Divergent leaf and fine root “pressure–volume relationships” across habitats with varying water availability

Abstract

Fine roots and leaves, the direct interfaces of plants with their external environment along the soil–plant–atmosphere continuum, are at the front line to ensure plant adaptation to their growing habitat. This study aimed to compare the vulnerability to water deficit of fine roots and leaves of woody species from karst and mangrove forests—two water-stressed habitats—against that of timber and ornamental woody species grown in a well-watered common garden. Thus, pressure–volume curves in both organs of 37 species (about 12 species from each habitat) were constructed. Fine roots wilted at a less negative water potential than leaves in 32 species and before branch xylem lost 50% of its hydraulic conductivity in the 17 species with available data on branch xylem embolism resistance. Thus, turgor loss in fine roots can act as a hydraulic fuse mechanism against water stress. Mangroves had higher leaf resistance against wilting and lower leaf-specific area than the karst and common garden plants. Their fine roots had high specific root lengths (SRL) and high capacitance to buffer water stress. Karst species had high leaf bulk modulus, low leaf capacitance, and delayed fine root wilting. This study showed the general contribution of fine roots to the protection of the whole plant against underground water stress. Our findings highlight the importance of water storage in the leaves and fine roots of mangrove species and high tolerance to water deficit in the leaves of mangrove species and the fine roots of some karst species.

Mangrove species favor succulence in leaves and fine roots, with an accentuated hydraulic segmentation in the roots, whereas most karst species promote sclerophylly and belowground segmentation.

Introduction

Water enters a plant mainly via its fine roots, that is, the roots of the first three orders (Guo et al., 2008; McCormack et al., 2015), and is released from leaves into the atmosphere in exchange for carbon dioxide via transpiration. Accordingly, fine roots and leaves represent the direct interfaces of a plant to its external environment through the soil–plant–atmosphere continuum (Tyree, 1997). Furthermore, they represent a bottleneck in the water transport pathway. Although the distance from the periphery of fine roots to stele and from leaf veins to stomata measures only a few hundred micrometers, it represents a large proportion of the total hydraulic resistance of a plant (Steudle and Peterson, 1998; Sperry et al., 2003). Nevertheless, while the water-relation characteristics of leaves have been extensively studied (Bartlett et al., 2014; Ding et al., 2014; Zhu et al., 2018), those of fine roots have been poorly investigated (Rodriguez‐Dominguez and Brodribb, 2020; Bartlett et al., 2022).

Controlling the permeability of the extravascular water diffusion pathways through the fine root cortex and the leaf mesophyll may help the plant cope with water stress. At the roots, soil water is absorbed through fine root epidermal cells, transported through the cortex via the symplastic and apoplastic transport pathways, and filtered by the endodermis, subsequently entering the root xylem network (Steudle and Peterson, 1998). In the leaves, water travels through the vein xylem network, through the mesophyll via the symplastic and apoplastic pathways, ultimately evaporating through the stomata. Microtomography imaging showed that, during water stress, root water uptake was disrupted by the formation of cortical lacunae (Cuneo et al., 2016). The cortex cells shrunk and suffered damage; consequently, the connectivity between the soil and the tine root, and between the epidermis and endodermis was disrupted. In leaves, desiccation induced mesophyll cell shrinkage and collapse (Ding et al., 2014). The mesophyll conductance decreased and severely reduced the overall leaf conductance (Scoffoni et al., 2017). During periods of water stress, high desiccation tolerance of the terminal organs can allow water transport and continuation of photosynthesis (Bartlett et al., 2014), whereas early wilting can protect more expensive organs from severe damage (Johnson et al., 2016; Klepsch et al., 2018). Different habitats may favor different water-relation strategies.

In general, water stress generates high water tension within the plant and can result in xylem embolism in the main water transport system—xylem vascular tissues (Tyree, 1997). Once embolized, vascular tissues—vessels and tracheids—may no longer transport water, which can potentially lead to plant death. As many plant species cannot repair embolized vessels, surviving trees rely primarily on new secondary xylem growth to transport water (Choat et al., 2019). Both embolism repair and xylem production require a high carbon investment (Brodersen and McElrone, 2013). Some studies have revealed that certain tree species across various habitats sacrifice their leaves and fine roots to protect their trunks and branches from severe hydraulic dysfunction, a mechanism known as hydraulic vulnerability segmentation (Tyree and Zimmermann, 2002; Johnson et al., 2016; Klepsch et al., 2018; Qin et al., 2021). However, other studies have refuted such findings (Tan et al., 2020; Zhang et al., 2021). One survey across topographic gradients in a tropical karst forest showed significant leaf-to-branch segmentation at the hilltops but not in the valleys (Zhang et al., 2021).

Previous research has mostly focused on hydraulic vulnerability segmentation between leaves and branches (Tyree and Zimmermann, 2002; Zhang et al., 2021; Levionnois et al., 2020), and researchers have increasingly analyzed the vulnerability of roots to water stress (Bourbia et al., 2021; Cuneo et al., 2021; Bartlett et al., 2022). However, fewer studies have analyzed the vulnerability segmentation of roots (Johnson et al., 2016), especially fine roots (Rodriguez‐Dominguez and Brodribb, 2020; Qin et al., 2021). A common garden experiment revealed minor differences in coarse root and stem embolism resistance (Peters et al., 2020), emphasizing the potential importance of fine root segmentation in protecting the xylem network against soil water deficits. Investigations of hydraulic vulnerability segmentation have mostly compared the water potential (WP) that induced xylem emboli and significant loss of conductivity/conductance across the studied organs (Johnson et al., 2016; Peters et al., 2020; Tan et al., 2020). However, mesophyll cell turgor loss reduces extraxylary hydraulic conductance in leaves and decreases the overall leaf conductance more than leaf vein emboli (Zhang et al., 2016; Scoffoni et al., 2017; Corso et al., 2020). Drought experiments on olive (Olea europaea) seedlings showed a substantial increase in the root hydraulic resistance relative to that of the whole plant under mild water stress, that is, before xylem embolism initiation (Rodriguez‐Dominguez and Brodribb, 2020). Turgor loss at the root cells is expected to reduce water conductance through the root cortex, as turgor loss does at the mesophyll cells in leaves. Thus, soil property-driven water deficits, such as those observed in karst and mangrove habitats, represent a suitable framework for assessing the contribution of fine root turgor loss to plant hydraulic vulnerability segmentation and adaptations to soil water stresses.

