Leaf vein topology confers water transport efficiency

doi:10.21203/rs.3.rs-2344990/v1

Download PDF

Research Article

Leaf vein topology confers water transport efficiency

https://doi.org/10.21203/rs.3.rs-2344990/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The evolution of xylem vessels and dense leaf vein networks in flowering plants enabled unprecedented increases in plant water transport and rates of CO₂ assimilation. We tested the hypothesis that independent of vein density, higher leaf vein topological efficiency (VTE), achieved with denser free vein endings, would reduce the extraxylary pathlength, further benefitting whole-leaf conductance, while reducing carbon investment, and releasing space for light capture. Our analysis across 52 phylogenetically diverse angiosperm species demonstrated that for a given vein density, high VTE conferred by dense free endings can shorten the extraxylary pathlength by up to 11%. Across species, high VTE was associated with high stomatal conductance, non-vein area fraction for light capture, and low leaf mass per area. Our findings identify leaf vein topological efficiency as an important measure of the use of leaf space and biomass, and a key factor influencing plant adaptation to historical and future environmental conditions.

Plant Physiology and Morphology

Leaf vein topology

minor veins

leaf mass per area

extraxylary water transport pathway

stomatal conductance

non-vein area

In land plants, leaves are the main sites of carbon fixation and biomass production, and these functions require hundreds of water molecules be transported to the leaf surface to fix a single molecule of CO₂. Yet, the pathlength between the soil and stomata remains a strong hydraulic resistance that must be overcome to maintain leaf irrigation, transpiration, and photosynthesis ^1–3. Thus, natural selection has favored the evolution of high-capacity transport, including efficient networks of xylem vessels, which contribute to organ and whole plant hydraulic conductance, especially in productive habitats ^4–6. A particularly strong bottleneck arises in the very terminus of the hydraulic network, when water exits the xylem conduits, crosses the bundle sheath cells, and traverses the leaf mesophyll to the sites of evaporation and photosynthesis ⁷. This short “extraxylary” pathway – through the root cortex and the leaf mesophyll – can account for up to 70 to 90 percent of the total plant hydraulic resistance ^8–11. In leaves, shortening this pathway by dense vasculature substantially reduces the leaf hydraulic resistance ¹², but increases the leaf xylem construction cost ^13,14. Therefore, natural selection should favor a specific venation design to achieve the highest hydraulic conductance and photosynthesis per unit carbon invested ^1,14,15. As the smallest veins in the leaf—the “minor veins”—constitute the bulk of the vein length, they should be arranged to minimize the extraxylary pathlength resistance per unit vein length ¹⁶. We aimed to identify vein topology that reduces the extraxylary pathlength and leaf biomass investment, increases photosynthetic gases uptake efficiency, and consequently leads to high physiological performance.

The role of optimizing vein topology may be especially important in species that already evolved high vein density. During the Cretaceous, the radiation and increasing dominance of angiosperms entailed, and may have been enabled by, dramatic increases in the length of vein per unit leaf area (hereafter “vein density”; VD), thus drove increases in hydraulic conductance and allowed high photosynthetic rates ^17,18. However, since ca. 120 Mya the evolution of high VD slowed in most angiosperm lineages, suggesting that increasing VD provided diminishing returns in fitness ¹⁷, a hypothesis supported by experiments showing that increases in VD had a saturating effect on hydraulic conductance in artificial leaves of a given thickness ¹⁶. Further, a study of phylogenetically and ecologically diverse species found that the optimal vein density should match the distance between veins with the vertical distance from vein to the evaporative surface; as such, for a given vein to evaporative distance, further increase in VD, would not lead to proportional hydraulic efficiency ¹⁸. In addition to providing diminishing returns in benefits, a high VD necessitates additional costs in space and biomass investment: while minor veins contribute less than 5% of the total leaf biomass ¹⁹, and do not meaningfully contribute to the leaf mass per area (LMA) ²⁰, the metabolic and construction costs of their thick-walled conduits and associated bundle sheaths may be substantial ¹⁴. Further, high vein density involves the displacement of photosynthetic tissues ²¹.

Yet, independent of VD, shifts in vein topology may shorten the extraxylary pathlength distance and thereby confer a higher transport efficiency contributing to gas exchange capacity and productivity. Such a benefit for characterizing vein topology would be in addition to those previously associated vein “loopiness”, i.e., the amount of connectivity per leaf area, focused on its benefits for redundancy of the network architecture, and potential for contributing tolerance against drought or freezing-induced xylem embolism, mechanical damage or herbivory ^12,22.

Notably, previous studies quantifying vein topology have focused on geometric parameters that did not completely isolate vein topological efficiency from VD ^12,23,24. We developed an approach to disentangle vein topology from VD by quantifying the relative benefit (extraxylary pathlength shortening) independent of vein length and projected area. We tested the hypothesis that angiosperms gradually acquired optimized vein topology, which also entails a greater space allocation to photosynthetic mesophyll, and enables greater stomatal conductance and photosynthetic rate, and less expensive leaves.

