transport, as it has been observed in the presence of DCMU (Sharkey & Raschke, 1981a; Schwartz & Zeiger, 1984; see also Roelfsema & Hedrich, 2005), and at low fluence rates (Zeiger, 2000; Shimazaki et al., 2007). The energy is thought to be supplied mostly from mitochondrial respiration (Shimazaki et al., 1982, 2007; Schwartz & Zeiger, 1984), although there is also evidence for energy supply from guard cell chloroplasts (Mawson, 1993a). Stomatal opening in response to weak blue light is greatly enhanced with a background of red light (Shimazaki et al., 2007), although red light is not essential (Sharkey & Raschke, 1981a). In additon to the Ci-driven response under red light, Shimazaki et al. (2007) proposed that guard cell chloroplasts translocate NADH and ATP into the cytocol under red light, which are then used for malate synthesis under blue light.

The above studies demonstrate a direct role for the products of guard cell photosynthetic electron transport in stomatal responses, which is particularly evident under red light, but also possibly plays a minor role in the specific blue-light response. Energy and/or redox power can be used for CO2 fixation (through either phosphoenolpyruvate carboxylase (PEPc) or ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco)), carbohydrate export or ion uptake (Gautier et al., 1991).

2. Role of guard cell chloroplasts in blue-light signalling and response

A recent comprehensive review by Shimazaki et al. (2007) discusses blue-light regulation of stomatal movement in great depth and I refer readers to this for full details. Briefly, blue light induces rapid and highly sensitive stomatal opening correlated with the phosphorylation of a plasma membrane H+-ATPase pump and increased H+ pumping, which results in the activation of voltage-gated K+ channels by membrane hyperpolorization (see Shimazaki et al., 2007 for details), along with the inhibition of s-type anion channels in Arabidopsis and V. faba (see Marten et al., 2007). Inhibition of blue light-induced opening with KCN suggests that ATP for proton pumping is supplied mostly from mitchondrial respiration (Schwartz & Zeiger, 1984; Assmann & Zeiger, 1987; Parvathi & Raghavendra, 1995). However, partial inhibition with DCMU (Mawson, 1993a) implies a role for guard cell photosynthetic electron transport in ATP supply, suggesting a possible metabolic co-ordination between photophosphorylation and oxidative phosphorylation in guard cells (Mawson, 1993b). H+ pumping results in K+ uptake correlated with malate2− synthesis and/or Cl uptake. Malate2− is the result mostly of starch breakdown (Outlaw & Manchester, 1979), as Arabidopsis mutants that do not accumulate starch lack a proper blue-light response (Lasceve et al., 1997). However, sucrose accumulation from starch breakdown as an additional osmoticum in blue light-stimulated opening in isolated V. faba stomata has also been demonstrated (Fig. 3; Tallman & Zeiger, 1988; Talbott & Zeiger, 1993).

Zeaxanthin (Zeiger & Zhu, 1998; Frechilla et al., 1999; Talbott et al., 2002) and phototropins (Kinoshita et al., 2001; Doi et al., 2004; Inoue et al., 2008) have both been suggested as the blue-light receptor. Support for zeaxanthin as the specific blue-light receptor came from experiments conducted on epidermal peels of Arabidopsis mutants lacking zeaxanthin (non photochemical quenching 1), which failed to respond to blue light (Frechilla et al., 1999), although these results could not be confirmed when experiments were conducted on whole leaves (Eckert & Kaldenhoff, 2000; Kinoshita et al., 2001). Furthermore, stomata in V. faba treated with an inhibitor of carotenoid biosynthesis maintained a blue-light response, ruling out zeaxanthin as the only blue-light receptor (Roelfsema et al., 2006).

Strong evidence for phototropins as blue-light receptors was provided by Kinoshita et al. (2001), who demonstrated (in both epidermal strips and intact plant material) that double mutants for phot1 and phot2 proteins (serine/threonine protein kinase) failed to respond to blue light. They established that phot1 and phot2 act redundantly as the blue-light receptors in stomatal responses to blue light, as single mutants showed a typical wild-type blue-light response (Kinoshita et al., 2001), which led to phototropins becoming widely accepted as the main blue-light receptor.

The magnitude of the blue-light response decreases from morning to afternoon (Doi et al., 2004), consistent with early morning stomatal opening, when light is enriched in the blue wavelengths (Assmann & Shimazaki, 1999), and also consistent with the theory of varying osmoregulatory pathways (Talbott & Zeiger, 1998), and changes in guard cell fluorescence transients through the day (Srivastava et al., 1998). It should be noted, however, that the stomatal response to blue light is not universal, with several species lacking blue light-induced stomatal opening. Stomata of the fern Adiantum capillus-veneris do not open in response to blue light, despite having functional phototropins and plasma membrane H+-ATPase (Doi et al., 2006). Additionally, facultative CAM plants displayed blue-light-specific stomatal opening in C3 but not in CAM mode (Lee & Assmann, 1992; Talbott et al., 1997).

