Species-Specific and Altitude-Induced Variation in Karst Plants’ Use of Soil Dissolved Inorganic Carbon

Abstract

Root-derived carbon sources supporting photosynthesis have been demonstrated to contribute to plant carbon gain in many laboratory experiments. However, it remains largely unknown whether and to what extent soil dissolved inorganic carbon (DIC) influences leaf photosynthesis in karst habitats characterized by alkaline soils with low water content. We explored this relationship by measuring the concentrations and carbon isotope signals (δ¹³C) of soil DIC, as well as the δ¹³C of water-soluble organic matter (δ_WSOM) in leaves of nine woody species across an altitudinal gradient in karst habitats. The δ_WSOM varied among species by 7.23‰ and deviated from the δ¹³C of photosynthates solely assimilated from atmospheric CO₂ (δ_A) by 0.44–5.26‰, with a mean value of 2.20‰. This systematical discrepancy (δ_A − δ_WSOM) could only be explained by the contribution of soil DIC to leaf total photosynthesis (f_{DIC_soil}). The average values of f_{DIC_soil} considerably varied among the nine species, ranging from 2.48% to 9.99%, and were comparable with or slightly lower than those of previous laboratory experiments. Furthermore, the f_{DIC_soil} of two species significantly increased with altitude, whereas another species exhibited an opposite pattern, suggesting a highly spatial heterogeneity of DIC utilization. The present study improved our understanding of how plants adapt to the alkaline–drought soil conditions of karst habitats and thus acquire additional carbon for growth.

1. Introduction

Karst is a distinctive topography developed in regions with soluble bedrocks, such as limestone, dolomite, and gypsum [1], and covers about 22 × 10⁶ km² of the earth’s surface [2]. The karst habitats are characterized by thin soil layers, high content of bicarbonate and high pH in soils, and less vegetation cover [3,4,5]. For plants growing in karst habitats, water demand for maintaining the basic functions is often restricted due to the shallow soil [6]. For example, drought stress can lead to the decline of stomata opening and photosynthesis and downregulated metabolic processes [7]. A prolonged drought will result in hydraulic failure and/or carbon starvation, increasing the risk of mortality [8,9]. On the other hand, the alkaline soil condition in karst habitats is unfavorable for plant uptake of nutrients, such as phosphate, ferrous iron, and zinc [10,11]. As a consequence, the vegetation productivity of the karst ecosystem is usually much less than that in nonkarst regions [12]. Although plants can adopt tolerance and/or avoidance strategies to cope with environmental stress [13,14], it does not fully explain why some species grow well in karst habitats. Hence, other physiological processes of karst plants are suspected to play an important role in regulating their growth under dry and alkaline soil conditions.

Recently, root-derived carbon sources supporting photosynthesis have reemerged as a hotspot in plant ecophysiological studies [15,16,17,18]. Research shows that ¹³C-enriched CO₂ or HCO₃⁻ labeling in the root zones can be transported upward through the transpiration stream [19] and thus affects the carbon isotope composition (δ¹³C) of leaf photoassimilates and aboveground tissues [20,21]. Although these results are obtained from the manipulated experiments inside a laboratory, it may be applied to karst plants grown in the field, given the high concentration of bicarbonate contained in karst soils. The additional carbon gain may also compensate for the decrease in the fixation of atmospheric CO₂ induced by drought stress [22,23]. However, to the best of our knowledge, no research has been reported to investigate plants’ use of dissolved inorganic carbon (DIC, mainly including dissolved CO₂ and HCO₃⁻) in karst habitats. Although many hydroponic experiments were designed to simulate karst soil conditions, for example, bicarbonate stress (usually more than 10 mM) [24,25], osmotic stress (simulating water scarcity) [26], and their interactions [26], their results do not necessarily reflect the utilization of soil DIC by plants due to the differences in the microenvironment of the root zones between soil conditions and hydroponic solutions [27]. Hence, there is a need to fill this knowledge gap between laboratory experiments and field trials.

Appropriate methods and models for quantification are critical to understanding plants’ use of soil DIC. Currently, two methods are available to estimate this utilization. Method (i) applies the high abundance ¹³C labeling and then calculates the ratio of soil DIC fixed in specific tissues or organs to the total carbon gain. The total carbon gain includes soil DIC fixation, apparent photosynthesis measured by commercial infrared gas analyzers, and respiration [20,28]. However, this calculation is not directly linked to photosynthesis. Method (ii) uses a two-source ¹³C labeling in combination with isotope mixing models to determine the contribution of root-derived DIC to leaf total photosynthesis [22,23,24,29]. Nevertheless, uncertainties remain associated with the photosynthetic ¹³C discrimination and post-photosynthesis processes. Additionally, both labeling methods depend on irrigation or hydroponic culture, which will inevitably change the soil conditions, especially in karst environments. The labeling experiment also requires a sealed environment and has a high cost, which does not apply to large-scale field experiments. For these reasons, a new approach needs to be developed to calculate the contribution of soil DIC to leaf total photosynthesis (see Section 2). It has been shown that the natural abundance of ¹³C has the potential to address a wide range of ecophysiological and biochemical questions [30]. For example, naturally occurring δ¹³C signals in plant tissues integrate plant–environment interactions over long periods [31]. In C₃ plants, δ¹³C of leaves is mainly controlled by photosynthetic ¹³C discrimination [32], which can be altered by stomatal control and the activity of the carboxylation enzyme, Rubisco. Furthermore, when other carbon sources are supplying the leaf photosynthesis, for instance, uptake of DIC from the xylem sap of the host by the mistletoe or utilization of root-derived DIC by some plants [26,33], the δ¹³C of leaves can also be modified.

