KANG Liang,LIANG Qiong-yueJIANG Qiang,YAO Yi-huaDONG Meng-mengHE BingGU Ming-hua

1 State Key Laboratory for Conservation and Utilization of Subtropical Agro-bioresources/College of Agriculture,Guangxi University,Nanning 530004,P.R.China

2 College of Pharmacy,Guangxi Medical University,Nanning 530021,P.R.China

3 Modern Agricultural Technology Research and Promotion Center,Baise 533612,P.R.China

Abstract As one of the top three tuber crops of the world,cassava is a staple food and feed crop for tropical and subtropical regions.Because of its high drought resistance and tolerance to nutrient deficiency,cassava is usually cultivated on hilly areas that are nutrient-poor.Nitrogen (N) is one of the significant factors affecting cassava yield.A double factorial (N level×genotypes)split-plot field experiment was conducted to analyze differences in yield and N accumulation of 25 cassava genotypes under low-N conditions to screen for cassava genotypes with high-N efficiency.The two-year field experiment showed that cassava yield and N accumulation are determined by specific genotypes,soil N levels,and year.Among these factors,soil N levels are the main factors that are responsible for differences in cassava yield.When yield and relative N accumulation under low-N conditions were used as screening markers,we identified an efficient and responsive genotype (SC10),and two inefficient and non-responsive genotypes (SC205 and GR5).The efficient and responsive genotype and the inefficient and non-responsive genotype can be used as study materials to further reveal the mechanisms for high-N efficiency in cassava.

Keywords: cassava,high-nitrogen (N) efficiency,genotype screening,low-N soil

1.Introduction

Nitrogen (N) is one of the most important mineral elements required for crop growth and development.In agricultural production,N fertilization is required to sustain the desired crop yield.Excessive use of N fertilizers,however,can drastically increase agricultural costs,delay crop maturation and lodging,and increase susceptibility to pests and diseases (Savaryet al.2012;Zakkaet al.2016;Xuet al.2017),ultimately resulting in a reduction in crop yield and quality (Magwazaet al.2017;Bisbiset al.2018;Kyriacouet al.2018).Also,residual nitrates in agricultural products can also affect human health (Sotoet al.2015).Excessive application of fertilizers promotes N loss into the environment,causing surface water eutrophication and groundwater degradation (Guoet al.2010;Mcallisteret al.2012;Liuet al.2017).

Cassava is one of the top three root and tuber crops of the world.According to the Food and Agriculture Organization(FAO) statistics,more than 100 countries and regions in the world are engaged in cassava cultivation.Cassava has become a widely cultivated food and feed crop in the tropical and subtropical regions and essential raw material and emerging energy crop for industrial processing.However,the application of N fertilizers is necessary for achieving high cassava yield,but the economic return diminishes with time due to low use efficiency.Therefore,improving N fertilizer use efficiency in crops is currently a significant goal of cassava production.Breeding of genotypes with high N use efficiency (NUE) that can tolerate low-N conditions is considered as a practical approach for sustaining cassava production without causing environmental pollution.

Many studies have indicated that NUE shows similar patterns in different cultivars (Koutroubas and Ntanos 2003;Zhanget al.2009;Gajuet al.2011;Kesselet al.2012).NUE in 135 cassava cultivars through field experiments was analyzed and NUE index decreases as follows: fresh weight of tuber roots＞number of tubers＞fresh weight of aboveground organs＞dry matter percentage＞branch number of the main stem＞plant height were found by Xiao (2016).The fuzzy membership function to identify five genotypes with high NUE and five genotypes with low NUE from 135 cassava genotypes was used (Xiao 2016).Differences in physiological traits of seedlings of 31 cassava cultivars under varying degrees of N deficiency were observed and cluster analysis to group the NUE of these cassava seedlings into three major categories was conducted (Guo 2013).Among these cultivars,two,namely,RoYong0 and SC205,showed unique physiological traits and thus can be used as materials for screening cultivars.

