agronomy

: Finger millet ( Eleusine coracana (L.) Gaertn.) is an annual allotetraploid that belongs to the grass family Poaceae subfamily Chloridoideae. Using less productive cultivars, biotic and abiotic stresses affect the yield and productivity of ﬁnger millet in Ethiopia. This research was aimed at investigating the acidity/Al tolerance of 328 ﬁnger millet accessions and 15 cultivars from Ethiopia and Zimbabwe. Prior to screening the accessions, optimization was performed on 15 cultivars and 15 accessions under three Al concentrations (0, 75, and 100 µ M), and, afterward, 100 µ M of Al concentration was selected as the threshold level. Root length (RL) and shoot length (SL) were recorded after 10 days of treatment. Accessions 215836, 215845, and 229722 and cultivars Urji, Bareda, and Axum were found Al-tolerant, while cultivars Tadesse, Padet, and Kumsa and accessions 212462, 215804, and 238323 were found Al-susceptible. ANOVA on RL indicated that the variance due to environment (42.3) was higher than genotypic variance (0.37). Whereas, the ANOVA on SL indicated the variance due to environment was not signiﬁcant, and genotypic variance (0.18) was higher than environmental (0.02). RL was highly affected due to Al stress, while no distinct and visible symptoms were observed on SL. Furthermore, the screening of 328 accessions under 100 µ M and the control resulted in Al-tolerant (n = 20), intermediate (225), and Al-susceptible (83). The results of the present study reveal that the presence of acid-tolerant accessions can be used as inputs for breeders to improve the productivity of ﬁnger millet in acidic areas.

In Ethiopia, finger millet is the sixth most important cultivated cereal crop after teff, wheat, maize, barley, and sorghum [7]. Grain of finger millet is rich in protein, minerals, dietary fiber, calcium, iron, and essential amino acids; it is also gluten-free, and has healthpromoting benefits such as hypoglycemic, anti-hypocholesterolemia, and anti-ulcerative effects [8]. Finger millet is often mixed with other grain crops such as sorghum, maize, or A total of 328 accessions representing various agro-climatic zones of Ethiopia and Zimbabwe were obtained from the Ethiopian Institute of Biodiversity (EIB, Addis Ababa, Ethiopia), and 15 cultivars were obtained from the Bako Agricultural Research Center (BARC, Bako, Ethiopia). All accessions were selected randomly from the gene bank and used in this study (Supplementary Table S1). Optimization was performed on selected 15 cultivars and 15 accessions. There are only 16 cultivars of finger millet in Ethiopia. We collected and used all cultivars except Diga-02, which failed to germinate and was omitted from the study. The 15 accessions were selected randomly from the 328 accessions. Similar size seeds and similar seed color (n = 15) from each accession were selected and surface-sterilized by soaking in 3% sodium hypochlorite solution for 5 min and rinsing thoroughly with water. Sterilized seeds of each accession were wrapped and germinated in tissue paper, and then moistened with distilled water in separated Petri dishes for 36 h under dark conditions for later use as the germinated seedlings in hydroponic experiments. Then, the seedlings were transferred to hydroponic nutrient solution and treated for about 10 days within the greenhouse adjusted to a temperature of 18 • C and a humidity level of 65% at the Swedish University of Agricultural Sciences (SLU, Alnarp, Sweden).

Hydroponics Experimental Setup
The basic assumption for setting up the equipment for the hydroponic screening was that the system should enable growth and development of seedlings while ensuring the seeds and later seedlings had maximal exposure to Al stress. This requirement can be realized only under submerged conditions, which demands a mechanism of aerating the seedlings within the nutrient solution. For this purpose, dense narrow holes were introduced into small centrifuge tubes (5 mL) in such a way that the holes did not allow finger millet seeds to pass through but allowed air bubbles in for aerating the seedlings in the tube. Continuous aeration was supplied by an aquarium air pump with an air stone. A rack-like plate to hold the perforated tubes was prepared from a jar plastic plate having wide holes capable of holding and submerging tubes in the nutrient solution ( Figure 1).