In karst habitats, altitude plays an important role in affecting the soil conditions and microclimate [34], which determine the distribution of plant species [35]. Species with different life forms may vary in functional traits, such as root uptake and xylem transport of water and DIC, thus influencing the extent of soil DIC used by plants. In addition, the concentration of soil DIC may change with altitudes, which probably acts on the proportion of soil DIC used by leaf photosynthesis. In this study, the primary objective was to investigate whether and to what extent soil DIC influences the leaf photosynthesis of karst plants at different altitudes. Based on our previous observation [26], we expected to find a large discrepancy between the measured δ¹³C of photosynthates and that predicted by a photosynthetic ¹³C discrimination model, which indicates the involvement of soil DIC in leaf photosynthesis. We hypothesized that the contribution of soil DIC to leaf total photosynthesis was species-specific due to the different physiological responses (e.g., leaf gas exchange) to the environmental factors. Our third hypothesis is the altitude exerts an influence on plants’ use of soil DIC, given the differences in soil conditions and microclimate induced by altitude.

2. Materials and Methods

2.1. Study Site

The field experiments were conducted in August 2016 on Mt. Sanmao, Nanming District, Guiyang, southwest China (26°33′57″ N, 106°44′56″ E), with an elevation between 1080 and 1315 m above sea level. In this region, the annual mean precipitation and mean air temperature during 2011 and 2015 were 1129.5 mm and 15.3 °C, respectively. The mean minimum temperature of the coolest month (January) was 4.6 °C, whereas the mean maximum temperature of the warmest month (July) ranged from 25 to 28 °C. Rainfall in August was decreased compared with that in July. The study site was a typical karst landform covered with a secondary forest. Soils under the vegetation examined were alkaline (pH 7.09–7.75), rich in organic matter and bicarbonate, and have low soil water content according to our preliminary survey at the same site [34]. Moreover, these soil properties changed along an altitude gradient. The bedrock was mainly carbonatite, which developed shallow calcareous soil. The thickness of the soil layer varied between 27 and 58 cm. The vegetation community was mixed with evergreen and deciduous broad-leaved forest, with the canopy height in the range of 0.8–7.6 m.

2.2. Experimental Design

Six plots were established along three altitudes: S1 and S2 on the lower altitude, S3 and S4 on the medium altitude, and S5 and S6 on the higher altitude (Figure 1). Each plot had an area of 20 × 20 m² and 60–80 m away from the neighboring plot at each altitude. Common species in each plot were recorded in

Table 1. Common species in plots of different altitudes. S1–S6 represented six plots established at three altitudes.

Altitudes	Sampling Sites	Elevation (m)	Species
			Tree	Shrub
Lower	S1	1115	L. lucidum, B. papyrifera	V. dilatatum, A. delavayana, R. cymosa
	S2	1123	L. lucidum, B. papyrifera	A. delavayana, R. cymosa
Medium	S3	1224	P. longipes, Z. serrata	V. dilatatum, A. delavayana
	S4	1229	L. lucidum, P. longipes, Z. serrata	A. delavayana, R. cymosa
Higher	S5	1289	P. longipes	V. dilatatum, A. delavayana, Z. armatum, R. biflorus
	S6	1292	L. lucidum	A. delavayana, Z. armatum, R. biflorus

. Collectively, four tree species (Ligustrum lucidum ait., Broussonetia papyrifera L., Platycarya longipes Wu., and Zelkova serrata (Thunb.) Makino) and five shrub species (Viburnum dilatatum Thunb., Ampelopsis delavayana Planch., Rosa cymosa Tratt., Zanthoxylum armatum DC., and Rubus biflorus Buch.-Ham. ex Smith) were selected for the subsequent measurement and sampling. Species in each plot included four replicates (individuals).

2.3. Leaf Gas-Exchange Measurements

Leaf gas exchange was measured in the morning (9:00–11:30 a.m.) of three sunny days (17–19 August). The net photosynthetic rate (A), stomatal conductance (g_s), transpiration rate (E), ratio of intercellular to the ambient partial pressure of CO₂ (c_i/c_a), and instantaneous water use efficiency (WUE_i) of the fully expanded leaves were measured using an infrared gas analyzer LI-6400 (Li-Cor, Lincoln, NE, USA), equipped with a clear topped 2 × 3 cm cuvette. The block temperature of the leaf chamber was set to be close to the air temperature. Near-ambient air at the height of 5 m above the ground was supplied to the leaf chamber at a rate of 500 μmol s⁻¹ in the sample cell of LI-6400. Leaves were placed inside the cuvette for 2–3 min until steady state, and then the gas exchange parameters were manually logged. In addition, dark respiration of leaves was measured after 30 min of dark adaptation. Three leaves were measured per tree or shrub. At the end of the measurements, leaf samples were immediately collected and stored in a cooler box for further determination.

