Changes in levels of enzymes and osmotic adjustment compounds in key species and their relevance to vegetation succession in abandoned croplands of a semiarid sandy region

Abstract Reclamation of cropland from grassland is regarded as a main reason for grassland degradation; understanding succession from abandoned cropland to grassland is thus crucial for vegetation restoration in arid and semiarid areas. Soil becomes dry when cropland is reverted to grassland, and enzyme and osmotic adjustment compounds may help plants to adapt to a drying environment. Croplands that were abandoned in various years on the Ordos Plateau in China, were selected for the analysis of the dynamics of enzymes and osmotic adjustment compounds in plant species during vegetation succession. With increasing number of years since abandonment, levels of superoxide dismutase increased in Stipa bungeana, first decreased and then increased in Lespedeza davurica and Artemisia frigida, and fluctuated in Heteropappus altaicus. Levels of peroxidase and catalase in the four species fluctuated; levels of proline, soluble sugar, and soluble protein either decreased or first increased and then generally decreased. According to a drought resistance index, the drought resistance of the four species was ranked in descending order as follows: S. bungeana > A. frigida > H. altaicus > L. davurica. The drought resistance ability of the different species was closely linked with vegetation succession from communities dominated by annual and biennial species (with main accompanying species of L. davurica and H. altaicus) to communities dominated by perennial species (S. bungeana and A. frigida) when soil became dry owing to increasing evapotranspiration after cropland abandonment. The restoration of S. bungeana steppe after cropland abandonment on the Ordos Plateau is recommended both as high‐quality forage and for environmental sustainability.


