P-fertiliser and rhizobial inoculation increased the concentration of mineral nutrients in the rhizosphere of two chickpea genotypes

ABSTRACT Co-application of fertilisers and biofertilisers increases solubility and concentration of nutrient elements in the soil. However, information on the effect of P fertiliser and rhizobium inoculation on the concentration of mineral elements in the rhizosphere of field-grown chickpea in contrasting soil types is lacking. We assessed the effect of phosphorus fertiliser, rhizobium inoculation, and rhizobium inoculation plus P fertilizer on the concentration of nutrients in the rhizosphere of 2 chickpea genotypes (ACC#1 & ACC#5) at clay and sandy soil sites in two consecutive years. ACC#5 accumulated higher amounts of rhizospheric nutrients than ACC#1 in both soil types despite the variable response to fertiliser treatment. Rhizobium inoculation plus P fertiliser resulted in higher accumulations of rhizosphere N (7-219%), P (20-348%), K (9-365%), Ca (2-155%) and Zn (15-259%) compared to the other fertiliser treatments suggesting an additive effect of P fertiliser and rhizobium inoculation. Moreover, the effect of fertiliser treatments was more pronounced in the clay than the sandy soil. Our investigation is the first to document the effects of fertiliser treatment, genotype, and soil type on the accumulation of rhizosphere nutrients in field-grown chickpea. However, we propose further studies, incorporating more cultivars and environments, before making practical recommendations.


Introduction
There is a huge and rising domestic demand for chickpea (Cicer arietinum) in South Africa, despite limited local production (Mpai and Maseko 2018) which is exacerbated by poor soil fertility.The most deficient mineral elements in smallholder cropping systems in South Africa include nitrogen (N), phosphorus (P), boron (B) and zinc (Zn) (Mohale et al. 2014).Furthermore, soils in this region are characterised by relatively low levels of organic carbon (Lusiba et al. 2017;Makonya et al. 2019;Maselesele et al. 2021), which aggravates the poor soil fertility (Musinguzi et al. 2013).However, chickpea adopts a variety of mechanisms that enhance the interception and/or solubility of nutrients in poor soils (Li et al. 2004;Pang et al. 2017).These include alteration of rhizosphere pH, secretion of phosphatases, modification of root morphology and/or architecture, production of organic acids, and secretion of protons and hydroxyl ions (Xue et al. 2016;Moloto et al. 2021).However, the responses might vary with plant cultivars (Zhu et al. 2019), and soil types (Motseo 2019).
Although the positive effects of fertiliser application on chickpea have been reported in South Africa (Madzivhandila et al. 2012;Ogola et al. 2013;Macil et al. 2017), Pfertilisers exhibit low efficiency in poor legacy P soils because biological availability of P is significantly decreased through adsorption, precipitation, or complexation (Simpson et al. 2011).Therefore, there is a need for alternative approaches that can improve the solubility and uptake of P (and other important nutrient elements), especially in the high P-fixing soils that are predominant in north-eastern part of South Africa (Lusiba et al. 2017).Recently, biochar increased accumulation of nutrients in the rhizosphere of chickpea in a pot experiment (Lusiba et al. 2021).Co-application of phosphate fertilisers with biofertilisers has also been shown to be beneficial (Bargaz et al. 2018;Wolde-meskel et al. 2018;Motseo 2019).
The co-application of fertilisers and biofertilisers, especially those containing substances that enhance the uptake of nutrients and nutrient use efficiency, increase the efficiency of fertilisers through increasing their solubility and concentration in the soil solution (Thonar et al. 2017).However, there is a dearth of information in literature on the effect of P fertilisation, rhizobium inoculation, and the application of P plus rhizobium inoculation on the concentration of mineral elements in the rhizosphere of field-grown chickpea in contrasting soil types.Therefore, we investigated the effect of application of P fertiliser, rhizobia inoculation, and co-application of P and rhizobia inoculation on the concentration of mineral nutrients in the rhizosphere of 2 chickpea genotypes in contrasting soil types.