Plant Materials and Germination Conditions
A total of 328 accessions representing various agro-climatic zones of Ethiopia and Zimbabwe were obtained from the Ethiopian Institute of Biodiversity (EIB, Addis Ababa, Ethiopia), and 15 cultivars were obtained from the Bako Agricultural Research Center (BARC, Bako, Ethiopia). All accessions were selected randomly from the gene bank and used in this study (Supplementary Table S1). Optimization was performed on selected 15 cultivars and 15 accessions. There are only 16 cultivars of finger millet in Ethiopia. We collected and used all cultivars except Diga-02, which failed to germinate and was omitted from the study. The 15 accessions were selected randomly from the 328 accessions. Similar size seeds and similar seed color (n = 15) from each accession were selected and surfacesterilized by soaking in 3% sodium hypochlorite solution for 5 min and rinsing thoroughly with water. Sterilized seeds of each accession were wrapped and germinated in tissue paper, and then moistened with distilled water in separated Petri dishes for 36 h under dark conditions for later use as the germinated seedlings in hydroponic experiments. Then, the seedlings were transferred to hydroponic nutrient solution and treated for about 10 days within the greenhouse adjusted to a temperature of 18 °C and a humidity level of 65% at the Swedish University of Agricultural Sciences (SLU, Alnarp, Sweden).

Hydroponics Experimental Setup
The basic assumption for setting up the equipment for the hydroponic screening was that the system should enable growth and development of seedlings while ensuring the seeds and later seedlings had maximal exposure to Al stress. This requirement can be realized only under submerged conditions, which demands a mechanism of aerating the seedlings within the nutrient solution. For this purpose, dense narrow holes were introduced into small centrifuge tubes (5 mL) in such a way that the holes did not allow finger millet seeds to pass through but allowed air bubbles in for aerating the seedlings in the tube. Continuous aeration was supplied by an aquarium air pump with an air stone. A rack-like plate to hold the perforated tubes was prepared from a jar plastic plate having wide holes capable of holding and submerging tubes in the nutrient solution ( Figure 1).

Nutrient Solution Culture and Treatment
The nutrient solution culture was prepared according to [22] and composed of 500 µM KNO 3 , 500 µM CaCl 2 , 500 µM NH 4 NO 3 , 150 µM MgSO 4 .7H 2 O, 10 µM KH 2 PO 4 , 2 µM FeCl 3 (III), and different concentrations of Al 2 (SO 4 ) 3 . In vitro-germinated seedlings (n = 10) of each accession with similar root lengths were transferred into the perforated tube, which was then arranged on plastic plate, and seedlings would be in full contact with the growth solution but would not be fully submerged. The control experiment was performed side by side with each treatment and composed of all the above nutrients except Al 2 (SO 4 ) 3 . The pH of the nutrient was adjusted to 4.3 by using 1 M HCl or NaOH and the solution was renewed every day (24 h) in order to refresh the detoxified solution and ensure continuous exposure of the seedlings to Al ions. The seedlings were treated for consecutive 10 days under hydroponic nutrient solution. After 10 days, root length (RL) and shoot length (SL) were measured from five seedlings per accession.

Screening of Accessions under Hydroponic Assay
To find the threshold level of Al tolerance in finger millet, optimization on different Al 3+ concentrations (0, 75, and 100 µM) was performed on 15 cultivars and 15 accessions in the hydroponic nutrient solution. The two Al concentrations (75 and 100 µM) were selected by considering the optimization protocol we developed previously [23]. After the threshold level of tolerance was decided (100 µM Al 2 (SO 4 ) 3 ), a large number of landraces (n = 328) were evaluated in the hydroponic system. Based on their RRL, the accessions were classified into three tolerance groups. Accessions grouped as Al-tolerant were those that had RRL ≥ 80%, whereas intermediates were between 80% and 20%, and susceptible were those below 20%.

Data Recording and Analysis
The root length (RL) of five seedlings per accession was measured from the base of the cotyledon to the tip of the roots, and shoot length (SL) was also measured from the base of the cotyledon to the tip of the shoot using a ruler. The normality of data collected from the hydroponic data was tested using R software. Analysis of variance (ANOVA) was performed using aov function in R software. Pairwise mean comparison was performed using Tukey test in R software. Root growth parameters such as relative root length (RRL) and relative shoot length (RSL) were estimated as described in [22]:

