Introduction

It is well known that macronutrients are of crucial importance for increasing crop yields. However, micronutrients, which influence plant nutrient balance, are also essential for sustaining crop productivity and food nutrition (Khoshgoftarmanesh et al. 2010; Losak et al. 2011; Ciampitti and Vyn 2013). Micronutrient deficiency and toxicity can progressively occur when only macronutrients are adequately applied. This is because of the synergistic or antagonistic interactions that occur between macro- and micronutrients in soils and plants, between ions being able to form a chemical bond such as precipitation and complex, and between ions whose chemical properties are similar to compete for sites of absorption, transport and function on root surfaces or in plant tissues (Fageria 2001). The interactions of N with macronutrients were widely studied, with the general consensus that N application stimulates phosphorus (P) uptake by plants (Bailey et al. 1997; Ma and Zheng 2016). Nevertheless, the relationships between N and other mineral nutrients were not well understood, especially in modern cultivars, with inconsistent literature reports. For example, Losak et al. (2011) demonstrated a nil-effect of N fertilization on copper (Cu) uptake, while a potential synergism of plant N and Cu uptake in maize hybrids has been repeatedly reported (Bruns and Ebelhar 2006; Ciampitti and Vyn 2013). In canola, it was shown that increasing N supply increased seed zinc (Zn) and Cu concentrations (Gao and Ma 2015). Plants use different strategies to absorb mineral nutrients from the soil, maintain their homeostasis within the cell to avoid the metal-induced cell damage (Cuypers et al. 2013), and translocate to various tissues for utilization in different metabolic processes (Fageria 2009; Alloway 2013). For example, Erenoglu et al. (2011) illustrated that high N supply is more effective in enhancing Zn translocation from roots to shoots of wheat than in increasing Zn uptake in wheat. Elevated levels of Cu, Zn and Mn micronutrients along with N uptake in maize was likely associated with the increase of enzymes containing these elements (Bruns and Ebelhar 2006). The discrepancy observed in different studies is likely associated with the variation in the uptake and internal use efficiency among different genotypes used and with the environments where the experiments were conducted.

It is also generally acknowledged that cropping practices affect soil mineral status by influencing soil physical, chemical and biological properties (Riedell et al. 2009; Khoshgoftarmanesh et al. 2010). Maize and other cereal crops gain the benefit from macronutrient bioavailability in soil, especially from the direct contribution of N when legumes are included in the rotation or from the enhanced seasonal N mineralization (Ma et al. 2003; Ma and Wu 2016). Crops grown in rotation may also profit by other available mineral nutrients. For example, Wei et al. (2006) found that bioavailable Zn and Cu in soil varied with a decreasing order: continuous clover > cereal-legume rotation > continuous maize after 18 years of cropping. Riedell et al. (2009) observed significant N rate and rotation interactions on macronutrient (N, P, K, Ca and Mg) and Zn concentrations, but not on iron (Fe) and manganese (Mn) in maize plant shoots, at the 12th fully expanded leaf (V12) growth stage. Xia et al. (2013) observed higher acquisitions of Fe, Mn, Cu, and Zn in shoot but lower in grain for maize intercropping with legume, compared with continuous maize cultures. The researchers considered that in the intercropping system, the lower micronutrient elements in maize grain were attributed to the reduced translocation from the vegetative tissues. In spite of this, nutrient content in plants are considered to be genetically controlled by crop genotypes (Rochester 2006; Ciampitti and Vyn 2013). For example, Ma and Dwyer (1998) illustrated that the higher yield of modern hybrid maize under both deficient and adequate N supply conditions is attributed to its genetically greater ability to take up soil minerals later during the grain-filling period, compared to the old hybrid with the same agronomic treatments.