| INTRODUC TI ON
Approximately 10%-20% of arid and semiarid areas throughout the world are changing dramatically as a result of desertification and grasslands degradation (Reynolds et al., 2007). There are two main causes of desertification: global factors, such as global climate change and rising global CO 2 concentrations, and local factors, such as chronic high levels of herbivory, overgrazing, grassland reclamation, and combinations of these factors (Belayneh & Tessema, 2017;Reynolds et al., 2007). To effectively restore degraded grassland resulting from reclamation, vegetation succession after cropland abandonment has been a key focus of scholars in the 21st century (Kawada, Wu, & Nakamura, 2015;Ward, Hoffman, & Collocott, 2014). In arid and semiarid areas, abandoned croplands usually revert to three community types: weeds, perennial grass with shrubs or native grass (Cai et al., 2018;Nie, Yuan, Kepner, Erickson, & Jackson, 2012;Reynolds et al., 2007;Wheeler, Archer, Asner, & Mcmurtry, 2007); therefore, interactions between shrub and grass species are crucial for driving vegetation succession of abandoned croplands (Howard, Eldridge, & Soliveres, 2012;Reynolds et al., 2007).
Bottom-up controls (resources such as water and nitrogen) and top-down controls (such as fire and herbivory) are commonly used to explain shrub and grass interactions (Cipriotti, Aguiar, Wiegand, & Paruelo, 2014). Niche separation, one of the bottom-up hypotheses, assumes that shrubs have deeper roots and use deeper soil water compared with grasses; therefore, deeper infiltration of soil water favors shrubs in arid areas (Cipriotti et al., 2014;O'Connor, Puttick, & Hoffman, 2014). The two-layer hypothesis states that the selective use of deep soil water increases the persistence of shrub species in arid areas, and the resource pool hypothesis proposes that shallow soil water benefits the growth of all plants and that deep soil water benefits only shrub species (O'Connor et al., 2014). Although most hypotheses suggest that water use strategies of grass species and woody species play key roles in the vegetation succession process in degraded grasslands, there are few studies that directly observe the long-term water use strategies of multiple woody and grass species.
Aerobic metabolism can provide energy for plant growth and development but also generates reactive oxygen species (ROS).
Normally, there is a dynamic balance between the intracellular generation and removal of ROS. When plants are under drought stress, the balance of ROS is disrupted, and cells are injured by excessive ROS (Fang & Xiong, 2015). To protect cells from the damage caused by excessive ROS, plants have developed a series of defense mechanisms that make use of antioxidant enzymes such as superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT) (Gill & Tuteja, 2010).
Superoxide dismutase is considered the first line of defense against ROS (Sayfzadeh et al., 2011). SOD is a major scavenger of superoxide anion radicals, and it can catalyze into hydrogen peroxide and oxygen (Scandalios, 1993). POD plays a key role in preventing oxidative damage; it can decrease H 2 O 2 accumulation and maintain cell membrane integrity (Chakhchar et al., 2015) because of its low substrate affinity (Weng, Cui, Liu, Zhang, & Shan, 2015). CAT is another effective enzyme that can break down H 2 O 2 (Chakhchar et al., 2015;Weng et al., 2015); it degrades H 2 O 2 and prevents the subsequent degradation of fatty acids by peroxisomes (Gill & Tuteja, 2010). Malondialdehyde (MDA) is the product of membrane lipid peroxidation and is an indicator of oxidative damage (Sedaghat, Sarvestani, Emam, & Bidgoli, 2017). MDA is also one of the main compounds in drought-damaged plants (Xu, Zhou, Wang, Han, & Li, 2008).
Another significant drought resistance strategy is osmotic regulation (Deligoz & Gur, 2015;Fang & Xiong, 2015). When exposed to drought, plants accumulate osmotic solutes (soluble sugar [SS] and soluble protein [SP]) to increase the liquid concentration of cells, reduce the osmotic potential to facilitate water absorption, maintain cell turgor, and ensure cell growth and metabolism (Morgan, 1984). When soil moisture levels change markedly, changes in osmotic adjustment compounds that help plants adapt to such changes may provide clues to understanding shrub and grass interactions.
The Ordos Plateau is a typical Chinese ecological transition zone that may be classified as an agriculture and animal husbandry ecotone, and because of the influence of overgrazing, land reclamation, and climatic change, the zonal vegetation (Stipa bungeana) has been severely degraded, and a semishrub species (Artemisia ordosica) has become dominant (Cai et al., 2018). There are some reports related to grass and shrub interactions, such as grass or shrub dominance are at alternative stable states in the processes of vegetation succession (Peng et al., 2013), climate and native grassland vegetation significantly influenced the community structures of grasslands with high abundance of shrub , cover of shrub patches significantly influenced the community structures, evapotranspiration and soil organic carbon of grassland ecosystems (Li et al., 2019;Wang et al., 2018;Zhou et al., 2018). Grassland degradation on the Ordos Plateau might have originated during the Yuan Dynasty (600-700 years ago), and during this long history, the plant community composition and soil physical and chemical properties have changed to varying degrees; therefore, the Ordos Plateau may be the ideal location for studying the interactions between shrubs and grass (Cai et al., 2018). Field observations have indicated that after abandonment, croplands reclaimed by semishrub dominated grasslands could be restored to native S. bungeana grasslands through long-term vegetation succession from bare land, to annual, biennial, semishrub, and then perennial species. During this successional process, at the first stage, when annual grasses dominate, the soil water content is relatively high, and at the last stage, when perennial grasses are dominant and community cover is greater, the soil water content is relatively low; therefore, in this semiarid environment, the dominant species in different stages will have different abilities to resist drought (Cai et al., 2018), and enzymes and osmotic adjustment compounds may play an important role in the abilities of species to maintain drought resistance. Plants resist drought through four basic mechanisms: drought avoidance, drought tolerance (DT), drought escape, and drought recovery (Fang & Xiong, 2015). DT refers to the ability of plants to maintain a normal level of physiological and biochemical activities while under water stress by adjusting a series of metabolic pathways to reduce damage from drought stress (Passioura, 1998). To cope with the damaging effects of drought, plants have developed several DT-associated strategies, which mainly include systems of antioxidant enzymes (such as SOD, POD, and CAT) and osmotic responses (such as changes in the levels of proline and SS) (Fang & Xiong, 2015;Ge et al., 2014). However, the role of these enzymes and osmotic adjustment compounds in plant species in this type of succession is still unclear.
We hypothesized that S. bungeana in later successional stages has higher drought tolerance than does Heteropappus altaicus in early successional stages. To test this hypothesis, we chose croplands on the Ordos Plateau of Inner Mongolia that had been aban-
The research site is in Ejin Horo Town, in the middle of Ejin Horo Banner, in abandoned cropland at 39°24-39°26′N and 109°50′-109°52′E on the Ordos Plateau. At the research site, annual average temperature is 6.2°C, the annual maximum temperature is 35.9°C (in July), and the annual minimum temperature is −20.3°C (in January).
Annual precipitation is 357.3 mm. Annual sunshine duration is 2,828.7 hr. Before the croplands were abandoned, the zonal vegetation was a warm-temperate S. bungeana steppe. Corn was planted in croplands; manure and urea were used, but no pesticide was used.
After abandonment, the abandoned croplands were succeeded by three community types: weeds, perennial S. bungeana with Artemisia frigida, and S. bungeana (Cai et al., 2018). Grazing was halted for vegetation restoration; grazing only happened occasionally and with very low grazing density.  POD, CAT, and MDA assays, the homogenates were centrifuged at 8,000 g at 4°C for 10 min. For the Pro assay, the homogenates were shaken in a boiling water bath (90°C) for 10 min, cooled and then centrifuged at 1,000 g at 25°C for 10 min. For the SP assay, the homogenates were centrifuged at 10,000 g at 4°C for 10 min. For the SS assay, the leaf samples that had been maintained at −80°C were ground into a powder in liquid nitrogen, then 0.1 g leaf tissue was homogenized with 1 ml distilled water and maintained in a boiling water bath for 10 min. After cooling, the mixture was centrifuged at 8,000 g at 25°C for 10 min to produce the supernatant.