Optimizing Threshold Level of Al-Toxicity on Finger Millet
We used morphological markers, RL and SL, to compare the Al tolerance of seedlings grown under control and Al-stress conditions. The dose-response experiment showed that finger millet accessions and cultivars grown under lower Al concentrations had higher RL than those treated with a relatively high level of Al concentration. In the control experiment, the highest RL was found in cultivar Tessema (2. Tadesse (0.10 cm), Padet (0.10 cm), and Kumsa (0.10 cm) were the least performing cultivars ( Figure 2). Overall, RL-based evaluation showed that the landraces perform better than the cultivars in Al-stress conditions ( Figure 2). Shoot length (SL) of the accession and cultivars grown under control (0 µM Al-concentration) varied from 0.46 cm (Urji) to 1.18 cm (215888 and 215911). At 75 µM Al-concentration, SL ranged from 0.66 cm (215897) to 1.07 cm (243642), and at 100 µM Al-concentration SL ranged from 0.45 cm (Meba) to 1.64 cm (213314) (Figure 3). The effect of Al stress on shoots of finger millet was not observed at Al concentrations of 75 µM or 100 µM.   Tadesse (0.10 cm), Padet (0.10 cm), and Kumsa (0.10 cm) were the least performing cultivars ( Figure 2). Overall, RL-based evaluation showed that the landraces perform better than the cultivars in Al-stress conditions ( Figure 2). Shoot length (SL) of the accession and cultivars grown under control (0 µM Al-concentration) varied from 0.46 cm (Urji) to 1.18 cm (215888 and 215911). At 75 µM Al-concentration, SL ranged from 0.66 cm (215897) to 1.07 cm (243642), and at 100 µM Al-concentration SL ranged from 0.45 cm (Meba) to 1.64 cm (213314) (Figure 3). The effect of Al stress on shoots of finger millet was not observed at Al concentrations of 75 µM or 100 µM.   Analysis of variance (ANOVA) on RL indicated significant differences between finger millet accessions grown at 0 µM, 75 µM, and 100 µM Al-concentrations (Table 1; Figure 4). ANOVA on RL indicated that the variance due to environment (42.3) was higher than genotypic variance (0.37) and variance due to replications (0.15). Whereas the ANOVA on SL indicated that the variance due to environment and replication was not significant, and genotypic variance (0.18) is higher than environmental (0.02) and replication variance (0.04). Analysis of variance (ANOVA) on RL indicated significant differences betwee millet accessions grown at 0 µM, 75 µM, and 100 µM Al-concentrations (Table 1  4). ANOVA on RL indicated that the variance due to environment (42.3) was high genotypic variance (0.37) and variance due to replications (0.15). Whereas the AN SL indicated that the variance due to environment and replication was not signific genotypic variance (0.18) is higher than environmental (0.02) and replication v (0.04). Key: DF Z = degree of freedom, MS = mean of squares, ns = not significant; *** significant at and ** significant at p < 0.01. Finger millet accessions and cultivars grown under the hydroponics display distinct Al-tolerance phases in different Al concentrations. A high phase tolera observed between 0 and 75 µM, slight tolerance at 100 µM, and an intolerance phas 100 µM. This indicates that low Al concentrations were not strong enough to crea Finger millet accessions and cultivars grown under the hydroponics displayed three distinct Al-tolerance phases in different Al concentrations. A high phase tolerance was observed between 0 and 75 µM, slight tolerance at 100 µM, and an intolerance phase above 100 µM. This indicates that low Al concentrations were not strong enough to create stress conditions on finger millet root, and high Al concentrations above 100 µM inhibit growth in all finger millet varieties without discrimination. Therefore, the 100 µM Al-concentration was selected as the threshold concentration for extensive screening activities due to its multiple advantages. Firstly, it allows for the distinguishing of the various tolerance classes (tolerant, intermediate, and susceptible) at the highest accuracy level (that is, p < 0.01, unlike the lower concentration levels). At Al concentrations above 100 µM, the growth of roots of all the varieties was greatly hampered to the extent that there were nearly no differences among them. Therefore, 100 µM was selected as the optimum Al concentration for the screening of 328 finger millet accessions.

