Root Traits Related with Drought and Phosphorus Tolerance in Common Bean (Phaseolus vulgaris L.)

Roots are key organs for water and nutrient acquisition and transport. Therefore, root phenes that are associated with adaptation to low phosphorus (P) environments could enhance top-soil exploration, while deeper allocation is important for acquiring water and mobile nutrients. The understanding of interactions among root phenes can help in the development of common bean (Phaseolus vulgaris L.) genotypes adapted to drought and low fertility through genetic improvement. Two experiments (pot and field) were conducted at the Agricultural Research Institute of Mozambique to assess the contribution of root phenes to common bean shoot biomass and grain yield under combined stress (drought and low P). The pot study assessed eight genotypes, with four treatments combining water regimes (drought and non-stress) and phosphorus levels (200 and 25) mg P kg−1 soil. In the field study, 24 common bean genotypes were also grown in high and low phosphorus (40 kg P ha−1 and without P application) under irrigation and limited water. The grain yield from fields under drought and P stress were correlated with the pot data on root traits. The response of root phenes to drought and phosphorus stress appeared to be related to the deep and shallow root systems, respectively. Deep rooted genotypes produced more total root biomass and high taproot lateral branching density, which resulted in high total root length under drought and low P stress, while shallow rooted genotypes had low total root biomass and less taproot lateral branching. Increased shoot biomass and grain yield under drought and low P was associated with higher mean values of taproot lateral branching density and total taproot length. Genotypes SER 125, BFS 81, FBN12111-66 and MER 22 11-28 showed a greater score of tap root branching density in the pot study with the highest grain yield in the field under low P and drought stress. Therefore, these can be recommended for use in low phosphorus and drought stress environment or serve as parents for improving phosphorus use efficiency and drought tolerance in common bean.


Introduction
Despite its importance, the yield of common bean (Phaseolus vulgaris L.) in developing nations is currently very low: only one third of what is achievable in developed nations, where high inputs are used and irrigation is available to reduce abiotic and biotic limiting factors [1][2][3][4]. However, more than half of the global production occurs in marginal lands where people have limited access to food [3,5]. In most of the bean production areas, drought and soil infertility are major constraints to production since access to irrigation and fertilizer is very limited. Therefore, the development of cultivars with superior adaptation to limited soil resources is an effective method for increasing yields in harsh environments [6,7].
While shoots have an important role in utilizing water and nutrients where these are limited, the root system is strictly responsible for the acquisition of these resources [8]. The importance of roots for adaptation to nutrient and water limitations is explained by the increases in root growth compared to shoot growth under edaphic stress [8,9], and it has been demonstrated that the root system has a significant effect on bean yield under water stress [10,11]. According to Kaeppler et al. [12], root phenotypes are composed of phenes, which are the important elements and are under quantitative genetic control, thus influenced by environmental interactions. Root phenotypes play an important role by participating in the improvement of crop yield under edaphic stress by increasing the metabolic efficiency of soil exploration and by allocating roots in soil domains where several limiting resources are mostly available [7,13].
The three principal resources that often interfere with plant growth are phosphorus (P), nitrogen (N), and water [7]. Phosphorus is typically more available in top soil due to the deposition of plant residues at the soil surface and low mobility in the soil [14,15]. Roots have an important role in water and nutrients absorption in the plants. Assessment of root traits of common bean by phenotypic analysis under drought stress showed the importance of variations in rooting patterns, which also include deep rooting systems that help to provide water from deeper soil layers [1,5,16]. Different ideotypes of root systems have been identified for better crop adaptation either for individual or combined abiotic stress conditions [17,18].
In Phaseolus vulgaris, root phenes that influence soil foraging depth comprise mainly of basal root growth angle [19,[20][21][22], basal root whorl number [20,21,[23][24][25], adventitious root abundance [21,23,24], and lateral root branching density [21]. The development of crop cultivars with enhanced productivity in limited water and nutrient availability is a fundamental strategy in addressing the challenges of low productivity [2,26]. The objectives of this study were, therefore, to assess the contribution of root systems to tolerance to low P and drought and to identify genotypes with greater response of lateral branching under suboptimal water and phosphorus supply.

