Introduction

Roots fulfil important functions for plants, including anchoring in the soil and providing access to soil nutrients and water. Plant roots are known to be an important factor determining the agricultural performance of crop plants. However, because root traits are difficult to assess and select for, their potential for crop plant improvement has as yet not been fully exploited and numerous details of factors and genes controlling root system traits remain underexplored (Lynch and Brown 2012; White et al. 2013; Rogers and Benfey 2015; Hochholdinger 2016; Koevoets et al. 2016; Bray and Topp 2018).

Root traits are also of vital importance for maize (Zea mays L.), which is one of the most important cereal grains grown worldwide (Shiferaw et al. 2011). A recent study using a machine learning program for trait analysis of 57 commercial maize hybrids concluded that root traits were most important for predicting yield (Tucker et al. 2020). Several root traits have been identified in maize that are relevant for the exploration of soil resources, particularly in resource-poor environments (Lynch and Brown, 2012; York et al. 2013). For example, drought tolerance is associated with an increase in rooting depth and water acquisition from the subsoil (Gao and Lynch 2016). Plants with improved root traits may contribute to relieve a major constrain for the production of maize in develo** countries, which is low soil fertility and high water requirement (Rusinamhodzi et al. 2011; Reynolds et al. 2015; ten Berge et al. 2019).

It is thought that selection for yield has indirectly selected also for root traits contributing to enhance maize yield (Hammer et al. 2009; Bray and Topp 2018). However, the genes regulating these traits are mostly unknown although a large number of QTLs associated with root traits have been identified in maize (Hund et al. 2011; Bray and Topp, 2018). However, so far only eight maize genes were identified that are involved in regulating root growth and development but their individual relevance for plant performance is as yet not clear (Hochholdinger et al. 2018). To understand and fully exploit the potential of roots for crop improvement, it would be important to compare near isogenic lines that differ principally in their root system while shoot growth and development should not be altered. To achieve this, we explore here an experimental approach based on changing the endogenous content of the plant hormone cytokinin by genetic engineering.

The hormone cytokinin is a well-known inhibitor of root elongation and branching (Werner et al. 2001, 2003; Chang et al. 2013, 2015). Cytokinin is degraded by cytokinin oxidases/dehydrogenases (CKX), which are encoded by small gene families in plants including maize (Schmülling et al. 2003). Enhanced degradation of cytokinin by enhanced expression of a CKX gene in roots caused the formation of a larger root system in Arabidopsis thaliana (Werner et al. 2010), barley (Ramireddy et al. 2018a,b), oilseed rape (Nehnevajova et al. 2019), rice (Gao et al. 2014) and chickpea (Khandal et al. 2020). Thus, it was shown repeatedly that a single dominant gene may be used to regulate a complex trait such as root system size. CKX transgenic plants with a larger root system were shown to respond less sensitive than the cognate wild-type plants to drought (Werner et al. 2010; Ramireddy et al. 2018a) underpinning the beneficial effect of a larger root system under water deficit (Comas et al. 2013; Gao and Lynch 2016; Klein et al. 2020). A surprising common feature of these plants had been the higher content of distinct micro- and macro-elements in their shoots. In particular, the concentration on zinc (Zn), a microelement missing in the diet of about 2 billion people, was found to be significantly increased in the seeds of CKX transgenic barley plants grown in the greenhouse and in the field (Ramireddy et al. 2018a,b). Consequently, it has been proposed that root enhancement might contribute to a sustainable solution for nutrient deficiencies (Werner et al. 2010; Ramireddy et al. 2018a; Gao et al. 2014). Supplemental Table S3 lists the primer sequences used in this study. The cDNA samples were used to determine CKX1 transgene expression and cytokinin primary response genes ZmRR1 (gene ID Zm00001d001865) and ZmRR2 (gene ID Zm00001d026594) levels by quantitative real-time PCR according to Cortleven et al. (2014).

Quantification of root system size and biomass

Maize seeds were germinated on soil and three days after germination seedlings were carefully lifted from the soil and cautiously washed to remove bound soil particles. Seedlings of similar size were transferred to a hydroponic system and cultivated for another seven days for root system size analysis, and 15 d for biomass quantification. For the hydroponic system 0.1 × Hoagland solution (1 mM KH2PO4, 0.5 mM KNO3, 0.4 mM Ca(NO3), 0.2 mM MgSO4, 0.1 mM FeNaEDTA, 0.01 mM H3BO3, 2 μM MnSO4, 0.2 μM ZnSO4, 0.2 μM CuSO4, 0.1 μM Na2MoO4 and 0.02 mM NaCI) was used (Krämer et al. 1996). 12 L nutrient solution per box was properly aerated and changed every second day. After harvest, roots and shoots were separated and their fresh weights were determined. Thereafter, samples were dried in an oven at 80 °C for 68 h and the dry weight was recorded. For root system size analysis by WinRHIZO™, roots were carefully lifted from the box and spread out in a root-positioning tray (20 × 30 cm) to minimize root overlap and scanned with a flatbed scanner (EPSON, EU-88, Japan). Greyscale images obtained in tiff format were analysed with WinRHIZO™ (Pro Version 2005a; Regent Instruments Inc., Canada). For quantification of root system size of soil-grown transgenic plants, soil-filled 30 cm diameter pots were used. After five weeks of growth in soil, plants were harvested and separated into shoots and roots. Roots were carefully washed to remove bound soil particles and aggregates. Samples were dried in an oven at 80 °C for 68 h and the dry weight was recorded.

Quantification of leaf and seed element content

Quantification and analysis of leaf and seed elements was performed as described in Ramireddy et al (2018a). Briefly, seeds of three independent transgenic lines and the NTC were germinated on filter paper in vitro. Three-day-old seedlings were transferred to the greenhouse into an unfertilized (type 0) soil supplied by the company Einheitserde (Sinntal-Altengronau, Germany). Composition of unfertilized soil was tested and certified by Institut Koldingen GmbH (Sarstedt, Germany) as described by Drechsler et al. (2015). Plants were grown further for four weeks by supplementing equal amounts of fertilizer solution every second or third day depending on soil moisture. The fertilizer solution was based on the composition of modified Hoagland solution (2 M KNO3, 1 M NH4NO3, 1 M KH2PO4, 2 M Ca(NO3)2 4 H2O,,2 M MgSO4 7 H2O, 100 μM Na-Fe-EDTA, 50 μM H3BO3, 50 μM MnSO4, 18.5 μM ZnSO4, 50 nM CuSO4, 50 nM CoCl2, 0.5 μM NaMoO4 and 2 mM MES). The solution was adjusted to pH 5.7 with 1 M KOH. Total leaves from four-week-old plants were dried for 72 h at 80 °C and grounded carefully. Then equal amounts of powder (1 g) were weighed into polytetrafluoroethylene tubes and digested with a HNO3 + H2O2 mixture in a pressurized microwave digestion system (MARS from CEM GmbH; Kamp-Lintfort, Germany). The concentrations of macro- and microelements were analyzed by inductively-coupled plasma optical emission spectrometry (ICP-OES, iCAP 6500 dual OES spectrometer; Thermo Fischer Scientific) with certified standard reference samples as control. The element content from seed samples was determined in a similar way as outlined above.