Screening Finger Millet Accessions
Finger millet accessions (n = 328) were rapidly screened after initially deciding the optimum Al concentration (i.e., 100 µM). There were observable differences among individuals within an accession such as variation in grain color and grain size, and to take this heterogeneity into account, each accession was evaluated systematically by recording data from similarly performing individuals. There were significant variations among accessions grown at 100 µM Al-concentration. The RRL of tolerant accessions ranged from 79.4% (245084) to 127.9% (215836), while the most susceptible accessions had an RRL of less than 10% (215804, 212462, and 238323) (Tables 2-6). There was a significant difference between the most extremely tolerant accession 215836 with 127.9% RRL and the most susceptible accession 215804 with 7.1% RRL ( Figure 5). Among the total accessions screened, 20 of them were better performing and grouped as Al-tolerant (Table 2), while 225 of them were grouped as intermediate (Tables 2-5), and 63 of them were highly susceptible to Al-stress and least perform and grouped as Al-susceptible (Tables 5 and 6). in all finger millet varieties without discrimination. Therefore, the 100 µM Al-concentration was selected as the threshold concentration for extensive screening activities due to its multiple advantages. Firstly, it allows for the distinguishing of the various tolerance classes (tolerant, intermediate, and susceptible) at the highest accuracy level (that is, p < 0.01, unlike the lower concentration levels). At Al concentrations above 100 µM, the growth of roots of all the varieties was greatly hampered to the extent that there were nearly no differences among them. Therefore, 100 µM was selected as the optimum Al concentration for the screening of 328 finger millet accessions.

Screening Finger Millet Accessions
Finger millet accessions (n = 328) were rapidly screened after initially deciding the optimum Al concentration (i.e., 100 µM). There were observable differences among individuals within an accession such as variation in grain color and grain size, and to take this heterogeneity into account, each accession was evaluated systematically by recording data from similarly performing individuals. There were significant variations among accessions grown at 100 µM Al-concentration. The RRL of tolerant accessions ranged from 79.4% (245084) to 127.9% (215836), while the most susceptible accessions had an RRL of less than 10% (215804, 212462, and 238323) (Tables 2-6). There was a significant difference between the most extremely tolerant accession 215836 with 127.9% RRL and the most susceptible accession 215804 with 7.1% RRL ( Figure 5). Among the total accessions screened, 20 of them were better performing and grouped as Al-tolerant (Table 2), while 225 of them were grouped as intermediate (Tables 2-5), and 63 of them were highly susceptible to Alstress and least perform and grouped as Al-susceptible (Tables 5 and 6).

