Ecological differentiation in Rumex crispus L. natural populations in metal mining areas

Background: Variations in phenotypic traits of various plants living in either normal or stressed environments have been well studied, but ecological responses of plants to long-term persistent toxic metal pollution have little been reported. In this study, in order to explore the effects of continuous metal pollution in soil on variation and differentiation in the plants, Rumex crispus L. populations exposed to different levels of long-term persistent toxic metal pollution were studied, and corresponding R. crispus populations that had not been exposed to pollution were used as controls. Results: Six phenotypic traits of R. crispus —root diameter, leaf area, leaf length, leaf width, leaf perimeter, and leaf length-to-width ratio—differed significantly among and within populations. Traits ranked in descending order of coefficient of variation were leaf area, leaf perimeter, root diameter, leaf length, leaf width, leaf length-to-width ratio. The average coefficient of variation was 46%. Phenotypic variation in R. crispus was much greater among populations (92.69%) than within populations (6.55%). The mean phenotypic differentiation coefficient (Vst) of 93.37% indicates that the interpopulation variability was the main source of phenotypic variation in R. crispus . Finally, root diameter was significantly positively correlated with metal factors, but leaf area, leaf length, and leaf aspect ratio were significantly negatively correlated with Pb, Zn, Mn, and Fe contents. Overall, underground growth is superior to aboveground growth in populations that have experienced long-term exposure to toxic metal pollution, and there were phenotypic differences between uncontaminated and contaminated populations. Conclusions: These results indicate that R. crispus adapts to the heterogeneous environment caused by toxic metal pollution through rich phenotypic variation, and ecological differentiation has occurred among different populations.

The impact of long-term persistent heavy metal pollution on organisms, in particular on genetic diversity, is currently an important issue in both pollution ecology and environmental biology.
Genetic diversity is an important part of biodiversity. Research methods for determining genetic diversity include morphology, cytology, biochemistry, and the use of molecular markers. Phenotypic shape is an indirect estimate of genetic diversity. Variation in phenotypic traits is of great significance for both adaptability and evolution, and studying phenotypes can help researchers more intuitively predict both population genetic structure and the number of evolutionary adaptations [4].
Understanding the effects of environmental factors on plants through the study of plant morphology can also reveal the way in which plants adapt to heterogeneous environments and the relationships between plants and natural factors. Wang [5] found that long-term persistent toxic metal pollution changes plants' morphological structure and promotes ecological differentiation. In recent years, research on phenotypic variation in natural populations has mostly focused on species affected by geography, climate, and physical and chemical properties of the soil [6][7][8][9]. Other research has included basic studies on phenotypic diversity in plant resources for conservation [2,10] and toxicity.
There are few reports on ecological differentiation caused by metal pollution.
The Yunnan Province of China is known as the "The Kingdom of Non-ferrous Metals," and toxic metal contamination in the soil is increasing because of mining activity [11]. The city of Gejiu has a wide variety of mineral resources, and heavy metal pollution in mining process is more complicated in this area than in other parts of the country. The complexity of the city's toxic metal pollution makes this region an ideal area for studying the relationship between plants and soil contaminated with toxic metals. Rumex crispus is a perennial wild plant that grows well in mining areas. Xue and Liu [12] found that R. crispus is resistant to a variety of heavy metals, including certain accumulations of Pb, Zn, and Cu, but a poor accumulation of Cd, which can be used for soil remediation and restoration in mining areas. However, the adaptation of R. crispus to persistent toxic metal pollution and the ecological implications of phenotypic variation for characterizing and indicating the long-term persistent toxic metal pollution environment has not been studied.
In this study, in order to explore the ecological differentiation and the ecological response of plants to long-term persistent toxic metal pollution, we investigated differentiation in Rumex crispus populations exposed to long-term toxic metal pollution and those that had not been exposed to pollution. We measured phenotypic traits of each R. crispus population and levels of metal in the soil and used nested analyses of variance, coefficients of variation, correlation analyses, and cluster analyses to analyze phenotypic variation within and among populations. We also explored correlations between plant phenotypic variation and metal factors to better understand plant strategies for adapting to polluted environments.