From the perspective of water availability and plant hydraulics, karst forests, especially epikarst trees, grow on a thin and patchy soil layer resting on leaky bedrocks despite abundant precipitation (Chen et al., 2010). Accordingly, water availability in karst habitats is spatially and temporally heterogeneous (Geekiyanage et al., 2019). Even during wet seasons, a major portion of the rainwater falling on hilltops and hillslopes runs off, and the other portion quickly infiltrates deeper horizons. The soil water content (WC) quickly decreases to levels comparable to soil WC in the dry season (Fu et al., 2016). In mangroves, roots uptake water and nutrients from saline soil in which osmotic pressure physically limits the water uptake by roots (Reef and Lovelock, 2015). Additionally, mangrove trees usually experience intense light, strong wind, and high temperatures, resulting in elevated evaporative demand. Consequently, mangroves constantly have high water tension in their hydraulic system (Ward et al., 2016). Notably, karsts and mangrove forests are similar in at least one aspect: water stress due to soil properties.

Plants growing in karst and mangrove habitats also face other stresses. High calcium concentration and associated pH in karst habitats constrain nutrient uptake by plants from the soil (Tyler, 1992). In mangroves, soil salinity is a major stress; moreover, the quasi-permanent flooding causes hypoxia, and strong winds weaken tree stability (Tomlinson, 2016). Comparing plants growing in karst and mangrove ecosystems against plants from a low-stress environment should highlight various mechanisms enabling plant adaptation to extreme environments.

The present study compared karst and mangrove leaf and fine-root traits for dozens of woody tree species, each from native habitats, against those of ornamental and timber-production species grown in a common garden that experienced minimal water stress. Such comparisons highlight how fine-root structure, resistance to turgor loss, and storage of internal water reserves of the fine roots and leaves contribute to plant adaptation to the environmental constraints in karst and mangrove forests.

Accordingly, the present study aimed to analyze pressure–volume characteristics of fine roots and leaves of woody species from mangrove and karst forests in northern subtropics, using common garden plants for comparison. We hypothesized that fine-root turgor loss occurs before stem xylem embolism and leaf turgor loss. Therefore, fine-root turgor loss might contribute to the hydraulic vulnerability segmentation of a woody plant, acting as a first position fuse to protect other organs from soil water stress in karst and mangrove plants. To this end, pressure–volume curves of the leaves and fine roots of 37 woody species from the three habitats were constructed, of which stem xylem vulnerability to embolism in 17 species was previously measured. The water-relation parameters and the trait correlations across organs and habitats were compared. Subsequently, a discussion is provided on how trait combinations could contribute to plant adaptation to their respective habitats.

Results

The root turgor loss point (TLP_root) was less negative than the leaf turgor loss point (TLP_leaf) for 32 of the 37 species studied; the pattern was stronger in mangrove plants than in karst and common garden plants (Figures 1 and 2). In contrast, intraspecies t tests demonstrated that the TLP_root of Bauhinia blakeana and Delavaya toxocarpa were significantly more negative than their TLP_leaf (P < 0.05). TLP_root was significantly less negative than the TLP_leaf for 10 of the 13 mangrove species, 4 of the 13 karst species, and 3 of the 12 common garden species (Supplemental Table S1). However, the correlation between TLP_leaf and TLP_root was not significant (Figure 1). Nine mangrove and eight karst species were used for xylem embolism resistance measurements conducted previously by our group (Supplemental Table S1). Among the mangrove species with vulnerability curve information, TLP_leaf and TLP_root were always less negative than the stem xylem WP at 50% and 12% loss of conductivity (P₅₀ and P₁₂, respectively, Figure 3). Across the eight karst species, four had less negative TLP_leaf values, while five had less negative TLP_root as compared to their stem P₅₀, although the TLP_leaf and TLP_root of these karst species were more negative than their stem P₁₂ (Fig. 3).

Figure 1.

Trait comparison and correlations between fine roots and leaves across the three habitats. (A) Comparison of water potential at turgor loss between fine roots and leaves (TLP_root versus TLP_leaf; respectively), and (B) correlation between Specific root length and specific leaf area, respectively from the three habitats. Error bars depict standard error. The TLP_leaf versus TLP_root 1:1 in (A) is drawn in a black solid line for water potential comparison. The significant correlation among mangrove species is depicted by a dashed line in (B). The reduced major axis correlation coefficients (R²) across the three habitats are indicated: mangrove (M), common garden (C), and karst (K). **Indicates 0.001 < P < 0.01, and nonsignificant correlations are labeled as ns (P > 0.05).

Figure 2.

Pairwise comparisons of the turgor loss point, capacitance before turgor loss (C₁), maximum osmotic potential (P_osm), and the bulk modulus (E_max) between leaves and fine roots across the three studied habitats. Common garden species are in (A), (D), (G), (J) panels, Karst species are in (B), (E), (H), (K), and mangrove species are in (C), (F), (I), (L). The paired t test P-values are shown at the bottom of the graph. C₁, represented by gray symbols (G–I), was magnified 20×.