Discussion
Among the abiotic factors, soil acidity is a major constraint for plant development and growth as well as the yield and productivity of crops. It has been estimated that over 50% of the world's potentially arable lands are acidic [13]. In this study, a hydroponic system was used to study the Al tolerance of finger millet accessions and cultivars under different Al concentrations. Hydroponic systems are suitable for early growth and seedling screening under submerged conditions. Previously published research on wheat, rice, and chickpea has used hydroponics to screen against Al stress by measuring root and shoot length [23,28]. Therefore, the present study also confirmed the suitability of using hydroponics while exercising an Al-tolerance study on finger millet. The morphological markers, RL and SL, were important traits to study Al tolerance as the primary response to Al stress occurs in the plant roots, with the Al-susceptible genotypes showing retarded root growth.
It is advisable to use seedlings with similar vigor and this is achieved by selecting seedlings with similar-sized endosperm, similar initial root length, and similar seed age to consider better performing individuals [25,29]. These accessions were sometimes comprised of two or more genotypes since there was a large variation in performance between individual plants of the accession. Furthermore, there were visually observable differences within an accession such as variations in grain color. To take this heterogeneity into account, an accession was scored based on its best-performing seedling. The use of the average performance of plants in representing an accession would have resulted in the rejection of many accessions because of poor average performance such that a single plant within the accession with an acceptable level of Al 3+ tolerance would be lost [25].
According to [22], RRL and RSL are morphological markers to study Al stress as they can eliminate genotype-specific differences in root growth and normalize comparisons between genotypes. Since RRL and RSL are the relative growth of the genotype in Al solution compared with its potential growth without Al, this parameter is a real measure of Al tolerance [22]. Short root length is considered to be the primary consequence of aluminum toxicity, resulting in a smaller volume of soil explored by the plant. Consequently, reducing its mineral nutrition and water absorption. Furthermore, it reduces cell membrane permeability and binds to the phosphate groups of the deoxyribonucleic acid, decreasing replication and transcription [15].
In this study, a hydroponic nutrient solution was employed to identify the threshold level of Al concentration in finger millet landraces and cultivars. Finger millet accessions and cultivars were evaluated at three Al concentrations including the control (0, 75, and 100 µM). At low Al concentrations, it is difficult to properly discriminate finger millet accessions and cultivars in relation to their Al tolerance. The reason could be that low Al concentrations (less than 75 µM) were not strong enough to create Al-stress conditions at finger millet roots. Similarly, at high Al concentrations above 100 µM, the Al stress inhibited growth in all finger millet accessions and cultivars, making it difficult to differentiate between the tolerant and susceptible groups. However, better discrimination among the genetic materials was observed at 100 µM, and it was selected and used as an optimum concentration level for the wider screening of 328 landraces.
Comparatively, the threshold level of Al tolerance in finger millet accessions was found higher than the Al tolerance of barley accessions, which had 30 µM [30], and maize accessions, which had a 20 µM threshold level of Al tolerance. Whereas, in line with the tolerance level of finger millet at 112.5 µM [23], chickpea accessions had Al-concentration thresholds of 110 and 120 µM [27,31]. The higher Al-tolerance level noted in finger millet might be because finger millet is a climate-resilient crop that is able to grow in marginal lands, which helps the crop to perform better than other crops in biotic and abiotic-stressprone environments [31]. Moreover, most of the accessions used in this study were collected from western and northern parts of Ethiopia, where soil acidity is predominant, and they developed a mechanism to tolerate this type of stress. Genotypes collected from acidic environments may accumulate mutations that adapt to acidic environments and develop rapid Al-tolerance mechanisms by activating genes responsible for the secretion of mucilage and organic acid anions when they are exposed to phototoxic forms of Al within minutes of exposure. Thus, due to natural selection, only the tolerant genotypes survive.
At the 100 µM Al-concentration screening, cultivars Tadesse, Padet, and Kumsa, as well as accessions 212462, 215804, and 238323, were the least performing (Al-susceptible). On the other hand, Urji, Bareda, and Axum cultivars, as well as 215836, 215845, and 229722 accessions, were relatively tolerant against Al stress. Accessions were found to be more tolerant against Al stress than cultivars. This indicates that landraces have a better Al tolerance compared to cultivars, implying that breeding activities have a significant effect on the stress tolerance, including on the Al tolerance of the crop.
In the present study, we did not observe any distinct and visible symptoms of Al toxicity in the SL of finger millet, which is in agreement with previous studies on pigeon pea using a 20 µM Al-concentration [32]. No significant effect of Al stress on SL was detected in our study due to the short exposure time in the hydroponic system.
The RRL considers control and treatment conditions. It allows for a comparison of accessions with a constant ranking according to their performance. The dose-response experiment on the wider number of accessions demonstrated that 20 (6.9%) of them were Al-tolerant, whereas 268 (93.05%) of them were ranked from low to medium tolerance. The majority of the accessions collected from Wellega and Gojam were found Al-tolerant, while those collected from the northern part of Ethiopia were found Al-susceptible. According to [21], acidic soil is prevalent in western Ethiopia. Accessions collected from soil-acid-prone areas were found Al-tolerant. Thus, their enhanced tolerance against Al concentrations was likely developed due to long-term exposure to soil acidity. Accessions identified as Al-tolerant in the hydroponic experiment often showed improved agronomic performance compared to Al-susceptible accessions [25][26][27]29]. Potential finger millet accessions identified here can be used as inputs for breeders to improve the Al tolerance of finger millet.

Conclusions
The results of the present study suggest that there are individual accessions that can better tolerate acidic soils and some of them are highly susceptible. Lower Al concentrations had no significant effect on the RL of most finger millet cultivars and accessions, while their growth starts to decline with an increasing Al concentration. At 100 µM Al-concentration, cultivars Tadesse, Padet, and Kumsa, as well as accessions 212462, 215804, and 238323, were Al-susceptible. Thus, these cultivars should not be recommended in areas where soil acidity is predominant. On the other hand, Urji, Bareda, and Axum cultivars, as well as 215836, 215845, and 229722 accessions, were relatively tolerant against Al and can be promoted in areas where soil acidity is highly prevalent. To confirm their performance, the accessions should be tested on multi-site fields by considering controlled and treated environments. Furthermore, association studies should also be considered to correlate field performance with genomic background. Transcriptomic analysis on the most tolerant and least susceptible should be tested by taking samples from different plant tissues (root, leaf, and stem) at different time intervals (0, 12, 24, 48, and 72 h). Finally, the Al-tolerant lines identified in this study should be used as inputs to finger millet breeding programs in relation to Al tolerance in Ethiopia, Zimbabwe, and elsewhere. If anyone is interested in studying Al tolerance on finger millet, we suggest that they include wild types for comparative analysis.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/agronomy13061596/s1, Table S1: List of finger millet accessions used in present study with their passport data.