2.4. Air and Soil Sampling

Ambient air at the height of 5 m above the ground was sampled once in each plot during the measurement of leaf gas exchange. The air sample was pumped into gas sampling bags (Dalian Delin Gas Packing Co., Ltd., Dalian, China) for further analysis. Soil air was sampled in the root zone of each plant. Three days prior to our sampling, two holes of 10–15 cm depth with 2 cm diameter were drilled into the soil. Subsequently, two soft gas-piping tubes with filters were inserted into the holes, and then soil air was sucked out. Finally, the end of the tubes was clamped with clips. These processes allowed a long time for soil air to equilibrate between soil porosity and the internal space of tubes. At the day of sampling, the litter layer was removed, and then soil air was pumped into a gas sampling bag with a volume of 30 mL through one of the tubes, waiting for further determination. Another tube was connected to LI-6400, and, thus, the CO₂ concentration of soil air was directly measured when it reached a peak value. Soil samples (approximately 150 g) were collected from the top layer of soil (0 to 20 cm) after removing the litter. They were immediately kept in sealed plastic bags and stored in a cooler box for further analysis.

2.5. Determination of Carbon Isotope Composition

The δ¹³C of water-soluble organic matter (WSOM) in leaves instead of bulk leaves was determined to represent the carbon isotope signals of photosynthates. Leaf WSOM has a high turnover rate [36,37,38] and mainly includes soluble carbohydrates, the hydrolysates of starch, and amino acids [39,40]. The extraction of WSOM referred to the protocols of [41]. Firstly, dried leaf material was ground, and 50 mg sample was put into a 2.5 mL centrifuge tube. After 1 mL of double-deionized water was added, the tube was agitated and then placed in a freezer for 1 h at 4 °C. Secondly, the samples were heated for 10 min at 95 °C, cooled to room temperature (RT), and centrifuged for 10 min at 12,000 g. Thirdly, an aliquot of 25 μL of the supernatant was transferred to a tin capsule and dried at 60 °C for 24 h. The δ¹³C of the WSOM was analyzed with an Isotope Ratio Mass Spectrometer (MAT-253, Thermo Fisher Scientific Inc., Waltham, MA, USA). Three leaf samples were measured per plant. The carbon isotopic ratio of the samples was calculated as δ¹³C (‰) = [(R_sample /R_standard) − 1] × 1000, where R_sample and R_standard are the ¹³C/¹²C ratio of the sample and Vienna Pee Dee Belemnite (VPDB), respectively. The standard deviation of repeated measurements for MAT253 was 0.1‰ for δ¹³C.

The δ¹³C of atmospheric and soil CO₂ was determined by TG-IsoPrime (GV. Instruments Ltd., Manchester, UK). The method for soil DIC extraction was slightly modified from that of [42]. Briefly, 100 g of soil was placed into a 500 mL glass bottle and then mixed with 200 mL of CO₂-free water. The mixture was homogeneously stirred for 5 min and then incubated at RT for 30 min. Subsequently, the supernatant was centrifuged for 15 min at 3500× g and then filtered through a 0.45 µm water-base membrane. The total alkalinity of the filtrate was detected by an Aquamerck alkalinity kit (MColortest, Merck KGaA, Darmstadt, Germany). At the same time, soil pH was measured using a pH meter (S210-B, Mettler Toledo Group Ltd., Zurich, Switzerland). The remaining filtrate was transferred to a 50 mL bottle and sealed with a rubber plug. The air inside the bottle was drawn by a vacuum pump. Subsequently, the filtrate of DIC was pretreated with an aliquot of 3 mL phosphoric acid to react with DIC and produce CO₂. The generated CO₂ was then processed with a custom-built vacuum extraction system, through which the CO₂ was purified. The δ¹³C of the CO₂ generated from soil DIC was then analyzed by MAT-252 (Finnigan, Bremen, Germany). The measurement precision for TG-IsoPrime and MAT252 were 0.1‰ and 0.01‰ for δ¹³C, respectively.

2.6. Quantification of the Utilization of Soil DIC by Plants

We accounted for two end members, namely atmospheric CO₂ and soil DIC, that might contribute to the constitution of newly formed photosynthates as well as its δ¹³C [26]. The δ¹³C of photosynthates (δ_A) solely assimilated from atmospheric CO₂ could be predicted with Equation (1) [43]:

$${\mathsf{\delta }}_{\mathrm{A}}=\frac{{\mathsf{\delta }}_{\mathrm{a}}-{\Delta }^{13}{\mathrm{C}}_{\mathrm{com}}}{1+{\Delta }^{13}{\mathrm{C}}_{\mathrm{com}}/1000}$$