In transgenic Bt crops [crops with the expression of an insecticidal crystalline σ-endotoxin (Cry1Ac) to control European corn borer (ECB)], researchers have speculated that Bt hybrids acquire more N from soils to synthesize Cry1Ac protein than non-Bt counterparts (Bruns and Abel 2003; Rochester 2006). The interaction of N and micronutrients in Bt crops were rarely reported, though the critical roles of micronutrients as functional components of proteins and activators of enzymes are well defined (Subedi and Ma 2009; Erenoglu et al. 2011; Chatzistathis 2014). A recent review on the effects of glyphosate-resistant (RR) traits and glyphosate herbicides on plant nutrition summarized that, although several adverse effects of glyphosate on micro-nutrition (i.e., Mn, Fe, Zn) in RR-soybean were reported, most studies indicated mineral nutrition in RR soybean crops is not affected by either the RR trait or by the application of glyphosate herbicides (Duke et al. 2012). Glyphosate, whether applied as a herbicide directly to the field, exuded from plant roots or released from decomposing plant tissues, is strongly absorbed on variable-charge surfaces, with low desorbability. This leaves little glyphosate in soil solution to be available for microbial degradation, interaction with trace metal cations, plant uptake, or offsite transport (Duke et al. 2012). There was no literature on macro- and micronutrient balance and interaction in maize with stacked transgenic traits (RR + Bt) under contrasting cropping practices (Ma et al. 2009; Subedi and Ma 2007; Yanni et al. 2010). In a previous report, we demonstrated increased P content of the transgenic maize in aboveground shoot and grain with increasing rates of N supply (Ma et al. 2016). Nitrogen nutritional status has been shown to influence the uptake of Zn and other minerals in wheat (Erenoglu et al. 2011), maize (Bruns and Ebelhar 2006), and canola (Gao and Ma 2015). Likewise, we hypothesized that the uptake of Zn, Mg, Cu and Fe in transgenic maize was also affected by N supply conditions and varied among cropping systems. The objective of this study was therefore to examine maize plant Zn, Mg, Cu, and Fe nutrition and its interaction with N nutrition in RR- and stacked RR + Bt-maize hybrids in response to various rotation systems and fertilizer N application rates.

Materials and methods

Experimental design

This study was imposed on a long-term rotation experiment, which was initiated on a Brandon loam soil (fine loamy, mixed, mesic Typic Endoaquoll) at the Central Experimental Farm of Agriculture and Agri-Food Canada in Ottawa, Ontario, Canada (45°22′N, 75°43′W), in 1992. The general field and weather data was reported previously (Ma et al. 2003, 2012, 2016). Here, we will provide a brief outline with information pertinent to the current study. The initial Ap horizon of the soil measured in 1991 had 21.4 g kg−1 organic carbon and 2.1 g kg−1 total N, and 15.6, 2.0, 0.5 and 0.3 meq/100 g of exchangeable Ca, Mg, K and Al, respectively, with a total CEC of 18.2 meq/100 g and a soil pH of 6.5. From 2008 to 2010, the initial rotation-by-N treatment plot (16 m long × 9.14 m wide) was split in half to host two maize hybrids. The experiment during this period was considered a randomized complete block design in a split-plot arrangement, with rotation-by-N treatment as the main plot and maize hybrid as the subplot (Ma et al. 2016). The three rotation treatments were composed of maize in biennual rotation with alfalfa (MA), soybean (MS) and continuous maize culture (MM). The experiment also contained alfalfa following maize (AM) and soybean following maize (SM) plots to form a phased rotation study. Each year, the maize crop received preplant N application of 0 (N0), 50 (N50), 100 (N100) and 150 (N150) kg N ha−1 respectively in all cropping systems, plus two additional rates of 200 (N200) and 250 (N250) kg N ha−1 in MM. No N fertilizer was applied on plots with alfalfa and soybean. But all plots, including the alfalfa and soybean plots, were fertilized with adequate P and K fertilizers according to soil test recommendations prior to maize planting. No other mineral nutrients were applied in any of the years, presuming that there were ample available forms of Cu, Zn, Mg, and Fe present in the soil, as no visual deficiencies were observed throughout the experimental period. The two glyphosate resistant (RR) maize hybrids, Bt-maize (cv. Pioneer 38N87) versus non-Bt near-isoline (cv. Pioneer 38N85), were planted at a density of 75 000 plants ha−1 in the 3rd to 4th week of May each year.

Data collection

Maize plant samples were taken two times during the growing season. The shoot samples (five plants/plot) were taken at the sixth fully expanded leaf (V6) growth stage (MM plots only in 2009, and all plots in 2010). At maturity, maize plant samples were taken and separated into stover and grain. All the samples were dried at 80 °C for 3–4 days to obtain constant weights. After the dry weights were recorded, all samples were coarsely ground, and then ground subsamples were ground again to pass a 1-mm sieve. The ground samples were digested by the Kjeldahl method and analyzed for N and P concentrations by a continuous flow injection auto-analyzer (QuikChem® 8000 Flow Injection Analyzer, Zellweger Analytics, Inc., Lachat Instruments, Milwaukee, WI, USA). Each subsample (0.30 g), after re-drying overnight, was put in a Teflon volumetric container with a mixture of 65% HNO3 and H2O2 (v:v = 7:3). The sample was digested at 170 °C for 40 min, using an ETHOS EZ Microwave Digestion System (Milestone Inc., Shelton, CO, USA) (Huang et al. 2004; Gao and Ma 2015). The concentrations of Zn, Mg, Fe, and Cu in the digested solution were measured using a flame atomic absorption spectroscopy (Varian SpectraAA-220, Varian Canada Inc., Mississauga, ON).