Some researchers have suggested that genotypes associated with high yield under high-N levels can be considered high-NUE genotypes,while low yield under high-N levels is low-NUE genotypes (Mollet al.1982;Ladhaet al.2005).Other researchers have suggested that genotypes that are tolerant of low-N conditions and can produce high yields under such conditions can be considered high-NUE genotypes (Heet al.2014;Quanet al.2017).The advantages of screening for high-NUE genotypes under high-N levels,with an emphasis of comparison between genotypes and an aim of pursuing higher yield under equally high-N conditions,are that the selected genotypes have good fertilizer tolerance and are easier to obtain high yields under high-N conditions (Bellet al.2013;Chuanet al.2013;Watmuffet al.2013;Zhanget al.2013).Its disadvantages are that it is difficult for researchers to truly differentiate between high-NUE and high-yield mechanisms (Quanet al.2017;EI-Sobky and Sayed 2017;Wanget al.2018).Screening for high-NUE genotypes under low-N conditions usually focuses on the level of response of crops to low-N conditions.The advantages of this method are that the differences in traits among various genotypes are specifically caused by N deficiency (Parket al.2017;Liang and He 2018).The disadvantage of this method is that the selected genotypes may not be suitable for agricultural production based on high-N fertilizers.However,enhanced environmental pollution has forced farmers to give up the concept of high-N fertilizer levels and utilize more environmentally friendly genotypes with low-N fertilizer requirements.Therefore,superior high-NUE cassava genotypes must have a strong low-N tolerance for efficiently absorbing and utilizing N under low-N conditions.In reality,the goal of many researchers is to increase NUE by the way of improving low-N tolerance of crops (Dinget al.2005;Diazet al.2006;Parket al.2017).To screen for cassava genotypes with high NUE under low-N conditions,we employed a double factorial(N level×genotypes) split-plot design.We screened the responses of 25 cassava genotypes to low-N levels through field experiments to select high-NUE cassava genotypes.This will provide a basis for further research on efficient N uptake and utilization mechanisms in cassava.

2.Materials and methods

2.1.Field screening experiments

The tested materials were 25 cassava genotypes.The names and number of these genotypes are shown in Table 1.The study site is located at the Xionghui Village,Nanyang Town,Nanning City,Guanxi Province,China (30°56′N,75°32′E).The two field experiments were performed under two different soil fertility conditions in 2014 and 2015.The soil in the two study sites was sandy loam,and the physiochemical properties of the soil are shown in Table 2.

Two different N treatments were set up: 0 kg ha-1(low-N treatment,N0) and 125 kg ha-1(normal N treatment,N1).A double factor split-plot design was used for field experiments,with 30 plots.The size of each plot was 1 m×1 m,10 plots lined up a row and cultivars of the same genotype were grown,and each line was a random distribution in the same solution.The main plot was the N treatment,and the subplots were genotypes.The same amount of potassium (K) and phosphorus (P) fertilizers were used for all plots (46 kg ha-1P2O5and 90 kg ha-1K2O).Urea was used as N fertilizer,superphosphate was used as P fertilizer,and potassium chloride (KCl) was used as K fertilizer.The P fertilizer was applied once,while the N and K fertilizers were applied thrice (40,30,and 30%),with the first and second applications conducted 30 and 60 d after sowing,respectively.At the Site 1 (chemical properties are shown in Table 2),cassava was planted and harvested on 28 March,2014 and 25 December,2014,respectively.Cassava was planted at the Site 2 and harvested on 1 April,2015 and 28 December,2015,respectively.Conventional field management methods were used during the entire growth period.

During the harvesting of cassava,the fresh weight of aboveground organs and total fresh weight of tuber roots of individual plants were measured.The fresh weight of a fixed amount of tuber root,stem,and leaf samples was collected and dried at 105°C for 30 min and then further dried until a constant weight was achieved at 65°C.The water and N contents of the samples were measured and converted to the total dry weight.The method of analyzingN content was as follows.After samples were digested by concentrated H2SO4-H2O2,these were cooled and the volume to made up to 50 mL with deionized water.A continuous flow analyzer (AAIII,SEAL Analytical,Germany)was used to measure N concentration in the digested samples (Liet al.2017).

Table 1 Names and sources of cassava genotypes used in this experiment

The various indicators were calculated as follows:

The total dry mass of the plant (t ha-1)=Tuber root dry weight (t ha-1)+Shoot dry weight (t ha-1)

N accumulation (kg ha-1)=N content of organ (kg t-1)×Dry weight of the organ (t ha-1)

where NUtE is N utilization efficiency.

2.2.Statistical methods

SPSS (version 22.0;SPSS Inc.,Chicago,IL,USA)and Duncan’s statistical method were used for multiple comparisons of the experimental data.

3.Results

3.1.Combined ANOVA for two years of data on yield and N accumulation

Table 3 shows the variations in cassava yield and N accumulation in 2014 and 2015.Cassava yield and N accumulation were significantly affected by genotype,soil N level,and year.The factors affecting yield in descending order were N levels＞genotype＞year.The factors affectingN accumulation in descending order were year＞N level＞genotype.