(1)

where δ_a is the carbon isotope composition of atmospheric CO₂, whereas Δ¹³C_com is the comprehensive discrimination against ¹³C including the diffusion of CO₂ across the boundary layer, intercellular air space, mesophyllcell, as well as the contributions of respiration and photorespiration. Δ¹³C_com is calculated using Equation (2) [44]:

$${\Delta }^{13}{\mathrm{C}}_{\mathrm{com}}=\frac{1}{1-t}\left[ā\frac{c_{\mathrm{a}}-c_{\mathrm{i}}}{c_{\mathrm{a}}}\right]+\frac{1+t}{1-t}[a_{\mathrm{m}}\frac{c_{\mathrm{i}}-c_{\mathrm{c}}}{c_{\mathrm{a}}}+\frac{c_{\mathrm{c}}}{c_{\mathrm{a}}}({b_3}^{\prime }-\frac{{\alpha }_{\mathrm{b}}}{{\alpha }_{\mathrm{e}}}e\frac{R_{\mathrm{d}}}{V_{\mathrm{c}}}-\frac{{\alpha }_{\mathrm{b}}}{{\alpha }_{\mathrm{f}}}f\frac{F}{V_{\mathrm{c}}})]$$

(2)

where c_a, c_i, and c_c denote the CO₂ partial pressure in the ambient air of the cuvette, intercellular air, and chloroplast, respectively; a_m is the summed ¹²C/¹³C fractionation during the dissolution of CO₂ and liquid-phase diffusion (a_m = 1.8‰); b₃′ is the ¹²C/¹³C fractionation of Rubisco (b₃′ = 29‰, [45]); e is the ¹²C/¹³C fractionation for day respiration (e = 0, [46]); and f is the ¹²C/¹³C fractionation during photorespiration (f = 11‰, [47]); R_d, V_c, and F are nonphotorespiratory CO₂ released in the dark, Rubisco carboxylation rate for the fixed atmospheric CO₂, and photorespiratory rate, respectively; α_b = 1 + b₃′, α_e = 1 + e, and α_f = 1 + f; t and ā are the ternary correction factor and the weighed fractionation for CO₂ diffusion across the boundary layer and stomata, respectively, which are calculated as [32]

$$t=\frac{{\alpha }_{\mathrm{ac}}E}{2g_{\mathrm{ac}}}$$

(3)

$$ā=\frac{a_{\mathrm{b}}(c_{\mathrm{a}}-c_{\mathrm{s}})+a_{\mathrm{s}}(c_{\mathrm{s}}-c_{\mathrm{i}})}{c_{\mathrm{a}}-c_{\mathrm{i}}}$$

(4)

where α_ac = 1 + ā, E is the transpiration rate; $g_{\mathrm{ac}}$ is the combined boundary layer and stomatal conductance to CO₂; a_b and a_s are the ¹²C/¹³C fractionation for CO₂ diffusion across the boundary layer and through the stomata (a_b = 2.9‰, a_s = 4.4‰), respectively; and c_s is the CO₂ partial pressure at the leaf surface. c_c was not directly measured by LI-6400, but it was calculated with Equation (5) [37]:

$$A=g_{\mathrm{i}}(c_{\mathrm{i}}-c_{\mathrm{c}})$$

(5)

where g_i is an internal CO₂ conductance. In this study, we assumed g_i of 0.5 mol m⁻² s⁻¹ for woody species according to [48]. V_c and F are calculated as [44]

$$V_{\mathrm{c}}=A+R_{\mathrm{d}}+F$$

(6)

where

$$F=\frac{{\Gamma }^{*}\left(A+R_{\mathrm{d}}\right)}{c_{\mathrm{c}}-{\Gamma }^{*}}$$

(7)

where $\Gamma$^* is the CO₂ compensation point in the absence of R_d, which was fitted using Equation (8) [49]:

$${\Gamma }^{*}=e^{(c-\frac{\Delta H\ast 1000}{(R\ast (273+T_{leaf}))})}$$

(8)

where c, ΔH, and R are the scaling constant (13.49), energy of activation (24.46 kJ K⁻¹ mol⁻¹), and molar gas constant (0.008314 kJ J⁻¹ mol⁻¹), respectively.

When accounting for the involvement of soil DIC in leaf photosynthesis, the carbon isotopic composition of photosynthate (δ_A′) originating from the assimilation of soil DIC is given by

$${{\mathsf{\delta }}_{\mathrm{A}}}^{\prime }=\frac{{\mathsf{\delta }}_{\mathrm{DIC}}-{\Delta }^{13}{\mathrm{C}}_{\mathrm{DIC}}}{1+{\Delta }^{13}{\mathrm{C}}_{\mathrm{DIC}}/1000}$$

(9)

where δ_DIC is the carbon isotope composition of soil DIC collected from root zones of each plant; Δ¹³C_DIC is the photosynthetic discrimination against ¹³C combining the carboxylation of DIC and the effects of respiration and photorespiration, which is expressed as

$${\Delta }^{13}{\mathrm{C}}_{\mathrm{DIC}}={b_3}^{\prime }-\frac{{\alpha }_{\mathrm{b}}}{{\alpha }_{\mathrm{e}}}e\frac{R_{\mathrm{d}}}{{V_{\mathrm{c}}}^{\prime }}-\frac{{\alpha }_{\mathrm{b}}}{{\alpha }_{\mathrm{f}}}f\frac{F}{{V_{\mathrm{c}}}^{\prime }}$$

