The Role of Deep Roots in Sorghum Yield Production under Drought Conditions

: Root function plays a vital role in maintaining crop production. However, the role of deep roots in yield production and their e ﬀ ects on photosynthetic performance in sorghum remain unclear. This study aimed to provide theoretical supports for establishing highly e ﬃ cient root systems of sorghum to achieve more yield under certain conditions. In this study, two sorghum ( Sorghum bicolor L. Moench) cultivars, Jiza127 and Jiza305, were cultivated in soil columns as experimental materials. Three treatments (no roots removed, CK; roots removed at 30 cm underground, R30; roots removed at 60 cm underground, R60) were carried out under drought conditions during the ﬁlling stage. The root bleeding intensity, endogenous substances in the root bleeding sap, photosynthetic characteristics, dry matter accumulation, and yield were measured. The results showed that R30 and R60 signiﬁcantly reduced yield in both sorghum cultivars, and the e ﬀ ect of R30 on yield was greater than that of R60. The contributions of roots below 30 cm to the yield of both sorghum hybrids were notably higher than those below 60 cm. R30 signiﬁcantly reduced the dry matter weights (DMWs) of leaves, stems, sheaths, and panicles. R60 signiﬁcantly reduced the DMW of panicles but had no signiﬁcant e ﬀ ect on the DMWs of leaves and stems. R30 signiﬁcantly reduced the photosynthetic level and PSII reaction center activity; however, the e ﬀ ect of R60 was not signiﬁcant. Although both R30 and R60 signiﬁcantly reduced root activity and the soluble sugar, amino acid, gibberellin (GA 3 ), and abscisic acid (ABA) contents of the root bleeding sap, some of the above indicators in R60 were signiﬁcantly higher than those in R30 during the ﬁlling stage, indicating that the deeper roots (below 30 cm) had a critical regulatory e ﬀ ect on the physiological processes of the aerial parts in sorghum, which resulted in a stronger e ﬀ ect on yield, especially under drought conditions. In brief, the deep roots of sorghum played a key role in yield production, but the roots in di ﬀ erent soil depths regulated yield production in di ﬀ erent ways. Our results indicate that deep roots of sorghum deserve consideration as a potential trait for yield improvement especially under drought conditions.


Introduction
Plant roots are important organs that absorb water and nutrients from the soil and are a center for the biosynthesis and transport of plant hormones such as abscisic acid (ABA) [1][2][3][4]. The morphological