3. PEPc activity, malate synthesis and starch breakdown

An alternative sink for the end products of guard cell photosynthetic electron transport is malic acid production via PEPc, and CO2 fixation (Willmer & Dittrich, 1974; Raschke & Dittrich, 1977; Schnabl et al., 1982; Willmer, 1983; Outlaw, 1990) using carbon skeletons provided by starch breakdown (Pallas & Wright, 1973; see also Asai et al., 2000). Outlaw & Manchester (1979) demonstrated a quantitative relationship between malate accumulation and starch loss. Light-stimulated increases in PEPc activity have been demonstrated together with increased NADP- or NAD-dependent malate dehydrogenase activity, which catalyses the reduction of OAA (Rao & Anderson, 1983; Scheibe et al., 1990), and malate accumulation has been correlated with stomatal aperture (Allaway, 1973; Pearson, 1973; Pearson & Milthorpe, 1974; Vavasseur & Raghavendra, 2005). It is widely accepted that guard cells contain high concentrations of starch and PEP carboxylase (Willmer et al., 1973; Willmer & Rutter, 1977; Raschke, 1977, 1979; Outlaw & Kennedy, 1978) compared with mesophyll cells (Cotelle et al., 1999) and many reports have suggested that this is the major or only pathway for CO2 fixation in guard cells (e.g. Willmer et al., 1973; Reckmann et al., 1990; see also Vavasseur & Raghavendra, 2005) into malate and aspartate (Ogawa et al., 1978). The importance of malate accumulation in light-induced stomatal opening (Asai et al., 2000) has been demonstrated using 3,3-dichloro-dihydroxyphophinoyl-methyl-2-propenoate (DCDP), an inhibitor of PEPc (Parvathi & Raghavendra, 1997). PEPc activity and malate formation have also been linked to CO2 response movements in stomatal guard cells (Outlaw & Lowry, 1977; Raschke, 1979; Schnabl et al., 1982; Hedrich & Marten, 1993; Hedrich et al., 1994; Cousins et al., 2007). Differences in stomatal opening between adaxial and abaxial stomata have been closely associated with differential starch hydrolysis, malate synthesis and K+ uptake (Pemadasa, 1983) as well as light wavelength (Wang et al., 2008), highlighting again the flexibility of stomatal osmoregulation and behaviour depending upon the environment. Further support for the importance of PEPc activity in stomatal opening comes from recent work conducted on PEPc-deficient mutants of the C4 dicot Amaranthus edulis, which showed reduced rates of both stomatal opening and final conductance compared with wild-type controls (Cousins et al., 2007). This recent work is in agreement with earlier studies on potato (Solanum tuberosum) plants which demonstrated greater rates of opening when PEPc was over-expressed and reduced rates of opening in plants with decreased amounts of PEPc (Gehlen et al., 1996). This work is discussed in greater detail in Section V.

Starch degradation to sucrose is also involved in stomatal opening. In dual light experiments, Tallman & Zeiger (1988) found substantial starch degradation under blue-light illumination, but only a small amount of K+ uptake. From these observations they suggested that, if starch was converted to malate (Schnabl, 1980), adequate uptake of K+ would be necessary to counterbalance the anion. These results are consistent with the observation that, in 10 µmol m−2 s−1 blue light, the rate of malate synthesis in V. faba guard cells was only 25% of their maximum (Ogawa et al., 1978). From such studies it was concluded that starch breakdown under blue light can also result in accumulation of sucrose rather than malate (Tallman & Zeiger, 1988).

4. Guard cell photosynthesis and sucrose production via the Calvin cycle

Research into guard cell photosynthesis and carbon metabolism has spanned several decades, but as yet there is no general consensus. There are conflicting reports in the literature concerning the capacity of photosynthetic carbon reduction in guard cell chloroplasts and its importance in stomatal function (see reviews by Shimazaki et al., 1989; Outlaw, 1989).

Early studies provided little evidence of Calvin cycle activity in guard cell chloroplasts (Outlaw et al., 1979, 1982; Outlaw, 1982, 1987, 1989; Tarczynski et al., 1989). Willmer & Dittrich (1974) showed that, in the epidermis of Tulipa and Commelina in the light, 14CO2 was fixed into malate and aspartate. This was validated later by Raschke & Dittrich (1977), who showed that neither radioactive 3-PGA nor Rubisco activity was present in epidermal peels of the same tissues when they were exposed to 14CO2. Subsequent experiments demonstrated that guard cell chloroplasts lacked ribulose-1,5-bisphosphate carboxylase (RuBPC) and ribulose-5-phosphate kinase (Ru5PK) activity (Outlaw et al., 1979) and other key enzymes (Outlaw et al., 1979; Schnabl, 1981) for the photosynthetic carbon reduction pathway (PCRP). This was in agreement with a lack of phosphorylated Calvin cycle products in V. faba guard cell protoplasts in the light (Schnabl, 1980). Screening 41 species using indirect immunofluorescence, Madavhan & Smith (1982) reported no evidence of Rubisco in the guard cells of C4 plants and only negligible detection in C3 plants, but appreciable amounts in one-third of CAM species surveyed. The size of the guard cell phosphoglycerate pool was unaffected by light, suggesting that PCRP is not involved in aperture regulation (Outlaw & Tarczynski, 1984). Reckmann et al. (1990) determined that only 2% of the solute required for stomatal opening was provided by Rubisco activity in Pisum sativum, and concluded that there was insignificant Rubisco activity, confirming the conclusion of Hampp et al. (1982) that photoreduction of CO2 by guard cells was absent.

In contrast to the above findings, the presence of Rubisco in guard cells of V. faba has been unequivocally shown with immunocytochemical localization (Zemel & Gepstein, 1985). Numerous studies have also shown that guard cells contain several of the other main Calvin cycle enzymes (Shimazaki & Zeiger, 1985; Gotow et al., 1988; Shimazaki, 1989). Zemel & Gepstein (1985) quantified Rubisco on a chlorophyll basis at 40–50% compared with mesophyll cells. Shimazaki et al. (1989) validated these figures, showing RuBPC activity in guard cells of the same species to be 40% of that in mesophyll chloroplasts (on a chlorophyll basis). However, they suggested that the low ratio of CO2 fixation to O2 evolution implied that the major proportion of ATP and reducing equivalents was used for reactions other than photosynthetic CO2 fixation. These authors also pointed out that in many previous studies values had been calculated on a cell rather than a chlorophyll basis, and that recalculation would significantly increase previously obtained values in line with their observations.

The results of Gotow et al.'s study (1988) contradicted earlier findings (Raschke & Dittrich, 1977) and showed that feeding radio-labelled CO2 to guard cell protoplasts under red light resulted in incorporation of radioactivity into 3-phosphoglycerate, ribulose bisphosphate (RuBP), fructose and sedoheptulose. Medium alkalinization indicating CO2 uptake and oxygen evolution by guard cell protoplasts was shown under white (Gotow et al., 1988) and red light (Shimazaki & Zeiger, 1987). A photosynthetic dependence of sucrose accumulation was illustrated using DCMU in epidermal peels of V. faba under red light by Poffenroth et al. (1992). However, reduced CO2 concentration triggered K+ uptake rather than sucrose accumulation under red light (Olsen et al., 2002).