(10)

where V_c′ is the rate of Rubisco carboxylation for DIC and given by

$${V_c}^{\prime }=A_{\mathrm{DIC}\_\mathrm{soil}}+R_{\mathrm{d}}+F$$

(11)

where A_{DIC_soil} is the net photosynthetic rate for soil DIC. Empirically, A_DIC is less than 10% of A in many species. Here, we assumed an initial value for A_{DIC_soil} (A_{DIC_soil0}) to be A/10. Although this assumption introduced some errors, it had little effect on the calculation of Δ¹³C_DIC and the contribution of DIC to leaf total photosynthesis. Considering bicarbonate was the dominant species in soil DIC under high pH (7.09–7.75), δ_DIC in Equation (9) is derived from the δ¹³C of bicarbonate minus the discrimination related to the conversion between bicarbonate and CO₂ (e_b′, 9‰ at 25 °C).

However, studies showed that root-respired CO₂ (C__R) can also dissolve in the xylem sap [50]. C__R is primarily reported to range from 0.01 to 10.98 mM and varies with species, height, season, etc. [51,52]. Therefore, we proposed a ratio of the concentration of soil CO₂ or DIC (C__DIC) to C__R in the xylem, f_s/x = C__DIC/(C__DIC + C__R). In the present study, we directly measured C__DIC and adopted the mean value of 5 mM for C__R. Thus, the δ¹³C of the mixed DIC in the xylem sap (δ_{DIC_xylem}) is calculated as

$${\mathsf{\delta }}_{\mathrm{DIC}\_\mathrm{xylem}}=f_{\mathrm{s}/\mathrm{x}}·{\mathsf{\delta }}_{\mathrm{DIC}}+\left(1-f_{\mathrm{s}/\mathrm{x}}\right) {\mathsf{\delta }}_{\mathrm{R}}$$

(12)

where δ_R is the carbon isotope composition of root-respired CO₂ dissolved in the xylem sap. We assumed that the respiration of living tissues came from the recently fixed C, such as WSOM, which would carry the enriched isotopic signal to roots due to post-photosynthetic fractionation (−2‰, [53]) and then result in −1–−3‰ of respiratory fractionation in roots [54,55,56]. Here, we used the intermediate value −2‰ for respiratory fractionation; thus, δ_R is calculated as δ_R = δ_WSOM + 4‰ − a_m.

Considering that both atmospheric CO₂ and soil DIC could be used for leaf photosynthesis, the mixture of photosynthates from two substrates determined the value of δ_WSOM, which is expressed as a two-end-member mixing model:

$${\mathsf{\delta }}_{\mathrm{A}}^{\prime }·f_{\mathrm{DIC}\_\mathrm{xylem}}+{\mathsf{\delta }}_{\mathrm{A}}\left(1-f_{\mathrm{DIC}\_\mathrm{xylem}}\right)={\mathsf{\delta }}_{\mathrm{WSOM}}$$

(13)

where f_{DIC_xylem} is the proportion of xylem DIC contributing to δ_WSOM; δ_A′ is recalculated with Equation (9) but in which δ_DIC is replaced by δ_{DIC_xylem}. Finally, the contribution of soil DIC to the total leaf photosynthesis (f_{DIC_soil}) is calculated as

$$f_{\mathrm{DIC}\_\mathrm{soil}}=f_{\mathrm{s}/\mathrm{x}}·f_{\mathrm{DIC}\_\mathrm{xylem}}$$

(14)

2.7. Statistical Analysis

The comparison of mean values between species and altitudes was determined by one-way ANOVA with a t-test. Linear regression was used to investigate the relationship between f_{DIC_soil} and δ_A-δ_WSOM. Pearson’s correlation was conducted to evaluate the relationships between physiological or environmental factors and f_{DIC_soil}. All statistical analyses were performed using SPSS statistical software for Windows 21 (SPSS Inc., Chicago, IL, USA). The tests were considered significant at the p < 0.05 level. All data were expressed as the mean ± 1SE.

3. Results

3.1. Leaf Gas Exchange

All gas-exchange parameters exhibited significant differences within nine species (

Table 2. Leaf gas exchange of nine species across six plots. A, net photosynthetic rate; g_s, stomatal conductance; E, transpiration rate; c_i/c_a, ratio of intercellular to ambient partial pressure of CO₂; WUE_i, instantaneous water use efficiency. Values represent mean ± SE. N = 4 for each species in the specific plot. Capital letters indicate significant differences (p < 0.05) among species.