The grain yield and shoot dry matter (DM), plant N, P uptake and their use efficiencies have been reported in our previous paper (Ma et al. 2016). The N data was retrieved here for the purpose of establishing the relationships of N uptake with other plant nutrient concentrations and content only. Mineral content was calculated by multiplying tissue DM by its mineral concentration. The nutrient harvest index (HI) of Cu, Zn, Mg and Fe, abbreviated as CuHI, ZnHI, MgHI and FeHI, was calculated as follows:

$${\text{Nutrient}}\,{\text{HI}} = \frac{{{\text{Nutrient}}\,{\text{content}}\,{\text{in}}\,{\text{maize}}\,{\text{grain}}}}{{{\text{Total}}\,{\text{nutrient}}\,{\text{content}}\,{\text{in}}\,{\text{maize}}\,{\text{shoots}}}}$$

Nutrient internal efficiency (IE) of Cu, Zn, Mg and Fe, abbreviated as CuIE, ZnIE, MgIE and FeIE, was then determined by dividing grain yield (kg ha−1) by the specific nutrient content in maize shoots, according to Ciampitti et al. (2013):

$${\text{IE}} = \frac{{{\text{Maize}}\,{\text{grain}}\,{\text{yield}}}}{{{\text{Total}}\,{\text{nutrient}}\,{\text{content}}\,{\text{in}}\,{\text{maize}}\,{\text{shoots}}}}$$

Statistical analysis

Three blocks each year over 3 years made up nine combinations of block by year (BlockYr). As this was an unbalance design in N levels for rotational maize and continuous maize, analysis of variance (ANOVA) was performed in two steps, with the MIXED procedure of SAS (SAS Institute Inc. 2010). First, the ANOVA was completed for the balanced maize rotation (maize cropping system-by-N treatment formed a factorial experiment), including MA, MS and MM at the first 4 N levels, with rotation, N rate and maize hybrid as the fixed effects and BlockYr and BlockYr*N*Hybrid in rotation as random effects. A separate ANOVA was then performed for the continuous maize MM, where two maize hybrids by 6 levels of N application were considered in a split-plot design, with hybrid and N rate as the fixed effects, BlockYr and BlockYr*N*Hybrid as random effects. The t test letter grouping was performed at the 95% confidence level, when significant treatment effects were shown from the ANOVA. The quantitative (linear and quadratic) relationships between N rates and the nutrient parameters were first tested by the ESTIMATE statement in the MIXED procedure. If it was significant, the linear and non-linear relationships between variables were then established by running the NLIN procedure. The contribution diagnosis by multiple regression analysis was performed using the SAS REG procedure, with a stepwise model at the 5% level of significance for variable entry and stay. The contributions of plant N concentration and content at the V6 stage were elaborated in accounting for the variations of the other mineral nutrient concentrations and content at maturity. All statistical analyses were performed at the 5% level of significance.

Results and discussion

Mineral nutrient concentrations in stover and grain

The ANOVA results showed that maize stover concentrations of Zn, Mg, Cu, and Fe were significantly affected by the interactions of rotation systems and fertilizer N rates (Table 1). Stover Cu concentration increased linearly with increasing N rates (Fig. 1a). Maize Cu concentration in the rotation system MA was significantly higher than in MS and MM. Stover Zn and Mg concentrations in MA and MS were not significantly different (Fig. 1b, c), and slightly decreased (for Zn) or increased (for Mg) with N rates. However, their concentrations in stover were quadratically responsive to N application rates. For example, concentrations were higher in MM than in maize grown in crop rotation when minor N (≤ 50 kg N ha−1) was applied, followed by a sharp decrease with increasing N rates to 150 kg N ha−1, thereafter revoked the decline at heavy N rates up to 250 kg N ha−1 (Fig. 1b, c). Stover Fe concentration was negatively related to N rates and significantly different among rotation systems, with the increasing order of: MA < MS < MM (Fig. 1d). Nevertheless, this difference tended to fade out at high N rates.