Plant Materials and Experimental Design
This experiment was conducted at the experimental base of Shenyang Agricultural University from 2015 to 2016. Average temperature and sunshine duration are similar in both experimental years ( Figure 1). The sorghum hybrids Jiza127 (J127) and Jiza305 (J305) were selected as experimental materials which are mainly cultivated in northeastern China. Before the test, PVC tubes (tube length 1 m, inner diameter 30 cm) were cut into two sections of the same size longitudinally, fixed tightly with pipe hoops (Figure 2), and placed in a 1 m deep soil pit. The original soil layer was placed into the PVC tubes to ensure the spatial distribution of the original soil layer. After that, the soil was compacted by watering. The tubes were arranged in two rows with a large ridge (ridge spacing 66 cm, row spacing 33 cm), and the planting density was 6 plants·m −2 . Five uniformly sized seeds were sown in each soil column on May 8. One seedling was left at the five-leaf stage. Surface soil samples were taken at 5 cm soil depth. Soil was silt loam with a pH of 7.0, organic matter content of 30.82 g kg −1 , alkali hydrolysable N of 114.52 mg kg −1 , available P of 78.33 mg kg −1 , available K of 102.92 mg kg −1 . At 45 cm soil depth, soil was loam with organic matter content of 14.82 g kg −1 , alkali hydrolysable N of 59.21 mg kg −1 , available P of 12.61 mg kg −1 , available K of 51.46 mg kg −1 . At 80 cm soil depth, there were organic matter content of 13.19 g kg −1 , alkali hydrolysable N of 45.35 mg kg −1 , available P of 12.53 mg kg −1 , available K of 75.85 mg kg −1 in the loam soil. Diammonium phosphate (2.46 g) was used as seed fertilizer, and urea (3.33 g) was applied to each column at the jointing stage based on local sorghum cultivation. Three treatments (no roots removed, CK; roots removed at 30 cm underground, R30; roots removed at 60 cm underground, R60) were carried out during the filling stage. There were 60 columns per treatment. When the roots were removed, the pipe hoop was opened, and the roots were cut with a saw at the desired depth. Meanwhile, the soil was not moved and the pipes were closed again. A 30-d drought period was started at the same time as the root removal, and the soil moisture was controlled to produce moderate drought conditions (the soil water content was 45%-50% of the maximum water holding capacity in the field). The soil columns were protected by a mobile rain shelter from the rain during the drought-stressed period. The soil water content was measured daily at soil depth of 10 cm using the soil water sensor ML2x (DELTA-T, United Kingdom). To ensure that the soil water content remained constant, an automatic equipment of drip irrigation was used to supplement the soil water to maintain the aimed soil water content. After root removal (from the beginning of the filling stage, August 17), plant aerial part and root bleeding sap from three plants in each treatment were collected every 10 d, and photosynthetic parameters and fluorescence parameters from ten plants per treatment, biomass and substances in root bleeding sap from three plants in each treatment were measured. The plants in all treatments were rewatered from 30 d until maturity and harvested on September 28 in the two years. All the parameters in the experiment except yield were collected in 2015.

Determination of Photosynthetic Parameters, Fluorescence Parameters and SPAD Values.
The net photosynthetic rate (Pn), stomatal conductance (Gs), and transpiration rate (Tr) of the second leaf from the top were measured using an LI-6400 portable photosynthetic system analyzer (LI-COR Inc., Lincoln, NE, USA) following the method of Chang et al. [36]. Ten plants were randomly selected for each treatment. Chlorophyll fluorescence parameters, including the initial fluorescence (Fo), maximum photosynthetic efficiency (Fv/Fm), photochemical quenching coefficient (qL) and electron transfer efficiency (ETR), were determined on the same leaf selected for determination of photosynthetic parameters by using a Junior-PAM fluorometer (Walz, Effeltrich, Germany) following the method described by Khoshbakht et al. [37]. The SPAD value (relative chlorophyll content) of the same leaf was determined with a SPAD chlorophyll meter; the upper, middle and lower parts of the leaf were measured, and the average value was calculated.

Collection of Root Bleeding Sap
Three uniform plants were selected in each treatment, and the plants were quickly cut with branch shears at the middle of the second stem node above the ground. After the incision was washed with distilled water and covered with preweighed absorbent cotton, the incision was placed in a valve bag and sealed tightly with waterproof tape. The root bleeding sap was collected from 6:00 am. to 18:00 pm., and the root bleeding intensity was determined by weighing (g·plant −1 ·12 h −1 ).

Determination of Osmotic Adjustment Substance and Hormone Content in Root Bleeding Sap
The soluble sugar content of the root bleeding sap was determined with the anthrone-sulfuric acid method described by Quan et al. [38]. The soluble protein content was measured with the Coomassie Brilliant Blue G-250 method described by Guzel and Terzi [39]. The free amino acid content was determined using the ninhydrin colorimetry method according to Sun et al. [40]. The ABA and gibberellin (GA 3 ) contents were determined using enzyme-linked immunoassay (ELISA). The kit was provided by the Crop Chemical Control Research Center of China Agricultural University.

Determination of Dry Matter Weight and Yield
Three plants were sampled from each treatment on the 10th, 20th, and 30th days from the initiation of root removal, and the dry matter weights (DMWs) of different organs (leaves, stems, sheaths, and panicles) were measured. The plant samples were killed at 105 • C for 30 min and dried at 80 • C to a constant weight. At harvest, 142 days after seeding, 10 plants were taken from each treatment, and the biological yield and grain yield were determined after air-drying.