Experimental Conditions
An experiment was conducted at Chókwè Research Station in pots of 30 cm diameter × 32 cm height filled with a sandy soil (physical and chemical properties shown in Table 1). Seeds of eight genotypes contrasting in root branching were grown in a system of stratified water and phosphorus [9]. Seeds were sterilized using 10% NaOCl for 3 min, rinsed twice with distilled water, pre-germinated in rolls of brown germination paper soaked in 0.5 mM CaSO4, and placed in a dark growth chamber for 48 h at 28 °C. After germination, four seedlings per genotype were transferred to each pot. The plots consisted of four pots per genotype per replicate, with a total of four replications. The seedlings were planted at a depth of 4 cm and plants were later thinned at the stage V3 (when the first fully trifoliate leaves expanded) to leave one plant per pot. Each pot was separated into two layers; 0 to 8 cm and 8 to 32 cm depths by an irrigation ring [9] as shown in Figure 1. A solid fertilizer containing the following in grams per pot: 3 g urea, 20 g KNO 3 , and 8 g Micromax granular micronutrients (6.0% Ca, 3.0% Mg, 12.0% S; 0.10% B 1.0% Cu, 17.0% Fe, 2.5% Mn, 0.05% Mo and 1.0% Zn), obtained from Omnia company was mixed thoroughly with the media using a mixer and applied to each pot. Phosphorus levels were maintained by using granular triple superphosphate at the rates of 25 mg P kg −1 of soil and 200 mg kg −1 of soil as low and high levels, respectively and was mixed thoroughly with the media. The water supply in the system was maintained by placing two irrigation rings in each pot, one in the middle of the pot and another on the top, which permitted separate irrigation of the two layers. The irrigation frequency depended on the soil moisture content in the pots. This moisture content was monitored by 30 cm length three-rod time domain reflectrometry (TDR) metallic probes with RG8 cable (CS610-L, Campbell Scientific Inc., Logan, UT, USA) connected to a TDR 100 (Campbell Scientific Inc., Logan, UT, USA) placed vertically, one in the top 0-8 cm and the other between 8-32 cm of each pot.
In the bottom layers, the moisture content was maintained at a constant level in all treatments to simulate normal field conditions, while in the water stress treatment irrigation was withheld 7 days (progressive water stress with no watering after 7 days of growth) in order to simulate terminal drought stress conditions. To avoid waterlogging, the top layer was irrigated first followed by watering the bottom layer to the saturation point. From 7 days after planting (DAP) until a day before harvest (33 DAP), drought stress pots were maintained between 50-60% field capacity, while control pots were kept at 85-95% field capacity.

Plant Material
For the pot study, a total of eight elite lines developed for multiple abiotic stress (drought-tolerant and low soil fertility) and contrasting in root architecture were selected, categorized into groups then planted; The criteria of categorical grouping was based on phenotypic characteristics of the root architecture from a previous pilot study. All common bean (Phaseolus vulgaris L.) seeds used in the studies were propagated from seeds originally obtained from the International Center for Tropical Agriculture CIAT, Cali, Colombia. Eight genotypes were tested in the pot experiment due to space limitations in the shelter. For the field experiments at IIAM-Chókwè Research Station a total of 24 lines, were used. Details of seed color, size and growth habit are described in Table  2.

Experimental Design
The pot experiment design was a randomized complete block design with four blocks (replicates). There were four treatments combining two water regimes (drought stress and well-watered) and phosphorus levels (high P: 200 mg P kg −1 of soil, and low P: 25 mg P kg −1 of soil as described: (i) Control: high water (well-watered) and high phosphorus in both layers; (ii) Stratified low phosphorus: high phosphorus in the top 0-8 cm, low phosphorus in the bottom 8-32 cm, and adequate water in both layers; (iii) Stratified low water (drought stress): low water in the top 0-8 cm, adequate water in the bottom 8-36 cm, and high phosphorus in both layers; (iv) Stratified low water (drought stress) and phosphorus: high phosphorus and low water in the top 0-8 cm and low phosphorus and adequate water in the bottom 8-36 cm layers.
For the field study, plants were grown in high and low phosphorus environments with full irrigation and limited water. The experiments were laid out in a split plot design and replicated 4 times, where phosphorus and water levels were each main factors. The high phosphorus (HP) level was maintained by applying 40 kg ha −1 of mineral fertilizer-triple superphosphate (TSP 46%) applied as basal fertilization 10 days before planting, and control-low phosphorus (LP) level without phosphorus fertilization, intended to simulate the real farmer's context in most bean production; while irrigation treatment had two soil moisture levels: well-watered (WW), where optimum soil moisture level was maintained by supplying water at 80-90% field capacity until the crop had reached pod formation stage), and water stressed (WS), where soil moisture was kept at 70-90% field capacity from planting up to 23 days after emergence and thereafter water was maintained below 50% field capacity, and supplemented when necessary until the crop had reached pod formation stage. The final soil moisture in the progressive waterstress treatment was on average 26% of field capacity. The sub-plot entries (twenty-four genotypes) were randomly assigned to each main plot. Each genotype was planted in three rows of 3.0 m length, inter-row spacing of 0.60 m and within-row plant spacing of 0.10 m. There were buffer plots surrounding the entire field as well as 6.0 m buffer zone separating the high and low phosphorus.