Figure 3.

Comparison of water potential at turgor loss in fine roots and leaves against branch xylem embolism resistance of karst (K) and mangrove (M) species. Branch embolism resistance was characterized by the water potential at 12% and 50% loss of conductivity (P_12branch and P_50branch, respectively) and compared against fine root (A and C) and leaf (B and D) turgor loss point (TLP) (TLP_root and TLP_leaf, respectively). The solid black line depicts the 1:1 relationship for water potential comparison. Significant reduced major axis correlations (^*P < 0.05, ^**P < 0.01, ^***P < 0.001) are depicted with the gray lines. The solid gray lines are the correlation lines for all species, and the dashed gray line is the correlation line across karst species. The data are means of a species ± standard error.

The average leaf bulk modulus (E_max) was 4.9 times greater than the root E_max (paired t test, P < 0.001; Figure 2 and Table 1). The root capacitance was 12.9 and 3.5 times higher than the leaf capacitance before and after turgor loss, respectively (both paired t tests, P < 0.001; Figure 2). Furthermore, root–leaf trait correlations were limited. For instance, SRL and specific leaf area (SLA) were negatively correlated only in mangroves (R² = 0.65, P < 0.01; Figure 1).

Table 1.

Comparisons of means ± standard errors of fine root and leaf traits across habitats

Habitats	Organs	TLP (−MPa)	Emax (MPa)	Posm (−MPa)	SWC (Relative)	C1 (mol.kg−1.MPa−1)	C2 (mol.kg−1 MPa−1)	SLA (10−3.m²/g)	SRL (m/g)	Cth (%)
Common Garden (12 Species)	Leaves	1.60 ± 0.06A,a	13.7 ± 1.7A,a	1.39 ± 0.07A,a	2.0 ± 0.1A	13.6 ± 2.1B,a	44.4 ± 5.2B,a	14.5 ± 1.5A
	Roots	1.36 ± 0.06A,b	4.7 ± 1.0A,b	0.90 ± 0.04A,b	2.7 ± 0.3B	48.1 ± 13.4,A,b	60.4 ± 15.6A,a		69 ± 16A	71.7 ± 2.1A
Karst (13 Species)	Leaves	1.74 ± 0.12A,a	19.4 ± 1.8B,a	1.50 ± 0.11A,a	1.6 ± 0.1A	7.4 ± 0.7A,a	32.5 ± 3.3A,a	11.5 ± 1.1A
	Roots	1.60 ± 0.09B,a	3.4 ± 0.3A,b	1.06 ± 0.05A,b	2.0 ± 0.2A	35.5 ± 4.1A,b	41.1 ± 8.0A,a		79 ± 15A	69.8 ± 1.5A
Mangrove (12 Species)	Leaves	2.38 ± 0.11B,a	18.2 ± 1.8A,B,a	1.99 ± 0.08B,a	2.9 ± 0.2B	10.7 ± 1.3A,B,a	27.8 ± 4.3A,B,a	7.3 ± 0.7B
	Roots	1.46 ± 0.06A,B,b	2.5 ± 0.3A,b	0.90 ± 0.03A,b	3.9 ± 0.3C	321.8 ± 21.4B,b	272.8 ± 39.1B,b		158 ± 36B	79.7 ± 1.2B

E _max, maximum bulk modulus; P_osm, maximum osmotic potential; C₁, capacitance before turgor loss; C₂, capacitance beyond turgor loss; and Cth, the fine root cortex relative thickness. For each organ, significant differences across habitats (phylogenetic analysis of variance, P < 0.05) are represented by uppercase letters. Within a given habitat, significant differences between organs (t-test, P < 0.05) are represented by lowercase letters.

Compared with the leaves of the common garden and karst plants, mangrove leaves had significantly more negative TLP and P_osm, higher leaf saturated water content (LWC), and lower SLA (Table 1). The TLP of the mangrove leaves was 0.78 MPa more negative, their LWC was 1.5 times higher, and their SLA was half that of the common garden leaves. The mangrove leaf capacitance C₁ was intermediate between the karst and common garden plants. Notably, karst leaves had significantly higher E_max than the common garden leaves (Table 1). A positive correlation between C₁ and SLA was observed in karst and mangrove leaves. The mangrove leaves delimited the upper-left boundary of the C₁–SLA relationship, and the karst leaves delimited the lower-right boundary (Figure 4).

Figure 4.

Correlations between turgor loss point, capacitance (C₁), and tissue construction—the SLA and specific root length—across the studied habitats. Trait correlations in leaves (A and B) and fine roots (C and D) were calculated using reduced major axis regressions. The SLA and TLP_leaf of 389 woody species from nine major biomes in China (Zhu et al., 2018) were plotted in the background of the panel (B) for comparison. Significant correlations (^*P < 0.05, ^**P < 0.01, ^***P < 0.001) are depicted with a thick solid line for cross-habitat relationships, while dashed gray lines depict those for mangrove (M) and dash-dot-dashed lines for karst species (K). None of the correlations were significant for the common garden species (C). Not significant correlations are presented as “ns” (P > 0.05).