It is now widely accepted that the Calvin cycle enzymes are present in guard cell chloroplasts; however, the debate over their activity, function and role in stomatal behaviour remains (Outlaw, 1996, 2003). Although guard cell photosynthetic carbon reduction has been shown in epidermal peels (Tallman & Zeiger, 1988; Poffenroth et al., 1992), guard cell chloroplasts (Shimazaki et al., 1982; Gotow et al., 1988; Shimazaki, 1989) and isolated guard cell pairs (Tarczynski et al., 1989), the contribution to osmotic requirements for stomatal opening ranges from 2% (Reckmann et al., 1990) to 40% (Poffenroth et al., 1992) (see Wu & Assmann, 1993). Many reports have suggested that rates are too low for any functional significance (Outlaw, 1989; Outlaw et al., 1982), whilst others have proposed the Calvin cycle to be a major sink for the products of photosynthetic electron transport (Cardon & Berry, 1992; Zeiger et al., 2002; Lawson et al., 2002, 2003). Evidence for guard cell production of sucrose has been obtained during red light-induced stomatal opening in V. faba, where no starch breakdown was observed and sugar import was ruled out as a result of the use of epidermal peels (Tallman & Zeiger, 1988; Talbott & Zeiger, 1993). Parvathi & Raghavendra (1997) also showed that Calvin cycle activity increased with application of DCDP, an inhibitor of PEPc activity, suggesting that this pathway may become important when PEPc is restricted.

5. Evidence for all four mechanisms in stomatal function

In conclusion, there is evidence for all four of the above mechanisms being involved in stomatal function. The blue-light stomatal response is believed to be mostly independent of guard cell electron transport, as stomatal blue-light responses have been observed in albino leaves (Karlsson et al., 1983; Roelfsema et al., 2006), with energy for the activation of a plasma membrane H+-ATPase supplied mostly by the mitochondria (Fig. 3; Schwartz & Zeiger, 1984; Assmann & Zeiger, 1987; Parvathi & Raghavendra, 1995), although there is also evidence for chloroplastic supply (Mawson, 1993a) and red-light enhancement (Shimazaki et al., 2007). There is support for both zeaxanthin and phototropins as the blue-light receptors, with phototropins the most widely accepted. Evidence exists for the direct use of ATP and/or reductants (produced by guard cell electron transport) in red light-induced stomatal opening (Fig. 3; e.g. Shimazaki et al., 1989; Tominaga et al., 2001; Olsen et al., 2002; Ritte & Raschke, 2003), as well as in sugar production by photosynthetic carbon reduction within the guard cells (Fig. 3; Talbott & Zeiger, 1996, 1998). Starch stored in the guard cells can be broken down into either malate2− (as a counterion for K+ uptake; see Fig. 3) (Willmer & Dittrich, 1974; Raschke & Dittrich, 1977; Outlaw & Manchester, 1979; Schnabl et al., 1982) or sugars, which act as osmotica for stomatal opening (Outlaw, 1982; Tallman & Zeiger, 1988; Poffenroth et al., 1992). It appears that guard cell chloroplasts can be involved in all four of the pathways described above, and that the pathway used is conditional on the species, time of day, and experimental protocols.

IV. Chlorophyll a fluorescence studies to examine guard cell photosynthesis

Chlorophyll a fluorescence is a powerful technique to probe and elucidate photosynthetic metabolism in guard cells. The early pioneering work of Zeiger and co-workers (Zeiger et al., 1980; Melis & Zeiger, 1982; Mawson & Zeiger, 1991) measuring Kautsky kinetics (Kautsky & Franck, 1943) in individual guard cells showed distinct features associated with Calvin cycle activity. The majority of early chlorophyll fluorescence work was restricted to epidermal peels (Ogawa et al., 1982), protoplasts (Outlaw et al., 1981; Goh et al., 1999), single guard cell pairs or the white areas of variegated tissue (Zeiger et al., 1980; Melis & Zeiger, 1982; Cardon & Berry, 1992). Cardon & Berry (1992) examined changes in steady-state chlorophyll fluorescence (F′) from guard cells in the white areas of intact leaves of Tradescantia albiflora under different CO2 and O2 concentrations and attributed changes to photochemical and nonphotchemical quenching. From these observations they concluded that both the carboxylation and oxygenation of RuBP were major sinks for the end products of photosynthetic electron transport. However, caution should be applied when interpreting steady-state fluorescence measurements as it is difficult to distinguish between photochemical and nonphotochemical quenching components (Baker, 2008). The report by Cardon & Berry (1992) was the first research to provide physiological evidence for Rubisco-mediated CO2 fixation and photorespiration and led the way for many subsequent studies.

Advances in fluorescence methodology (see Goh et al., 1999), with the development of the saturation pulse method of fluorescence quenching analysis (Bradbury & Baker, 1981; Schreiber et al., 1986) and pulse amplitude modulation (PAM) fluorimetry (Schreiber et al., 1986) in conjunction with technological developments in microfluorimetry (Goh et al., 1999) and high-resolution imaging (Oxborough & Baker, 1997), made it possible to parametrize measurements of chlorophyll fluorescence at the cellular (Oxborough & Baker, 1997; Goh et al., 1999) and subcellular levels (Baker et al., 2001). With such advancements, measurements of PSII operating efficiency (inline image, where inline image is the difference between maximum fluorescence in the light adapted state (inline image) and steady state fluorescence in the light (F′)) could be obtained for individual cells and protoplasts (Goh et al., 1999). The PSII operating efficiency estimates the efficiency at which light absorbed by PSII is used for the reduction of the plastoquinone QA, and can provide an estimate of the quantum yield of linear electron flux through PSII (Baker, 2008). Goh et al. (1999) first used such techniques to compare fluorescence quenching characteristics in guard and mesophyll cell protoplasts (in V. faba and Arabidopsis). Light induction curves displayed very similar characteristics, indicating similar functional organization of the thylakoid membranes, although guard cells were saturated at lower light intensities and mesophyll cell protoplasts had a higher capacity for photosynthetic electron transport. In the same study, anaerobic conditions suppressed photosynthetic electron flow in guard cells compared with mesophyll cells. The O2-dependent electron flow suggested a role for the Mehler-ascorbate peroxidase (MAP) cycle or a close metabolic coupling between photosynthetic electron transport and export of reducing equivalents via a 3-phosphoglyceric acid/dihydroxyacetone phosphate (PGA/DHAP) shuttle and oxidative phosphorylation in the mitochondria (Goh et al., 1999). However, again this work was restricted to measurements of protoplasts or white areas of variegated plant tissue and was not in agreement with later studies conducted on intact green material (Lawson et al., 2002, 2003).

