1. Introduction
In the current climate disaster scenarios, seasonal droughts induced by uneven rainfall and rapidly increasing population have highlighted food insecurity problems [
1,
2,
3,
4,
5]. To cope with the increased population and to alleviate hunger, the global grain yield production must be increased by 70% [
6,
7]; thus, the increasing demand for food conflicts with the shortage of food [
8]. In the face of such a great contradiction between supply and demand, many crops with strong resistance to environmental stress, including foxtail millet, are increasingly being planted and replanted [
9]. Foxtail millet (
Setaria italica L.), which has evolved from
Setaria viridis L., is widely cultivated in the world, with a long history of planting [
10], especially in northwest China [
11,
12]. Its grains are rich in a variety of trace elements and proteins [
13], and it is consumed by more than half of the world’s population, especially women and children [
14].
The growth of plants is strongly associated with the morphological and physiological characteristics of their roots [
15,
16,
17]. Environmental stresses can cause big changes in the development and form of plant roots, which are the interaction point between plants and soil [
18], and they significantly affect many important agronomic traits, including drought tolerance [
19]. According to previous reports, at the early growth stage of corn, the grain yield was found to be closely related to the root traits [
20]. The root system is the main organ used for obtaining underground resources and plays a key role in water and nutrient absorption [
6,
8,
21,
22,
23,
24,
25]. The root is usually the most sensitive organ for sensing and responding to miscellaneous soil stresses, such as soil drought, salinity, waterlogging, and nutritional deficiency [
15,
26,
27,
28]. Under abiotic stress, crop yield can be affected by the root phenotype through changes in the metabolic efficiency of soil exploration and the redirection of roots to the soil area with the highest amount of limited resources [
29,
30,
31]. In each crop species, investigating the association of different types of roots with the environment is the first stage of research; this includes studies on the adaptation of crop species with deep roots to drought-prone environments [
32]. Adapting to the intrinsic heterogeneity of soil resources, plants can modulate their root phenotype at different comprehensive levels [
20], change the relative input of aerial and root biomass at the individual level, and adjust their root structure or make changes at the organ level [
21,
33]. The root system promotes aerial growth, and the root tip is the most active part of the root system, which is essential for plant yield [
34]. Root morphology covers basal characteristics, such as root length, root length density, diameter, surface area, and volume, all of which affect the spatial structure of the root system [
35,
36]. Root length, surface area, and volume affect the spatial distribution of the roots, while the root diameter is related to intense soil penetration and drought tolerance. The relationship between the size of the aboveground plant parts and the size of the roots (S/R ratio) is critical to the plant water balance. A lower S/R is more beneficial to plants in a water-deficient environment and is an important indicator for evaluating plant drought tolerance [
37]. Some researchers have found a correlation between the phenotypic traits and drought resistance [
38]. Root system architecture (RSA) is a relatively complex quantitative trait which differs in different environments, and it is difficult to determine the phenotype of underground roots [
39,
40]. Thus, relatively few studies on millet root traits have been carried out to date. Applying the genetic information of root traits and root structure to the practice of millet breeding is expected to improve its resource efficiency and stress resistance [
27].
In millet, a good performance has been shown to be associated with drought resistance, but its yield remains lower than that of other cereals [
36]. Research on millet root-structure characteristics would not only improve our knowledge of its agronomic traits, but also enable plant breeders to enhance its tolerance to various stress conditions [
35]. The knowledge on root phenotypic diversity and the interrelationships among root traits provides guidance for follow-up research and will make better sense of these traits in order to formulate breeding strategies [
41]. It is extremely important to update local millet varieties with improved adaptation to environmental stress and climate changes. However, millet root morphological traits in different ecological regions have not previously been studied. Some early studies proved the practicability of semi-hydroponic phenotypic systems under a wide variety of environments, including the field [
42], and significant correlations in root traits were discovered between wheat seedlings grown on moist germination paper and field-planted wheat seedlings [
43]. The present study examined the variability of root morphological traits among 65 genotypes of foxtail millet from four ecological regions of China, cultivated in seed germination pouches. The objectives of this study were to (1) assess the variations in root system architecture among a set of millet germplasm and (2) to evaluate root trait correlations among the different ecological regions. The results will provide references for millet breeders.
2. Materials and Methods
2.1. Plant Materials
Sixty-five millet genotypes were selected from seven provinces (see
Figure 1 for the distribution and the average rainfall in 2020) that belong to four millet planting ecological areas in China (
Table S1). In the early maturing area of northeast China, the climate is the coldest, with long hours of sunshine during the growing season and low evaporation in the spring. Moreover, the area of the early maturing northwest is characterized by low rainfall with frequent drought seasons, high evaporation in spring, a high amount of wind and sand, and low soil fertility. In the middle and late maturing areas of northwest China, millet is not sensitive to temperature because these areas are characterized by a mild climate, drought, and a long sunshine duration. The summer maturing region is characterized by low latitude, high temperature, and rainy weather, as well as a small temperature difference between day and night and few sunshine hours.