Materials And Methods Sampling
Our study area was different metal mining areas in the city of Gejiu in southwestern China. Gejiu is located in a low-latitude subtropical high-prototype humid monsoon climate zone with an average annual rainfall of 2026.5 mm and an average annual temperature of 16.4 °C. In April 2019, seven populations around these mines that had experienced continuous toxic metal pollution were investigated, and two populations in the neighboring city of Mengzi that had not been exposed to pollution were selected as controls (Table 1). We first measured latitude and longitude with GPS and measured altitude. Depending on the size of the original population, 7 to 13 individuals from each population were randomly selected for inclusion in the study. Three basal leaves were selected from each R. crispus plant, and leaf traits were measured with a portable handheld laser leaf area meter (CL-203, American CID Corporation). The root diameter where the root is connected to the stem of each plant was measured with Vernier calipers. We took mixed soil samples at a depth of 0-30 cm near the collected plants and brought these samples to the laboratory for analyses.

Determination Of Soil Environmental Factors
The collected soil samples were first weighed to determine soil moisture, then air-dried in an airdrying room. When the samples were completely dry, they were ground up. Then they were mixed thoroughly on a colorless polyethylene film, passed through a 2 mm sieve and a 0.15 mm sieve, and put into Teflon bags until their indicators could be measured.
Determination of metal content of the soil: Air-dried soil samples passed through a 0.15 mm sieve were placed in a sample preparation mold, and metal contents were measured with X-ray fluorescence spectrometry (Vanta handheld portable X-ray fluorescence analyzer, Olympus, USA).
Each test was set up so that three parallel samples were measured simultaneously. Measurements were repeated three times, and the detection time was set above 120 s.

Calculation of the coefficient of variation and phenotypic differentiation coefficient
We calculated the coefficient of variation (CV) of phenotypic traits according to the following formula: CV = SD/Mean. In this formula, SD is standard deviation.
The phenotypic differentiation coefficient was used to further measure the degree of phenotypic differentiation among populations according to the following formula: V st = δ 2t/s / (δ 2t/s + δ 2s ). In this formula, V st is the phenotypic differentiation coefficient, δ 2t/s is the variance among populations, and δ 2s is the variance within populations.