Contribution of free-ending veinlets and vein topological efficiency to stomatal conductance

Vein topological efficiency (VTE), i.e., the vein-to-abaxial epidermis distance for actual leaves relative to modeled leaves (gridded vein systems) of the same vein density (VD), varied from − 1.2% ± 0.3 (high loopiness) to 11.0% ± 1.2 (more free ending veins) (Fig. 1). VTE was independent of vein density (Figure S1), but exhibited modest to strong association with FEV density metrics, such that high VTE was largely driven by high FEV relative length (Fig. 2a, Figure S2, Figure S3). Additionally, species with high VTE possessed narrow minor vein diameter (Fig. 2b) and the vein topology-driven extraxylary pathlength reduction correlated with high stomatal conductance (Fig. 3a). A VTE of 11% was associated with 38-fold higher G_smax (1.30 mol·m^− 1·s^− 1) than VTE of 0%. In addition to the link between VTE and extraxylary pathlength, the SEM distinguished two other potential links between VTE andG_smax; either via LMA or nVF (Fig. 4), which were negatively or positively associated with G_smax, respectively (Figure S4).

Influence of VTE on biomass investment and leaf photosynthetic area

High VTE was also associated with low biomass investment on leaf and low space allocation to veins. VTE correlated positively with the nVF and negatively with leaf mass per area; a VTE of 11% was associated with 32% higher non-vein area fraction (a relative increase of 57%, Fig. 3b) and 140 g m^− 2 lower LMA (a relative decrease of 85%, Fig. 3c, Figure S5) compared to VTE of 0%. The series of vein models with increasing FEV relative length demonstrated that the contribution of FEVs alone would not explain the observed gain in nVF with increasing VTE (Fig. 3b,d), unless the negative correlation between minor vein diameter and VTE was added to the model (Fig. 3b,e). Minor vein diameter reduction with increasing VTE resulted in similar slope between the modelled and the observed nVF, with approximately 10% nVF remaining unexplained.

Vein topological efficiency and safety trade-off

However, VTE was strongly correlated with vein geometrical metrics that mainly characterize vein redundancy: high VTE was associated with low vein loopiness, meshedness, areole circularity and high areole elongation ratio (Fig. 5). VTE was negatively correlated with leaf capacitance (C₁) for the pooled data from the two sites, and independent of turgor loss point (TLP) (Figure S6).

We demonstrated that across diverse species, abundant free-ending veinlets increased vein topological efficiency. Then, independent of vein density, high vein topological efficiency was associated with short extraxylary pathlength between the xylem and the leaf surface, enabling high maximal stomatal conductance, large mesophyll volume, and low carbon investment in leaf construction.

For a given vein density, an efficient vein topology significantly shortened the diagonal distance between the vein and the abaxial leaf surface. VTE was independent of vein density, and unlike vein density showed no significant phylogenetic signal, suggesting that the optimization of vein topology is labile and responsive to selection pressure during the angiosperm radiation, i.e., within and across basally and recently derived clades. For a given vein density, leaves may shorten the extraxylary pathway by constructing highly ramified venation and developing more free-end veinlets (FEVs). The additional benefit of a high FEVs in conferring a high VTE beyond a high vein density extends the previous work on the importance of FEVs as a contributor to vein density ²³. Further, the role of vein topology in hydraulic flow extends its previously described importance in providing vein redundancy to confer hydraulic safety to disruption by damage or herbivory ^25,26.

Our findings demonstrate that optimized vein topology similarly to high vein density can shorten the distance for water transport from xylem to the sites of evaporation, thus lowering the resistance to water transport across the mesophyll, and potentially increasing whole-leaf gas exchange. Across species, a high VTE in addition to a high VD was associated with high G_smax. Notably, while several studies have reported significant correlations between vein density, stomatal conductance and light saturated photosynthetic rate across diverse species ^27–30, while other studies have reported either weak or no correlation between these traits ^31–33. This implies the potential for decoupling between vein density and gas exchange, especially when considering species adapted across aridity gradients ³⁴. The plants in arid habitats are likely to exhibit contrasting soil-to-stomata pressure gradients (e.g., atmospheric vs soil aridity) as well as different stomatal responses to leaf water potential and vapor pressure deficit ^35,36. For plants considered in a given environment, the coordination between water supply (conductance) and demand (transpiration) is expected to maximize photosynthetic gas exchange, hence, the coordination between vein topology and stomatal conductance ³⁰.

