Gas exchange and hydraulics during drought in crops: who drives whom?

This article comments on: Wang X, Du T, Huang J, Peng S, Xiong D. 2018. Leaf hydraulic vulnerability triggers the decline in stomatal and mesophyll conductance during drought in rice (Oryza sativa). Journal of Experimental Botany 69, 4033–4045.

Drought stress is one of the largest threats to crop productivity and survival worldwide (Boyer, 1982;Ciais et al., 2005), hence the importance of unveiling the relationships between the different physiological mechanisms and traits that confer resistance in plants (McDowell et al., 2013). Water stress causes the decrease in leaf water potential (Ψ leaf ), which in turn causes the activation of turgor-related signals  and/ or hormonal signals. Abscisic acid (ABA) is considered the main plant hormone involved in the water stress response, although there is still debate as to whether the fraction of the total hormone pool involved in signalling is synthesized mostly in the roots (Dodd, 2005) or in the same leaf (McAdam et al., 2016). These hydraulic and non-hydraulic factors regulate stomatal but apparently also mesophyll conductances to control both transpiration (i.e. reduce hydraulic tension in the atmosphere-plant-soil continuum) and CO 2 supply for the optimization of gas exchange (Nadal and Flexas, 2018). These signals are coupled with the supply capacity of the hydraulic system, otherwise extreme water loss and/or hydraulic failure could lead to complete desiccation of the plant (Sperry, 2004;Hochberg et al., 2017). However, this general scheme of drought response may vary between plants depending on the degree of iso-or anisohydry (Martínez-Vilalta and Garcia-Forner, 2017). Signals induced by Ψ leaf also regulate leaf hydraulic conductance (K leaf ) (Coupel-Ledru et al., 2017), in tight coordination with gas exchange (Brodribb et al., 2014;Gleason et al., 2017). Decreases of K leaf are generally associated with hydraulic failures, such as embolism, but also with other forms of regulation (Hochberg et al., 2017). However, the relative importance and mechanisms of regulation of its componentsthe conductance within the xylem (K x ) and the outside-xylem conductance (K ox ) -during drought remain unresolved (Trifiló et al., 2016). If the drought worsens, the physiological effects on the leaves are incrementally increased, which may lead to the death of the leaf (e.g. full hydraulic failure, or 100% embolism; Martin-StPaul et al., 2017), and the whole plant may depend on the existence of safety margins among plant organs (Liu et al., 2015;Skelton et al., 2017;Rodriguez-Dominguez et al., 2018). Although the main processes that occur during drought are clear, knowledge of the general timescale of response and the importance of each parameter is limited because most studies do not monitor the same variables simultaneously, and few consider so many parameters during a prolonged drought as do Wang et al. (2018). So what do we really know about these inter-relationships and why is the work by Wang et al. important? Variability in the physiological responses of crops to drought stress There are very few interspecific studies on limitations to photosynthesis under drought, thus precluding broad generalizations. For instance, although a pattern has been suggested in which diffusion conductances limit photosynthesis under mild and moderate stress, while biochemical limitations appear only at the later stages (reviewed in Nadal and Flexas, 2018), some studies have found differences among species, especially regarding the relative importance of stomatal and mesophyll limitations (Galmés et al., 2007;Flexas et al., 2009;Galle et al., 2011) but also concerning the early appearance of biochemical limitations (Ennahli and Earl, 2005). Similarly, while it seems that a general coordination among both conductances occurs during drought, recent studies suggest that the nature of the relationship may be species-specific. In this sense, Flexas et al. (2013a) showed that the relationship between g s and g m varies across crops under well-watered and water-stressed conditions: although most of them show a tight coordination between these two conductances, some (e.g. poplar) did not show such relationship. Box 1. Limitations to net assimilation in relation to the vulnerability of its constraints (g s , g m , biochemistry and K leaf ) in different crops Response of limitations to photosynthesis -stomatal (SL), mesophyll conductance (ML) and biochemical (BL) limitations -to decreasing leaf water potentials (Ψ leaf ) in Oryza sativa (Wang et al., 2018), Olea europaea (data combined from Perez-Martin et al., 2009, andVarone et al., 2012) and Vitis vinifera (from El Aou-ouad et al., 2016). K leaf P 50 and P 80 are represented by red dashed and solid lines (data from Wang et al., 2018, for rice, and data combined from Torres-Ruiz et al., 2015, for O. europaea, and from Martorell et al., 2015. Yellow points in O. sativa represent the P 50 of g s , g m and electron transport rate (ETR) (each of them situated over the upper line of its limitation -SL, ML or BL, respectively -data from Wang et al., 2018). The blue dotted line represents the turgor loss point (data from Wang et al., 2018, for rice, and value from Hernandez-Santana et al., 2016, for O. europaea and from Martorell et al., 2015. The orange dotted line accounts for either K x P 50 in O. sativa (value from Stiller et al., 2003) or the Ψ leaf in which approximately 50% embolism occurs in the leaf midrib (based on optical measurements; data from Rodriguez-Dominguez et al., 2018, for O. europaea andfrom Hochberg et al., 2017 for V. vinifera).