Statistical Analysis
This experiment was conducted as a complete randomized design. Significance of main effects of root removal was determined using one way analysis of variance in software SPSS 18.0. Means were separated by Duncan's multiple range test at p < 0.05. SPSS 18.0 software was also used for regression analysis. The data are presented as the means ± standard deviation from all replications. Different characters indicate significant differences.

Effects of Root Removal on Photosynthetic Performance
The photosynthetic parameters and SPAD values of plants in the different treatments were ranked as follows: CK > R60 > R30 (Table 1). In most measurement periods, R60 showed no significant difference as compared to CK, while R30 showed significantly reduced photosynthetic parameters and SPAD values. The greater reduction of the photosynthetic parameters as compared to CK was measured in variety J127. At 30 d after root removal, the Pn, Gs, Tr, and SPAD values of R30 in J127 were reduced by 32.15%, 53.85%, 18.18%, and 25.56% compared with those of CK in J127, respectively. The relative decreases at 10 or 20 d after root removal were similar, but slightly less pronounced decreases as compared to those at 30 days.
Fo in different treatments was highest in R30, followed by R60 and then CK, and R30 showed a significantly increased Fo compared with CK at all measurement periods ( Table 2). In contrast, Fv/Fm, qL, and ETR were ranked as follows: CK > R60 > R30. At 30 d after root removal, both varieties in the R30 treatment showed significantly reduced Fv/Fm, qL, and ETR, which were decreased by 10.00%, 19.05%, 33.79% in J127 and 11.43%, 31.25%, and 17.85% in J305, respectively. R60 showed no significant difference from CK. The relative decreases at 10 or 20 d after root removal were similar, but slightly less pronounced decreases as compared to those at 30 days. No roots removed, CK; roots removed at 30 cm underground, R30; roots removed at 60 cm underground, R60; the net photosynthetic rate, Pn; stomatal conductance, Gs; transpiration rate, Tr; chlorophyll relative content, SPAD value. Different characters within the same column indicate significant differences at p < 0.05. The data are represented as the means ± standard deviation (n = 10).

Effects of Root Removal on Root Bleeding
Root bleeding intensity was significantly affected by root excision (Figure 3). R30 and R60 significantly reduced root bleeding intensity in J127 at 10 and 20 d after root excision. The root bleeding intensity was decreased by 42.69% and 35.30% in R30 and R60 at 10 d after root removal and 41.22% and 33.64% at 20 d after root removal compared with that in CK, respectively. In J305, R30 and R60 showed significantly reduced root bleeding intensity at 20 and 30 d after root removal. Compared with that in CK, the root bleeding intensity was decreased by 46.52% and 24.46% in R30 and R60 at 20 d after root removal and 52.60% and 33.99% at 30 d after root removal, respectively.  The content of osmotic adjustment substances in the root bleeding sap was significantly affected by root removal (Figure 4). R30 led to a significantly reduced soluble sugar content. R60 led to a significantly reduced soluble sugar content in J127 compared with CK at 10 and 20 d after root removal; the decreases were 25.71% and 12.98%, respectively. In J305, R60 also led to significantly decreased soluble sugar contents; the decreases were 20.94%, 17.33%, and 11.78% at 10, 20, and 30 d after root excision, respectively. Furthermore, R30 significantly reduced the soluble protein content of the root bleeding sap of J305 by 26.71%, 20.09%, and 31.55% at 10, 20, and 30 d after root removal, respectively. In contrast, R60 had no significant effect on the soluble protein content of the root bleeding sap at 10 and 30 d after root removal. R30 significantly reduced the amino acid content of the root bleeding sap by 47.41%, 34.63%, and 32.33% in J127 at 10, 20, and 30 d after root removal, respectively, and by 40.59%, 29.82%, and 30.23% in J305 at 10, 20, and 30 d after root removal, respectively. R60 significantly reduced the amino acid content in J127 by 30.94%, 20.65%, and 25.95% at 10, 20, and 30 d after root excision, respectively, and by 28.42% and 18.76% in J305 at 10 and 30 d after root excision, respectively.  At 10 d after root excision, compared with that of CK, the ABA content in the root bleeding sap of R30 was significantly increased, while that of R60 showed no significant difference from CK. R30 significantly reduced the ABA content in the root bleeding sap of J127 at 20 and 30 d after root removal. R60 significantly reduced the ABA content in the root bleeding sap of J127 at 20 and 30 d after root removal and that of J305 at 30 d after root removal ( Figure 5).