Shoot and Root Measurements
Shoot and root biomass were harvested between 35 and 42 days after planting (DAP) for the pot trials. Shoot tissue was dried at 60 °C until constant mass and weighed. Roots were harvested by root type and horizon. Root crowns with soil were extracted from pots, placed on 2 mm nylon screen mesh table, then gently washed using a lowpressure garden shower. The entire primary root was sampled for visual measurement of primary root lateral branching. The remaining root biomass from each horizon was stored separately and dried at 60°C until constant mass and then weighed for root biomass allocation assessment in each treatment.
The total root length (cm) of all genotypes was determined by calculating the specific root length from five root crowns randomly selected. The entire tap root was analyzed by imageG for root length, then dried and weighed to be used for extrapolation for the rest of the crowns for apparent total root length and tap root from the system.
Roots from the field experiment were phenotyped at flowering (approximately 40 days after planting) by excavating and washing root crowns of a subsamples o f three plants in each plot f o r visual evaluation. Root crown examination included counting of the number of basal root (BRN), basal root whorl number (BRWN), adventitious root abundance score (1 = none; 2 = 2-4 roots; 3 = than 4 roots), number of representative adventitious roots (greater than 1.5 mm diameter) (ARN), taproot lateral branching and taproot length as described [28]. Other measurements included basal root growth angle (BRGA) measured with a protractor, stem diameter and taproot diameter. To relate pot root architecture to field root architecture, comparisons were made in subsets of the genotypes grown in the pot experiment under drought, and low fertility conditions. Root architecture from pot crowns was correlated to seed yield to determine the influence of individual root phenes on seed yield in the field. Root distribution in the soil profile was determined from soil cores taken 2-3 days before root harvest (at flowering). Three subsamples of soil cores of three representative genotypes of each phenotypic class (dense and sparse lateral branches) were taken. Cores were taken in-row between two neighboring plants at a distance of 5 cm between cores using a stainless-steel soil cores (Giddings Equipment Company) with a 4.4 cm internal diameter and 60 cm length, which was inserted into the soil using dead blow sledgehammers and removed by hand. The cores were divided in depth increments of 10 cm, and depending on the depth of profile reached, 3-6 sections were sampled, washed, dried and the weight for specific root length determined.

Statistical Analysis
The statistical software Minitab software ( Version 16, State College, Pennsylvania, USA) was used for all data analyses. Prior to all statistical tests, the normality of the data was determined using the Shapiro-Wilk test. Where data did not meet normal distribution, a log transformation was used. Significant correlations and differences for all data analyses were considered at p ≤ 0.05 and at p ≤ 0.01 where noted. Pearson's correlation analysis was used for comparisons of phenes measured in pot study with the same phenes measured in the field for each treatment. These correlation analyses were performed using the genotype means for each phene in pot trial and in the field.

Primary Root Lateral Branching Effect on Shoot Biomass and Grain Yield
Between the two commercial varieties (Beniquez and Amadeus), both had lower scores for tap root branching, however, Beniquez showed the lowest grain yield among the genotypes, but with a relatively good score of taproot branching density, while Amadeus had high grain yield with the lowest score of tap root branching (Figure 3).

