• The observation that stomatal density (number mm−2) on herbarium leaves had decreased over the last century represents clear evidence that plants have responded to anthropogenic increases in CO2 concentration. The mechanism of the response has proved elusive but here it is shown that density responses to both CO2 concentration and humidity are correlated with changes in whole-plant transpiration and leaf abscisic acid (ABA) concentration.
• The transpiration rate of a range of accessions of Arabidopsis thaliana was manipulated by changing CO2 concentration, humidity and by exogenous application of ABA. Stomatal density increased with transpiration and leaf ABA concentration.
• A common property of signal transduction systems is that they rapidly lose their ability to respond to the co-associated stimulus. Pathways of water movement within the plant are connected and so variations in supply and demand can be signalled throughout the plant directly, modifying stomatal aperture of mature leaves and stomatal density of developing leaves. Furthermore, the system identified here does not conform to the loss of ability to respond.
• A putative mechanism is proposed for the control of stomatal density by transpiration rate and leaf ABA concentration.
The exchange of CO2 and water vapour between a leaf and the atmosphere is principally controlled by stomatal density (the number of stomatal pores per unit area) and their mean aperture. Globally this exchange is massive at c. 440 × 1015 g CO2 yr−1 by photosynthesis and 32 × 1018 g H2O yr−1 by transpiration (Hetherington & Woodward, 2003). The observation that stomatal density (numbers per unit area) on herbarium leaves had decreased over the last century (Woodward, 1987) represents clear evidence that plants have responded to anthropogenic increases in CO2 concentration. This contemporary increase in atmospheric CO2 concentration may stimulate photosynthetic uptake of CO2 (Woodward & Lomas, 2004). By contrast water loss from vegetation may be reduced through direct impacts of CO2 concentration on stomatal density and aperture (Hetherington & Woodward, 2003). Global runoff of water from vegetated lands (i.e. the difference between precipitation and evapotranspiration) has increased by 6% over the twentieth century, despite a decreasing trend in precipitation (Gedney et al., 2006). This is thought to be caused by a reduction in plant transpiration through the direct effect on stomata of increasing CO2 concentrations. These globally significant responses to CO2 will be caused by changes in both stomatal aperture and density (Hetherington & Woodward, 2003).
The mechanism by which CO2 controls aperture is the subject of very active research (Young et al., 2006). However, the mechanism by which CO2 concentration controls stomatal development is not known and there is limited evidence to indicate the signalling systems involved. The discovery of the HIC gene, in Arabidopsis thaliana, that disrupts the signal transduction pathway for the CO2 response of stomatal development (Gray et al., 2000) indicated genetic control, perhaps through changes in the permeability of the extracellular matrix of the stomatal guard cells. Further evidence for developmental control by signalling emerges from experiments demonstrating that the CO2 concentration around mature leaves determines the stomatal density of developing leaves (Lake et al., 2001; Miyazawa et al., 2006). However, any mechanism must account for the wide range of observed responses to CO2 by different species (i.e. increases or decreases in stomatal density, or alternatively no change in stomatal density) (Hetherington & Woodward, 2003). Arabidopsis shows promise for investigating the mechanism as different accessions of this single species show as wide a response to CO2 concentration as different species (Woodward et al., 2002). The control of stomatal density is the target response here as any changes in density may have direct impacts on water loss from plants, leaves and vegetation (Hetherington & Woodward, 2003).
After the completion of leaf expansion stomatal density remains constant, while aperture varies in the order of minutes in response to environmental conditions. The stomatal density of the developing leaf is influenced in some way by signalling from mature leaves (Lake et al., 2001; Miyazawa et al., 2006). This could conceivably occur by a change in whole-plant transpiration that in turn alters guard cell turgor pressure (Mott & Franks, 2001) and the mean stomatal aperture of the plant. The water status of the developing leaf would also need to change, at least locally, to the production of stomatal lineages, where the signal for the developing leaves will be a change in the rates of water demand and supply. The rate of transpiration would therefore determine the stomatal density of the developing leaf. Transpiration rate may be regulated by abscisic acid (ABA), a hormone known to reduce transpiration by inducing stomatal closure (Zhang et al., 1987) ABA is synthesized in different plant organs, is mobile in leaf tissue and can be rapidly translocated throughout the plant via the xylem stream, permitting function as a long-distance signal (Heilmeier et al., 2007). Furthermore, a glucose ester of ABA (ABA-GE) has recently been proposed as a potential long-distance signal (Jiang & Hartung, 2008). A role for ABA in the determination of stomatal numbers is unknown.
The aim of the study was to investigate the environmental control of stomatal density with the hypothesis that plant transpiration rate controls the stomatal density of developing leaves. The model plant Arabidopsis was used in which plant transpiration rate was manipulated by changes in humidity, carbon dioxide concentration and by the application of exogenous ABA.