Species	A (μmol m−2 s−1)	gs (mol m−2 s−1)	E (mol H2O m−2 s−1)	ci/ca	WUEi (μmol H2O mol−1)
L. lucidum	11.07 (0.72) C	0.14 (0.01) D	3.38 (0.20) CD	0.60 (0.01) C	3.30 (0.14) BCD
B. papyrifera	18.24 (0.92) A	0.28 (0.01) A	3.67 (0.19) BCD	0.67 (0.01) AB	4.98 (0.13) A
P. longipes	9.99 (0.59) C	0.19 (0.02) BCD	3.81 (0.16) ABCD	0.72 (0.02) A	2.71 (0.23) D
Z. serrata	13.91 (0.27) B	0.21 (0.01) BC	3.95 (0.25) ABCD	0.67 (0.01) AB	3.61 (0.22) B
V. dilatatum	10.58 (0.47) C	0.18 (0.01) CD	3.59 (0.11) CD	0.70 (0.01) AB	2.96 (0.13) CD
A. delavayana	13.39 (0.47) B	0.24 (0.01) AB	4.58 (0.26) A	0.70 (0.01) AB	3.11 (0.18) BCD
R. cymosa	9.81 (0.50) C	0.16 (0.02) CD	3.23 (0.22) D	0.65 (0.03) BC	3.10 (0.11) BCD
Z. armatum	14.07 (1.03) B	0.21 (0.02) BC	4.18 (0.27) ABC	0.66 (0.02) AB	3.42 (0.24) BC
R. biflorus	13.54 (0.41) B	0.21 (0.01) BC	4.46 (0.13) AB	0.68 (0.01) AB	3.04 (0.10) BCD

). However, no clear trend was observed between two life forms (tree and shrub). The maximum and minimum values of A were 18.24 ± 0.92 μmol m⁻² s⁻¹ in B. papyrifera and 9.81 ± 0.92 μmol m⁻² s⁻¹ in R. cymosa, respectively. The g_s of nine species ranged from 0.14 to 0.28 mol m⁻² s⁻¹, suggesting mild-to-moderate drought stress. In contrast, there were small variations of E, c_i/c_a, and WUE_i among nine species, with mean values of 3.91 mol H₂O m⁻² s⁻¹, 0.67, and 3.27 μmol H₂O mol⁻¹, respectively.

3.2. Characteristics of Soil DIC and CO₂

The average concentrations of DIC (or C__DIC) in the rhizosphere of four tree species were similar (around 9.26 mM), whereas in five shrub species, they exhibited large variations, ranging from 8.10 mM in R. cymosa to 10.50 mM in Z. armatum (Figure 2A). In contrast, the mean concentrations of soil CO₂ varied from 3919.91 ppm in P. longipes to 5279.83 ppm in B. papyrifera, higher than the mean value of 3454.42 ppm in all shrub species. According to Henry’s law, the partial pressure of CO₂ over a solution was proportional to the concentration of the CO₂ in the solution. The calculated quantity of CO₂ dissolved in soil water ([CO₂^*]) was in the range of 0.55 to 0.92 mM at pH 7 and 25 °C across nine species, much lower than the mean value of [CO₂^*] in xylem sap (or C__R) investigated in many species. Thus, it was unlikely that soil CO₂ could be fixed by plant roots due to a CO₂ gradient from roots to the soil. In addition, the interspecies differences in the δ¹³C of soil DIC and CO₂ were both less than 2‰ (Figure 2B). The mean values of soil DIC and CO₂ among nine species were −10.19‰ and −20.01%, respectively.

3.3. δ¹³C of Photosynthates

The δ¹³C of the newly formed photosynthates in leaves could be determined by measuring the isotope signals of WSOM, or predicted with Eqn. 1 that only accounted for the assimilation of atmospheric CO₂ (Figure 3A). δ_WSOM varied all among species by about 3–4‰ and was generally lower than δ_A in all species. The discrepancy between δ_A and δ_WSOM (δ_A-δ_WSOM) was species-specific. Figure 3B displayed the distribution of δ_A-δ_WSOM in leaves of nine species under varying altitudes. The highest frequency (approximate 30%) of δ_A-δ_WSOM was located in the range of 1 to 2‰. Furthermore, half of the δ_A-δ_WSOM was distributed between 2‰ and 6‰. This phenomenon clearly illustrated the fact that there were other carbon sources, for example, soil DIC, which participated in leaf photosynthesis. The photosynthates originating from the mixture of soil DIC and xylem [CO₂^*] carried more ¹³C-depleted signal (δ_A′, see Supplementary Data), thus resulting in δ_WSOM systematically lower than δ_A.

3.4. Interspecies Difference in f_{DIC_soil}

The potential contribution of soil DIC to leaf total photosynthesis (f_{DIC_soil}) considerably varied with species across the six plots (Figure 4). In tree species, the average values of f_{DIC_soil} in L. lucidum and B. papyrifera were 8.93% and 9.54%, respectively, significantly higher than those in P. longipes and Z. serrata (both less than 2.2%). By contrast, the mean values of f_{DIC_soi} within the shrub species varied from 2.48% to 9.99%, with the highest value in Z. armatum and the lowest value in R. biflorus. Furthermore, f_{DIC_soil} strongly correlated with δ_A-δ_WSOM (p < 0.001), with the coefficient of determination (R²) as 0.99. For the nine species, a higher value of δ_A-δ_WSOM generally indicated a higher value of f_{DIC_soil} (Figure 5). This tight relation further demonstrated the plants’ use of soil DIC in the karst habitats. Additionally, f_{DIC_soi} significantly correlated with A, c_i/c_a, and WUE_i (

Table 3. Relationships between physiological or environmental factors and f_{DIC_soil} analyzed by Pearson’s correlation. N = 108.