Influence of Root Class and Order on Grain Yield
Taproot laterals number by order (first and second order) was another phene assessed under low P and drought stress, and the results were quite similar to that of taproot branching density score. Although genotypes GFS 81, MER 2212-28 MHR311-17 had greater numbers of taproot lateral, the yield was relatively lower compared to the genotypes BIOF 2-106, SEN 52, and SER 125 that showed lower taproot lateral number with greater grain yield under low P and drought (Figure 4). The sensitive low P and drought genotypes; USMR 20, TARS LFR-1, SEF 16, and IBC 301-204, had lower taproot numbers with lower grain yield. Even though the genotype PR1217-16 had a high number of the taproot lateral, the phene did not have an effect in increasing grain yield. The commercial variety Beniquez increased the number of taproot lateral, but this did not translate into increased yield (Figure 4). Other root phenes with vital importance to low P and drought environments are basal root whorls number, basal root number and adventitious roots. Basal root whorls number was positive and highly significantly correlated with basal root number (r = 0.900 ***). Seven genotypes, SEN 52, SER 125, INB 841, SEQ342-87 and FBN1211 had higher numbers of basal whorls with higher grain yield under low P and drought environment. Two genotypes BIOF 2-106 and BFS 81 showed low basal root whorls' number with higher grain yield under low P and drought stress. The two commercial varieties, Amadeus 77 and Beniquez, showed similar number of basal whorls, but Beniquez had the lowest grain yield under low P and drought stress ( Figure 5).
The results on basal root number under low P and drought stress showed four genotypes SEN 52, SEQ 342 INB 841 and FBN 12 11-46 with high basal root number and also exhibited higher grain yield under low P and drought treatment. For the commercial varieties, Beniquez had relatively higher basal root number and the lowest grain yield, while Amadeus 77, had the lowest basal root number and a higher grain yield under low P and drought stress. The sensitive genotype TARSMST-1 had a lower number of basal root as well as lower grain yield under low P and drought stress. BFS 81, BIOF 2-106, SER 125, MER 2212-28 and MHR311-17 exhibited higher grain yield under low P and drought stress and low basal root number ( Figure 6).

Root Distribution from Soil Cores
The soil cores showed similar total root length for all phenes at harvest. Overall, the phenotype with dense taproot lateral branches had slightly more total root length in deeper horizons. A higher total root length was observed in the topmost 10 cm and last 10 cm division, which corresponded to 50 cm depth in plants grown under Low P and drought stress as well as high P and irrigated treatments. In general, plants grown under low P and drought showed a larger total root length in all segments of cores compared to high P and well-watered plants (Figure 7).