Table 2 The chemical properties of the tested soil

Table 3 ANOVA of tuber yield and nitrogen accumulation of two years separately and pooled

3.2.Yield,tuber root N content,and N accumulation

N levels and genotype significantly affected cassava yield and N accumulation (P＜0.05),indicating that assessment of N uptake efficiency (NUpE) of different genotypes under various N conditions can be used to screen high-NUpE and low-NUpE genotypes.The differences in cassava yield and N accumulation among various years were relatively significant.Therefore,for subsequent analysis,the experimental results of each year were assessed separately(Tables 4 and 5).

The mean tuber root yield under the same treatment conditions was higher in 2015 than that in 2014.In 2014,under N0 conditions,and the range of variation in tuber root yield of the 25 cassava genotypes was 1.79-17.36 t ha-1,with a mean value of 10.35 t ha-1.The top five genotypes based on yield in descending order were SC7＞TY1＞SC10＞SC8013＞SC5 (Fig.1).Under N1 conditions,the range of variation in tuber yield of the 25 cassava genotypes was 1.94-26.48 t ha-1,and the mean yield was 15.79 t ha-1.The top five genotypes based on yield in descending order were SC7,GR5,SC5,TY1,and SC201.In 2015,under N0 treatment,the range ofvariation in tuber yield of the 25 cassava genotypes was 2.21-17.83 t ha-1,with a mean value of 11 t ha-1.The top five genotypes based on yield in descending order were SC8002＞SC10＞TY1＞KU50＞SC7 (Fig.2).Under N1 treatment,the tuber yield of the 25 cassava genotypes varied from 4.75 to 31.79 t ha-1,with the mean value of 17.03 t ha-1.The top five genotypes by yield in descending order were SC6068＞SC8002＞SC10＞KU50＞TY1.In summary,genotypes in the top five in terms of yield in both years were TY1,SC10,and SC7 under the N0 treatment and TY1 under the N1 treatment.

Table 4 N utilization efficiency (NUtE) of cassava genotypes in 2014 and 2015 (kg kg-1)

Table 5 Path analysis of biological yield as determined by nitrogen (N) accumulation and utilization efficiency in cassava genotypes

Fig.1 Yield,tuber root nitrogen (N) content,and plant N accumulation of different cassava genotypes in 2014 (A,yield;B,tuber root N;C,total N accumulation).N0,low-N treatment,0 kg ha-1;N1,normal N treatment,125 kg ha-1.Bars mean SD (n=6 independent experiments).The bars with the same letter indicate no significant difference at the 0.05.

N accumulation in the entire plant in 2015 was higher than that in 2014 under the same conditions.In 2014,under the N0 treatment,N accumulation in the entire plant of the 25 cassava genotypes ranged from 6.84 to 60.08 t ha-1,with a mean value of 36.14 kg ha-1.The top five genotypes in descending order were SC124＞TY1＞SC7＞SC8013 and SC201 (Fig.1).Under N1 treatment,N accumulation in the entire plant of the 25 cassava genotypes ranged from 14.41 to 126.39 t ha-1,with a mean value of 80.25 kg ha-1.The top five genotypes in descending order were SC7＞TY1＞SC5＞SC124＞XX048.In 2015,under N0 treatment,N accumulation varied from 38.26 to 182.09 t ha-1,with a mean value of 116.09 kg ha-1.The top five genotypes in descending order were SC10＞SC8002＞SC6068＞TY1＞KU50 (Fig.2).Under N1 treatment,N accumulation varied from 81.18 to 358.72 t ha-1,with a mean value of 230.21 kg ha-1.The top five genotypes in descending order were SC6068＞TY1＞KU50＞SC10＞SC102.

In summary,the genotypes in the top five concerning N accumulation in both years were TY1,SC10,and SC7 under the N0 treatment and was TY1 under the N1 treatment.

N utilization efficiency (NUtE) in 2014 was higher than that in 2015 (Table 4).In 2014,the NUtE of 25 genotypes under N0 treatment ranged from 224.33 to 439.95 kg kg-1,with a mean value of 327.03 kg kg-1,and the top five highest NUtE genotypes in decreasing order were as follows:GR891＞SC5＞XX048＞SC10＞SC8013.The NUtE of the 25 genotypes under N1 treatment ranged from 161.0 to 326.90 kg kg-1,with a mean value of 228.81 kg kg-1,and the top five genotypes with the highest NUtE in decreasing order were as follows: SC10＞SC8＞SC9＞SC6＞GR5.In 2015,the NUtE of the 25 genotypes under N0 treatment ranged from 131.41 to 230.79 kg kg-1,with a mean value of 172.28 kg kg-1,and the top five genotypes with the highest NUtE in decreasing order were as follows: SC201＞SC7＞GR911＞GR5＞SC8013.The NUtE of the 25 genotypes under N1 treatment ranged from 116.44 to 168.22 kg kg-1,with a mean value of 137.98 kg kg-1,and the top five genotypes with the highest NUtE were arranged in decreasing order as follows:GR911＞NZ199＞SC201＞GR4＞SC5.In summary,the genotype in the top five in terms of NUtE in both years under N0 treatment was SC8013.Under N1 treatment,no genotype reached the top five for NUtE in both years.