The first study that simultaneously examined the PSII operating efficiencies (inline image) of guard and mesophyll cells in intact green tissue revealed guard cell photosynthetic efficiency to be 70–80% that of mesophyll chloroplasts (Lawson et al., 2002). However, electron transport rates for the two cell types could not be calculated because of uncertainties in the exact light absorption and the contribution of PSI fluorescence in guard and mesophyll chloroplasts. In the same study these researchers measured inline image at different CO2/O2 concentrations in guard cells of intact green leaves of Tradescantia albiflora and Commelina communis and confirmed that Rubisco was a major sink for the products of photosynthetic electron transport. Later this was confirmed in the guard cells of several other species, including the C4 plant Amaranthus caudatus (Lawson et al., 2003), and was consistent with the results of immunogold labelling studies, which found weak labelling of PEPc but significant Rubisco labelling in guard cells of Amaranthus viridis (Ueno, 2001). The fact that the same CO2/O2 response was observed in guard cells and in mesophyll cells suggests that a major proportion of the end products of electron transport are being used by Rubisco and the Calvin cycle. Guard cells contain 20–50-fold less chlorophyll than the underlying mesophyll (Willmer & Fricker, 1996) and therefore, at similar photosynthetic rates, extrapolation to the whole-cell level would result in much lower guard cell photosynthesis compared with the mesophyll. However, the small volume of guard cells (one-tenth that of the mesophyll) means that the guard cell CO2 assimilation rate could be one-third to one-tenth that of the mesophyll, and therefore guard cell chloroplasts could provide a significant energy source for these cells.

1. Alternative sources of energy

Although the aim of this review is to concentrate on guard cell chloroplasts and their possible role in stomatal function, guard cells also contain numerous mitochondria (Willmer & Fricker, 1996; Vavasseur & Raghavendra, 2005), about one-third the number of those in the mesophyll (Allaway & Setterfield, 1972), and several reports have suggested that these are the most important organelle in guard cells. Hampp et al. (1982) originally proposed that there was an absence of photoreduction of CO2 in guard cells but a high metabolic flux through the catabolic pathway. High respiration rates were observed by Raghavendra & Vani (1989), suggesting that ATP produced through oxidative phosphorylation was important for stomatal movements (Parvathi & Raghavendra, 1997).

Fumarase activity (which is involved in the tricarboxylic acid (TCA) cycle) has been shown to be high in guard cells of V. faba and P. sativum (Hampp et al., 1982; see also Outlaw, 2003), and trangenic tomato (Solanum lycopersicum) plants with considerable reductions in mitochondrial fumarate hydratase (fumarase) activity showed substantial reductions in stomatal aperture, resulting in CO2 limitation of photosynthesis (Nunes-Nesi et al., 2007). Application of inhibitors of photophosphorylation (DCMU) and oxidative phosphorylation (KCN) has shown that both mechanisms are used for light-induced opening, but depend on wavelength (Schwartz & Zeiger, 1984). Their relative importance alters when either of these pathways is restricted (Parvathi & Raghavendra, 1997), suggesting that both organelles (chloroplasts and mitochondria) play a role in stomatal function (Asai et al., 2000).

V. Linking stomatal behaviour to mesophyll photosynthesis

Stomatal conductance is well co-ordinated with mesophyll photosynthetic CO2 fixation (Wong et al., 1979; Farquhar & Wong, 1984; Mansfield et al., 1990). Numerous studies have demonstrated a strong correlation between photosynthesis and stomatal conductance under a variety of different light intensities and nutrient and CO2 concentrations (Radin et al., 1988; Hetherington & Woodward, 2003). This relationship causes, or is a consequence of, a constant Ci:Ca ratio (where Ca is the external CO2 concentration), which has been observed to remain constant over the long term (Wong et al., 1979, 1985), although short-term variations have often been apparent (Sharkey & Raschke, 1981; Morison, 1987). However, it should also be noted that this relationship has easily been broken in transgenic plants with various modifications to photosynthetic metabolism (e.g. Hudson et al., 1992; Lauerer et al., 1993; Stitt & Schulze, 1994; von Caemmerer et al., 2004; Cousins et al., 2007; Baroli et al., 2008). The close relationship between photosynthesis and stomatal behaviour led to the hypothesis that guard cell responses may be linked to mesophyll photosynthetic capacity via a mesophyll signal or that guard cell photosynthesis itself may provide a metabolite signal (Wong et al., 1979). Chloroplast ATP pool size was put forward by Farquhar & Wong (1984) as a possible metabolite, a theory that was built upon later by Buckley et al. (2003), whilst zeaxanthin has been put forward by Zeiger & Zhu (1998) in view of a close correlation between zeaxanthin concentration and stomatal apertures (in response to light and CO2; see Zhu et al., 1998).

The debate over whether guard cell chloroplasts and/or guard cell photosynthesis plays a direct role in the co-ordination of stomatal movements in relation to mesophyll photosynthetic CO2 demand remains unresolved. Recently, Roelfsema et al. (2002, 2006) have argued against a direct role for guard cell chloroplasts in red light-induced stomatal movements. These researchers used albino areas of variegated plant tissue of V. faba treated with norflurazon (nf; inhibits carotenoid biosynthesis), and showed stomatal opening in these two tissue types in response to blue light but not red (Roelfsema et al., 2006). They concluded that a lack of red-light response is consistent with intercellular CO2 concentration as the intermediate signal in the stomatal red-light response. Moreover, stomatal opening in response to red light was only apparent when light was applied to a large area of the leaf, and not when it was applied to individual guard cells, supporting a mesophyll-driven Ci response (Roelfsema et al., 2002).