Our findings suggest that a major benefit of increasing VTE is the resulting increase of lamina area for photosynthetic light capture. The areole outside the veins has a dense and thick layer of palisade mesophyll tissue, where the vast majority of light capture and photosynthesis in eudicot leaves takes place ³⁷, whereas the areas directly above the veins contain only a thin (or no) mesophyll layer and a few (or no) stomata. Compared to species with low VTE, species with high VTE had greater areole area fraction. However, FEVs alone did not suffice to explain the gain in nVF to the extent suggested by the correlation analyses, rather the negative scaling between minor vein diameter and VTE explained a large percentage of the nVF gain. Thus, the gain in nVF with increasing vein topological efficiency was largely driven by high VTE in species having thin minor veins. All else being equal, a high non-vein area fraction would contribute to a high photosynthetic rate and stomatal conductance per unit leaf area.

We demonstrated that high VTE was associated with low LMA. This finding suggests a co-selection of traits that would contribute to rapid gas exchange per leaf mass, i.e., high VTE and low LMA. Notably, increasing investment in VD or FEVs would not influence LMA; VD was statistically and mechanistically independent of LMA across over 350 angiosperm species ²⁰, and this is consistent with the minor veins contributing only ca. 5% of the total leaf mass ¹⁹. By contrast, several studies have reported significant correlations between LMA and other leaf traits, such as leaf and palisade mesophyll thickness ^38,39, and John et al. (2017) identified the leaf thickness and tissue density (thick cell walls) as the strongest driver of LMA, and both can increase the resistance to gas diffusion across mesophyll, limiting carbon uptake and assimilation.

Our analysis of model vein systems with increasing VTE showed that high VTE was associated with low loopiness and meshedness, metrics proposed by Price & Weitz (2014), or high loop elongation and low loop circularity proposed by Blonder et al (2018). Functionally, “loopy” vein systems allow for alternate water transport pathways (hydraulic “redundancy”) in case of embolism or damage to part of the vein network ^12,22. Conversely, excessive development of FEVs – low redundancy – may weaken resistance of the leaf to hydraulic failure, mechanical damage (force to punch) and herbivory (tannins) ^24–26. Our finding that increasing VTE would increase hydraulic delivery from veins to epidermis suggests the possibility of a trade-off in vein topology with respect to its contributions to efficiency and safety. Yet, the link between VTE and leaf hydraulic safety is still controversial. In fact, low VTE (high redundancy) species were associated with large-diameter minor veins, which may host conduits with large diameter that are more sensitive to implosion than narrow conduits ⁴⁰, disrupting water conduction before embolism ⁴¹. Plants with low VTE may compensate the vulnerability of large minor veins with high redundancy, leading to even high carbon investment in leaf veins. Further, VTE was weakly associated or independent of TLP which quantify the sensitivity of mesophyll cells during desiccation ⁴². Thus, the relationship between leaf hydraulic safety and vein topology is likely variable, depending on the influence of many other leaf structural and water relation traits.

From the perspective of leaf construction, the correspondence of high VTE with low LMA is consistent with rapid resource acquisition, hence high G_smax, while the correspondence of low VTE with high LMA and high area-based capacitance would reinforce stress tolerance. High capacitance serves the maintain the leaf water potential at mild value and has been marginally associated with low leaf conductance ⁴³. In addition, leaves with high LMA and low G_smax are commonly interpreted as “tough” (thick and/or dense) leaves conferring long leaf lifespan, but also slow leaf-mass-based rates of photosynthesis, whereas high G_smax is generally aligned with fast leaf-area-based rates of water transport and photosynthesis ^44,45. Vein topological efficiency would thus contribute to the set of traits that are related to area-based and mass-based rates of gas exchange, and thus potentially plant productivity.

Overall, the contribution of vein topology to gas exchange rates (via G_smax) would arise from multiple pathways including water flow pathways, leaf carbon economics, and spatial allocation of the mesophyll.

We developed a novel approach to quantify the effect of leaf vein topology for water transport outside the xylem, and its association with key structural and physiological features of angiosperm species of both basal and more recent divergences. The methods require only a basic light microscope and camera. Vein topology varied across species, with abundant free-ending veinlets leading to high vein topology efficiency, and the latter being associated with high maximal stomatal conductance, large leaf volume available for photosynthetic tissues, and low LMA. The shortening of the extraxylary pathlength through optimization of vein topology could result in increased hydraulic efficiency and photosynthesis relative to leaf construction cost, evolving repeatedly and probably contributing to the success of angiosperms in competitive habitats.

Study site and species

We selected 52 phylogenetically and ecologically diverse angiosperm species varying strongly in leaf venation architecture and vein density (VD; Table S1, Figure S7). All the 52 species are hypostomatic. To ensure a wide range of VD, and to test evolutionary shifts in vein topology, we sampled two groups of species: a first set from basally divergent groups occurring in a moist tropical forest, which was expected to have low VD ^17,46, and a second set from more derived groups native to a water-limited karst habitat, expected to have high VD ³⁴. Thus, we sampled thirty magnoliid species (23 trees, five lianas and two herbs), including a “living fossil” angiosperm tree, Takhtajania perrieri ⁴⁷, from the tropical montane forest of Anjanaharibe-Sud Natural Reserve in the Sava Region of Madagascar (14°44'29"S, 49°29'50"E), and 22 tree species from a subtropical evergreen-deciduous mixed karst forest in China (25°9’9”N, 108°2’23”E). The tropical montane forest grows between 875 m and 1950 m elevation, experiences mean monthly temperatures varying from 8.6 to 25.7°C, and receives 2000 to 3600 mm of annual precipitation ^48,49. The subtropical forest grows between 410 m and 640 m elevation on rocky karst hills with a mean annual precipitation of 1529 mm and a mean annual temperature of 19.3°C. The sampling and on-site measurements were undertaken from November 2017 to January 2018 in the Malagasy site and July to August 2019 for the Chinese site, coinciding with the growing season at each study site.