Table 1 The ANOVA p values of the fixed effects of rotation, N application rate and maize hybrid on mineral concentrations in stover and grain, total content in shoot (stover + grain) prior to harvest, harvest index (HI) and nutrient internal efficiency (IE), for the balanced rotation maize and for the unbalanced continuous maize (only P ≤ 0.05 are presented)
Full size table
Fig. 1
figure 1

Relationships of maize stover Cu, Zn, Mg and Fe concentrations at maturity as affected by rotation systems and N application rates on a Brandon loam from 2008 to 2010. Bars on markers are standard errors over nine combinations of block and year (n = 9). MA maize-alfalfa, MS maize-soybean rotation, MM continuous maize culture; the single, double and triple asterisks, are significant at P < 0.05, 0.01 and 0.001, respectively

Full size image

The response of stover Cu concentration to N rate was apparently similar to the change in stover N concentration, which was reported earlier (Ma et al. 2016). This similarity suggests that N and Cu nutrition in the stover retained a relative balance until maturity. This synergistic response could be the direct result of greater soil available Cu promoted by greater fertilizer N application (Wei et al. 2006). Soil pH and redox potential are known to be the most important parameters governing many processes which determine the fate of soil metals (Kabata-Pendias 2001). High N supply may have made Cu and Zn more available to the crop as N fertilization leads to the acidification of the soil (Divito et al. 2011). A small change in soil pH can significantly affect solubility and crop uptake of Zn, Cu, Mn, and Fe (Khoshgoftarmanesh et al. 2010). The enhanced root growth and greater release of organic acids from root exudates of highly N-fertilized maize may have indirectly promoted the formation of Cu2+, which could become available for uptake by plant roots (Fageria 2001; Riedell 2010). Soil acidification may have also increased Zn bioavailability, as illustrated in wheat (Cakmak 2008) and maize (Divito et al. 2011). In this study, the reason for the suppressive effects of N on stover Zn, Mg and Fe concentrations (except Mg in MA and MS), especially for MM, was less clear. This may be related to the increased remobilization and translocation of Zn, Mg and Fe from vegetative tissues to the grain, as evidenced by the increased HI values, with increasing N rates (Fig. 3b, d). We speculate that the increased grain yields under high N supply must be supported by higher rates of translocation of several mineral elements from shoots to the developing kernels, resulting in mineral concentrations being lower in the stover but higher in the grain (Riedell et al. 2009). It may also be attributed to the dilution effects of the increased DM production by N-application at a rate faster than the mineral uptake (Jarrell and Beverly 1981; Riedell 2010). In wheat, it was shown that 65Zn uptake by roots increased by threefold and its translocation from roots to shoots by up to eightfold, but plant growth was much less affected by increasing N supply (Erenoglu et al. 2011). In our study, the extremely lower DM and grain yields in MM compared to the other rotations indicated that a continuous maize culture over the 15 years since 1992 with nil (N0) or minor N (N50) application had substantially depleted the soil N. This led to the deterioration of soil quality with reduced soil organic matter, increased soil bulk density and poor soil structure (Gillespie et al. 2014), or the imbalance in plant nutrients (Ma et al. 2016). This might serve as an additional explanation for the greatly elevated Zn, Mg and Fe concentrations (especially for the extremely high stover Fe concentration in MM) at the same low N treatments (N0 and N50), followed by a sharp decline in stover Zn, Mg and Fe concentrations with increasing N rates, in MM, compared to those in MA and MS (Fig. 1b–d). Some earlier studies (Jarrell and Beverly 1981; Fageria 2001) have illustrated the synergistic and dilution effects of N fertilization on plant micronutrient concentrations. Ciampitti and Vyn (2013) also observed elevated shoot Mg and Cu concentrations, but no change in Zn and Fe concentrations as N rates increased from 0 to 224 kg N ha−1. During the growing season (at the V12 growth stage), Riedell et al. (2009) recorded a synergism of shoot N and Zn concentrations whereas a dilution effect of N fertilizer rate on Mg concentration in continuous maize. Mineral nutrient interactions are a complex issue. We speculate that the soil conditions on which the experiments were conducted, the maize varieties used, and the growth stage at which the samples were taken, all have exerted significant impacts on the observed differences in plant nutrient uptake and redistribution within the plants.