Fig.2 Yield,tuber root nitrogen (N) content,and plant N accumulation of different cassava genotypes in 2015 (A,yield;B,tuber root N;C,total N accumulation).N0,low N treatment,0 kg ha-1;N1,normal N treatment,125 kg ha-1.Bars mean SD (n=6 independent experiments).The bars with the same letter indicate no significant difference at the 0.05.

3.3.NUpE and NUtE

The NUE of crops includes NUpE and NUtE.The index of N accumulation was used to evaluate NUpE because it is difficult to calculate levels of N supplied by the medium(including available N in soil and N fertilizer).In this study,the NUtE was defined as the ratio of the total biomass to N accumulation for the whole crop.The contributions of the NUpE and NUtE to biological yield under two N levels in 2014-2015 in cassava were evaluated by path analysis(Table 5).The results showed that NUpE and NUtE both contributed to biological yield,and NUpE made a greater contribution than NUtE.

3.4.Classification of diverse N efficiency genotypes

To screen for high-NUpE cassava genotypes under low-N conditions,we divided the 25 genotypes into four types based on yield and relative N accumulation under the N0 treatment as described by Jhanjiet al.(2013) (Fig.3).Cassava genotypes with high tuber root yield and high relative N accumulation under the N0 treatment were defined as efficient and responsive (ER).Cassava with low tuber root yield and high relative N accumulation under the N0 treatment were defined as inefficient and responsive(IER).Cassava with low tuber root yield and low relative N accumulation under the N0 treatment were defined as inefficient and non-responsive (IENR).Cassava with high tuber root yield and low relative N accumulation under the N0 treatment were defined as efficient and non-responsive(ENR).

In 2014,there were seven ER genotypes (SC10,SC8013,KU50,SC9,SC124,SC201,and TY1),five ENR genotypes (SC7,GR4,SC5,XX048,and GR891),five IER genotypes (SC8,SC6,SC102,NZ199,and HY),and eight IENR genotypes (SC6068,SC11,FX01,GR5,GR3,GR911,SC205,and SC8002) (Fig.1-A).In 2015,there were seven ER genotypes (GR4,SC8002,XX048,GR891,SC10,SC11,and SC5),seven ENR genotypes(SC6,SC7,SC8013,KU50,TY1,SC201,and SC6068),four IER genotypes (FX01,SC8,GR3,and SC124),and seven IENR genotypes (SC205,HY,GR911,GR5,NZ199,SC102,and SC9) (Fig.1-B).Based on the results from both years,SC10 was classified as an ER genotype,SC8 an IER genotype,SC7 an ENR genotype,and SC205 and GR5 as IENR genotypes (Table 6).

4.Discussion

4.1.Contribution of NUpE and NUtE to biological yield

The NUE of crops is composed of NUpE and NUtE,but the specific contributions by the two components are unclear(Anbessaet al.2009;Beattyet al.2010;Bogardet al.2013).

Some researchers have shown that the contribution of NUtE to NUE is greater than that of NUpE.The contribution of NUtE in tomatoes was greater than that of NUpE (Abenavoliet al.2016;Lupiniet al.2017).In rapeseed,NUtE and NUE had a stronger correlation than that between NUpE and NUE(Stahlet al.2016).NUtE and NUE were significantly closely correlated under high- and low-N treatments,but NUpE did not show any significant correlation with NUE under low-N treatment (Gajuet al.2011).Conversely,some researchers have shown that the contribution of NUpE to NUE was greater than that of NUtE.NUpE and NUE in winter wheat showed an extremely significant positive correlation regardless of N levels,whereas NUtE only showed a significant positive correlation under high-N levels (Wanget al.2011).Similarly,the relationship between NUpE and NUE was more closely related to each other than NUtE in green beans (Akteret al.2016).Similar conclusions were reached in oilseed rape (Nyikakoet al.2014),maize(Mollet al.1982),winter wheat (Le-Gouiset al.2000),and soybean (Rotundoet al.2014).In contrast,many factors affect NUE in the field (Berryet al.2010).However,in general,NUpE more strongly influences NUE than NUtE under low-N conditions.The results of our experiments show that under the two N levels,the contribution of N accumulation was greater than that of NUtE (Table 5).Also,the application of N fertilizer increases the contribution of NUpE,but not that of NUtE.Therefore,selective breeding under low-N conditions should focus on N uptake capabilities in NUE cassava genotypes.