Branch and leaf sampling

Leaf samples were obtained from a one-meter-long branch which was cut from the upper sun-exposed canopy for each of five to six individuals per species before 9 am. These branch samples were double-wrapped in plastic bags and transported to an on-site laboratory, re-cut underwater and rehydrated for 2–4 hours covered with plastic bags.

Imaging and measurement of leaf venation, stomatal traits and cross-sectional anatomy

The fresh leaf area of three fully developed leaves per individual was measured using a calibrated digital camera (Olympus E-PM2, China) for the Malagasy samples and a leaf area meter (LI-COR Li 3000C, USA) for the Chinese samples. These leaves were preserved in Formalin – Alcohol – Acetic acid fixation solution (10: 50: 5: distilled water 35) ⁵⁰, and transported to the nearest research facility (University of Antananarivo, Madagascar for the tropical site, and Guangxi University, China for subtropical site). Minor vein density was measured from a 0.25 cm² section cut centrally in the leaf, avoiding major veins, which was digested in a 0.7% NaOH aqueous solution at 70°C until it became transparent ⁵¹, and when necessary, bleached, stained with a drop of 0.5% safranin for 5 min, washed with distilled water, and mounted on a microscope slide. Four images per leaf of approximately 1 mm² were taken at a resolution of 1.8 pixels·µm^− 1 (Leitz Orthoplan, Wetzlar, Germany) for the tropical site, and at 3.3 pixels·µm^− 1 (DM 3000 LED Wetzlar, Germany) for subtropical site, respectively. The calibration of both microscopes was verified using the same micrometer slide. Minor vein length, areole contours and length of free-ending veinlets (FEV) were manually drawn and measured using ImageJ (1.52p, National Institutes of Health, USA) ⁵². The minor vein density (VD) and non-vein fraction (nVF) were calculated by dividing the vein length and non-vein area (areole area) respectively by the area of the field of view. FEV relative length was calculated by dividing FEV length by the total minor vein length.

For the measurement of stomatal traits, small paradermal sections of epidermis were prepared by hand, bleached, washed with distilled water, mounted on slides with glycerol, photographed with a digital microscope (Tropical site; Subtropical site: DM 3000 LED Wetzlar, Germany), and measured for stomatal density (SD) and dimensions, that are, length (L), width (W), and pore length (p).

For the measurement of cross-sectional anatomy, transverse cross-sections were prepared by hand, double-stained with Safranin-Alcian blue, and imaged for measurement of leaf thickness, mesophyll thickness, the vertical distance between the minor vein and abaxial epidermis, and stomatal depth. Three images per leaf were taken using a 10× or 20× objective (C-mount 0.55x, respectively 3.3 and 6.6 pixels·µm^− 1), depending on the leaf thickness, and, for three stomata, using a 100× objective (33.5 pixels·µm^− 1) to measure the stomatal depth. The theoretical maximum stomatal conductance was calculated following Sack & Buckley (2016):

$${G}_{smax}=bmd{s}^{0.5} \left(1\right)$$

where, b and m are biophysical and morphological constants, respectively, and d and s represent stomatal density and stomatal size. b $=\frac{D}{v}$, where D is the diffusivity of water in the air and v is the molar air volume. $m=\frac{\pi {c}^{2}}{{j}^{0.5}(4hj+\pi c)}$, where c = the pore length to stomatal length ratio (p/L), j = the stomatal width to stomatal length ratio (W/L), h = the pore depth to stomatal width ratio (l/W).

Vein topological efficiency

Our approach to quantify vein topology efficiency (VTE) based on combining information of vein density with anatomical measurements of leaf cross-section images is illustrated in Fig. 6. For each leaf vein image, the edges around leaf veins were traced manually using a modified ImageJ polygon tool equipped with a drawing pad (Bosto 16 HDK, China), and the traced image was converted into a black (vein) and white (non-vein lamina) raster, from which the vein area was measured – hereafter referred to as the “real leaf” image. Total vein length was measured manually using a modified ImageJ polyline tool. Keeping the vein length and non-vein area fraction (nVF), the “real leaf” image was reconfigured into a simple square mosaic vein distribution image (hereafter “modeled leaf” image) with the exact same vein area fraction and vein length as the real leaf image.