These obervations are in good agreement with earlier studies by Karlsson (1986), who showed that lowering the atmospheric CO2 concentration had a similar effect to red light and enhanced the blue-light response. It should, however, be mentioned that stomata in the albino portion of the leaf cannot be considered to be completely indicative of responses in green tissue (Scarth & Shaw, 1951; Lawson et al., 2002) as their movements are much slower than those in green areas (Scarth, 1932). However, Scarth (1932) also noted under red light that stomata located near the green tissue tend to open further than those at greater distances, adding support to the theory of an indirect effect of red light on stomatal movements through the action of mesophyll photosynthesis. Further evidence for a CO2-mediated red light-induced stomatal opening response has been provided by the Arabidopsis high temperature 1 (HT1) mutant which carries a mutation in the gene encoding a protein kinase (Hashimoto et al., 2006). These mutants lack both a guard cell CO2 response and a red-light response, but respond to blue light, supporting the notion that red light-driven stomatal opening is promoted by reduced Ci, although to corroborate this it would be necessary to present stomatal conductance against Ci rather than CO2. Additional support for Ci-driven stomatal opening has been provided in Nicotiana tabacum in which a MAP kinase gene (NtMPK4) involved in the activation of anion channels was silenced. These plants did not close in elevated atmospheric CO2 and showed a reduced response to red light (Marten et al., 2008).

A recent publication by Mott et al. (2008) has suggested (as have many other studies; see above) that most stomatal responses to light and CO2 occur in response to an unknown mesophyll-generated signal. Epidermal peels of Tradescantia pallida, V. faba and P. sativum showed no stomatal response to light or CO2, but when T. pallida and P. sativum peels were grafted back onto mesophyll (either their own corresponding mesophyll or that of a different species), stomatal responses were restored, although this was not the case for V. faba. The authors argued against a direct effect of mesophyll-driven changes in Ci, as increasing ambient CO2 from 120 to 540 µmol mol−1 did not induce stomatal closure, whereas darkness resulted in complete closure, but Ci was only 200 µmol mol−1. In agreement with this, a further recent study reported that abaxial stomata of Helianthus annuus were more sensitive to light transmitted through the leaf (self-transmitted light) than to direct illumination, highlighting a photosynthesis-dependent involvement in stomatal responses to light. This was attributed to an unknown photosynthetic metabolite and not a Ci-driven effect, as Ci was maintained at a constant value (Wang et al., 2008).

Several studies have also argued against a direct effect of Ci on stomatal-driven responses to red light; for example, stomata were shown to respond to light even when Ci was held constant (Messinger et al., 2006; Lawson et al., 2008; Wang et al., 2008). Additionally, stomatal responses to Ci and Ci responses to light are believed to be too small to account for the large changes in stomatal conductance that are often observed in response to light (Sharkey & Raschke, 1981b). Furthermore, red-light responses have been documented in several studies conducted in epidermal peels (Tallman & Zeiger, 1988; Olsen et al., 2002) and protoplasts (Raschke & Dittrich, 1977; Goh et al., 1999) isolated from the mesophyll.

Evidence for a direct role of guard cell chloroplasts in red light-induced stomatal opening has been reported, although this may be species dependent. Recently, Doi & Shimazaki (2008) examined stomatal responses in the fern Adiantum capillus-veneris to CO2 in darkness and found the stomata to be unresponsive to low or high CO2 concentrations but to open in response to red light. The fact that they observed a synergistic effect of red and far-red light on stomatal opening, and greater sensitivity when light was applied directly to the lower surface along with a lack of response to Ci, led these authors to conclude that opening in this species is driven by photosynthetic electron transport in guard cell chloroplasts. This is probably driven by K+ uptake, as CsCl (a K+ channel block) inhibited the response (Doi & Shimazaki, 2008). In the same experiment, Arabidopsis was used as a control and showed ‘typical’Ci responses in the dark, highlighting the possibility that different species may use alternative signalling pathways and mechanisms. In contrast to these findings, the stomata of an Arabidopsis mutant that lacks a functional SLOW ANION CHANNEL-ASSOCIATED 1 (SLAC1) gene, which encodes a plasma membrane anion channel, were found to fail to close at a high CO2 concentration in the dark and in the light in one study (Negi et al., 2008), and to open in the light and close more slowly in the dark in another (Vahisalu et al., 2008). These studies support the involvement of ion transport mechanisms in light-dependent stomatal movements, that are not dependent solely on Ci-driven responses.

To address stomatal behaviour in relation to mesophyll photosynthesis, Messinger et al. (2006) suggested that the balance between photosynthetic carbon reduction by Rubisco and electron transport capacity was the key mechanism linking stomatal response to light and CO2 concentration. This work was based on alterations in the amount of ATP and/or zeaxanthin resulting from a change in the balance of guard cell electron transport (and energy states of the thylakoid membrane) in relation to photosynthetic carbon reduction, determined by light and Ci. Variations in the concentration of zeaxanthin in turn would alter guard cell aperture in response to blue light and CO2 (Zeiger & Zhu, 1998; Zhu et al., 1998; Zeiger et al., 2002). As with zeaxanthin, ATP should increase with light and decrease with Ci with increased Calvin cycle activity. Increased ATP in guard cells could be used in the cytosol for proton pumping at the plasmalemma (Tominaga et al., 2001), or alternatively the energy may be utilized to produce sucrose as an osmoticum for stomatal opening through photosynthetic carbon reduction (Talbott & Zeiger, 1993). This work also suggested that there are at least two mechanisms by which stomata respond to CO2, one dependent on photosynthesis, and the other photosynthetically independent (Messinger et al., 2006). Multiple CO2 response mechanisms have previously been suggested by Assmann (1999). However, studies conducted on transgenic plants have demonstrated similar conductances in wild-type and trangenic plants (see next section), despite the latter having an alteration in the balance of electron transport rates relative to carboxylation capacity (von Caemmerer et al., 2004; Baroli et al., 2008).