Regarding the fine roots, the mangrove plants had remarkably higher C₁, C₂, and SRL than the karst and common garden plants (Table 1). Furthermore, mangrove fine roots exhibited significantly higher root saturated water content (RWC), 6.7 times and 4.5 times higher capacitance before and beyond the turgor loss, respectively, and 2.3 times higher SRL than the common garden plants (Table 1). On the contrary, the fine roots of the karst plants had significantly more negative TLP than the common garden roots, despite an average difference of 0.24 MPa. High RWC and C1_root were strongly correlated with a high root cortex thickness ratio across habitats (Supplemental Figure S1). The negative correlations between TLP_root and SRL, and positive correlation between TLP_root and C₁, were significant only for karst fine roots (R² = 0.33, P < 0.05; R² = 0.61, P < 0.01, respectively; Figure 4).

In the principal component analysis (PCA) diagram, the overlaps between the functional spaces of the two groups represents the functional similarities between them. Thus, the small overlap in functional space occupied by the leaves and fine roots in the PCA diagram of the six variables resulting from pressure–volume curves indicated that their multivariate functional spaces were largely different (Supplemental Figure S2). Notably, the associated functional diversity indices between the two organs were quite similar. The contrast in root and leaf capacitance was exceptionally high in mangrove species, and the functional space occupied by the mangrove roots was elongated parallel to the capacitance axis (Supplemental Figure S2). The functional space occupied by the karst leaves was elongated parallel to the bulk modulus axis, coinciding with the importance of bulk modulus among the karst leaves.

Discussion

This study showed that the fine roots wilted at a less negative WP than the leaves for most woody plants from the three habitats examined. The findings demonstrated divergent water-relation strategies adopted by the leaves and fine roots of mangrove and karst species, wherein mangrove plants had high resistance against wilting and succulent leaves with low SLA. Furthermore, their fine roots exhibited high water storage, extremely high capacitance, and a high cost-efficiency for soil exploration (high SRL). Conversely, karst plants exhibited a high foliar cell bulk modulus and promoted sclerophylly, and their fine roots wilted only with advanced dehydration.

Belowground hydraulic segmentation in karst and mangrove species

On average, 32 of the 37 species examined had a more negative TLP_leaf than TLP_root. This pattern was stronger in mangrove plants than in the karst and common garden plants. In principle, the relative vulnerability of terminal organs—high relative hydraulic resistance under water deficit—referred to as the hydraulic vulnerability segmentation, is a strategic adaptation to water deficits (Johnson et al., 2016). The lack of belowground segmentation in two-thirds of the karst species examined was equivalent to the lack of leaf segmentation observed in karst valleys (Zhang et al., 2021). The karst trees growing in valleys have quasi-permanent access to soil water and thus require less adaptation to water deficits, contrarily to the slope and epikarst trees (Geekiyanage et al., 2019; Zhang et al., 2021); however, in mangroves, salt is quasi-permanent, even in a diluted state. Mangrove trees prefer a low salinity to survive (Burchett et al., 1984); however, evaporation and transpiration can elevate salinity levels and thus decrease soil osmotic potential. Therefore, a circuit breaker is required at their roots to prevent water loss, justifying the belowground segmentation in most of the mangrove species (Figure 5).

Figure 5.

Synthesis of hydraulic safety segmentation in leaves and fine roots during water stress, and water loss prevention at high salinity spots at the fine roots of mangroves.

In habitats where water uptake is limited by soil properties, such as the karsts and mangroves, plants must implement mechanisms to prevent water leakage from the roots to the soil. During water stress, fine roots can disrupt the link between the water transport system and the soil by disrupting intercellular exchange, decreasing aquaporin activity (Javot, 2002), and forming cortical lacunae that disconnect the root epidermis from the endodermis (Cuneo et al., 2021, 2016), In addition, the shrinkage induced by cortical lacunae can detach the root epidermis from the soil (Cuneo et al., 2021). However, cortical lacunae and root shrinkage cannot isolate mangrove roots from the substrate because saline water can fill any gap and lacuna. Instead, decreasing root turgor has been reported to parallel decreasing aquaporin activities and loss of root conductance (Kim and Steudle, 2007). Turgor loss at mild WP in mangrove roots can temporarily interrupt the exclusive cell-to-cell radial pathway that links the stele with the soil, thus preventing water loss (Javot, 2002). In addition, root cortical lacuna formation in terrestrial species damages the root cells, permanently disrupting the connection between the vascular system and the soil (Cuneo et al., 2016). Experimentation on grapevines showed that fine roots with cortical lacunae did not completely recover their hydraulic conductance upon rehydration (Cuneo et al., 2021). Instead, cell turgor is recoverable, and mild wilting (loss of turgor) only temporarily disrupts the link between the stele and the soil, forming a circuit breaker similar to that observed in leaves (Zhang et al., 2016; Scoffoni et al., 2017) and contributing to the hydraulic vulnerability segmentation.

A good circuit breaker should open or close the circuit with low energy costs and minimal damage to the system. Xylem embolism, turgor loss, and cortical lacunae (in fine roots) reduce root and leaf hydraulic conductance (Scoffoni et al., 2017; Cuneo et al., 2021). Furthermore, xylem embolism of coarse roots and branches may not efficiently protect the trunk xylem against embolism, as the coarse roots were not significantly or only moderately more vulnerable to embolism than the stem (Peters et al., 2020), and distal branches are generally more resistant to embolism than the trunk (Bouche et al., 2016; Johnson et al., 2016). Alternatively, decreasing leaf turgor during dehydration significantly reduced leaf conductance before embolism initiation (Brodribb and Holbrook, 2006). In addition, complete turgor loss occurred before the WP reached a threshold beyond which severe damage to the leaf tissues occurred (Bartlett et al., 2016). Therefore, turgor loss-induced segmentation occurred before the embolism-induced hydraulic segmentation. In addition, leaf hydraulic conductance measurements showed turgor recovery at a WP far more negative than that required for the embolism repair (Trifilo, 2003), and leaf water status recovered quickly upon rehydration without any sign of embolism repair. Compared to xylem embolism and root cortical lacunae formation, early and easily recoverable turgor loss protects plants growing in habitats with frequent water stress.