To build the modelled leaf, we estimated the length of each grid segment (i.e., vein segment) (S) and the thickness of each grid segment (i.e., vein diameter) (Fig. 6-1B). Given that the vein density is the vein length per unit leaf area, the first step was to identify the leaf area associated with each grid segment (the area delimited by the broken green lines in Fig. 6-1B). Geometrically, this area represents half of whole areole. Therefore, S of the model can be calculated from VD of the real leaf using the Eq. (2):

$VD=\frac{2S}{{S}^{2}}=2 \times {S}^{-1}$ , so, $S=2 \times {VD}^{-1}$ (2)

Once the one-pixel thick grid was created, the n closest pixels were assigned as veins, where n is the number of vein pixels in the real leaf image. The thickness (T) of the vein on the model can also be calculated by solving the second-degree Eq. (2), the geometrical principle of this equation is detailed in Supporting Information Note S1. The pixels representing veins were determined as being within a distance T/2 from the nearest vein center.

${\left(\frac{T}{S}\right)}^{2}-2\frac{T}{S}+{V}_{f}=0$ (3), thus, $\frac{T}{S}=1-\sqrt{1-{V}_{f}}$, the other solution of the equation is unrealistic.

For both real and modeled images, the distance of each areole pixel to the nearest vein was computed using the distance function from the raster R-package ⁵³. Then, a buffer zone of 250 pixels from the edge of each image was deleted to remove the edge artifact. Indeed, areole pixels close to the edge were often farther from the nearest vein than they are in reality. The vein-to-lower epidermis distance was extracted from the cross-section image of the corresponding leaf, and the hypotenuse of the horizontal distance from the raster and the vein-to-epidermis distance from the leaf cross-section was calculated. To avoid the effect of outlier regions, we calculated the 95% quantile of the hypotenuse length (VS₉₅) for each image. The difference between VS₉₅ calculated from the model-vein and its corresponding real-vein image divided by the VS₉₅ of the real-vein image represents the influence of the native vein topology on the extraxylary pathlength, and is thus referred to as the vein topological efficiency (VTE; %). A positive VTE indicates that the arrangement of veins shortens the linear hydraulic pathlength through the mesophyll. Notably, this theoretical linear hydraulic pathlength is a proxy corresponding to the paths of the bulk flow of liquid water through and around mesophyll cells and diffusion of vapor in the airspaces ⁷. We assumed the pathlength from vein to epidermis a proxy for flow to the stomata after Brodribb et al. (2007)Brodribb et al. (2007), without accounting for variation in stomatal density. We suggest that this assumption is justified given the typical correlation between stomatal density and vein density, as well as the conservation of stomatal arrangements in areoles ²³.

Pressure-volume parameters

Two to four fully grown leaves were chosen from each individual and dehydrated on the bench while their leaf mass (10^− 3g or 10^− 4g digital balance; MT203T and MT204T, Shanghai, China) and water potential (Ψ_leaf) (Scholander pressure chamber; PMS 1505 D-EXP, Corvallis, OR, USA) were repetitively measured. Successive measurements were separated by a lag of 5 to 30 minutes to allow the leaves to dehydrate. Pressure-volume curves were constructed by plotting the reciprocal pressure (1/Ψ_leaf) against one minus relative water content (1 − RWC), where RWC was calculated as (fresh mass - oven-dry mass)/(saturated water mass - oven-dry mass) (Schulte & Hinckley 1985). Fresh leaf area of the species from the tropical and subtropical site was measured using a calibrated digital camera (Olympus E-PM2, China) and a leaf area meter (LI-COR Li 3000C, USA), respectively. Leaf mass was then determined by oven-drying the leaves at 70°C for at least 72 hours ⁵¹. RWC-Ψ_leaf data were fitted with an exponential function after Schulte & Hinckley (1985) using the nls function in the base R-package ⁵⁴. The turgor loss point (TLP) is the water potential at the intersection between the non-linear and the non-linear portion of the curve (1/Ψ_leaf) = f(1 − RWC). The hydraulic capacitance C₁ (mol·MPa^− 1·m^− 2) was calculated following Blackman & Brodribb, (2011)Blackman & Brodribb, (2011):

$${C}_{1}=\frac{\varDelta RWC}{\varDelta WP}\times \frac{SWM}{LA\times M}$$

Where, ($\frac{\varDelta {R}{W}{C}}{\varDelta {c{\varPsi }}_{\mathbf{l}\mathbf{e}\mathbf{a}\mathbf{f}}})$ (MPa^− 1) was calculated from the slope of the initial non-linear portion of the pressure-volume curve, SWM (g) is the absolute saturated water mass, LA (m²) is the fresh leaf area and M (g.mol^− 1) is the water molar mass.