1. Progress using transgenic and mutant plants

The choice of an ideal experimental system is critical when attempting to determine the role of mesophyll or guard cell photosynthesis in stomatal function and in the past has relied on epidermal peels or guard cell protoplasts. Such systems are often criticized because of possible mesophyll contamination (Weyers & Travis, 1981; Outlaw, 1983). However, the removal of the mesophyll could prevent any mesophyll signalling and may induce other mechanistic responses (Lee & Bowling, 1995; Lawson et al., 2002). In the late 1970s, Outlaw and co-workers introduced the technique of dissecting individual cells (Outlaw, 1980; Hampp & Outlaw, 1987; Outlaw & Zhang, 2001), highlighting its advantage in controlling contamination. Improved molecular and transgenic techniques have provided modern powerful tools (Webb & Baker, 2002) with which to address many questions regarding photosynthesis in relation to stomatal function and have already provided some invaluable information (see above).

Early studies conducted on transgenic plants with impaired photosynthesis revealed some surprising results. The effects of Rubisco concentration on photosynthesis were studied independently by several groups, all of which studies suggested little effect on stomatal behaviour (Quick et al., 1991; Stitt et al., 1991; Hudson et al., 1992). Specifically, these studies showed similar stomatal conductance values in transgenic plants compared with wild-type controls, despite a severe reduction in photosynthesis and higher Ci concentrations. Studies on tobacco (Nicotiana tabacum) plants with reduced phosphoribulokinase (Paul et al., 1995) or Rieske FeS protein (Price et al., 1998) also showed no effect on stomatal conductance, and Price et al. (1998) concluded that ‘stomata are not strongly reliant on photosynthetic electron transport for setting conductance’. However, work on transgenic antisense PEPc plants supported a role for malate and PEPc activity in guard cells, with delays in stomatal opening responses in potato with reduced PEPc activity, whilst over-expressors showed accelerated opening (Gehlen et al., 1996). These findings are supported by recent work on Amaranthus edulis mutants deficient in PEPc, which show both reduced rates of opening and also reduced final stomatal conductances (Cousins et al., 2007). Stomata in plants with 12% wild-type fructose-1, 6-bisphatase (FBPase) activity showed significantly faster opening responses and higher final conductances with increasing irradiance, despite lower photosynthetic rates and elevated Ci. However, this was dependent upon humidity and external CO2 concentration (Muschak et al., 1999).

However, the aims of the above studies were not specifically to determine stomatal responses. Certain assumptions were made: firstly, because Cauliflower mosaic virus (CAMv) promotors were used in the majority of the studies, it was assumed that all cells were antisensed in a similar manner, and, secondly, it was assumed that the observed response was equivalent to steady-state stomatal conditions. To directly address the influence of reduced photosynthetic capacity using antisense technology, von Caemmerer et al. (2004) used high-resolution chlorophyll fluorescence imaging to show for the first time that photosynthetic efficiency was reduced to a similar extent in the guard cells as in the mesophyll cells in tobacco plants with reduced concentrations of Rubisco. Decreasing Rubisco activity resulted in an imbalance between chloroplast electron transport and the photosynthetic carbon reduction capacity, which in turn could lead to an increased amount of ATP and/or conversion of xanthophyll pigments to zeaxanthin. Increased nonphotochemical quenching was observed in the antisense plants, which could be interpreted as an increase in the amount of zeaxanthin. Both ATP and zeaxanthin have been implicated as playing a key role in stomatal opening responses. However, a step change in irradiance revealed similar stomatal responses in terms of opening rates and final conductances in antisense Rubisco plants and wild-type controls, despite significantly lower photosynthetic rates in the former. This led to elevated internal CO2 concentrations within these plants, which initially was interpreted as a reduced sensitivity to Ci (von Caemmerer et al., 2004). However, manipulation of Ci through changes in Ca resulted in stomatal closure, suggesting that stomata may response to Ca and not Ci (von Caemmerer et al., 2004). The overall conclusion from this work was that neither mesophyll nor guard cell photosynthesis was necessary for stomatal opening responses.

The fact that the majority of studies used white light or a mixture of red and blue light does not rule out the possibility that blue light-stimulated opening, independent of photosynthesis, overrides any mesophyll/guard cell signal (Talbott & Zeiger, 1993). To resolve this issue, a recent study of red-light responses by Baroli et al. (2008) distinguished between antisense Rubisco tobacco plants with 10–15% wild-type Rubisco activity, which have major reductions in the carboxylation capacity of photosynthesis, and antisense tobacco plants with impaired rates of electron transport via reductions in the cytochrome b6f complex. No changes in stomatal opening were observed in either of the transgenic plants in response to a step change in red light at ambient CO2 concentrations, leading to the conclusion that this response was independent of guard or mesophyll cell photosynthesis. The fact that no phenotypic stomatal responses were observed despite a decrease in sucrose concentration also strongly suggests that something other than sucrose acts as osmoregulator during opening. However, recent work conducted on antisense sedopheptulose-1,7-bisphosphatase (SBPase) tobacco plants has shown a minor regulatory role for photosynthetic electron transport in response to red light (Lawson et al., 2008). A step change in red illumination resulted in an increased rate of stomatal opening which was not observed under a blue/red light mix. ATP concentrations in the antisense SBPase plants may be increased because of the reduced ATP consumption by the Calvin cycle. These authors suggested the possibility of increased stomatal opening under red light, as a result of an increase in ATP available for proton pumping (Tominaga et al., 2001). However, Baroli et al. (2008) reported little effect of reduced ATP on red light-induced stomatal opening in transgenic tobacco plants with reduced cytochrome b6f complex (Baroli et al., 2008).

Other suggested functions of chloroplasts in guard cells in regulation of stomatal behaviour include the production of reactive oxygen species such as H2O2 (possibly via Mehler activity at PSI) which may play a role in abscisic acid (ABA) signal transduction (Zhang et al., 2001). A role for ascorbic acid (Asc) redox state has also been postulated in guard cell regulation. Plants with increased guard cell Asc redox state (through increased expression of dehydroascorbate reductase (DHAR)) exhibited a reduced concentration of guard cell H2O2 and consequently higher stomatal conductances (Chen & Gallie, 2004).

As highlighted above, the use of transgenic and mutant plants has provided significant information regarding guard cell mechanisms (von Caemmerer et al., 2004; Baroli et al., 2008; Lawson et al., 2008), sensory molecules (e.g. Eckert & Kaldenhoff, 2000) and signal transduction cascades (Inoue et al., 2008), as well as ion uptake and ion channel regulation (Serna, 2008), all of which play key roles in stomatal sensitivity and behaviour.