Turgor loss, cortical lacunae, root shrinkage, and xylem embolism are physiological mechanisms capable of disrupting root hydraulic conductance. Future research analyzing the triggers, timings, WP threshold, and potential recovery mechanisms across natural and controlled environments should further our understanding of water loss prevention mechanisms during water stress.

Adaptation to mangrove habitat

The results showed that mangrove leaves had significantly lower TLP_leaf and P_osm.leaf and thus a higher resistance to wilting than the common garden and karst plants. The TLP_leaf of the studied mangrove species ranged within the more negative half of the TLP_leaf global range (Bartlett et al., 2014; Zhu et al., 2018). The WP in mangrove leaves is always low owing to osmotic regulation against seawater and reaches its minimum during midday and rainless periods (Lin and Sternberg, 1992; Jiang et al., 2017). Thus, the high resistance to wilting of mangrove leaves constitutes an adaptation to the high temperatures, winds, solar radiation, and salinity; the former four factors contribute to the high evaporative demand that mangrove leaves experience.

The mangrove species maintained a significantly higher leaf capacitance C₁ than the karst plants. Their succulent leaves were associated with the lowest SLA and the highest LWC among the three habitats in the study. The positive correlation between SLA and C₁ was observed in the mangrove plants and delimited the upper-left boundary of the SLA–C₁ area, suggesting that leaf capacitance has been implemented to the maximum allowed by leaf biomass investment. Thick endoderm layers and specialized thick spongy tissues have been observed in the leaves of many mangrove species, particularly in Avicennia marina. Both structures are the main contributor to the high leaf capacitance and water storage of mangrove species (Nguyen et al., 2017). Avicennia marina is globally the most widespread mangrove species and is highly tolerant to salinity (Morrisey et al., 2010). In theory, capacitance buffers the WP by delaying the pressure changes during high transpiration, resulting in a longer time constant τ (McCulloh et al., 2019). High capacitance is favorable for mangroves to compensate for the thin atmospheric boundary layer due to coastal winds. Besides, decoupled branch xylem water supply and leaf hydraulics suggested a substantial contribution of leaf water storage to transpiration of mangrove trees (Agduma et al., 2022). However, water release from leaf tissues substantially reduces the leaf osmotic potential unless it has sufficient water storage. The high LWC of the mangrove leaves allowed a high E_max despite the high C₁. A high LWC is equally important for diluting the salt content of the mesophyll tissues.

Furthermore, low SLA is generally associated with leaf sclerophylly, but succulent leaves also have a low SLA (Read et al., 2006). Hence, the higher SWC of the mangrove species leaves in the present study indicates leaf succulence. In addition, low SLA and leaf thickness should mechanically strengthen the leaves (Read et al., 2006), possibly serving as an adaptation to wind exposure. Theoretically, the associated leaf thickness confers additional stiffness to mangrove leaves, given the three-power contribution of thickness to a rectangular beam’s moment of inertia. Empirically, wind exposure experiments on 39 tropical seedlings showed less wind damage and less leaf drop in species with high LMA–low SLA (Butler et al., 2012). The high resistance to wilting, large water storage, leaf succulence, and low SLA of mangrove species make them suitable for the strong wind, high salinity, and high evaporative demand in coastal habitats.

Mangrove fine roots combined high capacitance with high SRL, but their TLP_root and P_osm.root were modestly negative. Such trait combinations coincide with the primary adaptation mechanisms to salinity. Mangrove species could maintain the root WP at a moderate value by synchronizing various mechanisms such as (1) the salt exclusion mechanism in roots, (2) forage for water and nutrients in heterogeneous substrates (Reef and Lovelock, 2015), and (3) water status buffer by water storage. With the endodermis preventing the apoplastic transport and fine root turgor loss disrupting the cell-to-cell transport pathway, salt absorption at highly saline locations is largely inhibited.

The high SRL allows the fine roots to explore large soil volumes with less biomass and resource investment (Fitter, 2002). The fast resource acquisition and the short life span of high SRL (Eissenstat and Yanai, 1997) are particularly advantageous to the mangrove species growing under highly heterogeneous substrate conditions. Mangrove soil salinity varies both temporally and spatially, and mangrove fine roots are likely to uptake water from soil microsites with lower salinity (Reef and Lovelock, 2015). Thus, a high capacity for soil exploration increases the probability of finding spots with low salinity.

The high capacitance of mangrove fine roots reflects an ability to fast release stored water during water stress. Accordingly, it can buffer the xylem WP when the roots fail to absorb enough water (Meinzer et al., 2003; Scholz et al., 2007). While the thick and porous root cortex observed in the mangrove roots is vital for gas transportation (Burchett et al., 1984), the strength of the correlation between RWC, C₁, and the relative cortex thickness justifies its contribution to the root water regime. This supports the contribution of the root cortex to both water storage and root respiration, as observed in flooded Poaceae (Yamauchi et al., 2021). Therefore, water storage and capacitance in the fine roots and leaves help mangrove trees maintain vital water transport under high soil salinity conditions.