Testing the mechanism underlying VTE and non-vein fraction

To clarify the mechanistic association between VTE and non-vein fraction, we tested two hypotheses:

All else being equal, gain in non-vein area is due to the absence of a connection between FEV tips and the vein at its opposite side. We selected a leaf model with maximum redundancy and created a series of models with increasing number of FEV, from 1 to 13 FEV, as higher numbers of FEVs resulted in VTE > 11% which exceeded the range of the observed VTE. VTE and non-vein fraction of each model was calculated and plotted against empirical values.

The gain in non-vein area is related to thinner veins along with increasing numbers of FEV. A series of models with increasing number of FEV was created as above and the VTE of each element was calculated. For each element of the series, the thickness of the vein was set based on the empirical correlation between VTE and vein diameter. Then, their VTE and non-vein fraction was plotted against empirical values.

This series of modelled leaves were also used for analyzing the correlation between VTE and other leaf vein topological metrics in the published literature: the areole elongation rate and circularity using the LeafVeinCNN program ^26,56 and the vein loopiness (mm^− 1) and meshedness (unitless) was calculated using LeafGUI program ^25,57.

Statistical analysis

Statistical analyses were performed using R 3.5.3 ⁵⁴. The leaves of two species from Peperomia genus were highly succulent, they were ouliers with extremely low vein density and long VS₉₅ distance, and they were excluded from all correlation analyses. Correlations were fitted using the sma function from smatr package ⁵⁸, and structural equation model (SEM) analyses were performed using the lavaan package ⁵⁹. Different hypothetical SEM models were fit and compared to evaluate the importance, placement, and effect size of different paths linking FEV properties, VTE and stomatal conductance. Key paths that were tested included direct linkage between vascular traits and LMA and/or stomatal conductance (e.g., VTE ◊ G_Smax), indirect linkage between vascular traits and stomatal conductance via biomass economics (e.g., VTE ◊ LMA ◊ G_Smax), as well as indirect linkage via pressure-volume parameters. Model fitting was simplified by manual step-wise path deletion beginning from the weakest path. At each step, the root mean square error of approximation (RMSEa) and the Akaike Information Criterion (AIC) were recorded. The most parsimonious model (lowest RMSEa and AIC values) was selected. Phylogenetic information for the studied species was extracted from the GBOTB.extended database using the v.Phylomaker R-package ⁶⁰, Blomberg’s K phylogenetic signal and phylogenetic ANOVA for cross-habitat comparison were calculated using the phytools package ⁶¹. Phylogenetic independent contrasts were calculated using the ape package ⁶². Data were log transformed to normalize the distribution of residuals when required. The customizations made to the polygon and polyline tools in ImageJ, and all R-code, are available from the first author upon request.

Competing interests: The authors declare no competing interests.

Acknowledgments

CKF was financially supported by the National Natural Science Foundation of China (31470469) and the Bagui Scholarship (C33600992001) of Guangxi Zhuang Autonomous Region. AA received funding from the China Scholarship Council. The authorization for the sample collection in Madagascar was issued by the 'Ministère de l'Environnement et du Développement Durable' – Madagascar – Research authorization: 262/17-MEEF/SG/DGF/DSAP/SCB.Re of 2017, 24^th October. We are grateful to the Madagascar National Park MNP, the institution managing the Anjanaharibe-Sud Nature Reserve, and Mr. Jo, the forest ranger who served as our field guide in Madagascar.

Author contributions

AA and KFC designed the study. AA, KFC, YZL, SMG and LS wrote the manuscript. AA and RT collected the samples and analyzed the species from Madagascar. AA, MYN and SDZ analyzed the species from China. The image and data analyses were done by AA. All authors have reviewed and approved the manuscript.