Factors	fDIC_soil
	r	p
A	0.287	0.003
gs	−0.122	0.210
E	−0.096	0.325
ci/ca	−0.486	<0.001
WUEi	0.350	<0.001
C_DIC	0.053	0.586
δDIC	−0.100	0.306
δWSOM	−0.155	0.105

).

3.5. Variations in f_{DIC_soil} at Different Altitudes

The impact of altitude on f_{DIC_soil} was investigated in one tree (L. lucidum) and two shrub species (V. dilatatum and A. delavayana), which all appeared at three altitudes (Table 1; Figure 6). In L. lucidum and A. delavayana, f_{DIC_soil} tended to increase from lower to higher altitude (p < 0.05), whereas in V. dilatatum, the pattern was opposite.

4. Discussion

4.1. Isotope Evidence for the Utilization of Soil DIC by Karst Plants

Two issues needed to be addressed in field trials: (1) whether the natural abundance of ¹³C signals of soil DIC could be detected in leaves, and (2) what the major carbon source supplying the plant roots was. A few studies showed that even if the plants were irrigated with ¹³CO₂ or H¹³CO₃⁻ solution, the C isotope signals were too weak to be detected [20,57]. This is probably due to a low concentration of the labeled species dissolved in the solution (usually less than 1 mM), resulting in less ¹³CO₂ or H¹³CO₃⁻ taken up by roots as the diffusion against the concentration gradient from roots to the soil was minimal [15]. However, in the karst habitats, the high concentration of soil DIC made it possible to be fixed by plants. In addition, labeling experiments usually have a single type of carbon source [16,17,20], whereas, in field conditions, the major form of carbon source depends on soil pH, humidity, temperature, root respiration, microbial activity, etc. [58,59]. For this reason, we measured both the concentrations of soil DIC and CO₂ (Figure 2). The result suggested that soil DIC rather than soil CO₂ was the main source supplying the roots. However, direct evidence is needed to support this idea, especially in field conditions. Then we examined whether soil DIC could affect the δ_WSOM. The experimental data confirmed our speculation that δ_WSOM largely deviated from the predicted δ¹³C of photosynthates (δ_A) solely assimilated from atmospheric CO₂ across nine species (Figure 3), consistent with our previous finding [26]. As the values of δ_A-δ_WSOM were all positive and half of them were larger than 2‰ (Figure 3B and Figure 5), the isotope measurement error (less than 0.1‰) and diurnal variations in δ_WSOM (less than 1‰) [60] are not sufficient to explain the discrepancy. When combined with the very negative ¹³C signal of soil DIC (δ_A′, see Supplementary Data), this large and positive deviation of δ_A from δ_WSOM suggests the participation of soil DIC in leaf photosynthesis. Therefore, both atmospheric CO₂ and soil DIC contributed to the newly assimilated carbon (i.e., leaf WSOM). The contribution of soil DIC (f_{DIC_soil}) to leaf total photosynthesis could be quantified with a two-end-member mixing model (Equation (13)). We further showed that δ_A-δ_WSOM was closely linked to f_{DIC_soil} among all species (Figure 5). That is, the more the soil DIC was involved in photosynthesis, the more δ_A deviated from δ_WSOM, and the larger the proportion of soil DIC contributed to leaf WSOM.

However, uncertainties remain concerning the estimation of f_{DIC_soil}. For example, we considered the tissues-respired CO₂ dissolved in the xylem sap in the calculation of f_{DIC_soil}. Although we assigned a mean value of 5 mM for the C__R, species might differ in C__R due to their different physiological statues and anatomical characteristics [15,52,61]. We also assumed that the source of respired CO₂ came from the newly assimilated carbon and used the value of δ_WSOM to infer the δ¹³C of CO₂ respired by living tissues (δ_R). Nevertheless, recent studies [62,63,64,65] showed that vegetation pools respired carbon of a wide range of ages in many species. For instance, the respired carbon in stems and roots was, on average, older than 1 year [65,66]. When plants were exposed to stress, respiration almost completely relied on the old carbon [64]. Therefore, in this study, there was a high probability that the investigated species used reserved carbon for respiration under the condition of frequent water limitation in karst habitats. To the best of our knowledge, the δ¹³C of reserved carbon has a close relationship with that of respired CO₂ [56]. Our unpublished data showed that the maximum range of δ¹³C in the storage carbon of some karst-adaptable species was 4.13‰ during the growing season, which implied a similar or less variation in the range of δ¹³C for respired CO₂ in stems and roots. Some studies also showed that seasonal or annual variability of δ¹³C in respired CO₂ in stems and roots ranged within 2.0–4.5‰ [67,68,69]. In this study, taking L. lucidum for example, a change of 4.5‰ in δ¹³C of leaf WSOM only led to 0.5% of the change in f_{DIC_soil}, indicating a limited influence of the source of carbon used for respiration on the quantification of plants’ utilization of soil DIC.