Grain Cu and Fe concentrations at maturity were not affected by the rotation systems (Table 1), but were positively responsive to N rates in all cropping systems (Fig. 2a, d). In this study, the concurrent augmentation of Cu concentrations in stover and grain, along with increasing N, affirmed the synergism of plant N and Cu nutrition. Grain Fe concentration responded to N application in a manner in contrast to its stover concentration (Fig. 1d). However, this divergence would not neutralize the dilution effect of N application on stover Fe concentration (Fig. 1d), as Fe harvest index indicated a very low proportion of grain Fe to total Fe content (Fig. 3d). Grain Zn concentrations were negatively related to N rates in MS and MM but not in MA (Fig. 2b). Similar dilution effect of grain Zn concentration was noted by Feil et al. (2005), but the nil-effect of N rates was reported in other studies (Bruns and Ebelhar 2006; Riedell et al. 2009; Ciampitti and Vyn 2013). Increasing N supply resulted in a greater Zn translocation from roots to shoots and retranslocation from flag leaves to grains, but with minimal impact on grain yield in wheat (Erenoglu et al. 2011). Grain Mg concentration was not affected by N rates, regardless of rotation systems (Fig. 2c). The lack of N effect on grain Mg concentration was reported by Feil et al. (2005) and Bruns and Ebelhar (2006), but increased Mg concentration by N treatment was noted by Ciampitti and Vyn (2013). Even though there were somewhat conflicting findings, N fertilizer supply had smaller effects on mineral nutrient concentrations in grain rather than in shoots (Feil et al. 2005; Bruns and Ebelhar 2006). As a natural buffer, soil controls the transport of chemical elements and substances to the atmosphere, hydrosphere, and biota (Kabata-Pendias 2001). In most cases, soils, especially those of fine-textured loamy or clayey loam textures, contain large amounts of total Fe, Zn, Cu and Mg, and the complex interactions of soil physical, chemical and biological conditions and agronomic practices influence the deficiencies or toxicities of soil mineral nutrients (Khoshgoftarmanesh et al. 2010; Alloway 2013). There appeared to be sizable plasticity and large variations in nutrient content and functional relationships as well as a large genetic potential for increasing grain mineral nutrient concentrations of wheat and other cereal crops (Jaradat 2017). Accordingly, balanced plant nutrition is the best way to avoid problems resulting from interactions with other nutrients (Blaylock 2008). Therefore, additional research is necessary to investigate the appropriate ratio or balance among different nutrient elements as affected by genotypes and agronomic practices.

Fig. 2
figure 2

Relationships of maize grain Cu, Zn, Mg and Fe concentrations at maturity as affected by rotation systems and N application rates on a Brandon loam from 2008 to 2010. Bars on markers are standard errors over nine combinations of block and year (n = 9). MA maize-alfalfa, MS maize-soybean rotation, MM continuous maize culture; the single, double and triple asterisks, are significant at P < 0.05, 0.01 and 0.001, respectively

Full size image
Fig. 3
figure 3

Relationships of maize Cu, Zn, Mg and Fe content in shoot (stover + grain) at maturity harvest and nutrient harvest index as affected by rotation systems and N application rates on a Brandon loam from 2008 to 2010. Bars on columns are standard errors over nine combinations of block and year (n = 9), values on the bars are nutrient harvest index (ratio of the content in grain to that in shoot); MA maize-alfalfa, MS maize-soybean rotation, MM continuous maize culture

Full size image

In this study, significant maize hybrid effects on grain Zn and Mg concentrations were observed (Table 1), with higher grain Zn and Mg concentrations in the non-Bt hybrid than the Bt hybrid. Specifically, averaged over the 3 years, grain Zn concentrations in the non-Bt hybrid were 18.4 mg kg−1 for maize in rotation and 19.1 mg kg−1 for the MM, respectively compared to 17.4 and 18.1 mg kg−1, in the Bt maize. The grain Mg concentration in the non-Bt hybrid was 1.03 ± 0.003 versus 0.99 ± 0.003 g kg−1 in the Bt-hybrid for the balanced rotation maize. Zn and Mg are important in plant metabolism, particularly for the activation of enzyme systems closely involved in plant physiological processes and protein metabolism (Chatzistathis 2014). Significantly greater N content was reported earlier in grain and shoot in Bt-maize than in non-Bt maize under continuous culture MM (Ma et al. 2016). The higher grain Zn concentration appeared to match the higher N content in Bt-maize, likely due to a greater rate of remobilization (than of uptake) of Zn from the vegetative tissues (stover) to the grain (Erenoglu et al. 2011). Therefore, lower Zn concentration in stover was measured at maturity (Fig. 1b). However, we cannot explain why Bt maize accumulated less Mg (lower Mg concentrations) than its non-Bt near-isoline. Perhaps this was associated with the complex interactions between the mineral nutrition and DM production/yield formation of these hybrids (Ma et al. 2016). No difference in Cu concentration was observed between the two maize hybrids.