Study location
A 2-year field experiment was conducted at two locations: the University of Limpopo's Research Farm, Syferkuil (23°50' S and 29°40' E, and 1230 m asl), and the University of Venda's experimental Farm, Thohoyandou (22°58' S and 30°26' E, and 595 m asl)..The two sites, approximately 150 km apart (in a straight line), are in the Limpopo Province of South Africa.Syferkuil has a semi-arid biome and experiences maximum and minimum temperatures that range from 4 to 20 °C in the winter season, and Thohoyandou is characterised by sub-tropical climatic conditions, with average temperatures in winter ranging from 12 to 24 °C.Both sites receive negligible rainfall during the winter season (Tadross et al. 2006;Makonya et al. 2019).Syferkuil has a predominantly loamy sand with a slightly alkaline pH while the soil at Thohoyandou is deep well-drained clay soil, with slightly acidic pH, classified as Rhodic Ferrosols (Fey 2010).Therefore, the sites were selected due to their contrasting soil types and temperature regimes.

Experimental design and management
The field experiment was conducted in two consecutive winter chickpea growing seasons (year 1 and year 2).The treatments, each with three replicates in a randomised complete block design, consisted of a factorial combination of four fertiliser levels: zero control [C], 90 kg P ha −1 [P], rhizobium inoculation [R], 90 kg P ha −1 plus rhizobium inoculation [R + P] and two desi chickpea genotypes (ACC#1 and ACC#5).Each experimental plot was 2 m × 1.5 m.Seeds for the inoculated plots were soaked in a mixture of Bradyrhizobium japonicum powder (containing 5 × 10 8 cells/gram) and water at the rate recommended by the manufacturer (Stimuplant, South Africa), and air-dried in the shade before planting.Uninoculated plots were planted before the inoculated ones to prevent contamination of the uninoculated seeds with the inoculum.Superphosphate fertiliser (10.5% P) was band applied according to the treatments, based on previous studies (Madzivhandila et al. 2012;Ogola et al. 2013).
The experimental plots were evenly irrigated immediately after sowing to promote uniform germination, emergence and crop establishment.Supplementary irrigation was provided during the cropping season when necessary.The experimental plots were kept weed-free by manual weeding.Cypermethrin (pyrethroid) was sprayed in all the plots weekly between flowering and early podding stage, at a rate of 10 ml per 20 L of water, to prevent pod borer (Helicoverpa armigera) damage.

Determination of soil chemical properties
Ten soil samples were randomly collected from 0 to 20 cm soil depths at each site before planting.Each sample was sieved (2.0 mm) to remove macrofauna, stones, and large roots, and air-dried, and samples from each location were pooled and mixed thoroughly into a composite sample.The composite samples from each site were used for the determination of the initial soil chemical properties such as pH, cation exchange capacity (CEC), soil organic carbon (SOC), total N, and extractable P, K, Mg, Ca, S, Zn, Mn and B.
Soil pH (1:2, soil: water) and SOC were determined using a glass electrode pH metre (Sparks 1996) and the Walkley and Black method (Nelson and Sommers 1982), respectively.Total N was determined using the Kjeldahl method (Bremmer and Mulvaney 1982), and available P was extracted using the Bray I method (Bray and Kurtz 1945).Potassium, magnesium and calcium were determined using the ammonium acetate extraction procedure (Du Plessis and Burger 1964), while Zn, Mn and B were determined using EDTA titration method (Trierweiler and Lindsay 1969), and soil S was determined as described by the Fertiliser Society of Southern Africa (FSSA 1974).

Determination of rhizosphere mineral nutrients
The concentration of macro (N, P, K, Mg and Ca) and micro (Zn) nutrients in the rhizospheric soil was determined at 50% flowering.Three stands of chickpea plants in the innermost rows in each plot were carefully dug out (the entire root system) from the soil, taking special care to ensure that the rhizospheric soil was intact.Rhizospheric soil was considered as the soil that was attached to chickpea roots (as well as soil up to 2 mm from the roots).The soil was air-dried at room temperature, sieved, and put into plastic zipper bags.Total N was determined using the Kjeldahl method (Bremmer and Mulvaney 1982) while the concentrations of P, K, Ca, and Mg were determined using the citric acid method (Du Plessis and Burger 1964).Zn was extracted using a di-ammonium ethylene-diaminetetraacetic acid (EDTA) solution (Trierweiler and Lindsay 1969).