2.2. Plant Culture
An incubator cultivation experiment was implemented from December 2020 to May 2021. Millet seeds were sterilized with 3% NaClO for 20 min and germinated on moist germination paper and kept at 16 °C for 10 h in the dark, after which they were placed at 24 °C, with 30,000 Lx, and an environmental air humidity of 65% for 14 h. The seeds were then cultivated in seed germination pouches (160 × 125 mm2) with 40 mL of distilled water. Four evenly located positions were selected at the groove of the germination paper, and two seeds were planted at each position. The port was sealed with a paper clip to reduce water loss.
Each genotype was cultivated in one pouch, and there were four replications of the pouches per genotype. Ten genotypes were placed on each shelf within an incubator. On the seventh day of the growing season, two leaves were visible; the paper clips were then removed, and four seedlings were established. The rest of the distilled water was poured out, and 40 mL of 25% Hoagland’s nutrient solution was added; the position of the pouches was randomly changed daily. After a week, a fresh nutrient solution replacement was made, and the seedings were harvested on day 21 for the assessments.
2.3. Root Trait Measurements and Calculations
Leaf number, root number, root width, maximal root depth, and maximal root length were measured manually prior to the harvest. The root width, maximal root depth, and maximal root length of each plant were measured using a ruler, and the root angle was measured with a protractor.
Root width is the maximal gap between the primary and the lateral roots, and maximal root depth is the maximal vertical length from the root top to the root tip. Maximal root length is the longest root of a single plant. After manual measurements, the roots were separated from the shoots, and the roots were divided into 5 cm sections from the base for scanning. There were 3 sections in each plant. Subsamples of every 5 cm section in each pouch were placed in distilled water in a plastic tray (22 cm × 31 cm × 2 cm), and they were scanned on a grey scale at 150 dpi using a desktop scanner (Epson Perfection V800, Long Beach, CA, USA). The images were analyzed using the WinRHIZO software (Regent Inc., Quebec, QC, Canada) to determine the length, diameter, surface area, and volume of the roots. The shoot samples were deactivated in a baking oven for 20 min at 105 °C, after which all the root and shoot samples were dried at 80 °C for 48 h, and the dry weights of the roots and shoots were determined.
The first 5 cm layer of roots and the remaining part of the roots of each root system were defined as the “top-root layer��? and the “sub-root layer��?, correspondingly. The functional traits were calculated by the following formulae:
Detailed descriptions of the 48 traits, including 17 global traits (three shoot traits and 14 root traits) and 31 local traits, are shown in
Table 1.
2.4. Statistical Analysis
The data processing of the 48 traits (
Table 1) was performed using Excel 2019. A distribution map with average rainfall in 2020 (
Figure 1) was performed in ArcGIS (version 10.8, Environmental Systems Research Institute, Inc, Redlands, CA, USA). A descriptive statistics analysis was performed for minimum, maximum, medium, mean, and standard deviation; one-way analysis of variance (ANOVA,
p < 0.05) was carried out with the genotypes as the variables, and the Pearson correlations of 18 traits with CV ≥ 0.2 (
p < 0.05) were performed using SPSS Statistics (version 23, IBM, New York, NY, USA). The data of four replications of each genotype were calculated as means and then classified by province and ecological region, respectively. The mean values of root length in three layers and the root dry mass of each province were selected and ordered with the total root length in ascending order; the mean data of each ecological region were processed by the same method. The data of the two groups were used to draw double Y-axis column plots. The root lengths of the different layers and the biomass dry mass were plotted as stacked histograms, and the significant differences were noted. The selected 18 root traits of the genotypes were plotted as box charts (grouped by layer, ecological region, and province). The boxes were based on the median values of the defined traits by the first and third quartiles, and the individual data points that fell outside the whiskers (1.5 times the interquartile range from the median) were considered as outliers. A General Linear Model multivariate analysis was performed for the genotype and ecological region when
p < 0.01. The 18 selected traits were used for principal component analysis (PCA) for both the global and the local traits to identify the determinants of the root morphological variability across the genotypes. The PCA comprehensive values were ordered in an ascending order with the ecological regions and drawn into a column plot. Cluster centers of the six identified groups were generated by a dendrogram of the agglomerative hierarchical analysis. All the figures were plotted using Origin Pro 2021b graphing packages (Origin Lab, Northampton, MA, USA).
