Leaf hydraulic conductance is coordinated with leaf morpho-anatomical traits and nitrogen status in the genus Oryza

Highlight The range of leaf hydraulic conductance across the genus Oryza is caused by leaf morpho-anatomical traits and leaf N status.


Introduction
Leaf hydraulics is the major bottleneck of the overall plant hydraulic system, and therefore the fundamental factor restricting gas exchange and biomass production (Sack et al., 2003;Sack and Holbrook, 2006). The efficiency of water transport through the leaf to the evaporating surface of the mesophyll is quantified by leaf hydraulic conductance (K leaf ), which is generally expressed on a leaf area base (Sack and Holbrook, 2006). Decreases in K leaf usually cause leaves to become less hydrated (corresponding to a low leaf water potential), a response often associated with stomatal closure and, consequently, reduced CO 2 assimilation (Sperry, 2000;Johnson et al., 2009). This reduction occurs partly because CO 2 and water exchange between leaves and air share a common pathway through stomatal pores. The coupling of stomatal conductance (g s ) to CO 2 and water vapour leads to strong coordination between g s and K leaf (Sack et al., 2003;Brodribb et al., 2005;Sack and Holbrook, 2006). Furthermore, owing to the tight coupling between g s and photosynthetic rate (A) in C 3 species (Wong et al., 1979), a positive relationship between K leaf and A is reported (Brodribb et al., 2005;Franks, 2006;Brodribb et al., 2007;Flexas et al., 2013b).
Previous studies have found that K leaf varies greatly between species, ranging 65-fold from the lowest to highest value (Sack and Holbrook, 2006). Interspecific variation in K leaf reflects differences in the morpho-anatomy of leaves, as well as pathways through the outside xylem to evaporation sites. In plants, leaf vein systems, as distinct water transport systems, vary greatly in arrangement, density, vascular bundle features, and xylem conduits within the bundles . In the past two decades, increasing numbers of studies have focused on the relationship between K leaf and venation architecture, expressed as vein length per area (VLA). Positive and negative relationships between K leaf and VLA have been reported (Nardini et al., 2012;Sack and Scoffoni, 2012;), although no relationship was found in another study (Flexas et al., 2013b). Most of these studies, however, were conducted with woody species, and very few focused on cereal crops such as rice.
Outside the xylem, there are three main pathways for water flow: apoplastic, symplastic, and transcellular (Sack and Holbrook, 2006). Several early studies suggested that water exits the xylem mainly through the apoplastic pathway, because of high resistance in the symplastic and transcellular pathways. However, many recent studies have shown that aquaporins have a positive effect on water transport across the membranes of bundle sheath and mesophyll cells (Martre et al., 2002;Sack et al., 2004). Dye and cell pressure probe experiments also suggest that the symplastic and transcellular pathways play a vital role in water transport in plants (Murphy and Smith, 1998;North et al., 2013). Furthermore, the distance that water travels from veins to stomata (D s ), which is usually expressed as the distance between veins and stomata (D m ) in the leaf cross section (Brodribb et al., 2007;North et al., 2013), has been suggested to be an important trait affecting K leaf . Although water movement in mesophyll tissues is now widely recognized, how leaf mesophyll architecture contributes to water flux in the mesophyll and water evaporation at the cell wall surface remains unclear (Sack and Holbrook, 2006;Flexas et al., 2013b;North et al., 2013).
N is a vitally important element for plants, and it profoundly influences leaf anatomical and functional traits (Rademacher and Nelson, 2001;Lee et al., 2011). Previous studies have shown that leaf N promotes A by increasing Rubisco content and CO 2 diffusion conductance (Imai et al., 2008;Franks et al., 2009). However, the correlation of leaf N content per leaf area with K leaf remains to be investigated. Studying the interactions between leaf N status and K leaf may help determine the effects of N on rice leaf morpho-anatomical traits associated with K leaf and CO 2 movement in leaves.
Oryza spp. are distributed worldwide, and they exhibit a wide range of phenotypes. This diversity is an important resource (Giuliani et al., 2013) that is being utilized to improve rice yield and other agronomic traits, particularly in unfavourable environments. In the present study, four cultivated and seven wild cultivars in the genus Oryza were investigated with the aims of: (i) identifying the variation in leaf morpho-anatomical traits and K leaf ; (ii) investigating whether leaf morphoanatomical traits and leaf N status influence K leaf ; and (iii) determining the relationship between K leaf and gas exchange.