| Experimental design
In each 20 m × 20 m leaf-sampling plot, quadrats covering 5 m × 5 m for semishrub species and 1 m × 1 m for grass species were established. To ensure accurate measurements of biomass, leaves were not collected in these quadrats for analyses of enzyme activity, MDA, or osmotic adjustment compounds. To obtain the aboveground biomass, the aboveground parts of every species were harvested separately, taken to the laboratory and dried to a constant weight at a temperature of 80°C (Cai et al., 2018).
In each 20 m × 20 m leaf sample plot, intact soil cores were collected randomly using a cutting ring (volume of 100 cm 3 ) from five soil depths (0-5, 5-10, 10-20, 20-30, and 30-40 cm) after removing any rocks and litter. After collecting the soil samples, we immediately measured the fresh weight (FW), and then, the samples were taken to the laboratory and oven-dried at 105°C to a constant weight to measure the dry weight (DW). Soil water content (SWC) was calculated as

| Statistical analysis
To evaluate the drought resistance ability of a species, principal component analysis (PCA) was used to develop an index (Wold, Esbensen, & Geladi, 1987). Averaged data from three replicates for each of the Within each of three replicates, samples were collected from three subplots; data from three subplots were averaged as one datum point for each replicate. A statistical analysis was performed by twoway ANOVA. If significant differences were found, Duncan's test was used to determine mean differences between treatments (p < .05) (Kabacoff, 2015). The relationships among SOD, POD, CAT, MDA, Pro, SS, and SP were examined using Pearson's correlation analysis. All statistical analyses, including the test for homogeneity of variance, were performed using SPSS Statistics 17.0 (SPSS Inc.).

| Changes in the levels of enzymes, MDA, and osmotic adjustment compounds in different species and their correlations
In general, F values were significant for the responses of SOD, POD, CAT, MDA, Pro, SS, and SP to species, years since abandonment and their interactions (Table 1).
Different species showed different patterns with increasing number of years since abandonment. In S. bungeana, SOD increased, Pro increased and then decreased, SS decreased and then increased, and SP showed a decreasing trend. In L. davurica, SOD decreased and then increased, while Pro, SS, and SP increased and then decreased or slowly decreased. In A. frigida, SOD decreased and then increased, while Pro, SS, and SP increased and then decreased. In H. altaicus, SOD fluctuated, while Pro, SS, and SP increased or decreased and then increased. POD, CAT, and MDA fluctuated in these four species (Table 2, Table A3, and  (Table 3).

| Change in levels of enzymes, MDA, and osmotic adjustment compounds in different plant functional types
Levels of POD, CAT, and Pro were significantly higher for grasses than they were for semishrubs (p < .05), while there were no significant differences between grasses and semishrubs in SOD, MDA, SS, or SP. The (1) F I G U R E 1 Changes in MDA, Pro, SS, and SP (mean ± SE). Each symbol in lines represents the mean of three replicates; results of multivariable comparison are shown in Table A3. Abbreviations are shown in Table 1 average levels of POD, CAT, and Pro for grasses were 689.13, 4.16, and 2.73 times higher than those of semishrubs, respectively ( Figure 2).

| Change in aboveground biomass
With increasing number of years since abandonment, aboveground biomass increased for S. bungeana, increased and then decreased for L. davurica and A. frigida, and declined for H. altaicus (Table 4).

| Comprehensive drought resistance index (D)
The contribution rate was over 90% for the first two principal components, which contained almost all information from the measured parameters. The eigenvalue of the first principal component was 4.582, its contribution rate was 65.5%, and the corresponding characteristic vectors of SOD, POD, CAT, and SS were large. The eigenvalue of the second principal component was 1.809, its contribution rate was 25.8%, and the corresponding characteristic vectors of SP and MDA were large (Table 5, Table A1).
Based on the standardized characteristic vector, the linear combinatorial equations of the first two principal components and the seven parameters were obtained: where X 1 , X 2 , … X 7 were SOD, POD, CAT, MDA, Pro, SS, and SP, respectively.

| Coefficients of variation of different parameters of the four species
For the seven parameters measured, the coefficients of variation (CVs),  (Table 6).