Discussion
In this study we observed that taproot lateral branching and taproot lateral number are phenes under some degree of genetic control and variation exists for taproot phenotype in common bean genotypes. In addition, tap lateral branching density and taproot lateral number under controlled environment (pot study) had a significant relationship with shoot biomass, total root biomass and taproot diameter under low phosphorus and drought stress in pot, and these were also correlated with yield grain under stressed environments.
Root development and shoot development are interdependent. According to Polania et al. [29], shoot growth supplies the roots with carbon and certain hormones, in return root growth supplies the shoot with water, nutrients as well as hormones. Increased grain yield through better plant growth in harsh conditions (drought and low fertility) is achieved if the root system is able to supply water and nutrients without absorbing too much photo assimilate from the shoot [30]. Futhermore, associations between roots and fungi, especially mycorrhizal fungi are also an avoidance mechanism of drought [31,32]. Through its hyphae, mycorrhizal fungi are able to extend to soil particles and penetrate in smaller soil pore spaces allowing mycorrhizal plants to access water that is normally inaccessible to non-mycorrhizal plants [33]. This improves drought resistance by increasing water uptake and yield under limited soil moisture [34]. It has also been observed that mycorrhizal fungi when associated with common bean root systems, impove P acquisition [15,35]. They do so by colonising the root system, thus connecting the crop and soils; improving the efficiency with which the root system draws P from the soil. The external hyphae permits the roots to tap large amounts of soil, thereby absorbing P from non-labile sources [36], passing it to the plant in exchange for organic elements released in soil by roots [15,35]. Through this relationship, roots absorb P which would otherwise be inaccessible to them from P deficient soils [35].
In this current study, a positive correlation between shoot biomass with taproot length, total root length, taproot lateral branching and grain yield was observed, while a there was a negative association with taproot biomass and total root biomass. There was also a positive relationship between tap root biomass with total root biomass, tap root length, taproot lateral branching, basal root number, basal root whorls number and grain yield. These positive associations between phenes under pot study and field measurements are of importance in phenotyping process as a tool for screening and selection in breeding programs [37].
Asfaw et al. [38], also reported that interdependency among numerous root phenes are of practical interest in breeding programs where multiple selection of more than one trait would be convenient, since a selection for one trait may influence improvement or deterioration in an associated root phene [8]. Miguel et al. [20] and Walk et al. [25], found that under some environmental conditions, increased adventitious rooting system can decrease allocation of resources to the growth of the taproot and lateral roots emerging from basal roots. This agrees with our pot study, where we observed negative relationship between shoot biomass with taproot biomass and total root biomass under low phosphorus and drought conditions. Negative associations were also observed between total root length with taproot lateral branching and basal root number (Table 3). These tradeoffs in resources allocation to distinct root classes during early stage of vegetative growth have effects for subsequent vegetative growth and reproduction [37].
Although no previous study has explicitly explored the utility of taproot lateral length and lateral branching for plant performance under edaphic stress, results from the present study suggest that an increased taproot lateral branching and length provide benefits under multiple environments. Genotypes that exhibited higher scores of taproot lateral branching in the field such as SER 125, BFS 81, FBN12111-66 and MER 22 11-28, combining with higher grain yield under low P and drought were ranked as deeper rooted and suited to environments where water is limiting. In contrast, genotypes INB 814, SEN52, BIOF 2-06 and SEQ342-87 had relatively low scores of taproot branching density, but with better yield under low P and drought, and were classified as shallow rooted and suited to environments where P is limiting. The vigor of the taproot associated with many lateral branching of these genotypes, permitted plants to access available water in the deeper horizon and cumulatively foraging for available nutrients through the lateral branching. Although taproot is difficult to evaluate in the field, a longer taproot with higher lateral branching density may be related to greater rooting depth in mature plants [37,39]. Lynch [40], Lynch and Wojciechowski [41] and Wasson [42] reported a positive association between rooting depth and access to water in the deep soil profile under drought.
Previous studies by Miguel et al. [20] and Rangarajan [21] on the utility of greater basal root whorls number under P uptake stress have been supported in this study, in addition to showing that more basal whorls number and greater basal root number are associated with increased grain yield in non-stress conditions. The results of this study indicated that genotypes SEN 52, SEQ 342 INB 841 and FBN12 11-46 produced more basal root whorls number as well as basal root number and higher yield under low P and drought stress. Basal root has been reported to have mixed effects on performance under low P and drought stress, by exploring both shallow and deeper horizons under conditions of high root mortality [8]. Therefore, basal root whorl number and basal root number are phenes of capital importance for resources acquisition under edaphic stress environment, as it was supported by other studies [20].
Furthermore, our results supported that plants under water stress increased total taproot length at deeper profiles, as some studies have demonstrated, that the basic difference between shallow and deeper-rooted genotypes is expressed in the respective stress conditions imposed [43,44]. In addition, some of the reported studies clearly showed that the presumed relation between rooting traits and drought tolerance in some way overlap by escape mechanisms related to the phenology of the genotypes [5,45].
Drought is known to affect soil penetration resistance, and thus influence root development and branching [46]. As with the results from field study from soil cores, we observed differences between deeper-rooted phenotypes and depth of rooting. Plants with dense taproot lateral under water deficit localized relatively greater root length below 40 cm compared to the plants with sparse taproot lateral branching. This is in agreement with what some studies have shown supporting that a small volume of roots in deeper layers where water is not limiting, would be enough to fully supply water to the plants when the topsoil is dry [47], thereby improving water uptake. Our interpretation is that the severe the drought the plant faces, the more roots tend to be allocated at deeper profiles as a result of survival mechanism of foraging water.

Conclusions
Common bean genotypes differed for total root biomass, taproot biomass, taproot length and tap root lateral branching under low phosphorus and drought stress. Phosphorus and drought tolerance in common bean are related to a stronger development of the root system which plays an important role in water and nutrient absorption. Although no previous study has explicitly explored the utility of taproot lateral length and lateral branching for plant performance under edaphic stress, results from the present study suggest that an increased taproot lateral branching and length provide benefits under multiple environments, and this is a major contribution in seedling traits associated with drought stress. Thus, genotypes with greater taproot branching density and length, and high yield were identified as deeper-rooted and suited to water stressed environments, while genotypes with fewer lateral branching density were identified as shallow-rooted and suited to low phosphorus availability environments and genotypes with poor root system (lower taproot branching density and lower yield) were not adapted to either low P or drought stress.