The negative correlation between SRL and SLA was significant only for the mangrove species. SRL and SLA are analogous traits that express the potential for resource exploration per given biomass investment. A literature review synthesized that significant leaf–root trait correlations are less common, and most studies failed to identify any significant correlation between SLA and SRL (Weigelt et al., 2021). While SLA is strongly related to carbon and nutrient economics (Wright et al., 2004), a low SLA is associated with a slow metabolism and long lifespan. Recently, a study aligned SRL with the plant–mycorrhiza collaboration axis and independently to the fast–slow resource acquisition axis (Bergmann et al., 2020). The independence of SRL to the fast–slow spectrum justified the lack of correlation between SRL and SLA within karst and common garden species, and the significance of the relationship within mangrove species denoted the complexity of root–leaf trait coordination. Therefore, resource use, uptake, and investment coordination across fine roots and leaves require further large-scale investigation.

Adaptation to karst habitat

Compared to the common garden plants, the karst species had a significantly higher leaf bulk modulus, representing the inverse of the relative water discharge per unit drop in WP (Schulte and Hinckley, 1985). Finite element models revealed that the bulk modulus was strongly correlated with the membrane elasticity modulus, strongly contributing to the critical pressure for cell collapse independent of the cell size and membrane thickness (Ding et al., 2014). This indicates that the high leaf E_max observed in the karst species resulted from the reinforcement of the cell wall, which reduced membrane permeability. Thus, the low permeability of the membrane combined with the low LWC explained the low capacitance before and after TLP_leaf. Fu et al. (2019) reported that tropical karst woody species had a thicker epidermis than neighboring tropical, nonkarst woody species. Therefore, the strong cell wall and thick epidermis strengthened the karst leaves and promoted sclerophylly. Karst habitats in southwestern China provide the ecological conditions to promote sclerophylly, such as low nutrient and water availability, as well as a pronounced seasonality (Chen et al., 2010; Fu et al., 2016). Following a drought-rewatering cycle, the water content of sclerophyllous leaves recovered completely within few minutes, while that of non-sclerophyllous leaves recovered faster, although the recovery of the latter was not always complete (Salleo et al., 1997). In a habitat with frequent water stress, a full recovery of the water status of the plant is vital. Thus, sclerophyllous species were well-adapted to karst habitats.

The roots of karst species had a higher resistance to turgor loss than those of the common garden plants and mangrove species. The positive correlation between TLP_root and C₁ and the negative correlation between TLP_root and SRL suggest that karst species with highly negative TLP_root sacrificed fine root capacitance in favor of resistance to turgor loss and soil exploration. This finding coincides with the extensive soil exploration of the roots observed in karst species in another study (Nie et al., 2014). The correlation between TLP_root and SRL, which referred to the link between root wilting resistance and construction in their fine roots, left less room for the contribution of hydraulic capacitance to the relationship. It is suggested here that the strategies adopted by the karst species involved strong fine roots but a low capacity to control the transient variation of the fine root water status.

The thin soil layer in karst hilltops and hillslopes leaves few functional alternatives for species that rely on soil water. When comparing the functional advantages of fast against slow root metabolism, fast but ephemerous roots may not be advantageous as small soil volume can be conquered. Conversely, strong and long-lived roots may permit the resource uptake as soon as the edaphic conditions become favorable; however, these favorable conditions are often temporary (Fu et al., 2016). Thus, wilting resistant fine roots are better suited for karst habitats to ensure a permanent occupation of the limited soil volume while awaiting ephemeral favorable conditions for the water and nutrient uptake.

Conclusions

This study characterized the water-relation traits of leaves and fine roots for 37 woody species growing across three different habitats in southern China. Generally, fine roots favored storage and capacitance, whereas leaves favored resistance to wilting and higher cell bulk modulus. More specifically, we found that the karst and mangrove species have adopted different strategies to better adapt to their habitats. The loss of turgor at the fine roots can break the link between the soil and the root, thereby preventing water leakage during periods of water deficit in mangrove species. In addition to high leaf resistance to wilting, high capacitance of both leaves and fine roots and increased capacity for soil exploration were other notable assets of the mangrove species analyzed. The karst species were characterized by mechanical adaptation associated with high leaf bulk modulus and root wilting resistance. Therefore, they adopted sclerophylly, while mangroves promoted succulence. The diversity in functional trait combinations in the fine roots and leaves across habitats showed that these terminal organs are vital loci for adaptation to different environments. The present findings can be incorporated into function-based species distribution models and forest sensitivity analyses to study the effect of climate change.

Materials and methods

Study sites and plant species

Study sites were chosen within a narrow climatic, latitudinal, and altitudinal range. The plant materials were collected between mid-August and early November from three different habitats in the northern marginal tropics: 12 species of the common garden plants were collected from the campus of Guangxi University, Nanning, Guangxi, China. Thirteen tree species were collected from a karst forest under the administration of the Experimental Center of Tropical Forestry, Chinese Academy of Forestry, located in PingXiang, Guangxi. Mangrove plant materials were collected from two locations in Hainan, China, about 50-km apart: eight species from the DongZhaiGang mangrove natural reserve and four from another mangrove reserve in WenChang Bay. The selected karst and mangrove species are among the most abundant in their respective habitats. The common garden location and species were selected to enable comparison with mangrove and karst species. The selected common garden combines climate and soil properties similar to those of native lowland forest habitats in this study area. Common garden plants were selected from different families across the angiosperm phylogeny. Two common garden species (Teck: Tectona grandis and eucalypt: Eucalyptus grandis) are popular timber trees planted in southern China. Three species (Bougainvillea glabra, Calliandra haematocephala, and Ceiba speciosa) are common ornamental plants grown in tropical cities, and all others naturally grow in the moist forests of southern China. The geographical locations and climatic conditions of the studied sites are summarized in Table 2, and details about the studied species are included in Supplemental Table S2.

Table 2.

Location and climates of the three study sites: MAT and MAP are the mean annual temperature and precipitation, respectively. The wet season lasts from May to October at all locations. The climate data was downloaded from the China Meteorological Data Service Center website (www.data.cma.cn).