Mineral nutrient content and harvest index

Maize total mineral nutrient content in shoot (stover + grain) at maturity were significantly affected by rotation systems and/or N application rates (Table 1). Total Cu, Zn and Mg content were generally in the order of: MA > MS > MM, with average values of 48.1, 36.9 and 24.2 g Cu ha−1, 295, 223 and 210 g Zn ha−1, and 15.5, 12.9 and 8.6 kg Mg ha−1, respectively for MA, MS and MM (Fig. 3a–c). This was mainly attributed to the much larger DM production in MA and MS than in MM (Ma et al. 2016), as their concentrations in stover and in grain did not follow the consistent order (Figs. 1, 2). Crop rotation influences mineral nutrient availability in soils by the residual effects of preceding crops and by root exudates on soil physical and chemical properties (Riedell et al. 2009; Khoshgoftarmanesh et al. 2010). Bruns and Ebelhar (2006) speculated that the higher Zn, Cu, and Mn concentrations in maize ear leaves due to increasing N supply were likely attributed to the elevated levels of these mineral containing enzymes. Riedell et al. (2009) observed a non-significant effect of N inputs on shoot Ca, Zn, and Mg concentrations for maize in crop rotations but higher P and Zn concentrations in the MM system with no N inputs. They speculated that root length may have been reduced in the no N fertilizer MM system resulting in lower shoot Ca and Mg concentrations due to less exploitable area by lateral roots. In contrast, an increased shoot Zn concentration may simply reflect the higher soil extractable Zn concentrations under the no N inputs MM systems. Plant species and growing conditions contribute to the divergent influence on trace element status in plants (Kabata-Pendias 2001). Our data demonstrated that maize grown in crop rotations, especially following alfalfa, not only had a higher uptake of N but was also conducive to plant Cu, Zn and Mg acquisition. This was in agreement with Xia et al. (2013) who found that compared with a continuous culture, maize following legumes had higher Cu and Zn content in shoots. In addition to a significant difference in DM between rotation systems, the closer association between soil available micronutrients and organic matter rather than with available N, P and CaCO3 (Wei et al. 2006) indicated a significant and direct influence of soil organic matter on bioavailability of soil mineral nutrient sources. In our study, after the 15-year continuous maize culture, the MM soils had the lowest organic matter (16.1 ± 0.5 g kg−1) (Chan et al. 2013), and the maize Cu, Zn and Mg content in MM were also the lowest. Nevertheless, total Fe content was conversely lower in MA than in MS and MM (Fig. 3d), evidently supported by its lowest Fe concentration in stover (Fig. 1d). This was inconsistent with Riedell et al. (2009) who used the canonical discriminant analysis to show a no-difference in plant Fe accumulation between continuous maize, 2-year maize-soybean rotation and 4-year maize-soybean-wheat-alfalfa rotations. This discrepancy was likely due to the possibility that the length of the crop rotation and the frequent cropping to alfalfa (every other year in our study) exerted a different impact on soil Fe bioavailability compared to that of the 4-year rotation reported by Riedell et al. (2009).

In this study, no severe European corn borer (ECB) infestation was observed in any year (2008–2010). We did not observe maize hybrid (Bt- vs. non-Bt maize) effect on plant total mineral nutrients (Table 1) or on DM production (Ma et al. 2016).

The mineral nutrient HI was affected by the interaction of rotation systems and N rates (Table 1). Averaged across rotation systems and N rates, HI values were 0.21, 0.46, 0.53 and 0.10 for Cu, Zn, Mg and Fe, respectively. This suggested a dramatic difference in remobilization and/or importance of Cu, Zn, Mg and Fe in grain production. Similar mean mineral HI values were recorded by Ciampitti and Vyn (2013) for Cu (0.17), Zn (0.44), Mg (0.53) and Fe (0.14), and earlier by Karlen et al. (1988) at a high yield level (16.3 Mg ha−1). Mineral nutrient harvest index is a measure of retranslocation or utilization efficiency of the absorbed mineral from plant vegetative parts to the harvestable grain. Large differences in mineral nutrient HI values are expected as each mineral nutrient has its specific function within the plant. For example, Cu plays important structural and functional roles in the physiological processes involving Cu-dependent enzymes that affect photosynthesis and lignification of cell walls. Similarly, the absorbed Fe that is incorporated into heme-proteins is involved in the formation of lignin and suberin within the plant cell (Alloway 2013). These structural minerals would have remained in the plant vegetative parts, causing the crop to exhibit small HI values. In addition, the maize crop took up more than 50% of its total Fe and Zn content after flowering, compared to the 73% of total Cu and Mg content accumulated at the vegetative stage (Ciampitti and Vyn 2013). In this study, the CuHI counteracted to N rates whereas ZnHI, MgHI and FeHI positively reacted to N increase (Fig. 3), following the same trends in Ciampitti and Vyn (2013). The notably larger increases in ZnHI, MgHI and FeHI in the continuous maize rather than in rotation maize (Fig. 3) were the consequence of the non-proportional increase in DM and grain production. This therefore caused the extremely lower HI values in the N0 and N50 plots, as compared to those of maize in crop rotation at the same N levels (Fig. 3). Therefore, mineral nutrient accumulation appeared to be strongly dependent on maize grain yield as well as the soil nutrient status (Ciampitti and Vyn 2013).