| Effects of enzyme activity, MDA, and osmotic adjustment compounds
Studies have shown that when plants suffer from drought stress, the antioxidant enzyme activity increases, but when the degree of stress exceeds a certain threshold, the antioxidant enzyme activity decreases (Chakhchar et al., 2015;Ge et al., 2014). Excess active oxygen species caused by water stress can cause membrane F I G U R E 2 Changes in MDA, Pro, SS, and SP (mean ± SE) in two plant functions (grass and semishrub). Each bar represents the mean of three replicates; bars with different lowercase letters are significantly different from each other with different plant functions at p < .05 (Duncan's test). Abbreviations are shown in Table 1 lipid peroxidation. MDA can be an indicator of oxidative damage (Sedaghat et al., 2017). Many studies have shown that when plants are subjected to water stress, MDA levels increase (Jia, Sun, Li, Li, & Chen, 2015), and low MDA is associated with drought resistance, which can improve plant growth (Bacelar et al., 2007). In addition, water stress can increase the accumulation of osmotic regulators (Jia et al., 2015), but there have also been several studies that have found no increase in levels of osmotic regulators (Deligoz & Gur, 2015).
In our study, antioxidant enzyme activity varied among durations since abandonment and among species, indicating that the species may encounter different degrees of drought stress and that soil water content generally declines after abandonment because different species may consume different amounts of soil water (Table A4). The roots of the perennial grass in this study were mainly located in the 0-10 cm soil layer, and soil bulk density decreased with increasing time since abandonment, which was unfavorable for the use of water for deep-rooted species.  Abbreviations are shown in Table 1 TA B L E 6 Coefficients of variation for four species indicating that MDA was linked to the increases in antioxidant enzyme activity and antioxidant ability, the elimination of ROS, and the decreased membrane lipid peroxidation damage (Esfandiari, Shekari, Shekari, & Esfandiari, 2007;Ge et al., 2014). Levels of Pro and SP showed a clear correlation, and the CVs of Pro and SS were large, indicating S. bungeana was sensitive to water stress.
Pro may play a different role in drought resistance mechanisms; it can act as a scavenger of free radical species to protect cells from oxidative damage (Girija et al., 2002) or as a reserve for plant assimilation of nitrogen and carbon after stress (Silveira et al., 2003). Decreases in Pro levels in our study may be due to its role in the transition between drought resistance mechanisms (Silva, Ferreira-Silva, Viégas, & Silveira, 2010).

| Drought resistance of the four species and implications
The four species differed in their resistance to drought. The various parameters showed that the four species had different mechanisms of drought resistance. S. bungeana is a perennial grass species with a root system mainly distributed in the 0-10-cm soil layer. A. frigida is a perennial grass species or occasional small semishrub species with a deep root system that reached the 1-m soil layer and was mainly distributed in the 30-cm soil layer. L. davurica is a semishrub species with a deep root system that reached the 1.3-m soil layer and was mainly distributed in the 30-cm soil layer. H. altaicus is a perennial grass species with a deep root system that reached the 1-m soil layer and was mainly distributed in the 0-20-cm soil layer. Antioxidant enzyme activity was higher in S. bungeana than in the other species, which indicated that S. bungeana had strong drought resistance.
The Pro level was higher in A. frigida than in other species, which indicated that A. frigida was more tolerant to water stress and had higher drought resistance (Siddiqui et al., 2016 (Cai et al., 2018). At the initial stage after cropland abandonment, community coverage and transpiration were low, and soil water was relatively high, especially in the deep layer in sandy soil. The soil gradually improved from the first stage to the last stage after cropland abandonment as fine clay particles accumulated in the surface soil layer, and seepage of rain to deep soil decreased.
Therefore, the soil water content increased in the shallow soil layers and decreased in the deep soil layers. This change is beneficial for shallow-rooted species but not for deep-rooted species, as suggested by the two-layer hypothesis (Cipriotti et al., 2014). This pattern was also evident in this study; the aboveground biomass of S. bungeana increased, that of H. altaicus decreased and that of L. davurica was relatively stable, with a decreasing trend for that of A. frigida (Table A4). The results of this study clearly indicate why the four species differ in drought resistance abilities, mainly because of their long-term adaptation to different soil water environments, and our results were also consistent with the two-layer hypothesis (Cipriotti et al., 2014).
Based on the trends of soil changes and drought resistance for the studied species, recommended practices for restoring S. bungeana steppe for high forage value and favorable environmental effects include reducing soil disturbance, accelerating the accumulation of fine clay particles, improving the soil structure, and increasing the soil water in the shallow layer.

| CON CLUS IONS
Based on analyses of enzymes, MDA, osmotic adjustment compounds, and drought resistance, we found that the four studied species

ACK N OWLED G M ENTS
This work was supported by the National Natural Science Foundation of China (grant number 41330749).

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
Liu Yang and Yuanrun Zheng designed the work and the experi-