Plant materials
Four rice (O. sativa L.) cultivars and seven wild cultivars in the genus Oryza (Table 1; these were provided by the National Key Laboratory of Crop Genetic Improvement) were investigated in a pot experiment at Huazhong Agricultural University, Wuhan, China. Three hills of seedlings were grown in 15.0 l pots filled with 13.0 kg soil. N, P, and K were applied as basal fertilizers at a rate of 3.0 g, 1.95 g and 1.95 g per pot, respectively. There were three pots per cultivar. Values are mean ± SD; ***, P < 0.001.
Throughout their growth, plants were well watered and a water depth of at least 2 cm was maintained. Pests were controlled using chemical pesticides.

Gas exchange measurements
To avoid the effect of fluctuation in outdoor air temperature, light intensity, and humidity on gas exchange measurement, measurement was done between 9.30 and 15.30 in an environmentally controlled room with an air temperature of 27.8 ± 2.1°C, a photosynthetic photon flux density (PPFD) at the leaf surface of 1200 ± 47 μmol m -2 s -1 (artificial light source), and relative humidity of 77.4 ± 5.3%. Measurements were taken on newly and fully expanded leaves of three plants for each cultivar after they were acclimated for ~1.5 h. Gas exchange and chlorophyll fluorescence were simultaneously measured using an LI-6400XT portable photosynthesis system equipped with a leaf chamber (LI-COR, NE, USA). Leaf temperature during measurements was maintained at 28°C. In the leaf chamber, PPFD was maintained at 1500 μmol m -2 s -1 , and leaf-toair vapour pressure deficit at 1.1-1.4 kPa; CO 2 concentration was adjusted to 400 μmol m -2 s -1 with a CO 2 mixture. After equilibration to a steady state, A, g s , steady-state fluorescence (F s ), and maximum fluorescence (F m ʹ) were recorded. The actual photochemical efficiency of photosystem II (Φ PSII ) was calculated as follows: Electron transport rate (J) was calculated as follows: where α is the leaf absorptance and β represents the distribution of electrons between PSI and PSII.
Light response curves were determined under low O 2 concentration (<2%) for estimating α and β. The gas exchange system was immediately switched to low O 2 concentration (<2%) without removing the leaves from the chamber. Simultaneous measurements of light response curves and chlorophyll fluorescence were then performed. During the measurements, chamber conditions were the same as those described above, except that PPFD was controlled across a series: 2000,1200,800,400,250,200,150,100,50,20, and 0 µmol m -2 s -1 . After reaching a steady state, parameters of gas exchange and chlorophyll fluorescence were simultaneously recorded. The values of αβ and daytime respiration rate (R d ) were calculated as the slope and intercept, respectively, of the linear regression of A on PPFD·Φ PSII /4 (Yin et al., 2009). Our αβ values are consistent with the values estimated from the slope between Φ PSII and Φ CO2 with varying light intensity under non-photorespiratory conditions (O 2 < 1%) (Supplementary Figure S1).
The variable J method described in Harley et al. (1992) was used to calculate mesophyll conductance of CO 2 (g m ) and CO 2 concentration in the chloroplast (C c ). C c was calculated as follows: where Γ* represents the CO 2 compensation point in the absence of respiration. Г* is related to Rubisco-specific factor, which is relatively conserved under given temperature conditions (Bernacchi et al., 2002;Warren and Dreyer, 2006). In the present study, a Г* value of 40 μmol mol -1 , typical for Oryza plants, was taken based on the studies of Franks et al. (2009) and Giuliani et al. (2013). Then, g m was calculated as follows: where C i represents the intercellular CO 2 concentration.
Leaf hydraulic conductance K leaf was measured using the evaporative flux method (Sack et al., 2002;Brodribb et al., 2007;Guyot et al., 2012;Sack and Scoffoni, 2012). Three to nine leaves of each cultivar were excised in water and placed under conditions favourable to transpiration (i.e. PPFD of 1200 µmol m −2 s −1 and air temperature of 28°C) with the petiole attached to a potometer. When leaves reached a transpirational steady state, the transpirational flux rate (E) was recorded. The leaf area was then measured using a leaf area meter (LI-Cor 3000C, LI-COR, NE, USA) and leaf length and width were measured quickly using a plastic ruler. The leaves were detached and cut into small sections, immediately followed by leaf water potential (Ψ leaf ) measurement using a WP4C Dewpoint PotentiaMeter (Decagon, Pullman, WA, USA). K leaf was calculated as follows:

VLA and leaf thickness
Three leaves per cultivar were cleared in 20% aqueous NaOH after their widths were recorded. Three sections of leaf lamina of ~5.0 mm length were excised from the middle portion of each leaf, stained, and mounted in glycerol for the determination of vein density. According to Scarpella et al. (2003) and Smillie et al. (2012), rice vascular bundles can be categorized into three types based on their size: midrib, large veins, and minor veins. In the present study, the numbers of major veins (sum of midrib and large veins) and minor veins, and the inter-vein distance (IVD, distance between two minor veins), were recorded using a microscope at 40× magnification. The proportion of minor vein length was calculated as the percentage of minor vein length per area (VLA minor ) over VLA. The leaf thickness (T leaf ) was measured using a DTG03 digital thickness gauge (Digital Micrometers Ltd, Sheffield, UK).

Leaf N content per leaf area
After Ψ leaf measurement, leaves were oven-dried at 80°C to constant weight, and ground using a mixer mill homogenizer (MM400, Retsch, Germany). Approximately 5.0 mg was used to measure N content per leaf area using an NC analyzer (IsoPrime100 IRMS, Isoprime Ltd, UK).

Statistical analyses
One-way analysis of variance (ANOVA) and multiple regression analysis were applied to assess the significance of cultivar effect with SAS 9.2 (SAS Institute Inc., USA). Regression analyses between parameters were performed using SigmaPlot 12 (SPSS Inc., Chicago, IL, USA). All regressions were fitted by both linear and power models, and the model with higher regression coefficient was selected.

Differences in leaf morpho-anatomical traits and K leaf across cultivars
There were very large variations in leaf morpho-anatomical traits in the genus Oryza (Table 1). The differences were 6.9fold in leaf area (ranging from a minimum of 18.4 cm 2 in Rhi to a maximum of 127.3 cm 2 in Lat), 4.6-fold in leaf length, and 5.8-fold in leaf width. With respect to leaf veins, VLA, VLA major , and VLA minor were significantly different across cultivars. There was a 2.3-fold difference in VLA (minimum in Rhi and maximum in Wcr). The difference in leaf mass per area (LMA) was 1.8-fold (minimum in Rhi and maximum in Lat), and the difference in K leaf was 3.7-fold (minimum in Rhi and maximum in Lat).