Effects of Root Removal on Dry Matter Accumulation and Yield
Compared with CK, the dry weight of leaves, stems, sheaths, and panicles in J127 decreased by 33.10%, 46.88%, 43.08%, and 24.63% under R30 at 30 d after root removal, respectively (Table 3). R60 had no significant effect on the dry weights of leaves and stems in J127 but significantly reduced those of sheaths and panicles by 30.11% and 13.15% compared with CK, respectively. In J305, R60 had no significant effects on the dry weights of the leaves, stems, and sheaths but significantly decreased the dry weight of the panicles, which was reduced by 14.27% compared to that of CK. The relative decreases at 10 or 20 d after root removal were similar, but slightly less pronounced decreases as compared to those at 30 days. No roots removed, CK; roots removed at 30 cm underground, R30; roots removed at 60 cm underground, R60. Different characters within the same column indicate significant differences at p < 0.05. The data are represented as the means ± standard deviation (n = 3).
Both R30 and R60 significantly reduced grain yield ( Table 4). The two-year average grain yield of R30 and R60 decreased by 53.46% and 36.04% compared with that of CK in J127 and by 48.16% and 33.74% in J305. The biological yield was also significantly affected by root removal. The average biological yield of R30 and R60 in J127 decreased by 19.41% and 9.40% compared with that of CK and by 31.42% and 19.86% in J305, respectively. The yield trend of the two-year experiment was consistent. No roots removed, CK; roots removed at 30 cm underground, R30; roots removed at 60 cm underground, R60. Different characters within the same column indicate significant differences at p < 0.05. The data are represented as the means ± standard deviation (n = 10).

Correlation of Root Bleeding with Grain Yield and Pn
A significant positive relationship (R 2 = 0.6072, p < 0.01) was found between the Pn and root bleeding intensity, indicating that the increasing root bleeding intensity had a positive influence on Pn.
Similarly, there was a significant positive relationship (R 2 = 0.5862, p < 0.01) between grain yield and root bleeding intensity, indicating that the increasing root bleeding intensity had a positive influence on grain yield ( Figure 6). Figure 6. The correlation between root bleeding intensity and grain yield, Pn (net photosynthetic rate). The data was from two cultivars and included every treatment and replicates; p < 0.01 represents extremely significant correlation.