Site	Habitat type	Latitude	Longitude	Elevation (m)	MAT (°C)	MAP (mm)
Campus of Guangxi University	Common Garden	22°50′38″N	108°17′31″E	120	21.7	1,294
PingXiang County	Karst	22°7′22″N	106°44′33″E	246–444	22.0	1,470
Dong Zhai Gang Mangrove Protected Area	Mangrove	19°57′8″N	110°34′38″E		24.8	1,700
WenChang Bay	Mangrove	19°33′5″N	110°50′31″E		24.4	1,749.5

Sample collection

Leaf and fine root samples (<3 mm in diameter) were collected from 06:00 to 08:30 h on clear days and from 06:00 to 09:30 h on rainy days in nonmangrove habitats (from early September in the karst and late September in the common garden). In mangrove sites, sample collection dates were chosen to satisfy low tide and light precipitation conditions from 06:00 to 10:00 h (October 31 to November 1, 2019). Branches were harvested and immediately wrapped in plastic bags containing wet paper towels to maintain humidity and prevent the sample from drying out. Samples from the karst forests were transported, processed, and measured at a local Experimental Center of Tropical Forestry. In contrast, mangrove samples were stored in polystyrene boxes with ice bags and transported to the Guangxi University within 2 days. The proximal ends of branches were recut underwater, with the recut end kept submerged in a bucket, and leaves were covered with a black plastic bag to release the water tension.

Root samples were excavated using the “mothering method,” that is, following fine roots from their tips to the mother tree or vice versa (Guo et al., 2008). When the mothering method was not possible, the surrounding species were identified, and the structure of the collected roots was identified according to previously collected sample dimensions, colors, aspects, and odors. Fine root samples were collected at depths <20 cm, taking great care to minimize damage due to digging and manipulation. Samples were enclosed in plastic bags along with saturated soil and stored at ∼4°C for <3 days before measurement. The purpose of this low-temperature storage was to reduce root respiration and decay.

Pressure–volume curve

To build the pressure–volume curve of a leaf, it was cut from the supporting branch, and its leaf area was measured using a leaf area meter (Li-3100C; Lincoln, NE, USA). It was immediately wrapped in an aluminum foil to control transpiration during the stabilization process. The WC and WP were measured repeatedly and successively using a 10⁻⁴ g precision scale (Mettler Toledo, MLT204T; Shanghai, China) and a pressure chamber (PMS 1505 D-EXP, Corvallis, OR, USA), respectively. A wet paper tissue was kept inside the pressure chamber to prevent evaporation during measurements. If the initial WP of any leaf was above −0.2 MPa, it was excluded from the analysis and replaced with another. Between each measurement, the wrapping was opened, and the leaf was dried on a bench for 15–30 min, with the exact duration depending upon the dehydration rate of the leaf sample. A WP difference of 0.2–0.4 MPa between two successive measurements was targeted. Curve patterns were monitored after each measurement, and it was ensured that at least four meaningful points beyond the TLP had been acquired before oven-drying the sample.

For roots, the studied species covered a wide diversity of root architecture, which posed challenges for the WP measurement of a fixed root order sample as some roots were too large while others were too small to fit the pressure chamber clamp. Finer roots were expected to provide better approximations of the finest root behavior under water stress than coarser roots. Therefore, the root samples used for the pressure–volume relationship measurements were selected based on the lowest root order that fit the pressure chamber, that is, from the 5th to the 15th order and approximately 1 mm in diameter at the cutting edge.

Mud and other particles were carefully separated from the roots underwater, and any excess surface water was gently removed using a dry paper towel. Then, the sample was wrapped in aluminum foil. The aluminum wrap was kept closed during measurement to protect the roots from damage due to manipulation. Successive measurements of WC and WP pairs were taken, as was done for the leaves.

Following the series of measurements for both leaves and roots, the wraps were opened, and the samples were oven-dried at 70°C for 72 h and then weighed to obtain the oven-dry mass. The SLA was calculated as the leaf area ratio by oven-dry mass.

The pressure–volume curve (Supplemental Figure S3) was fit using a personalized R-package based on the pressure–volume equation proposed by Schulte and Hinckley (1985), which allows for the analysis of the curve sensitivity by resampling and bootstrapping the WC–WP pairs. For each curve, the average, and the confidence intervals of the pressure–volume parameters were calculated from 100 bootstraps. The pressure–volume parameters include the WP at turgor loss (TLP), the capacitance before and after the wilting point (C₁ and C₂, respectively; Equation 1), the bulk modulus at full turgor (E_max; Equation 2), the saturated water content (SWC), and the osmotic potential at full turgor (P_osm) (Supplemental Table S3). The latter significantly correlated with TLP (Supplemental Figure S4). The capacitance was standardized by the oven-dried sample mass and represented in mol.kg⁻¹.MPa⁻¹ to allow the comparability of roots and leaves as per the methodology of Roddy et al. (2019). The capacitance is the absolute variation of the tissue WC per unit drop in WP (Equation 1). It quantifies the rate at which water can be discharged from the organ during water stress. The bulk modulus is the inverse of the relative water discharge per unit drop of WP (Equation 2). A high modulus means that the organ loses a small percentage of its WC during dehydration. While C₁ and E_max are seemingly inversely related, Equations 1 and 2 suggest that organs may maximize both C₁ and E_max by implementing high water storage:

$$C=\frac{d\mathrm{WC}}{d\mathrm{WP}},\mathrm{ }$$ (1)

$$E=\frac{d\mathrm{WP}}{d\mathrm{WC}}\mathrm{ }\mathrm{WC}.\mathrm{ }$$ (2)