Relationships among leaf morpho-anatomic traits, leaf N, and K leaf
Across all cultivars, K leaf was positively correlated with leaf area (r = 0.80, P < 0.01), leaf length (r = 0.62, P < 0.05), and leaf width (r = 0.66, P < 0.05) (Fig. 1). No significant correlation was observed between K leaf and VLA, VLA major , or VLA minor . However, a positive correlation (r = 0.86, P < 0.01) between the proportion of minor vein length and K leaf was observed (Fig. 2). In addition, K leaf was positively correlated with LMA (r = 0.83, P < 0.01), IVD (r = 0.92, P < 0.01), T leaf (r = 0.67, P < 0.05) (Fig. 3), and leaf N content per leaf area (r = 0.86, P < 0.01) (Fig. 4). IVD and T leaf were positively correlated with leaf N content per leaf area, while VLA was independent of leaf N content per leaf area (Fig. 5). In order to identify the direct effects of leaf N content per leaf area on K leaf , a multiple regression analysis was performed between K leaf and leaf N content per leaf area, T leaf , and IVD. Our results show that K leaf tightly correlated with N content per leaf area (P = 0.015) compared with T leaf (P = 0.673) and IVD (P = 0.052).

Relationship between K leaf and gas exchange
There were very large variations in A, g s , and g m in the genus Oryza (Supplementary Figure S2). The g m estimated by a combination of gas-exchange and chlorophyll fluorescence methods showed a linear relationship with the value estimated from the A-C i curve-fitting method (Supplementary Figure S1). Across all cultivars, a positive correlation (r = 0.63, P < 0.05) was found between A and K leaf (Table 2; Supplementary Figure S2). A was closely related to total CO 2 diffusion conductance (g t ) (r = 0.85, P < 0.01), g s (r = 0.86, P < 0.01) and g m (r = 0.73, P < 0.01). The g t was positively correlated with both g s (r = 0.92, P < 0.001) and g m (r = 0.92, P < 0.001). There was a strong relationship between g s and g m . K leaf was positively correlated with g t (r = 0.88, P < 0.01), g s (r = 0.75, P < 0.01), and g m (r = 0.77, P < 0.01).

Relationship between A and K leaf
Improving photosynthesis is central to improving crop yield. In C 3 plants, an important determinant of photosynthesis is the CO 2 concentration in the chloroplast. (Evans and Von Caemmerer, 1996;Flexas et al., 2008;Franks et al., 2009;Flexas et al., 2013a). Previous studies have shown correlations between A and K leaf across a wide range of species (Brodribb et al., 2007;Flexas et al., 2013b). In the present study, A was correlated with K leaf in the genus Oryza (Table 2). During photosynthesis, CO 2 must move from outside the leaf through the stoma to the sub-stomatal internal cavities, and from there to the site of carboxylation inside the chloroplast though leaf mesophyll (Evans et al., 2009;von Caemmerer and Evans, 2010;Flexas et al., 2012). Opening the stomata would benefit photosynthesis in the presence of sufficiently high intercellular CO 2 concentration. However, maintaining open stomata depends on leaf water supply capacity, which is determined by K leaf . Under normal conditions, K leaf is limited by leaf anatomy (Sack et al., 2003;Sack and Holbrook, 2006).