McCulloh, K. A., Sperry, J. S. & Adler, F. R. Water transport in plants obeys Murray’s law. Nature 421, 939–942 (2003).
Anfodillo, T., Carraro, V., Carrer, M., Fior, C. & Rossi, S. Convergent tapering of xylem conduits in different woody species. New Phytol 169, 279–290 (2006).
Olson, M. E. et al. Universal hydraulics of the flowering plants: vessel diameter scales with stem length across angiosperm lineages, habits and climates. Ecol. Lett. 17, 988–997 (2014).
Sack, L. & Frole, K. Leaf structural diversity is related to hydraulic capacity in tropical rain forest trees. Ecology 87, 483–491 (2006).
Brodribb, T. J., Feild, T. & Jordan, G. J. Leaf maximum photosynthetic rate and venation are linked by hydraulics. Plant Physiol. 144, 1890–1898 (2007).
Brodribb, T. J. Xylem hydraulic physiology: The functional backbone of terrestrial plant productivity. Plant Science 177, 245–251 (2009).
Buckley, T. N., John, G. P., Scoffoni, C. & Sack, L. How does leaf anatomy influence water transport outside the xylem? Plant Physiol. 168, 1616–1635 (2015).
Cochard, H., Nardini, A. & Coll, L. Hydraulic architecture of leaf blades: where is the main resistance? Plant Cell Environ 27, 1257–1267 (2004).
Sack, L. & Holbrook, N. Leaf hydraulics. Annu. Rev. Plant Biol 57, 361–81 (2006).
Scoffoni, C. et al. Outside-xylem vulnerability, not xylem embolism, controls leaf hydraulic decline during dehydration. Plant Physiol. 173, 1197–1210 (2017).
Liu, X.-R. et al. Water transport from stem to stomata: the coordination of hydraulic and gas exchange traits across 33 subtropical woody species. Tree Physiol 39, 1665–1674 (2019).
Sack, L. & Scoffoni, C. Leaf venation: structure, function, development, evolution, ecology and applications in the past, present and future. New Phytol 198, 983–1000 (2013).
Beerling, D. J. & Franks, P. J. The hidden cost of transpiration. Nature 464, 495–496 (2010).
McKown, A. D., Cochard, H. & Sack, L. Decoding leaf hydraulics with a spatially explicit model: principles of venation architecture and implications for its evolution. Am. Nat 175, 447–460 (2010).
Koçillari, L. et al. The widened pipe model of plant hydraulic evolution. Proc. Natl. Acad. Sci. U.S.A. 118, e2100314118 (2021).
Noblin, X. et al. Optimal vein density in artificial and real leaves. Proc. Natl. Acad. Sci. U.S.A. 105, 9140–9144 (2008).
de Boer, H. J., Eppinga, M. B., Wassen, M. J. & Dekker, S. C. A critical transition in leaf evolution facilitated the Cretaceous angiosperm revolution. Nat Commun 3, 1221–1232 (2012).
Zwieniecki, M. A. & Boyce, C. Evolution of a unique anatomical precision in angiosperm leaf venation lifts constraints on vascular plant ecology. Proc R Soc Lond B Biol Sci 281, 20132829–20132829 (2014).
John, G. P. et al. The anatomical and compositional basis of leaf mass per area. Ecol Lett 20, 412–425 (2017).
Sack, L. et al. Leaf mass per area is independent of vein length per area: avoiding pitfalls when modelling phenotypic integration (reply to Blonder et al. 2014). J Exp Bot 65, 5115–5123 (2014).
Chapin, F. S., Matson, P. A. & Vitousek, P. M. Principles of terrestrial ecosystem ecology. (Springer New York, 2011). doi:10.1007/978-1-4419-9504-9.
Katifori, E. The transport network of a leaf. C R Phys 19, 244–252 (2018).
Fiorin, L., Brodribb, T. J. & Anfodillo, T. Transport efficiency through uniformity: organization of veins and stomata in angiosperm leaves. New Phytol 209, 216–227 (2016).
Blonder, B. et al. Structural and defensive roles of angiosperm leaf venation network reticulation across an Andes–Amazon elevation gradient. J Ecol 106, 1683–1699 (2018).
Price, C. A. & Weitz, J. S. Costs and benefits of reticulate leaf venation. BMC Plant Biol 14, 234 (2014).
Blonder, B. et al. Linking functional traits to multiscale statistics of leaf venation networks. New Phytol 228, nph.16830 (2020).
Brodribb, T., Feild, T. & Sack, L. Viewing leaf structure and evolution from a hydraulic perspective. Funct. Plant Biol. 37, 488–498 (2010).
Sack, L. et al. How do leaf veins influence the worldwide leaf economic spectrum? Review and synthesis. J Exp Bot 64, 4053–4080 (2013).
McElwain, J. C., Yiotis, C. & Lawson, T. Using modern plant trait relationships between observed and theoretical maximum stomatal conductance and vein density to examine patterns of plant macroevolution. New Phytol 209, 94–103 (2016).
Scoffoni, C. et al. Hydraulic basis for the evolution of photosynthetic productivity. Nat. Plants 2, 16072 (2016).
Zhao, W.-L., Chen, Y.-J., Brodribb, T. J. & Cao, K.-F. Weak co-ordination between vein and stomatal densities in 105 angiosperm tree species along altitudinal gradients in Southwest China. Funct. Plant Biol. 43, 1126 (2016).
Gleason, S. M. et al. Weak coordination among petiole, leaf, vein, and gas-exchange traits across Australian angiosperm species and its possible implications. Ecol Evol 6, 267–278 (2016).
Xiong, D. L. & Flexas, J. From one side to two sides: the effects of stomatal distribution on photosynthesis. New Phytol 228, 1754–1766 (2020).