Branch xylem embolism resistance and hydraulic vulnerability segmentation

Branch xylem embolism resistance traits of the karst species were derived from a previous study conducted in the same locality (Zhang et al., 2021). Although further information on the stem xylem embolism of six mangrove species can be found in previous studies (Jiang et al., 2017), the methods used have been recently criticized (Chen et al., 2020). Accordingly, branch xylem embolism resistance of nine mangrove species was measured in this study using bench-drying and flow measurement, identical to that used for karst species. Briefly, branches were first collected in the early morning, bagged in plastic bags, and transported to the laboratory. Subsequently, they were left to dehydrate to different WPs. Subsequently, a leaf close to the segment of interest was detached, and its water potential was measured using the PMS pressure chamber as above. Then, branch segments were cut at least twice underwater and connected to a flowmeter (Bronkhorst L13, Ruurlo, The Netherlands) for flow rate measurement. Subsequently, emboli were flushed at a pressure higher than 200 kPa, and the maximum flow rate was measured using the flowmeter. The percentage flow difference before and after the emboli were flushed represents the percentage loss of conductivity (PLC). The WP and PLC pairs of multiple branches were assembled, plotted, and fit a sigmoidal equation (Equation 3):

$$\mathrm{PLC}=\frac{100}{1+\exp \left(a\left(\mathrm{WP}-b\right)\right)},\mathrm{ }$$ (3)

where a and b are the curve parameters.

The hydraulic vulnerability segmentation of each species was estimated by the difference between the leaf or fine root TLP and the branch WP, corresponding to a 12% or 50% loss of conductivity.

Fine-root structure and anatomy

The SRL was measured following the sampling recommendations of McCormack et al. (2015). First-order fine roots were cut from the root network, spread on a microscope slide, and scanned with a high-resolution flatbed scanner (Epson V800, Jawa Barat, Indonesia) at 3200 dpi, with a ruler for calibration. The root lengths were measured manually using the ImageJ software (Fiji 1.52p, National Institute of Health, USA). After scanning, the samples were oven-dried at 70°C for 72 h and weighed using a 0.01 mg resolution scale (Sartorius Secura225D-1CN, Beijing, China). SRL was calculated as the ratio between the total root length and dry mass.

For each plant, the root diameter, cortex thickness, and cortex thickness ratio, that is, the root cortex thickness divided by the root diameter, were measured from first-order roots without apparent secondary growth (Supplemental Figure S5). Freehand transversal sections were prepared from fresh samples and optionally stained with safranin-alcian blue. Sections were imaged on a digital microscope (DM 3000 LED Wetzlar, Germany) at the highest magnification that displayed the entire section on a single frame. The measurements were performed using ImageJ software.

Statistical analysis

All statistical analyses were performed using R software v. 3.5.0 (Vienna, Austria). Paired t tests were performed to analyze the trait differences between the two organs. The phylogenetic information regarding the studied species was extracted from the GBOTB.extended tree (Zanne et al., 2014; Smith and Brown, 2018) using the PhyloMaker R-package (Jin and Qian, 2019). Subsequently, phylogenetic analysis of variance was performed using the Phytools package (Revell, 2012) to compare trait differences across habitats (Supplemental Table S4). The Levene-test function from the car R-package (Fox and Weisberg, 2019) was used to compare the trait variance in each organ across habitats. The prcomp function from the stats R-package was used to perform PCA using six primary parameters derived from the PV curves of leaves and fine roots (Bartlett test, P < 0.001; Kaiser–Meyer–Olkin index = 0.57 for the PCA of the three habitats). The PCA coordinates of each organ/site were extracted, and the functional spaces were delimited using the kde2d function from the MASS package (Venables and Ripley, 2002) to illustrate the functional morphospace occupied by each organ/habitat. Furthermore, the functional richness of the two organs was computed using the FD package (Laliberté et al., 2014). Notably, this study analyzed correlative and noncausal relationships. Thus, all trait coordination and tradeoffs were evaluated using the sma function from the smatr package (Warton et al., 2012).

Supplemental data

The following materials are available in the online version of this article:

Supplemental Table S1. Safety and cross-organ safety comparison across the 37 studied species.

Supplemental Table S2. List of the studied species and their growth characteristics by habitat and collection site.

Supplemental Table S3. Additional traits from the pressure–volume relationship and carbon economic parameters across the 37 studied species.

Supplemental Table S4. Contribution of the phylogeny and the habitats to the leaf and fine root traits among the studied species.

Supplemental Figure S1. Correlation between the relative cortex thickness, the root capacitance (C1_Root), and the root saturated water content of mangrove (M), karst (K), and common garden species (C).

Supplemental Figure S2. Functional similarities and differences between leaf and fine root water relationship parameters across the studied species revealed by PCA.

Supplemental Figure S3. Illustration of pressure–volume curves of leaf and fine root of Michelia alba.

Supplemental Figure S4. Relationship between the maximum osmotic potential and the water potential at turgor loss of the roots and leaves across and within habitats.

Supplemental Figure S5. Representative root anatomical structure.

 

A.N.A.A. and K-F.C. designed the study, analyzed the data, and wrote the manuscript. A.N.A.A., S.W., and B-N.L. performed the experiments on the roots and the leaves in the common garden and mangroves and on the karst roots. X.J. did the xylem embolism resistance measurement in the mangroves. Y-D.Q. contributed to the measurements of mangrove species. F-S.T. and Q-W.Z. measured the traits of the leaves of the karst species. All authors read and approved the manuscript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Kun-Fang Cao (kunfangcao@gxu.edu.cn).