Mineral nutrient internal efficiency (IE)

Like nutrient use efficiency (NUE), nutrient internal efficiency (IE) is also an index of the nutrient being used for grain production. In this study, the nutrient IE was interactively affected by rotation systems and N rates (Table 1). Similar to IE of N, reported earlier by Ma et al. (2016), IE of Cu was negatively related to the increased N rates (Table 2), with the efficiency in MS consistently greater than that in MA (on average, 198 vs. 159 kg grain g−1 Cu). The similar manner of NIE and CuIE in response to N rates, once again, suggests the synergism of plant N and Cu nutrition and balance for DM and grain production in this study. Notwithstanding, ZnIE and FeIE were positively related to N rates (except for FeIE in MA) and were significantly higher in MA and MS than in MM (Table 2). The positive relationship between ZnIE and N rates was due to the cumulative effects of increasing grain yield while diluting grain Zn concentration with N rate increase, especially in MM (Fig. 2b). However, the distinguishing FeIE amongst N rates mainly reflected the difference in grain yields, as there was no response of plant Fe uptake to N rates in this study (Table 1 and Fig. 3d). In other words, the remarkably lower grain yield in MM than that in MA and MS resulted in significantly lower FeIE in continuous maize relative to maize in crop rotation.

Table 2 The least square means (LS-means) of nutrient internal efficiency (IE) of Cu, Zn, Mg and Fe (CuIE, ZnIE, MgIE and FeIE), as affected by rotation systems, N application rates and maize hybrids
Full size table

The maize MgIE in crop rotation was not responsive to N application but it could be described by a downward curvilinear relation with N rates in continuous maize (Tables 1, 2). This indicates that the grain increment induced by N rate increase was proportional to the increment of plant Mg content in rotational maize but less so in continuous maize. Overall, the internal efficiency was 185, 26.8, 0.53 and 4.96 kg g−1 for Cu, Zn, Mg and Fe, respectively, which were comparable with those reported by Ciampitti and Vyn (2013). In addition, in this study, CuIE was about ten- to hundred- fold that of Zn, Mg and Fe, echoing its similar lower fold plant uptake as compared to Zn, Mg and Fe (Fig. 3). The observed synergism in this study between N and Cu nutrients, but not between N and Zn, Mg or Fe was also an explanation of the much higher use efficiency of Cu than of Zn, Mg and Fe in grain production. We speculate that under high N supply conditions, especially in MM, translocation of photoassimilates from the vegetative tissues to the grain was somehow limited, resulting in HI values much lower at the zero or low N supply than at the high N supply. Enzymes, containing Fe and Zn as cofactors involved in the remobilization from the source to the sink, were also affected, leading to a smaller proportion of these minerals in the grain. As earlier reported for NIE and PIE (Ma et al. 2016), hybrids did not have a significant effect on mineral nutrient IEs in this study (Tables 1, 2). Clearly, the information about mineral nutrient IE and HI presented here could serve as useful references for estimating the crop mineral nutrient level needed per unit of yield production and the mineral nutrient removal in the grain under different management practices (such as rotation systems and N rates). This is potentially helpful in diagnosing mineral nutrient deficiencies and imbalances with macronutrients such as N.

Interrelationships of N and other mineral nutrition

The interrelationships of N and mineral nutrition were examined two ways in this study. First, we established the stoichiometry between N and other mineral nutrients at the V6 stage as well as at maturity. The multiple regression analysis was then performed to relate plant mineral nutrition at maturity to plant N nutrition at the V6 stage. This will determine the contribution of the early plant N which accounts for the variation of plant mineral nutrition at maturity. The linearly positive relations were demonstrated between plant N and Cu, and between plant N and Mg concentrations at the V6 stage (Fig. 4a). This stoichiometry was, however, not reserved at maturity (data not shown), due to the on-going concentrative or dilution effects of N application on plant Cu or Mg concentration from V6 to maturity. However, on the nutrient uptake basis, closer and linear relations were established between N and each of the analyzed mineral content at the V6 stage (Fig. 4b). These relations were preserved in the grain nutrient content while for the shoot, only Cu and Mg content showed a positive association with N at maturity (Fig. 5a, b). The disappearance of associations between N and Zn or Fe content at maturity probably implies that different relationships between N and Zn or Fe content existed among the three rotation systems. There existed divergent effects of N supply on stover Zn and Fe concentrations between continuous maize and rotational maize from the V6 stage to maturity (Fig. 1b, d). Likewise, the observed associations between N and each of the mineral content in grain were probably attributed to the relatively analogous tendencies of their concentrations in response to N input amongst rotation systems, even though significant difference between rotation systems existed (Table 1 and Fig. 2). Above inferences were based on the fact that DM accumulation in this study was positively and linearly responsive to fertilizer N increase, regardless of rotation systems (Ma et al. 2016).