VI. Stomata in relation to water use/manipulation of behaviour

One of the important outcomes of understanding how guard cells function is the potential to engineer drought-tolerant plants. This prospect has received increasing attention from the wider scientific community, with several reports published recently suggesting that stomatal metabolism may hold the key (Nilson & Assmann, 2007). For example, maize (Zea mays) plants with increased amounts of NADP-malic enzyme (ME), which converts malate and NADP to pyruvate, NADPH and CO2, had altered stomatal behaviour. ME-transformed plants had decreased stomatal conductance, showing signs of drought avoidance associated with guard cell malate metabolism. A negative aspect of this drought-tolerance engineering was that, following exposure to drought, the development of necrosis was more rapid in leaves from plants with the highest ME expression (Laporte et al., 2002). Masle et al. (2005) reported the isolation of a ‘transpiration efficiency gene’, ERECTA, which acts on cell expansion and cell division, amongst other processes, resulting in modification of leaf diffusive properties and mesophyll capacity for photosynthesis, leading to greater water use efficiency, in Arabidopsis. Increases in drought resistance have also been reported in Arabidopsis mutants, with alterations or disruptions of guard cell membrane transporters (Klein et al., 2004), calcium-dependent protein kinases (Ma & Wu, 2007), and the expression of aquaporin genes (Cui et al., 2008) and genes involved in ABA biosynthesis, expression or sensitivity (Jakab et al., 2005; Wang et al., 2005; Yang et al., 2005). Such studies are not restricted to Arabidopsis; for example, over-expression of the stress-responsive gene SNAC1 (STRESS-RESPONSIVE NAC1) enhanced drought tolerance in rice (Oryza sativa) (Hu et al., 2006).

In the attempt to produce plants with increased water use efficiency or drought tolerance, genetic engineering or mutations provide an opportunity to alter not only stomatal physiology and function (Nilson & Assmann, 2007) but also anatomical features, such as stomatal density and size (originally proposed in the 1970s; Jones, 1976, 1977) and amounts of leaf cuticular wax (Aharoni et al., 2004). A recent review by Wang et al. (2007) highlights the importance of, and recent progress made in, identifying genes controlling stomatal density or patterning, and how such genetic manipulations may increase plant water use efficiency.

Altering the stomatal density does not automatically alter stomatal conductance (Fig. 4; Lawson, 1997; Weyers & Lawson, 1997; Lawson & Morison, 2004). Figure 4 shows a model of predicted stomatal conductance with changes in various stomatal characters. From this model it is obvious that stomatal aperture has the greatest control over stomatal conductance, with stomatal density being secondary. An example of this can be found in experiments conducted on Arabidopsis over-expressing the STOMATAL DENSITY AND DISTRIBUTION 1 (SDD1) gene, resulting in plants with a 40% reduction in stomatal density, and the Arabidopsis sdd1-1 mutant (Berger & Altmann, 2000), in which stomatal density is increased to 300% of that of wild type. Under growth conditions, no differences in stomatal conductance or assimilation rate were observed in the over-expressers and the sdd1-1 mutants compared with wild type. Lower stomatal density was compensated for by an increase in aperture and, conversely, reduced stomatal aperture compensated for increased stomatal density (Bussis et al., 2006). It should be mentioned that, although mutants may be identified as ‘drought resistant’ or with ‘increased water use efficiency’, such traits may not be evident when they are grown in competitive environments. Basco et al. (2008) recently reported that Arabidopsis ABA oversensitive mutants, which display enhanced stomatal closure, could not compete with wild type for water when the plants were grown together. Such findings also have significant implications for screening protocols when attempting to identify mutants (Basco et al., 2008). It is also important to note that screening plants for increased water use efficiency should include measurements of photosynthetic performance in relation to stomatal behaviour, as reduced stomatal conductance can decrease water use but also limit photosynthetic carbon assimilation.

Figure 4.

Predicted sensitivity of stomatal conductance (gs) to changes in pore dimension and frequency within empirically derived ranges. Effects on gs of adjusting each anatomical character within its estimated range were calculated following equations of Lawson (1997), Weyers & Lawson, (1997) and Lawson & Morison (2004). The analysis uses typical ranges of values derived from observations of Phaseolus vulgaris (Lawson, 1997): stomatal aperture, 0–15 µm; stomatal density, 35–65 mm−2; pore length, 33–40 µm; pore depth, 15–25 µm. Values within each range were used to calculate stomatal conductance using the following equation: 1/rs = (d + 2c)/(Dw × As × SD), where rs is stomatal resistance, d is pore depth (mm), c is an end correction (see Weyers & Meidner, 1990), Dw is water diffusivity in air (mm2 s−1), As is pore area (mm2), and SD is stomatal density (mm−2). The vertical lines represent the gs obtained using the median values for each variable and was calculated at 346 mmol m−2 s−1.

In conjunction with advances in molecular biology, substantial progress has been made in technology and methodology. The use of thermal imagery (Jones, 1999, 2004; Jones et al., 2002) in combination with chla fluorescence (Chaerle et al., 2007) has the potential to determine instantaneous water use efficiency, and is not only a potential screening tool allowing determination of both photosynthetic performance and stomatal behaviour but also a powerful approach to elucidating correlations between stomatal behaviour and photosynthetic capacity. An example of combined chlorophyll fluorescence imaging and thermography is shown in Fig. 5. Extreme treatments, in which stomata in one area of the leaf were blocked with grease to prevent stomatal conductance and a vein was severed to prevent uptake of DCMU, were used to show the relationship between PSII operating efficiency and stomatal conductance. Images of inline image showing spatial and temporal resolution of PSII operating efficiency were compared with thermal images of leaf temperature, which is modulated by stomatal behaviour and other environmental factors (i.e. in general, the greater the stomatal conductance the greater the evaporative cooling of the leaf and the lower the leaf temperature). From the images it is apparent where stomatal behaviour is influencing PSII operating efficiency and vice versa.

Figure 5.