Materials and Methods
Plant material and growth conditions
A range of Arabidopsis thaliana (L.) Heynh. accessions previously established as exhibiting decreases and increases of stomatal density under elevated CO2 were selected (Colombia, Col-0; C24; Cape Verde Islands, Cvi; Kashmir, Kas-1; Rschew, Rld-2; and Landsberg erecta, Ler) together with a range of mutants disrupted in hormonal signal pathways (abi1, aba1, aux1, ein2; see Table 1) supplied by NASC, Nottingham, UK. Plants were grown in controlled environment chambers (Conviron BDR16, Winnipeg, Canada; Fitotron SG2352/FM/HFL, Sanyo Gallenkamp, Loughborough, UK) under an 8 h photoperiod, with a 22°C : 18°C day : night temperature regime and an irradiance of 150 ± 20 µmol m−2 s−1. The impact of humidity on stomatal development was achieved at two humidities, 45% and 85 ± 5%. Carbon dioxide concentrations were 400 ppmv, 800 ppmv and 3000 ppmv for the CO2 enrichment treatments. Plants were sown, stratified for 2 d at 4°C and germinated in either ambient (400 ppmv) or elevated (800 ppmv, Col-0 only at 3000 ppmv) CO2-controlled environment growth cabinets and at the treatment relative humidity. Seedlings were transplanted at the four-leaf stage into 8-cm pots containing Arthur Bowyers multipurpose compost (Williamsons, Sheffield, UK). Plants were watered regularly to maintain moist soil conditions. An additional experiment, using Col-0 alone, was also run with the same environmental conditions, with relative humidity at 65%.
Table 1. Arabidopsis accessions used in abscisic acid (ABA) experiments and their characteristics
Stomatal density (SD) was determined by making high-precision surface activated dental putty (Coltene Whaledent, Altstätten, Switzerland) impressions of both leaf surfaces. Cellulose varnish was then applied to the dental impression, mounted on slides and viewed through a light microscope (Olympus CX40RF200). Three rosette leaf insertions, 11–13 (mature leaves – analysed for comparison of effect of applying ABA – see later), and insertions 17–19 (newly developed leaves) of five replicate plants per treatment were analysed. All data presented correspond to insertions 17–19 (new leaves).
Abscisic acid treatment
Abscisic acid (50 µm) (racemic ± mix, Sigma, UK) was applied daily to leaf insertions 6–13 for a period of 14 d with a soft brush, or bottom fed to roots with 10 ml ABA (50 µm) each day. Controls were either untreated or given a deionized water application. Stomatal measurements were made on three rosette leaf insertions 11–13 (mature leaves) and insertions 17–19 (newly developed leaves) of five replicate plants per treatment. The ABA concentration was measured by radioimmunoassay (RIA) after Quarry et al (1988), modified by Taylor et al. (1996) and J. Thoebold (pers comm). The ABA extraction method of the 3000 ppm CO2 experiment differed from that of Quarry et al. (1988), following that of Overy et al. (2005).
Whole-plant transpiration was measured according to the method of Jones (1999) using a calibrated infrared thermometer (Model 68; Fluke, Everett, WA, USA) with a c. 3 mm diameter sensing area from a distance of 15 cm and a preset emissivity value of 0.95, providing a mean temperature of all visible leaves to a resolution of 0.1°C. Wet and dry leaf replicas were placed adjacent to the measured plants and were used for calculating leaf boundary layer conductance.
Reduced major axis (RMA) regression was calculated according to the routine in Sokal & Rohlf (2001) coded in Microsoft Excel.
Results and Discussion
Stomatal density is the controller of gaseous exchange and as such is the primary measurement in our analyses. Stomatal index (SI = SD/SD = ED; the ratio typically used to indicate changes in guard cell production relative to epidermal cell production – ED) was not measured here as SI changes when either SD or ED remain constant and ED or SD change. This does not necessarily indicate changes in the ratio of divisions that produce stomata and epidermal cells. Guard cell size is relatively unchanging but epidermal cell size can change significantly after division, so that the initial guard cell to epidermal cell ratio (just after division) is lost as leaf development progresses.
Stomatal density responses to CO2 doubling were quantified (Fig. 1) in controlled environment conditions for five wild-type accessions of Arabidopsis (Columbia, Col-0; C24; Cvi; Kas-1, and Rld-2). Stomatal density significantly increased with CO2 concentration in C24, did not change in Cvi and decreased for Col-0, Kas-1 and Rld-2 (95% confidence limits), providing the range of responses seen across species (Woodward et al., 2002).