Statistical analysis
The data were subjected to a 3-way analysis of variance (ANOVA) using STATISTICA version 10 (StatSoft Inc., 2010) to test the effects of fertiliser treatment, chickpea genotypes and location on the concentration of rhizosphere nutrients.Where significance was detected (p ≤ 0.05), treatment means were compared using Fisher's Least Significant Difference.The data from the two years were analyzed separately.The ANOVA results are presented in Table S1.

Soil properties
Soil at Thohoyandou had a low pH (6.06) and contained 24% sand, 16% silt and 60% clay while that from Syferkuil was slightly alkaline (7.52) and made up of 87, 2, and 11% sand, silt and clay, respectively (Table 1).Although soils from both sites had SOC below the critical level (Musinguzi et al. 2013), the SOC was much lower at Syferkuil compared to Thohoyandou (Table 1).In contrast, extractable P (more than 3-fold), K (about 2fold), Ca (about 2-fold) and Na were higher at Syferkuil compared to Thohoyandou (Table 1).The clay soil at Thohoyandou was characterised by slightly higher ammonium-N and nitrate-N content (23.5 & 3.4 mg kg −1 , respectively) and CEC (10.5%) compared to Syferkuil (19.8 mg kg −1 ammonium-N, 2.4 mg kg −1 nitrate-N, and 9.3% CEC) but the concentration of Mg and S was almost similar at the two sites (Table 1).Thohoyandou appeared to have higher soil micronutrients (Cu, Mn, Zn, and B) content compared to Syferkuil, except for Fe, (Table 1).

Accumulation of mineral nutrients in the rhizosphere
The genotype (G) × fertiliser (F) × location (L) interaction affected the rhizosphere concentrations of N, P, K, Ca, and Zn in both years (Figures 1 and 2).In contrast to the other nutrients, the accumulation of Mg in the rhizosphere was not subject to any interactive effects (Table S1) but the main effects of genotype (Figure 3), fertiliser treatments (Figure 4) and location (Figure 5) were significant.
Moreover, ACC#5 accumulated higher rhizospheric N than ACC#1 at all fertiliser levels (including the zero control) at both sites in each year, except at 90 kg P ha −1 at Thohoyandou in year 2 (Figures 1 and 2(a)).Except for R + P (year 1) and R / R + P × ACC#1 (year 2), the concentration of N in the rhizosphere was greater at Thohoyandou than Syferkuil at all fertiliser levels in both genotypes in both years (Figures 1 and 2(a)).Phosphorus Rhizobial inoculation (32-51%), P (50-193%) and R + P (117-258%) increased the concentration of P in the rhizosphere of both genotypes at Syferkuil and Thohoyandou in year 1, except for an unexpected 11% decrease due to inoculation in the rhizosphere of ACC#5 at Thohoyandou (Figure 1 Similarly in the second year, the increase in rhizosphere P due to inoculation was higher in ACC#1 at Syferkuil (127%) compared to Thohoyandou (50%) while in ACC#5 there was little difference in the response  between the two sites (Figure 2(b)).With R + P, the increase in rhizosphere P was consistently greater at Syferkuil compared to Thohoyandou irrespective of genotype in the first year (Figure 1(b)) but in the second year the increase was greater in ACC#1 at Syferkuil compared to Thohoyandou (182% vs 89%) while in ACC#5 greater increase was observed at Thohoyandou than at Syferkuil (56% vs 47%) (Figure 2(b)).

Potassium
Consistent with the effect of fertiliser treatments on the accumulation of N & P in the rhizosphere, R (26-101%), P (98-162%) and R + P (203-365%) increased the accumulation of K in the rhizosphere of ACC#1 and ACC#5 at both locations in year 1 (Figure 1(c)).A similar trend was observed in year 2, except for ACC#5 at Syferkuil and ACC#1 at Thohoyandou where the effect of rhizobial inoculation was non-significant (Figure 2(c)).However, the increase in rhizosphere K due to inoculation and P application was consistently greater at Syferkuil compared to Thohoyandou in contrast to R + P which elicited greater response at Thohoyandou than Syferkuil in both genotypes (Figure 1(c)).Moreover, with application of P and R + P, ACC#5 accumulated higher rhizosphere K compared to ACC#1 at both sites, but no difference was observed between the two genotypes without fertiliser application and with rhizobial inoculation (Figure 1 (c)).In contrast to the first year, no difference in rhizosphere K was observed between the two genotypes at all fertiliser levels at Syferkuil while at Thohoyandou, R and R + P elicited greater response in ACC#5 compared to ACC#1 in the second year (Figure 2(c)).