de Boer, H. J. et al. Apparent overinvestment in leaf venation relaxes leaf morphological constraints on photosynthesis in arid habitats. Plant Physiol 172, 2286–2299 (2016).
Rodriguez-Dominguez, C. M. & Brodribb, T. J. Declining root water transport drives stomatal closure in olive under moderate water stress. New Phytol 225, 126–134 (2020).
Carminati, A. et al. Stomatal closure prevents the drop in soil water potential around roots. New Phytol 226, 1541–1543 (2020).
Li, L., Ma, Z., Niinemets, Ü. & Guo, D. Three key sub-leaf modules and the diversity of leaf designs. Front. Plant Sci. 8, 1–8 (2017).
Griffith, D. M., Quigley, K. M. & Anderson, T. M. Leaf thickness controls variation in leaf mass per area (LMA) among grazing-adapted grasses in Serengeti. Oecologia 181, 1035–1040 (2016).
Alonso-Forn, D. et al. Revisiting the functional basis of sclerophylly within the leaf economics spectrum of oaks: Different roads to Rome. Curr Forestry Rep 6, 260–281 (2020).
Sperry, J. S. Evolution of water transport and xylem structure. International Journal of Plant Sciences 164, S115–S127 (2003).
Zhang, Y. et al. Xylem conduit deformation across vascular plants: an evolutionary spandrel or protective valve? New Phytol (2022) doi:10.1111/nph.18584.
Ding, Y.-T., Zhang, Y.-X., Zheng, Q.-S. & Tyree, M. T. Pressure-volume curves: revisiting the impact of negative turgor during cell collapse by literature review and simulations of cell micromechanics. New Phytol 203, 378–387 (2014).
Fu, X. et al. Coordination and trade-offs between leaf and stem hydraulic traits and stomatal regulation along a spectrum of isohydry to anisohydry. Plant Cell Environ 42, 2245–2258 (2019).
Wright, I. J. et al. The worldwide leaf economics spectrum. Nature 428, 821–827 (2004).
Gorné, L. D. et al. The acquisitive–conservative axis of leaf trait variation emerges even in homogeneous environments. Ann. Bot 129, 709–722 (2020).
Feild, T. et al. Fossil evidence of Cretaceous escalation in angiosperm leaf vein evolution. Proc Natl Acad Sci USA 108, 8363–8366 (2011).
Schatz, G. E., Lowry, P. P. & Ramisamihantanirina, A. Takhtajania perrieri rediscovered. Nature 391, 133–134 (1998).
Goodman, S. A floral and faunal inventory of the Réserve spéciale d’Anjanaharibe-Sud, Madagascar. (Field Museum of Natural History, 1998). doi:10.5962/bhl.title.3148.
Hudson, P. J., Razanatsoa, J. & Feild, T. Early vessel evolution and the diversification of wood function: Insights from Malagasy Canellales. Am. J. Bot. 97, 80–93 (2010).
Ruzin, S. E. Plant microtechnique and microscopy. (Oxford University Press, 1999).
Perez-Harguindeguy, N. et al. New handbook for standardised measurement of plant functional traits worldwide. Aust. J. Bot 61, 167–234 (2013).
Rueden, C. T. et al. ImageJ2: ImageJ for the next generation of scientific image data. BMC Bioinform 18, 529 (2017).
Hijmans, R. & van Etten, J. Raster: Geographic data analysis and modeling. R Package Version 517, 2–12 (2014).
R Core Team. R: A language and environment for statistical computing. (2019).
Blackman, C. J. & Brodribb, T. J. Two measures of leaf capacitance: insights into the water transport pathway and hydraulic conductance in leaves. Funct. Plant Biol. 38, 118 (2011).
Xu, H., Blonder, B., Jodra, M., Malhi, Y. & Fricker, M. Automated and accurate segmentation of leaf venation networks via deep learning. New Phytol 229, 631–648 (2021).
Price, C. A., Symonova, O., Mileyko, Y., Hilley, T. & Weitz, J. S. Leaf extraction and analysis framework graphical user interface: segmenting and analyzing the structure of leaf veins and areoles. Plant Physiol 155, 236–245 (2011).
Warton, D. I., Duursma, R. A., Falster, D. S. & Taskinen, S. smatr 3- an R package for estimation and inference about allometric lines: The smatr 3 - an R package. Methods Ecol. Evol. 3, 257–259 (2012).
Rosseel, Y. lavaan: An R package for structural equation modeling. J. Stat. Softw 48, 1–36 (2012).
Jin, Y. & Qian, H. V.PhyloMaker: an R package that can generate very large phylogenies for vascular plants. Ecography 42, 1353–1359 (2019).
Revell, L. J. phytools: an R package for phylogenetic comparative biology (and other things): phytools: R package. Methods Ecol. Evol. 3, 217–223 (2012).
Paradis, E. & Schliep, K. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics 35, 526–528 (2019).

SupportingInformation20221205nolink.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Leaf vein topology confers water transport efficiency

Status:

Version 1

Abstract

Figures

Introduction

Results

Contribution of free-ending veinlets and vein topological efficiency to stomatal conductance

Vein topological efficiency and safety trade-off

Discussion

Conclusion

Online Methods

Study site and species

Branch and leaf sampling

Imaging and measurement of leaf venation, stomatal traits and cross-sectional anatomy

Vein topological efficiency

Pressure-volume parameters

Statistical analysis

Declarations

Acknowledgments

Author contributions

References

Supplementary Files

Status:

Version 1