Chlorophyll fluorescence (a, b) and thermal (c, d) images of a sycamore leaf fed with 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) through the transpiration stream. An area of stomata was blocked on one half of the leaf by applying a patch of grease, and a major vein was severed on the other half. The patch increased leaf temperature (c), and reduced the quantum efficiency of photosystem II (PSII) photochemistry (inline image, where inline image is the difference between maximum fluorescence in the light adapted state (inline image) and steady state fluorescence in the light (F′) (a). After DCMU feeding (b, d), inline image was reduced and leaf temperature was increased. However, DCMU was not distributed where the vein was severed so inline image remained high and the leaf temperature was lower. Under the patch, there was little transpiration and DCMU uptake, and therefore inline image remained high even though the CO2 supply was limited, indicating that photorespiration was the sink for the products of electron transport Scale bars represent 20 mm (unpublished data of T. Lawson, J. I. L. Morison and N. R. Baker).

VII. Concluding remarks and future direction

Stomatal research over the past few decades has revealed a complicated network of osmoregulatory and signalling pathways in guard cells (e.g. Li et al., 2006). It appears that these highly plastic cells have the capability to alter mechanisms of response depending upon environmental growth and experimental conditions, complicated further by time of day and pretreatments (Zeiger et al., 2002), all of which appear to be species dependent. Such flexibility gives stomata the necessary capability to maintain a regulatory role in plant water status and photosynthetic capacity. This review has concentrated on guard cell chloroplast photosynthesis and in particular Calvin cycle function, a highly controversial topic (Outlaw, 1989, 2003), with evidence for and against functional guard cell photosynthetic regulation of stomatal behaviour. Recent research conducted on antisense SBPase plants suggests guard cell photosynthesis and/or carbon reduction may play a role in stomatal responses to red light (Lawson et al., 2008). However, at the same time, work on antisense Rubisco and b6f plants casts doubt on any role for guard cell photosynthesis, including the production of ATP, in red light-induced opening (e.g von Caemmerer et al., 2004; Baroli et al., 2008). Discrepancies in results and conclusions regarding the role of guard cell chloroplasts in stomatal function are probably attributable to the unique plasticity of guard cells, which can make interpretations difficult, with often opposing conclusions in different laboratories in which research was conducted under different conditions (see Zeiger et al., 2002). Stomatal research in the future should therefore take into account the time of day experiments are conducted, the conditions under which the plants are grown and the type of material used, as all of these factors can impact on stomatal responses, signalling pathways, and solutes required for osmoregulation of stomatal aperture.

The current transition towards using mutants and transgenic plants along with the identification of gene trap lines (Galbiati et al., 2008) opens a new window of opportunity to pursue different avenues of research to answer some of the many questions that still remain regarding guard cell metabolism. To date, attention has focused on photosynthetic pathways in guard and mesophyll cells, and to a certain extent the oxidative phosphorylation pathway has been neglected. Transgenic plants and mutants provide an ideal opportunity to determine the role of this pathway in stomatal sensory and response mechanisms. The development and discovery of guard cell specific promoters (see Yang et al., 2008) will allow manipulation of guard cell metabolism without disruption of mesophyll photosynthetic metabolism. Such systems will hopefully provide a probe that will help to fully elucidate the link between mesophyll photosynthesis and stomatal conductance. Microarray and proteomic technology allows gene expression patterns involved in signal transduction pathways to be identified and assessed under different environmental conditions and stresses (Coupe et al., 2006). Leonhardt et al. (2004) have demonstrated the power of microarray technology comparing the expression profiles of guard and mesophyll cells. They noted that, when leaves were sprayed with ABA, there was repression of many of the enzymes involved in guard cell metabolism, including a decrease in PEPc transcript, which agrees with earlier work reporting decreased PEPc activity under drought (Kopka et al., 1997). Transcriptomic analysis can also identify transcription factors that are necessary for stomatal movement mediating stomatal responses to light and darkness (Gray, 2005; see review by Casson & Gray, 2008).

To date, most stomatal research has concentrated on plant species very familiar to stomatal biologists, but there are still numerous gaps in our knowledge regarding stomatal behaviour in CAM and grass species. C3/CAM intermediates may provide an ideal opportunity to uncover light and CO2 responses as well as induction of specific genes or signalling pathways. Stomata respond to numerous environmental stimuli, yet most studies are conducted in isolation. There is a desperate need to determine the hierarchy of stomatal responses, and the influence of combined factors on stomatal behaviour, response and signalling mechanisms.

Although significant advances in the understanding of guard cell function and stomatal responses have been made over the last century, many gaps in our knowledge remain regarding guard cell metabolism and its role in stomatal behaviour. The use of antisense techniques in conjunction with guard cell-specific promoters, and modifications of guard cell chloroplast metabolism, coupled with in situ measurements of photosynthetic performance, stomatal function and responses to various stimuli, may provide the key to ascertaining the roles of these highly conserved organelles.


I would like to thank several colleagues for their contributions, ideas and discussion over the course of my past and current research. In particular Dr James I. L. Morison and Professor Neil R. Baker are acknowledged for the opportunity to work in their laboratory imaging chlorophyll fluorescence in guard cells. My initial interest in the stomatal regulation of gas exchange was inspired during work for a PhD with Dr Jonathan Weyers (Dundee University), at which time I was encouraged to carry out my first research on this subject. Professors Christine Raines and Susanne von Caemmerer have provided many useful discussions. Dr Tanja Hofmann is gratefully acknowledged for critical reading of the manuscript. I would also like to thank three anonymous reviewers for their comments, which have greatly enhanced the manuscript. Work on chlorophyll fluorescence imaging in guard cells was funded by a BBSRC grant awarded to Dr James I. L. Morison and Professor Neil R. Baker, whilst recent work on transgenic plants was funded by the University of Essex.



II.Osmoregulation in guard cells16
III.Role of guard cell chloroplasts in stomatal function18
IV.Chlorophyll a fluorescence studies to examine guard cell photosynthesis22
V.Linking stomatal behaviour to mesophyll photosynthesis23
VI.Stomata in relation to water use/manipulation of behaviour26
VII.Concluding remarks and future direction27