Calcium
Rhizobia inoculation increased rhizosphere Ca only in ACC#5 (Syferkuil in both years and Thohoyandou in year 2) (Figures 1 and 2(d)).Similarly, P and R + P increased rhizosphere Ca in ACC#5 (both sites) in year 1 (Figure 1(d)) and in both genotypes and sites in year 2 (Figure 2(d)).

Zinc
Rhizobium inoculation, P application and the combined rhizobia inoculation and P application increased rhizosphere Zn in ACC#1 and ACC#5 at both locations in the first and second years, except at Thohoyandou in the second year where rhizobial inoculation did not increase rhizosphere Zn accumulation in either genotype (Figures 1 and 2(e)).Moreover, in ACC#1 the increase in rhizosphere Zn due to rhizobial inoculation was greater at Syferkuil compared to Thohoyandou (54% vs 42%) while in ACC#5 the increase was greater at Thohoyandou (86%) than at Syferkuil (47%).In addition, P & R + P resulted in greater increase in  rhizosphere Zn in both genotypes at Syferkuil compared to Thohoyandou, except in ACC#1 in the second year where the increase was greater at Thohoyandou compared to Syferkuil (150% vs 128%).
Magnesium ACC#5 accumulated greater rhizosphere Mg compared to ACC#1 at both sites in year 2 (475 mg kg −1 vs 420 mg kg −1 at Syferkuil and 512 mg kg −1 vs 430 mg kg −1 at Thohoyandou, respectively).Moreover, rhizosphere Mg varied with locations with greater concentrations observed in Syferkuil (644 mg kg −1 in year 1 and 491 mg kg −1 in year 2) compared to Thohoyandou (489 mg kg −1 and 432 mg kg −1 in the first and second year, respectively)despite the initial concentrations at the two sites being similar (Table 1).The combined application of rhizobium and P fertiliser increased rhizosphere Mg by 19% (76 mg kg −1 ) at Syferkuil in year 2 while R & P did not have any significant effects (Figure 3(a)).At Thohoyandou, in contrast, R, P and R + P increased rhizosphere Mg in both years with greater increases being observed in R + P treatments (Figure 3  (b,c)).