Fig. 4
figure 4

Relationships of A maize Cu, Zn, Mg and Fe concentrations with plant N concentrations, and B maize Cu, Zn, Mg and Fe content with plant N content (on an area basis) at the V6 stage on a Brandon loam, based on data from continuous maize plots only in 2009, and all plots in 2010 (n = 118). The triple asterisks are significant at P < 0.001 level

Full size image
Fig. 5
figure 5

Relationships of maize Cu, Zn, Mg and Fe content with plant N content in a grain and b shoot (stover + grain) at maturity on a Brandon loam, based on data from overall plots from 2008 to 2010 (n = 252). The triple asterisks are significant at P < 0.001 level

Full size image

The close interrelations between mineral nutrient uptake at maturity and plant N nutrition at the V6 stage of maize (Table 3) indicated that the variations in final plant mineral nutrition can partially be interpreted by early plant N concentration (Ncon) and content (PlantN) at the V6 stage. This also suggests the potential of early N nutrition in maize as an in-season diagnostic indicator of mineral nutrient deficiency or imbalance. Even if the relative contribution of PlantN and Ncon to dependent variables varied from one to another, the variations in 11 out of the 16 dependent variables were explainable mainly by PlantN (Table 3). This was because PlantN integrated both plant N concentration and DM, and the direct measurement of plant nutrient absolute amount is, in general, better than the relative quota (concentration) to assess plant nutrient status, if we don’t take into account its practicability in the field (Jarrell and Beverly 1981; Bailey et al. 1997). Specifically, the variations of total content and the stover concentrations of Cu, Zn and Mg at maturity were mainly accounted by PlantN rather than by Ncon. For example, the partial R2 showed that plant N content at the V6 stage accounted for 48% of the variation in total Cu content, of which 46% was explained in the grain, and 38% was in the stover at maturity (Table 3). This corresponded to its solid relations with plant N content held from V6 to maturity (Figs. 4b, 5b). The variations in plant total Fe content at maturity were mainly interpreted by Ncon rather than PlantN, and is supported by the non-responsiveness of its content but responsiveness of stover Fe concentration to N rates (Figs. 1d, 3d). The variations of maize grain mineral concentrations were more intricate in this study. It was significantly related to Ncon for Cu and Mg but not to PlantN for Fe, while a null association was observed between grain Zn concentration and N nutrition at the V6 stage (Table 3). This was probably because grain nutrition was prone to be governed by factors other than nutrient sources (such as environment) that affect the nutrient remobilization and transfer. In spite of the above established associations by the ‘two-stage’ evaluation used in this study, more periodic sampling is warranted to identify the optimal sampling stage and critical values or sufficiency ranges of mineral nutrients, and the norms of the N: mineral ratio needed to develop the in-season mineral nutrient diagnosis in maize.

Table 3 Multiple regression analysis of mineral content in shoot (TotalCu, -Zn, -Mg and -Fe) and in grain (GrainCu, -Zn, -Mg and -Fe), and of mineral concentrations in stover (StoverCucon, -Zncon, -Mgcon and -Fecon) and in grain (GrainCucon, -Zncon, -Mgcon and -Fecon) at maturity as a function of plant N concentrations and total N content at the V6 stage
Full size table

Conclusions

This study revealed that rotation systems and N fertilizer rates interactively or separately affected maize plant Zn, Mg, Cu, and Fe concentrations and content at crop maturity. Plant Zn, Cu, and Mg content were generally greater in maize grown in crop rotation than in continuous maize, with positive responses to N application rates. In this study, the presented mineral concentrations and total content, coupled with the earlier report on grain and dry matter (Ma et al. 2016), demonstrated that fertilizer N application synergistically affects the plant Cu nutrition, whereas there was a dilution effect on plant Zn and Mg concentrations. The uptake of maize plant Cu, Zn and Mg nutrients at maturity was associated with plant N content at the V6 stage while plant Fe content was mainly explainable by early plant N concentration at the V6 stage. This indicates the possibility of in-season mineral nutrient diagnosis with early stage plant N nutrition. In this study, there was no maize genotype effect on mineral concentration and content, nutrient harvest index and internal use efficiency in grain production, except for Zn and Mg concentrations in grain.

Our data presented here, as well as the contrasting observations reported in the literature, has not yet completely explained the complex interactions between N inputs and crop rotations on plant Cu, Zn, Mg and Fe nutrients. Additional efforts are needed to understand the interactions between macro- and micro-nutrients under the umbrella of different growth stages and cropping systems. The integration of data on soil fertility, plant nutrition and crop ecophysiology is warranted to reasonably explain these interactions and to develop more efficient strategies to improve overall maize nutrient uptake and balance of essential macro- and micronutrients in contrasting environments and cropping systems.