Discussion
Root activity reflects the strength of root metabolic capacity and has an important role in promoting the formation of yield [41,42]. Munns et al. [43] reported that a decline in osmotic adjustment substances increased the osmotic potential of plant root cells, resulting in a decrease in the water absorption capacity of the root system. In this study, R30 and R60 significantly reduced root bleeding intensity and osmotic adjustment substance content, indicating that roots below 30 cm still had strong water absorption capacity under drought conditions during the filling stage. Meanwhile, some indicators in the root bleeding sap of R60, such as soluble protein and soluble sugar, were significantly higher than those of R30 during the filling process, indicating that deeper root systems (below 30 cm soil depth) hid high physiological activity and strong regulatory effect on the aboveground parts. With respect to signaling substances, root excision significantly reduced the GA 3 content in the root bleeding sap. This result indicates that the reduction of GA 3 synthesis in the root system should lead to a decrease of the available GA 3 in the aerial part. GA 3 is an important hormone in plants that can promote plant growth and material accumulation [44]. These results suggested that the decrease in synthesis and output of GA 3 may be one of the reasons for the reduction in aboveground dry matter, and this effect was more obvious in R30. The changes in ABA were more complicated due to root excision and plant senescence. On the one hand, the change trend in ABA content in the root bleeding sap of CK reflects a response to drought conditions and senescence during the filling stage. A similar change trend of ABA was also found in the results of Abid et al. [45]. The changes in ABA content in R30 and R60 may indicate a stress response in the root system in the early stage due to root excision. This damage and loss in the root system decreased the ability of the roots to synthesize ABA, which significantly reduced the ABA content exported from the roots to the shoots. Lower ABA content in the leaves could reduce plant drought resistance [46], ultimately leading to a reduction in yield. The correlation of the root bleeding intensity with Pn and grain yield also implied that higher root activity could better support the development of photosynthetic potential that ultimately resulted in higher yield.
Song et al. [47] and Qi et al. [48] reported that deep roots play a critical role to improve root vitality, maintain root nutrition and water supply to the aerial part, promote the production of photosynthetic materials and grain filling, and increase yield. In the present study, regardless of R30 or R60, the yield of sorghum was significantly reduced by root removal, reflecting the important role of deep roots in yield production. Root excision at 30 cm underground significantly reduced leaf chlorophyll content (SPAD value), photosynthetic level, Fv/Fm, qL, and ETR and increased Fo. An increased Fo and a decreased Fv/Fm indicate that the PSII reaction center is damaged [49]. These results revealed that root excision directly reduced the function of the PSII reaction center, and photosynthetic activity and electron transfer were affected. As a result, the photosynthetic capacity was reduced, and the yield was significantly reduced. In addition, the decreased photosynthetic capacity reduced the overall biomass of the aerial parts, which manifested as decreased dry weights of the sorghum leaves, stems, sheaths, and panicles. However, the effect of root excision at 60 cm underground on the photosynthetic parameters was not significant, indicating that R60 had little effect on photosynthetic performance. From the perspective of root system distribution, the root system was shown to be mainly distributed in the upper 0-30 cm of topsoil, which contained up to approximately 95% of the total number of roots in cotton and maize, respectively [50,51], indicating that roots of taproot and fibrous root crops are mainly distributed in shallow layers, and the surface root system was directly related to photosynthesis [52]. However, in this study, the contribution of roots below 60 cm accounted for more than 30% of the yield, indicating that deeper root activity, which was reflected by the root bleeding intensity and component content, played a critical role in yield production under drought conditions during the filling stage. Schittenhelm et al. [53] indicated that root depth was also a significant feature of drought resistance in sorghum. Common farming fertilization is mainly concentrated in the plow layer, and drought also usually occurs in the cultivated layer, which reduces nutrient availability. Although the root system can penetrate deeper soil depths to obtain water under drought conditions, nutrients are relatively scarce in these deeper layers, leading to a degree of spatial dislocation between nutrients and water. As the root system tends to grow toward fertilizer and water, the spatial distribution of the root system changes accordingly; thus, the role of deep roots will be more obvious under drought conditions.
In brief, R30 and R60 significantly reduced the biological and grain yield of sorghum, especially grain yield. R30 mainly reduced yield by reducing photosynthetic level and root activity, while R60 had no significant effect on photosynthetic characteristics. Although root activity was significantly reduced in R60, it was still significantly higher than that of R30, indicating that the greater vitality of deeper root systems (soil depths below 30 cm) was the reason for their greater contribution to yield, especially under drought conditions. In conclusion, the deep roots of sorghum play a crucial role in yield production, but the roots in different soil depths regulate yield production differently. Our results support the hypothesis that deep roots of sorghum greatly contribute to yield production and thus merit investigation as a potential agronomic trait for more yield under drought conditions.