Discussion
Fertiliser treatments (R, P & R + P) increased the accumulation of nutrients in the rhizosphere, but the response varied with genotypes and location/soil types (Figures 1 and 2), apart from Mg which was not subject to any interactive effects (Table S1).Despite the variable response to fertiliser treatments, ACC#5 consistently exhibited greater nutrient accumulation in the rhizosphere compared to ACC#1 irrespective of soil type (Figures 1 and 2) probably due to a more superior rooting system, in terms of biology and chemistry, (for example, fine roots with small diameter and greater capacity to solubilise nutrients through secretion of exudates) compared to ACC#1 (Eissenstat 1992;Dakora and Phillips 2002).The higher accumulation of rhizosphere Mg in ACC#5 compared to ACC#1 at both sites in the second year further demonstrates the superior performance of ACC#5 compared to ACC#1.The variable response in the rhizosphere nutrients to fertiliser treatments we observed in the two genotypes is in line with previous findings (Zhu et al. 2019).Consistent with our findings, Motseo (2019) reported genotypic differences in the concentration of nutrients that accumulated in the rhizosphere of desi-type chickpea grown in soil with different texture and macromicronutrients.
The fertiliser treatments (P and rhizobial inoculant, individually or in combination) generally enhanced the concentrations of the nutrient elements in the rhizosphere at both locations in the two cropping seasons (Figures 1 and 2) probably by (i) enhancing the secretion of citrate, tartrate and acetate thus increasing the mobilisation of nutrients, especially soil P, from Al-P and Fe-P (Shen et al. 2002), and facilitating the dissolution of soil-Zn and soil-Fe from minerals (Chairidchai and Ritchie 1990), as well as improving the activity of microorganisms in the soil (Nuruzzaman et al. 2005), (ii) decreasing the pH in the rhizosphere resulting in increased accumulation of nutrients (Bolan et al. 1997;Zhang et al. 2015), (iii) increasing the rhizosphere nutrients by synthesising siderophores that chelate oxide-bound and particlebound nutrients (Crowley 2006) as well as synthesising phytohormones such as auxins, cytokinins and gibberelins, lumichrome, riboflavin (Schutz et al. 2018), or (iv) a combination of two or more of the mechanisms.However, we did not investigate these mechanisms, and the effect of soil types and chickpea genotypes thereof, in the current study and this needs further investigation.Nonetheless, our study is the first to document the interactive effects of fertiliser treatment, genotype, and soil type (location) on the accumulation of rhizosphere nutrients in field-grown chickpea.Moreover, our study is probably the first to show that the application of Bradyrhizobium japonicum increases K in the rhizosphere.
Where the fertiliser effect was significant, rhizobium inoculation in combination with P application (R + P) resulted in greater accumulation of the nutrient elements in the rhizosphere compared to the other fertiliser levels irrespective of soil type and genotype (Figures 1-3) suggesting an additive action of rhizobium inoculation and fertiliser P (Wolde-meskel et al. 2018).The additive effect of rhizobium inoculation and phosphorus fertiliser on chickpea yield has been reported previously (Wolde-meskel et al. 2018) but to the best of our knowledge our study is the first to show the additive response of nutrient accumulation in chickpea rhizosphere to rhizobium inoculation and P fertiliser.However, we propose further studies, incorporating more chickpea cultivars and environments, to investigate the mechanisms underpinning the response of rhizosphere nutrients to P and rhizobium inoculation.
(b)).A similar trend was observed in the second year, albeit with non-significant difference between the zero control and R in ACC#1 at Syferkuil (Figure2(b)).However, the relative increase in rhizosphere P due to fertiliser application at the two sites varied with genotypes.In ACC#1, application of P resulted in greater increase in the accumulation of rhizosphere P at Thohoyandou compared to Syferkuil (193% vs 103%) while it was the other way round in ACC#5 (50% vs 99%) in the first year (Figure1(b)) but the response was reversed in year 2 (Figure2(b)).In contrast, the increase in rhizosphere P with rhizobial inoculation in ACC#1 was greater at Syferkuil (51%) compared to Thohoyandou (32%) while in ACC#5, R increased rhizosphere P at Syferkuil but decreased it at Thohoyandou (Figure1(b)).

Figure 2 .
Figure2.The interactive effect of genotype × fertiliser treatment × location in the concentration of rhizosphere (A) nitrogen, (B) phosphorus, (C) potassium, (D) calcium, (E) zinc in year 2. Data is mean values ± se (n = 12).Different letters indicate significant differences between genotype, fertiliser and location by Fisher's least significant difference (p < 0.05).R stands for rhizobium inoculation, P for 90 kg P ha −1 and P + R for rhizobium inoculation plus P fertiliser.

Figure 3 .
Figure3.The effect of fertiliser on the concentration of rhizosphere magnesium at (A) Syferkuil in year 2, (B) Thohoyandou in year 1, (C) Thohoyandou in year 2. Data is mean values ± se (n = 12).Different letters indicate significant differences by Fisher's least significant difference (p < 0.05, p < 0.05 and p < 0.01, respectively).R stands for rhizobium inoculation, P for 90 kg P ha −1 and P + R for rhizobium inoculation plus P fertiliser.

Figure 4 .
Figure 4.The effect of fertilizer on the concentration of rhizosphere magnesium at Syferkuil in 2018, year 2 (a) and Thohoyandou in 2017, year 1 (b) and 2018 (c) Data is mean values ± se (n = 12).Different letters indicate significant differences by Fisher's least significant difference (p < 0.05, p < 0.05 and p < 0.01, respectively).+Rh stands for rhizobium inoculation, P for 90 kg P ha −1 and Rh+P for rhizobium inoculation plus P fertiliser.

Figure 5 .
Figure 5.The effect of site on the concentration of rhizosphere magnesium in 2017, year 1 (a) and 2018, year 2 (b).Data is mean values ± se (n = 12).Different letters indicate significant differences by Fisher's least significant difference (p < 0.001 and p < 0.01, respectively).

Table 1 .
Initial Physical and chemical properties of the soil at Syferkuil (loamy sand) and Thohoyandou (clay) sites year 1.