Phenotypic variation in Rumex crispus
The phenotypic traits of R. crispus differed significantly among populations (Table 3). Leaf area, leaf length, leaf width, leaf perimeter, and root diameter were largest in the WC population, followed by the CC population. The CC population had the thickest roots, followed by the WC population. The SH population had the smallest and most narrow leaves, the shortest perimeter, the largest leaf aspect ratio, and the thinnest roots. The leaf shape for this population was closer to elliptic. The Y population had the shortest leaf length and smallest leaf aspect ratio, and the leaf shape was closer to round.
The coefficient of variation quantifies the degree of variability in a trait. The larger the coefficient of variation, the more discrete the trait ( Table 4). The average coefficient of variation among the six phenotypic traits was 46% (range = 23-81%). The traits, listed from greatest to least variation, were on the order leaf area, leaf perimeter, root diameter, leaf length, leaf width, and leaf length-to-width ratio. The least variation was in the leaf length-to-width ratio (CV = 23%), and the largest was in leaf area (CV = 81%), which was about 3.5 times the leaf length-to-width ratio. We found great differences   Phenotypic Differentiation And Sources Of Variation The variance components for phenotypic traits within and among populations differed markedly (Table 5). The average variance component percentages for the six phenotypic traits among and within populations were 93.17% and 6.09%, and the remaining 0.74% were from individuals. The mean phenotypic differentiation coefficient was 93.84%, and the phenotypic differentiation coefficient of each phenotypic trait was greater than 80%. Correlations between phenotypic traits of Rumex crispus and soil environmental factors Root diameter was positively correlated with all metals in the soil (p < 0.05; Table 6). Leaf area, leaf length, and leaf length-to-width ratio were negatively correlated with Pb, Zn, Mn, and Fe. Leaf perimeter was negatively correlated with Pb, Zn, and Mn (p < 0.05), but positively correlated with Sn, Cu, and As (p < 0.05). As and leaf length were also positively correlated (p < 0.05). Correlations between traits and metals, listed from greatest to least, were on the order root diameter, leaf length-to-width ratio, leaf length, leaf perimeter, leaf area, and leaf width. Leaf phenotypic traits were negatively correlated with pH (p < 0.05) and positively correlated with organic matter, total phosphorus, and total nitrogen (p < 0.05). Root diameter was negatively correlated with total nitrogen (p < 0.05). Correlations between traits and physical and chemical properties of the soil, listed from greatest to least, were on the order leaf length, leaf area, leaf perimeter, leaf width, root diameter, and leaf aspect ratio.
The comprehensive correlation between phenotypic traits and soil factors was the sum of the absolute values of the significant correlations of each trait, listed from greatest to least: organic matter, Zn, Mn, Pb, Fe, pH, total nitrogen, total phosphorus, As, Sn, Cu, and soil moisture. Metal content related to phenotypic traits, listed from largest to smallest, was on the order Zn, Mn, Pb, Fe, As, Sn, and Cu. Relations between physical and chemical properties of the soil and phenotypic traits, listed from largest to smallest, were on the order organic matter, pH, total nitrogen, total phosphorus, and soil moisture.

Clustering Of Populations Based On Phenotypic Traits
The nine R. crispus populations clustered based on the six phenotypic traits. After dividing by the Euclidean distance of 2.5, we divided the nine R. crispus populations into three groups (Fig. 1) not strictly clustered by geographic distance. The two populations not exposed to pollution, SH and XG, were a group.