Relationship between K leaf and leaf morpho-anatomical traits
Across a large variation in leaf area, we observed a positive correlation between K leaf and leaf area (Fig. 1), as was also observed in Acer and Quercus spp. (Nardini et al., 2012). However, our results were contrary to those of Simonin et al. (2012), who showed, by summarizing published data, that K leaf was independent of variations in leaf area. There are two reasons for the discrepancy between our results and those of Simonin et al. (2012). Firstly, our results were derived from the genus Oryza, which has a homologous hydraulic architecture, and the relatively expanded (leaf area and leaf thickness increasing) leaf needs to evolve stronger water transportation ability, because vein xylem conductivity tends to increase with leaf size. However, the result reported by Simonin et al. (2012) was derived from a wide range of plant species with a multiplicity of leaf hydraulic architectures, masking the effects of leaf area and leaf thickness on K leaf . Secondly, the large variation in leaf area in the present study was contributed by Lat (Table 1; Fig. 1), which caused a significant correlation between K leaf and leaf area. In the present study, a strong positive correlation was observed between K leaf and LMA (Fig. 3). If LMA is considered as the sum of the mass of different leaf tissues per unit of leaf area, variation in LMA occurs via changes in leaf tissue composition. Blonder et al. (2011), on the basis of a mathematic model, hypothesized that high VLA results in high LMA. However,  contested this by compiling a large database, reporting that, in fact, vein xylem and sclerenchyma accounted for <10% of leaf volume per area and thus did not contribute strongly and directly to either leaf thickness or leaf density . In the present study with the genus Oryza, no relationship between VLA and LMA was observed (Supplementary Figure S3). Additionally, especially within species, LMA correlates with T leaf , which is derived from layers of mesophyll cells. Our result indicates that the variation in LMA resulted from changing proportions of mesophyll tissue rather than from changes in VLA in monocots.
There are conflicting reports on the relationship between K leaf and VLA (Scoffoni et al., 2011;Carins Murphy et al., 2012;Flexas et al., 2013b). In the present study, we found that K leaf was not correlated with VLA, VLA major , or VLA minor . However, K leaf significantly increased with an increasing proportion of minor vein length in the genus Oryza (Fig. 2). In monocots, the water in major veins, as in minor veins, exits into the surrounding tissue, instead of into minor veins. Minor veins have a large surface area for exchange of xylem water with the surrounding mesophyll, and a short distance through which water travels outside the xylem (Sack and Holbrook, 2006). These results suggest that K leaf in the genus  Oryza may be driven by the cross-sectional conductivity of veins and outside xylem conductance (McKown et al., 2010;Sommerville et al., 2012).
After leaving the xylem, water must pass through liquid and gas phases before it reaches the sub-stomatal cavities.
The water must first move through the bundle sheath, which is made up of parenchymatous cells wrapped around the veins, to mesophyll cells, and then diffuse into the intercellular airspace; or directly diffuse to intercellular airspace. Finally, the water escapes into the atmosphere via stomatal pores. The distance travelled by the water within leaves has been quantified in several ways (Brodribb et al., 2007;Noblin et al., 2008;North et al., 2013), such as by measuring D m and IVD. Brodribb et al. (2007) reported that K leaf had a strongly negative relationship with D m (in monocots IVD = 0.5D m ) across species with a wide range of habitats and leaf structures. Furthermore, the relationships between K leaf and IVD depend on water travel pathways and the water vapour concentration gradient between the intercellular airspace and atmosphere. In fact, the water in leaves turns into water vapour at mesophyll cell walls exposed to intercellular air space (Sack and Holbrook, 2006). Thus, if the liquid water supplement in leaves is not a limiting factor, an increase in K leaf may occur via an increase in the mesophyll cell wall area exposed to the intercellular airspace. Indeed, Nardini et al. (2012) reported that K leaf was enhanced by an increase in mesophyll porosity (the fraction of leaf mesophyll volume occupied by intercellular air space) under high irradiance.
The value of mesophyll porosity is relatively stable within the genus Oryza (Giuliani et al., 2013). In other words, the volume of intercellular air space per leaf area depends on the proportion of mesophyll tissue in leaves. In rice, it has been shown that the proportion of mesophyll tissue in leaves is related to IVD (Smillie et al., 2012) and T leaf (Sack et al., 2003). Early studies hypothesized that K leaf in thick leaves should decline with increasing pathway length outside the xylem. However, experimental results show that K leaf correlates with T leaf across species, and across sun and shade leaves within a given species (Sack et al., 2003;Zhang and Cao, 2009). This is because thicker leaves have more parallel flow pathways outside the xylem. Here we demonstrated that increases in IVD and T leaf benefit K leaf in the genus Oryza (Fig. 3).