4.2.Screening of high-NUE cassava genotypes

Fig.3 Distribution of cassava genotypes with relative nitrogen (N) accumulation and tuber yield at low N (A,2014;B,2015).IER,inefficient and responsive;ER,efficient and responsive;IENR,inefficient and nonresponsive;ENR,efficient and nonresponsive.

Table 6 Classification of cassava genotypes for nitrogen (N) efficiency

Although a large number of high-NUE screening indicators have been used,the lossless,stable,and quickly acquired high-NUE indicators are much more popular (Sadras and Lemaire 2014).For example,Wang (2013) employed cameras to acquire rice canopy images and analyzed the green channel minus red channel (GMR) value and canopy cover (CC) value to determine the level of N deficiency in rice.Erdleet al.(2013) found that theR2of R760/R730(ratio of two near-infrared wavelength reflectance) with dry matter (DM) reaches 0.89,which means R760/R730 can be used for the lossless and rapid classification of rice varieties.In addition,N content in various tissues (Ecarnotet al.2013),N accumulation (Caoet al.2013),activity of enzymes associated with N metabolism (Robinsonet al.2007;Ben Slimaneet al.2013;Konget al.2016),and photosynthesis indicators (Chenet al.2014;Panget al.2014) have been considered by researchers as high-NUE classification indicators.However,N content changes with the developmental stages of crops,whereas the activity of enzymes associated with N metabolism and photosynthesis indicators are related to sampling time and location.Therefore,the aforementioned indicators mentioned above are not suitable for distinguishing genotypes.The classification of cassava into high-NUE genotypes should be based on yield and the ability to tolerate low-N conditions.Our experiments followed the screening methods of Jhanjiet al.(2013),in which low-N yield was taken as thexcoordinate,and relative N accumulation was taken as theycoordinate.Low mean N yield and relative N accumulation in cassava genotypes were used to divide cassava according to NUE.Although the field screening results are the closest to actual production,environmental factors largely affect the stability of the results (Workuet al.2007).Different soil fertility levels may result in variations in screening findings (Sadras and Lemaire 2014).The present study conducted two years of field experiments to screen for cassava genotypes that have a high yield and can tolerate deficient soil conditions from 25 genotypes.We observed significant genotypic differences in yield and N accumulation in the 25 cassava genotypes,which may consequently be utilized in directly identifying genotypes with greater differences in NUE in existing germplasm resources.After combining the data from the two years,we finally identified a genotype with high yield and low-N resistance,SC10,which was defined as a high-NUE and low-N tolerance genotype.This genotype is suitable for propagation in nutrient-deficient hilly areas because it could produce high yield under low-N conditions.A genotype with low yield and less responsiveness to low-N conditions,SC205,was defined as an IENR genotype (Table 6).This genotype is not recommended for cultivation in infertile soils.SC8 was defined as an IER genotype,whose tolerance of low-N conditions deserves further in-depth research.SC7 was defined as an ENR genotype,which has the potential for good yield under high-N conditions and is suitable for cultivation in nutrient-rich soils.We also detected significant differences in yield and N accumulation among cassava genotypes between in 2014 and 2015,which may be due to differences in soil fertility.Also,organic matter,total N,total P,and total K levels in Site 1 were lower than those in Site 2.

5.Conclusion

Assessment of 25 cassava genotypes showed extremely significant differences in NUpE and NUtE,N treatment,and interaction between the two factors.Among these factors,soil N levels are the main factor causing differences in cassava yield.We finally selected SC10 as a high-NUE genotype and SC205 as a low-NUE genotype based on yield and relative N accumulation.Path analysis showed that the contribution of NUpE in cassava to NUE is greater than that of NUtE.Therefore,selective breeding under low-N conditions should focus on N uptake capabilities among NUE cassava genotypes.

Acknowledgements

This research was provided by the Guangxi Natural Science Foundation,China (2014GXNSFAA118077 and 2018GXNSFDA281056) and the Guangxi Postgraduate Innovative Education Research Program,China(YCBZ2017013 and YCSW2018039).We are grateful to Dr.Li Kaimian of Chinese Academy of Tropical Agricultural Sciences for his help in collecting cassava varieties.Gratitude is also expressed to all anonymous reviewers for suggestion on the manuscript.