4.2. Interspecies Difference in Plants’ Use of Soil DIC

An assessment of f_{DIC_soil} between two plant life forms, namely trees and shrubs, showed no clear trend (Figure 4), indicating that the variation of f_{DIC_soil} was mainly resulted from the interspecies differences. In most species, the values of f_{DIC_soil} were comparable with those reported in the laboratory experiments [24,25,26], suggesting that soil DIC was easily accessible to plants in the karst habitats. In this study, nine species were all native, most of which were deciduous trees or shrubs. Although species differed in height, biomass, ages, etc., our measurements and sampling were all conducted with newly expanded leaves and within the root zones of the same soil layer. As leaves and roots control the uptake of atmospheric CO₂ and soil DIC, respectively, the species-specific variation in f_{DIC_soil} might be related to leaf gas-exchange traits (Table 2) and soil conditions (Figure 2).

f_{DIC_soil} positively correlates with A (p = 0.003) and WUE_i (p < 0.001) but negatively with c_i/c_a (p < 0.001). At first sight, it was surprising that f_{DIC_soil} increased with an increase in A, which was inconsistent with the results reported by our previous study [26]. For instance, when plants were confronted with moderate or severe water limitation, it would drastically reduce A and g_s [7,29] and thus increased the proportion of root or soil-derived DIC to support photosynthesis [26,29]. The fact was that the change of A was not proportional to that of A_{DIC_soil} or E [26], thus arithmetically promoting the value of f_{DIC_soil}. In this study, the level of g_s indicated mild drought stress in many species (Table 2). Among these species, higher A corresponded to higher g_s and E, which benefited the long-distance transport of DIC in the xylem sap [70,71]. Similarly, higher WUE_i also implied a higher contribution of soil DIC to leaf photosynthesis, although the correlation between E and f_{DIC_soil} was not significant (Table 3). The negative correlation between f_{DIC_soil} and c_i/c_a corresponded to the reverse change of f_{DIC_soil} and the contribution of atmospheric CO₂ to leaf total photosynthesis (1-f_{DIC_soil}) because higher c_i/c_a usually suggested higher Δ¹³C_com (Equation (2)) and lower A and E [32]. In addition, there was no significant correlation between f_{DIC_soil}, C_{_DIC}, and δ_DIC. This could be explained by less variation of these indicators in comparison with that of leaf gas-exchange parameters among species (Figure 2; Table 2).

4.3. Effect of Altitude on Plants’ Use of DIC

As the altitude increases, the number of tree species reduces and that of shrub species increases (Table 1), suggesting a change in the plant community. In this study, only three species presented in all plots and thus were available for evaluating the effect of altitude on f_{DIC_soil} (Figure 6). Among these species, f_{DIC_soil} of L. lucidum and A. delavayana increased from the lower to higher altitude, whereas in V. dilatatum, f_{DIC_soil} decreased along the altitudes. The impact of altitude on f_{DIC_soil} was a little bit confusing. However, altitude did not directly affect f_{DIC_soil}, but through imposing influences on site-characteristic related microclimate and soil conditions [35,72]. For example, the altitude could constrain daily mean temperature and vapor pressure deficit [72], and produce considerable variations in soil moisture and nutrient availability [35]. However, in the present study, the air temperature and relative humidity recorded by LI-6400 did not linearly change with altitude (see Supplementary Data) but exhibited a high degree of spatial heterogeneity. The reasons might be that the altitude gradient was not large enough, and some occasional factors, e.g., wind, shading, and vegetation coverage, could redistribute these resources, e.g., light and vapor [72,73]. Our previous study reported decreasing patterns for soil water content, organic matter, and some nutrients with the altitudes in the same area and same season of 2015 [34]. However, the increasing trend of f_{DIC_soil} along the altitude gradient in L. lucidum and A. delavayana was opposite to that in V. dilatatum (Figure 6), implying that the differences in soil conditions could not solely explain the variation of f_{DIC_soil}. Therefore, we speculated that all these variabilities were combined to influence the leaf gas-exchange and subsequent estimation of f_{DIC_soil}. Future studies may choose a large range of environmental gradients and exclude the influences of some occasional factors.

5. Conclusions

In the present study, we aimed to investigate whether soil DIC could be assimilated through leaf photosynthesis in the karst habitats, and whether this process, if taken place, varied between species and altitudes. Firstly, we showed that all plots in the study sites had a high content of soil DIC in the root zones of nine species. This implied a high possibility that soil DIC could be fixed by karst plants through photosynthesis. Secondly, we observed that there were large discrepancies between the measured and predicted δ¹³C of newly formed photosynthates (δ_A-δ_WSOM), and this systematic difference could not be explained by measurement errors or diurnal variations in δ_WSOM alone. Therefore, we accounted for the involvement of soil DIC in photosynthesis and thus deviating δ_WSOM from δ_A. Thirdly, we applied a two-end-member mixing model to estimate the contribution of soil DIC to leaf total photosynthesis (f_{DIC_soil}). The values of f_{DIC_soil} largely differed among species and some of which were influenced by altitudes. The present study improved our understanding of how plants adapted to alkaline soil conditions of karst habitats and acquired additional carbon for growth.