Discussion
Ecological differentiation in Rumex crispus after exposure to persistent toxic metal pollution The phenotypic traits of the R. crispus population contaminated by long-term toxic metal pollution in this mining area were quite different from those of populations that had not been exposed to pollution ( Fig. 1). Phenotypic variation in plants is inseparable from their genes and the environment in which they grow. Rich morphological variation in plants is a manifestation of their ability to adapt to heterogeneous environments [2]. Plant leaves are more sensitive to different environments than other traits. Leaves are crucial to photosynthesis, respiration, and water exchange. They are the starting point for material circulation and energy flow. Leaves can directly affect the basic behavior and function of plants and can be a main indicator of plant genetic variation, which in turn reflects the survival strategies plants use to adapt to environmental changes [13,14].
The six phenotypic traits we studied (leaf area, leaf length, leaf width, leaf perimeter, leaf length-towidth ratio, and root diameter) across the nine R. crispus populations differed significantly among populations, which indicates that this plant has abundant phenotypic variation. The coefficient of variation measures the degree of variation in a trait. A larger coefficient of variation indicates greater variation and more abundant diversity. The average coefficient of variation for the six phenotypic traits was 46%, and variation in all phenotypes ranged from 26% to 51%, which indicates that phenotypic variation in R. crispus populations was great, and these plants are highly adaptable to heterogeneous environments.
The six phenotypic traits of R. crispus, listed from greatest to least coefficient of variation, were on the order leaf area, leaf perimeter, root diameter, leaf length, leaf width, and leaf length-to-width ratio. The leaf length-to-width ratio had the least variation within populations (CV = 24%) and the most stability, but differences among populations were significant in analyses of variance, which indicates that the trait is variable among populations but stable within populations. Leaf area had the largest variation (CV = 81%), which shows that it has the least stability but high plasticity and thus is susceptible to the influence of individual development and environment. It differed significantly within and among populations.
Reasons for ecological differentiation in Rumex crispus populations after exposure to continuous toxic metal pollution A total of 93.17% of the variation in the six phenotypic traits measured in these nine R. crispus populations originated from interpopulation variation, which was much higher than intrapopulation variation (6.09%) or intraindividual variation (0.74%). The mean phenotypic differentiation coefficient was 93.84%, which indicates that phenotypic variation among populations accounted for 93.84%, within-population variation accounted for 6.16%, and the contribution among populations was 15 times that within populations. Once again, interpopulation variation was the main source of phenotypic variation in R. crispus. The greater genetic differentiation among populations and the low genetic diversity within populations could be attributed to genetic segregation among different populations, which is likely due to decreased gene flow caused by severe habitat fragmentation [15][16][17]. This could also explain why phenotypic differentiation has occurred among R. crispus populations in the mining area despite the limited geographic distribution of these populations.
Adaptive mechanisms of Rumex crispus exposed to persistent toxic metal pollution The phenotypic characteristics of R. crispus populations growing in the mining area are the result of long-term adaptation to soil toxic metal pollution. R. crispus growing in the mining area generally has only one main root and almost no fibrous roots, and its main roots are thicker. The root diameter is mostly greater than 1 cm. The coefficient of variation for root diameter was also greater, at 45%, which indicates that variation in root diameter is high among different populations. The correlation between root diameter and soil metal factors was the strongest (p < 0.05; Table 6), which means that the more severe the toxic metal pollution in the soil, the thicker the root. R. crispus is similar to Arabis alpina L., which has larger root diameters in locations with high levels of toxic metals [18]. The increase in root diameter may be due to the thickening of the root epidermis to reduce the absorption of toxic metal elements.
The roots of R. crispus growing in the mining area are generally long. The average root length of the plants in the area was as long as 20 cm, and the longest root was 30 cm. This is different from eggplant, which inhibit root elongation under high concentrations of heavy metals [19], but similar to Sedum alfredii, the concentrations of Zn of S. alfredii were positively correlated with root length [20].
This may have to do with the resistance of different plants to different heavy metals, and R. crispus is a plant that is resistant to a variety of heavy metals [12]. Therefore, the roots of the plant in the mining area were long and thick, which contributes to good growth despite continuous metal pollution.
The leaf area, leaf length, and leaf aspect ratio of R. crispus were negatively correlated with Pb, Zn, Mn, and Fe (p < 0.05), which is similar to the response of Arachis hypogaea L. to heavy metals-the increase in concentrations of Zn, Cu and Cd leads to a decrease in leaf area [21]. Overall, the phenotypic traits of R. crispus show that underground growth is better than aboveground growth for populations that experience long-term persistent toxic metal pollution. Under conditions of such pollution, plants tend to have a reduced leaf area, and underground growth leads to the morphological development of plants in the direction of dry biochemicals [22,23]. R. crispus in the polluted area needs a longer root system to absorb water to adapt to physiological dehydration caused by heavy metal pollution, and the smaller leaf area may also be to cope with physiological water shortage, reduce transpiration and improve adaptability to metal pollution. This xeromorphic adaptive trait variation has also occurred in plants in other metal-contaminated areas such as Arabidopsis arenosa from a lead zinc waste heap in southern Poland [24],and this trait variation may be heritable. It is through morphological variation in the direction of dry biochemicals that R. crispus can survive in this mining area, which has long been polluted with toxic metals.

Conclusions
In a heterogeneous environment characterized by toxic metal pollution in a mining area, Rumex crispus has rich phenotypic variation within and among populations, and a large degree of phenotypic differentiation has occurred among populations. Morphological differences between populations growing in polluted and unpolluted areas are also obvious. R. crispus adapts to long-term continuous toxic metal pollution by increasing underground growth and reducing aboveground growth. Finally, ecological differentiation has occurred among different populations.

Declarations
Acknowledgments Figure 1 Cluster analysis of nine Rumex crispus populations based on variation in phenotypic traits.