Box 2. Interrelationships between stomatal and hydraulic conductance in different crops
The graph shows the relationships between stomatal (g s ) and leaf hydraulic (K leaf ) conductances and the magnitudes of each for the same crop species considered in Box 1: Oryza sativa (mean data from Wang et al., 2018), Olea europaea (data combined from Fernandes-Silva et al., 2016;Hernandez-Santana et al., 2016) and Vitis vinifera (data combined from Pou et al., 2012Pou et al., , 2013El Aou-ouad et al., 2017). Lines represent quadratic polynomial fittings for each species and shaded areas are their 95% confidence intervals.
In two rice cultivars, Wang et al. show that there is strong coordination between K leaf , g s and g m during their decrease under drought. Indeed, a similar sequence of events can also be observed for olive when combining data from several studies (Box 1), although olive seems to operate along a wider range of Ψ leaf . On the other hand, this early decline in all three conductances is not observed in grapevine, where the decline of K leaf (P 50 ) occurs at the latest stages of water stress, after a previous progressive and strong decrease in photosynthesis, mainly due to limitation by stomatal conductance. The three examples displayed in Box 1 suggest different possibilities regarding limitations to photosynthesis and coordination of conductances across species.
The species-dependent coordination between stomatal and K leaf responses to drought could indicate different strategies regarding water conservation and safety of transport (see Box 2). As shown by Wang et al., rice presents a tight coordination between K leaf and g s ; in fact, the decrease of g s is mainly attributed to K leaf . This has also been shown in woody crops Rodriguez-Dominguez et al., 2016). On the other hand, no such coordination has been observed in soybean (Locke and Ort, 2014). On a broader phylogenetic scale, clearer differences emerge; for example, g s presents a higher sensitivity to Ψ leaf in ferns compared to coexisting angiosperms (Brodribb and Holbrook, 2004). In ferns, stomata closed before any significant drop in K leaf , whereas in the angiosperms studied there was a tighter coordination between g s and K leaf . This was also observed when studying the different responses of g s and K leaf not to drought but to varying light intensity . Indeed, the differences in P 50 for g s and K leaf may be more related to phylogeny than to ambient conditions as no common pattern in P 50 was observed in co-occurring tree species (Liu et al., 2015). In the case of the drought-induced g m -K leaf relationship, significant variability has been reported even at the clone level (Théroux- Rancourt et al., 2015). Some degree of plasticity in these relationships has also been seen in grapevines, where K leaf presented a decreasing P 80 as summer progressed (Martorell et al., 2015). Moreover, even the mechanistic basis for the decline in K leaf (i.e. the relative importance of K ox and K x ) may be species-dependent (Trifiló et al., 2016). All these examples of interspecific variation hinder disentanglement of the factors limiting photosynthesis and transpiration under water stress.

Box 3. Variables and hypothetical relationships controlling physiological drought response in crops
Diagram showing the potential interrelations between water potential and leaf conductances under mild to moderate drought stress conditions. Solid lines indicate positive relationships between variables, whereas broken lines indicate negative relationships. The dotted broken line indicates the hydraulic disconnection between leaf and stem due to embolism. Left diagram (a) follows the hypothesis of the safety valve function of stomata to prevent hydraulic failure (K leaf mostly constituted by leaf xylem conductance, K x ). In this scenario, stomatal and mesophyll conductance (g s and g m , respectively) are reduced to keep K leaf within the safety margin to avoid hydraulic disconnection from the stem. Right diagram (b) reflects a hypothesis that can be derived from the suggestion by Wang et al. of outside-xylem conductance controlling K leaf , which in turn triggers the decline of both g s and g m . In this case, cavitation would be of little magnitude because K leaf would be governed mainly by K ox . Notice the double-arrowed blue line linking g m and K leaf in both diagrams; this accounts for the coordinated nature of these two conductances (Flexas et al., 2013b), which could emerge from a common structural basis (Xiong et al., 2017), rather than by one being directly affected by the other.
Role of, and relationships among, water conductances during drought: universal or species-specific?
Many theories have considered the stomata as the safety valves preventing hydraulic dysfunction under mild to moderate water stress conditions (Hochberg et al., 2017 and references therein), considering leaf xylem hydraulic vulnerability as the main component of leaf hydraulic vulnerability. However, results from Wang et al. challenge these theories. The fact that the K leaf P 50 was achieved before the g s and g m P 50 s suggests that, in rice, the stomata do not function as a safety valve and therefore either: (i) if K leaf =K x , leaf xylem cavitated before stomata closed; or (ii) if K leaf =K ox , outsidexylem hydraulic vulnerability protected against xylem failure instead of stomata (see Box 3 for a depiction of these two possibilities). The first hypothesis is unlikely as the xylem vulnerability P 50 reported by Stiller et al. (2003) is about -2.0 MPa. On the other hand, although Wang et al. measured K leaf without distinguishing K x from K ox , the second hypothesis may be more likely: indeed, Trifiló et al. (2016) and Scoffoni et al. (2017) showed that outside-xylem hydraulic vulnerability explains 75 to 100% of K leaf decline before reaching the turgor loss point in most of the species studied. However, this hypothesis cannot be considered confirmed yet, at least for all vascular plants, as measurements performed using new techniques (such as the leaf optical vulnerability; Brodribb et al., 2016) that allow the simultaneous measurement of K x and g s (Hochberg et al., 2017) provide new evidence supporting the hypothesis that K leaf is mainly driven by K x . Nonetheless, these two hypotheses are not necessarily irreconcilable; in fact, they may represent species-or even genotype-specific strategies for plants coping with water stress along the iso-anisohydric spectrum (Tombesi et al. 2014;Coupel-Ledru et al., 2017).
In summary, until methodological limitations are improved, and more experiments are carried out monitoring the multiple interrelated variables that act during drought for multiple species, a very interesting debate where (at least) two major hypotheses are possible will continue. The work by Wang et al. (2018) adds important new data and ideas to this debate.