4. Discussion
Recently, the interest in root structure affecting crop yields to improve food security has been increasing [
30,
32,
44,
45]. In many ecosystems, air dehydration precedes soil dehydration [
46]. Usually, the shoots of the plant are first exposed to reduce the water potential, after which the environment around the root system becomes increasingly arid [
47,
48]. In wheat, root biomass has been shown to decrease at high soil temperatures [
49], and a deeper and more effective root system has been shown to be essential for absorbing more water from the soil and for significantly increasing wheat yield under drought stress [
39,
50,
51]. Root growth is dynamic and spatially and temporally heterogeneous [
52]; therefore, the changes in root structure depend on the environment [
42]. The germination and the seedling growth of plants are necessary stages for the growth of plant populations [
53,
54]. During seedling development, plant functional traits can rapidly respond to altered environments; for example, the depth of the plant roots affects the ability of the ecosystem to recover from drought stress [
55]. Deep roots link the deep soil and groundwater to the atmosphere, further affecting the hydrological cycle and climate [
40]. Water and nutrient acquisition are closely related to these measured root traits [
56] of MRD [
57], RL, RDM, and RLD in the three layers [
26,
44,
58]. Therefore, it is very important to study the root system to improve the stress resistance and yield of plants [
59].
In the present study, the traits of millet from the four ecological regions showed different performances. Millet cultivars from the early maturing area of northwest China had the greatest RL, RDM, SDM, and TDM (
Figure 2). RL is an indicator of a plant’s yield potential [
60]. Root size, length, and architecture could improve water uptake, which were associated with increased drought tolerance in crop plants [
61,
62]. The RL and RDM were about two-fold higher than those in the cultivars from the early maturing area of northeast China. A moderate water deficit can promote the increase in root and shoot dry mass [
63]; meanwhile, a plant will pass on the characteristics of a larger root system to the next generation and obtain higher yields [
37,
64]. The correlation coefficient between the RN and RL of the same root type, as well as among the RL, RV, and RA in hydroponics [
65], was high, which is in line with those of previous studies [
66]. A strong positive correlation was observed between RLR and the root ratio (RSR and RVR,
Figure 6a,b), which illustrated that the root in the top and sub-layers had similar trends of RL, RA, and RV (
Figure 4a,c,d). The high correlations between RDM, RSM, and RL indicated that root changes affect the ratio of aboveground and underground biomass, which is an important manifestation of a plant’s response to environmental stress [
67].
The root system architecture (RSA) is systematically shaped by the diversity of characteristics of a species and the evolutionary processes [
65]. The longer RL and the greater RA indicate that the ability of a root to obtain nitrogen from the soil profile increased [
68]. This study came to the same conclusion: millet cultivars from the early maturing area of northwest China had the largest RL among all the regions (
Figure 3), which was significantly positively correlated with RA (R = 0.901,
p < 0.01,
Table S2). SRL reflects the absorption capacity of roots per unit of biomass input and is also a measure of root thickness, and thus, it is considered to be one of the most important traits affecting the exploitation and uptake of water and nutrients, taking into account the economic cost of root establishment [
69,
70,
71]. In a field experiment, it was found that genotypes with lower SRL, from a 0–60 cm depth, produced higher yields, suggesting that thicker roots in the topsoil represent a favorable feature for obtaining deep water in soil [
72,
73]. Millet cultivars from the early maturing area of northwest China had low SRL (
Figure S1c), which could be associated with high yield. Future studies will validate these genotypes in the field conditions.
Root diameter is a powerful predictor of penetration strength in hard soils, and greater RD may result in greater hydraulic conductivity [
74]. A comparison of the findings with those of other studies confirms that RD was negatively correlated with root density (RLD and RWD,
Figure 6c,d). The cultivars from different ecological regions showed some variation, reflecting the potential for their ecological and climatic linkage and the need for further research. Mild water stress can expand the root system and improve RWD and RLD. The development of longer roots and higher dry weights in the upper soil saves irrigation water [
63]. Moreover, water deficiency leads to the redistribution of carbon during root growth, as well as a decrease in RLD in the topsoil layer, but an increase in RL. Thus, RLD is the key factor for assessing the amount of nutrients and water absorbed by the root system per unit volume and is an important symbol of the drought-tolerant varieties [
75]. Drought treatments can increase the density of roots in deep soil; however, the pearl millet with deep roots cannot grow well in a moist environment [
37,
76].
Because the growth of a field plant is irreversible, the root morphology and topological structure encode traits derived from “morphological memory��? and provide clues to unlock the secrets of the plant–environment interaction [
30,
77,
78]. The exploration of 65 millet genotypes at the seedling stage did not allow us to classify them by root diameter due to the smaller differences in this factor. However, there are still some limitations in transforming the phenotype data obtained from hydroponic systems at the early stage into field conditions in the breeding programs. Modern root imaging technology and root modeling simulations could provide efficient tools to assist in root phenotyping studies in different environments.