Effects of leaf N status on K leaf
N significantly influences rice leaf anatomy, structure, and function (Lee et al., 2011). In the present study, leaf N content per leaf area had a significant positive effect on K leaf (Fig. 4). Increased IVD and T leaf under high N supplementation (data not shown) facilitates water evaporation at the cell wall surface, and this response could be one of the reasons   1.00*** 0.92*** 0.94*** g s 1.00*** 0.73** g m 1.00*** a *, P < 0.05; **, P < 0.01; ***, P < 0.001.
why K leaf increased with increasing N content per leaf area in leaves. Moreover, water flux across bundle sheath and mesophyll cells travels through either apoplastic, or cell-tocell pathways, or both (Sack and Holbrook, 2006). In the cell-to-cell pathway, water molecules diffuse either across the plasma membrane or through plasmodesmata. Water channels, plasma membrane-intrinsic aquaporins (PIPs), play an important role in this process (Maggio and Joly, 1995;Pou et al., 2013). Several studies have shown that PIP expression varies with N supply (Clarkson et al., 2000;Guo et al., 2007). It is reasonable to speculate that in rice PIPs are regulated by leaf N content per leaf area.

Relationship between K leaf and g m
Inside leaves, K leaf and g m are two traits which play central roles in determining gas exchange and plant performance (Sack and Holbrook, 2006;Flexas et al., 2013b). However, very few studies have focused on their coordination; rather, the two traits have been studied independently in the past two decades. Recently, by summarizing the published data, Flexas et al. (2013b) reported that K leaf was correlated with g m . In the present study, we found coordination of K leaf and g m in the genus Oryza, which provides further evidence that water and CO 2 diffusion in the leaf share common pathways (Table 6). Many studies have found that g m correlates with certain leaf structural traits in some species, particularly with the mesophyll cell surface area exposed to intercellular airspace per leaf area (S m ) (Flexas et al., 2008;Evans et al., 2009;Flexas et al., 2012). This correlation occurs because increasing S m provides more pathways in parallel for CO 2 diffusion. In fact, the mesophyll surface exposed to the intercellular airspace is the site at which water changes from liquid to vapour via evaporation. Cell wall thickness has been recognized as another important limiting factor for CO 2 diffusion in the leaf. Interestingly, thick mesophyll cell walls may increase the extra-xylem apoplast path length, thereby increase K leaf . Further, membrane PIPs are known to facilitate transmembrane water transport as well as CO 2 transport. For instance, Otto et al. (2010) reported a trade-off between water and CO 2 permeability through membranes, depending on the proportion of PIP1 and PIP2 present. Similar responses of g m and K leaf to various environmental factors, including temperature, light, leaf N status ( Fig. 4;  Supplementary Figure S4), and leaf water status, provide another line of evidence for their relationships (Flexas et al., 2013b). However, the relative effects of leaf structural traits on g m and K leaf , and the coordinated dynamics of g m and K leaf under various environmental conditions, needs to be clarified in the future.
In conclusion, there were significantly positive relationships between K leaf and LMA, leaf area, proportion of minor vein length, IVD, T leaf , and leaf N content per leaf area in the genus Oryza, but K leaf was independent of VLA. High K leaf was associated with high A, g s , and g m . Our results indicate that leaf morpho-anatomical traits and leaf N content per leaf area had significant effects on K leaf , and suggest that more detailed anatomical and structural studies are needed to elucidate the impacts of leaf feature traits on K leaf and gas exchange in grasses.

Supplementary material
Supplementary data can be found at JXB online.
Supplementary Figure S1. Relationship between αβ values obtained using Yin's method (Yin et al., 2009) and the Φ PSII and Φ CO2 slope method; and between g m values estimated from a combination method with gas-exchange and Chl fluorescence, and the A-Ci curve-fitting method (b).
Supplementary Figure S2. Relationship between K leaf and A, g t , g s , and g m .
Supplementary Figure S3. Relationship betweenVLA and LMA in the genus Oryza.
Supplementary Figure S4. Relationship between leaf N concentration and both g s and g m .