The main objective of this study was to develop a controlled-environment assay to help study the effects of genotype, nodulation, pH, and calcium carbonates on IDC. Ideally, a single combination of specific levels of the previously mentioned stress factors could be used to maximize differences in IDC symptom differences between iron-efficient and inefficient genotypes. This would make it possible to increase the number of replicates and develop a simple “control” and “treatment” experimental design to assess IDC resistance among genotypes. The significant three-way interactions between genotype, nodulation, and CCE% are critical as they indicate potential differentiation between efficient and inefficient genotypes in IDC resistance using this assay. The two genotype comparisons in this study (Check Genotypes vs. NIL Genotypes) displayed their greatest pairwise differences when the CCE was adjusted to 20% (regardless of pH). However, optimal nodulation status and response variable behavior were not the same between genotype comparisons. For example, the greatest difference between Check Genotypes was found for RDM (Fig. 5a) and SDM (Fig. 5b) when the plants were unnodulated and the CCE adjusted to 10% or 20%. However, only visual score had a significant interaction with genotype in the NIL comparison. The greatest differences between NIL genotypes occurred in the visual score when nodulated at 20% CCE (Fig. 6). These findings indicate that optimal treatment conditions and the best response variable to measure for IDC severity can vary for different genotype comparisons when using this method. The optimal experimental parameters are not universal for all genotype comparisons. In our experiments, it is possible that the differential genetics of the Check Genotypes comparison (polygenic) and the NIL Genotypes comparison (single locus) may have influenced the optimal experimental parameters to elicit differences in IDC phenotypes, due to the genes and biological processes influencing the respective comparisons.
As expected, increased pH worsened IDC symptoms in the Check Genotypes for visual score, SPAD value, height, RDM, and SDM (Table 2) but was surprisingly not a significant source of variation in the NIL experiment (Table 3). Varying effects of pH between the Check Genotypes and the NIL Genotypes suggests the existence of multiple types of resistance to iron deficiency. While the check genotypes are affected by both pH and carbonate sources of iron stress, the NIL genotypes overcome iron stress due to pH but not increased carbonates. Iron reduction in the rhizosphere is the primary mechanism of iron acquisition in Strategy I plants (Marschner and Romheld, 1994). It is possible that the NIL genotypes are better able to reduce Fe3+ to Fe2+ in the rhizosphere than the check genotypes, and are thus less affected by higher soil pH per se.
It was expected that a significant interaction would be found between pH and CCE%, however no significance was found for this interaction in either experiment for any measured traits. In fields with calcareous soils, high soil pH often results in the release of bicarbonates (Hansen et al., 2003; Inskeep and Bloom, 1987). Bicarbonate release potentially has a greater effect on IDC severity than pH per se, but in field conditions this is nearly impossible to decouple. Further dissection of the relationship between carbonates and pH using this assay could help to improve our understanding of IDC in soybean.
Iron deficiency symptoms for all traits occurred with the treatment of calcium carbonates in both the Check Genotypes Experiment (Table 2) and the NIL Experiment (Table 3). While symptoms showed increased severity for the check genotypes with increases in CCE%, no significant differences were found in any traits in the NIL Experiment among CCE levels higher than 10%. It is possible that the NIL genotypes were better able to maintain iron homeostasis with increasing carbonate levels. Alternatively, they may be immediately overwhelmed with the addition of 10% CCE and maintain their physiological limit at higher levels of CCE. This could be discerned by further increasing the percent CCE to see if symptoms become more severe or are still maintained, although in field conditions CCE rarely exceeds 30%.
As an often-overlooked factor in IDC studies, a discussion on symbiosis is warranted here. Inoculation with rhizobia significantly improved all measured traits in the Check Genotypes Experiment (Table 2) and NIL Genotypes Experiment (Table 3). This is unsurprising when considered alone, as nodulation and nitrogen fixation by soybean has been well documented to improve yield more than nitrogen fertilizer (Beard and Hoover, 1971; Sorensen and Penas, 1978; Sogut, 2006).
Nodulation was the only factor that was a component in all significant interactions in this study (Table 1). Of particular interest and importance for IDC are the significant two-way interaction of nodulation with CCE and the significant three-way interaction between nodulation, genotype and CCE. For all significant interactions of nodulation with CCE, the addition of calcium carbonates exacerbated IDC symptoms. However, unnodulated plants had more severe IDC symptoms than nodulated plants as CCE% increased, similar to findings by Soerensen et al. (1988). It was assumed that nitrogen deficiency was not the cause of this observation because differences between efficient and inefficient genotypes did not occur until CCE% was increased. With advances in molecular techniques, a better dissection of the relationship between nodulation and IDC stress may be timely.
Investigating genotypes with small differences in resistance to IDC in a controlled environment is novel. Previous controlled environment IDC studies have focused on genotypes with sizeable differences in IDC resistance (Coulombe et al., 1984; Dragonuk et al., 1989; Lin et al., 1998; O’Rourke et al., 2007; Peiffer et al., 2012). While the NILs used in this study did not have a significant interaction with genotype and pH or CCE% in any of the individual symptoms of IDC, the interaction of genotype × inoculation × CCE level was significant with visual score. While the efficient and inefficient NILs had similar visual scores at many treatment levels, the differentiation found when nodulated and at 20% CCE and when unnodulated and at 10% CCE followed the expected pattern of the inefficient genotype having a higher visual score indicating less resistance to IDC. The one-point difference in visual score found when nodulated and at 20% CCE in this study is close to the 1.5-point difference for the NIL pair found by Merry et al. (2019) in field conditions.
The significant difference in SDM between the NIL genotypes is intriguing. The NIL pair used in this study is highly isogenic (F11-derived) and developed to express differences in resistance to IDC at a single genetic locus, thus any physiological differences between the NIL genotypes should be considered as a potential mechanism for IDC resistance. Because most studies focus on resultant phenotypes after IDC stress, investigating differences in pre-stress plant physiology to identify potential iron efficiency traits may add novel information to this area of study.
It is important to note that the method described here is practical for studying IDC response differences among a relatively small number of genotypes. However, it is unlikely that further refinements could be made to successfully score a large number of genotypes for IDC resistance, as is routinely done in a breeding program, using this method. There are several reasons for this conclusion. First, the number of replicates required to accurately rate IDC severity makes space requirements limiting as the number of genotypes increases. Second, the time required to apply nutrient treatments and measure traits on a large number of genotypes in controlled conditions is far more expensive than visually rating plants in an IDC field nursery or taking automated measurements with drone images from field data (Dobbels and Lorenz, 2019). Lastly, a “universal” optimal treatment for a large set of genotypes would not give accurate IDC resistance ratings using the approach described here. This is indicated by the different optimal combination of inoculation and CCE that were identified for the Check Genotypes and NIL Genotypes for inducing IDC symptoms. While these limitations make examining many genotypes impractical, this method can still be utilized for advancing our understanding of the genetics and physiology of IDC through comparisons of efficient and inefficient genotypes, as well as decoupling the effects of pH and CCE on IDC.