The impacts of changing humidity and ABA application were assessed for Col-0, to provide further information on the stomatal density responses to the imposed environmental conditions. Stomatal density increased with transpiration rate (Fig. 2, Table 2), irrespective of whether changes in humidity, CO2 concentration or ABA application took place. It is interesting to note that growth of Col-0 at a very high CO2 concentration of 3000 ppmv increased both the transpiration rate and stomatal density (Fig. 2, no. 11). Monteith (1995) postulated that transpiration rate drives a change in stomatal aperture and the relationships shown here suggest that transpiration and leaf ABA concentration also drive a change in stomatal density of leaves developed under the three treatments (CO2, ABA and relative humidity (RH)).
Arabidopsis hormonal mutants abi1, ein2 and aux1 (Table 1) in the Col-0 background were also tested to determine the impact of changes in disrupted hormonal (auxin, ethylene and ABA signalling) activity on the stomatal density response to transpiration rate (Fig. 3, Table 3). A comparison of slopes (Figs 2 and 3) shows no change in the response from that observed with Col-0 (Fig. 2). The three treatments exert no significantly different effects on stomatal density of the Col-0 wild type and the mutants. However, there is a unifying influence on the relationships between transpiration and stomatal density (Fig. 2), where stomatal density increases as the transpiration rate increases, extending also to accessions that show increases in stomatal density with CO2 (Figs 1 and 3).
The ABA-insensitive mutant (abi1) shows a reduction in stomatal density as the rate of transpiration decreases, with increasing humidity (Fig. 3, no. 2, Table 3). However, the stomatal density response is less than that observed by the wild-type Col-0 (Fig. 2, no. 2).
To investigate further a mechanistic link between transpiration rate and stomatal density, another mutant with the potential to influence the transpiration/stomatal link, the ABA-deficient aba1 in the Ler background (Table 1), was analysed for changes in density under the same treatments (ABA, CO2 and RH; Fig. 4, Table 4). The Ler mutant includes a mutation in the erecta gene that has been shown to have a major effect on stomatal development and water-use efficiency (Shpak et al., 2005; Masle et al., 2005). However, the Ler mutant responds in the same manner as Col-0 (Ler closed symbols, solid line, Fig. 4; Col-0 dashed line, Fig. 4– taken from Fig. 2), suggesting that the erecta mutation does not form part of the stomatal density response to CO2. The ABA-deficient mutant, aba1, behaves differently from the parental accession Ler (Fig. 4) when CO2 and RH are increased (open symbols). This mutant has a reduced control of stomatal aperture regulation (a ‘wilty’ mutant; Koornneef et al., 1982), which results in a disruption to the stomatal density response to transpiration. The application of exogenous ABA to mature leaves and roots of the deficient mutant restores the ability to regulate stomatal aperture and also restores the response of stomatal density (Fig. 4, nos 11 and 13) to be the same as Ler (Fig. 4, no. 2). These results support the hypothesis that the stomatal density response to increases in CO2 and RH are driven by the regulation of transpiration and stomatal aperture, and again implicates ABA in the overall response. This also indicates that signalling from mature to developing leaves can be achieved by changing the transpiration rate of mature leaves, alone.
The mechanism by which transpiration rate and stomatal aperture change the stomatal density of developing leaves is not known. However, the observations presented here indicate a necessary connection between whole-plant transpiration and stomatal aperture and the sites of stomatal and epidermal cell development in new leaves. Therefore, the direction of lineages from early protodermal cells to either stomatal guard cells or epidermal cells (Geisler et al., 2000) will change in response to the whole-plant balance of water supply and demand.
There is evidence that all stomata of the plant respond in a similar manner to environmental changes. For example, changes in stomatal conductance, in the short term of hours and in one part of a leaf (Mott & Franks, 2001) or part of a plant canopy (Whitehead et al., 1996) can control stomatal conductance in another part. This is seen when one part of a leaf or plant is shaded, with the stomata responding in the same manner in the unperturbed region (Mott & Franks, 2001; Whitehead et al., 1996). Stomatal changes on these short time-scales will be caused entirely by changes in stomatal aperture and in response to changes in water supply and demand. Such changes can be very local to stomatal guard cells, in the order of hundreds of µm and are probably caused by changes in epidermal turgor without necessarily changes in bulk leaf water potential (Cochard et al., 2002). Evidence for connectivity between transpiration rate and changes in stomatal density is provided by Luomala et al. (2005) who found no change in stomatal density in Scots pine needles at elevated CO2, but that an increase in temperature resulted in a decrease of both stomatal density and vascular bundle size. Reduced vascular bundle size would result in reduced transpiration rate. This is in agreement with responses of Arabidopsis results shown here (Figs 2, 3 and 4).
Fig. 5 (Table 5) shows the response of changes in stomatal number to leaf endogenous ABA concentration, with a positive relationship in the parental accessions Col-0 and Ler (closed symbols), including Col-0 at very high CO2 (square symbol). This relationship is absent in the aba1 mutant (open symbols) unless ABA is added directly to either leaves or roots (Fig. 5, nos. 11 and 13). Furthermore, the responses of aba1 at low humidity (45%; Figs 4 and 5, nos 14 and 17, dashed circle) show a greater decrease in stomatal density than the responses at high humidity (85% RH, Figs 4 and 5, nos 15 and 16), illustrating an overriding effect of humidity on transpiration rate despite the application of ABA to leaves or roots.
The response of stomatal aperture to CO2 concentration is greater at higher humidities (Talbott et al., 2003) and the responses of stomatal density to humidity and CO2 concentration also predict that at high humidities the stomatal density response to CO2 should be greater than at low humidities. At an RH of 45% stomatal density of Col-0 decreased by 13% with CO2 doubling (Fig. 2) but at an RH of 65% the stomatal density response to CO2 enrichment was greater, decreasing by 32% (data not shown). These observations provide further evidence for a connection between changes in transpiration rate and stomatal density.
The central nature of guard cells in the CO2 response of stomatal development also follows from work on the hic (high in carbon dioxide) mutant of Arabidopsis (Gray et al., 2000). The HIC gene is only expressed in guard cells and encodes a negative regulator of the response of stomatal density to CO2. Work on the HIC gene is one of the few cases where an environmental signal is connected with the programme of stomatal development.
Recent work (Wang et al., 2007) proposes that environmentally sensitive mitogen-activated protein kinases (MAPKs) may act downstream of ERECTA receptor-like protein kinases, in the pathway for stomatal development (Bergmann et al., 2004; Shpak et al., 2005), to modify quantitatively the responses of stomatal numbers to the environment. Evidence that MAPK pathways are activated by drought and low humidity (Ichimura et al., 2000) indicates a potential method by which changes in water status in the developing leaf could directly control stomatal differentiation. A putative mechanism (Fig. 6) describes the relationships between transpiration rate, leaf ABA concentration and stomatal density, indicating how these two factors might influence downstream signals to effect stomatal development in response to changes in CO2 and humidity. A possible connection between ABA and environmentally sensitive MAPKs is unknown, but would allow continuous feedback in monitoring and responding to environment. Feedback can also occur through the stomatal density of mature leaves, altering transpiration rate (in conjunction with environmental conditions) to produce a change in stomatal density of new leaves. As transpiration rate is an inherent function, this would allow continued monitoring without incurring a loss of sensitivity.
A number of consequences emerge from these new observations. Negative to positive responses of stomatal density to CO2 enrichment between accessions and species can be accommodated, when environmental conditions are otherwise constant. If humidity in controlled environments differs between treatments then this will influence the response of stomatal density to CO2 concentration. An example would be the case when ambient humidity of a controlled environment is strongly influenced by plant transpiration. Enrichment of CO2 reduces stomatal aperture and if the transpiration rate also decreases then the humidity could decrease. For Col-0 a 20% reduction in relative humidity decreases the CO2 response of stomatal density by 19%. However, the idea of a CO2 response of stomatal density originated from observations that stomatal density increases with altitude, in parallel with decreases in CO2 concentration (Woodward, 1986). This altitudinal response of stomatal density is not universal (Woodward, 1986), indicating that other environmental conditions beyond CO2 concentration are the most critical determinants of stomatal density and where plant transpiration is most critical (Johnson et al., 2005). When precipitation increases with altitude and radiation decreases, reducing the gradient for transpiration, then stomatal density could decrease with altitude. Such scenarios allow for both increases and decreases in the stomatal density response to CO2 found in both controlled environments and in the field.
It is suggested that a common property of signal transduction systems is that they rapidly lose their ability to respond to a given stimulus (Hoa et al., 2007). Pathways of water movement within the plant are connected and so variations in supply and demand can be continuously signalled throughout the plant directly modifying stomatal aperture, and therefore transpiration rate, of mature leaves and stomatal density of developing leaves. We propose that a primary driver of stomatal density responses is the ABA regulation of stomatal aperture and as ABA may be rapidly metabolized in leaves (Jiang & Hartung, 2008), the system we identify need not conform to the loss of ability to respond. The next new challenge is to elucidate the interactions involved in the proposed mechanism.
We thank the Royal Society for funding J.A.L., C. Bennett for technical assistance, J. Thoebold for RIA. W. J. Davies, J. E. Gray, A. M. Hetherington and J. C. McElwain for comments on an early version of the manuscript and D. Cameron for comments on this manuscript.