Soybean (Glycine max (L.) Merrill) is an important crop cultivated for its rich protein content (Wu et al., 2015). However, in China, soybean production faces a shortfall in domestic soybean production, relying heavily on imports (>80%). Southwestern China, a key soybean production area, has undertaken numerous efforts, including soybean breeding, to improve soybean yields (Yang et al., 2022; Zhang et al., 2022). Breeding programs aimed at increasing seed yield have significantly improved yields (Todeschini et al., 2019; de Felipe et al., 2020), primarily through increased biomass in China (Jin et al., 2010; Yang et al., 2022), United States (Cafaro La Menza et al., 2017) and South America (Todeschini et al., 2019). Increases in harvest index have contributed to soybean yield gains worldwide (He et al., 2016; Todeschini et al., 2019; Tamagno et al., 2020; Feng et al., 2022). Changes in yield components, such as seed number and seed size, have also been associated with increased soybean seed yields in China (Wang et al., 2016; Qin et al., 2017; Zhang et al., 2022), United States (Kumudini et al., 2001), and South America (de Felipe et al., 2016). Genetic gains in soybean seed yield range from 0.4 to 2.2% y⁻¹ worldwide and about 2.0% y⁻¹ in southwestern China (Yang et al., 2022). Improved lodging resistance has also helped boost yields during soybean breeding (Jin et al., 2010; Kumudini et al., 2001; Rogers et al., 2015; Milioli et al., 2022). However, the impact of soybean breeding on P and N utilization efficiencies in southwestern China remains unclear.

Nutrient uptake, especially nitrogen (N) and phosphorus (P) is vital for crop growth and productivity (Jat and Bijay-Singh, 2014; Salvagiotti et al., 2021; Meng et al., 2022). Both P and N are involved in leaf photosynthesis, essential for crop growth, and high accumulation of these nutrients is important for achieving high seed yields (Tamagno et al., 2017) and efficient nutrient utilization (Meng et al., 2022). High nutrient accumulation is needed to accumulate high biomass (Cafaro La Menza et al., 2017), associated with high seed number and/or seed size and, thus, seed yield (He et al., 2019; Meng et al., 2022). Thus high-yielding soybean cultivars have high N (Salvagiotti et al., 2008; Gaspar et al., 2017) and P accumulation (He et al., 2017b, 2019), requiring increased P and N uptake through root inputs and/or modifications to root structure characteristics, such as increased adventitious root density and shallow root angle (He et al., 2017b; Lynch, 2019). Recent studies have shown that N accumulation increased with seed yield improvement during soybean breeding in Argentina (de Felipe et al., 2020) and United States (Donahue et al., 2020). However, the relative contributions of increased biomass and nutrient concentration to nutrient accumulation during soybean breeding are poorly understood. Enhancing nutrient utilization efficiencies (seed yield/total nutrient accumulation) under low fertilizer inputs can enhance yields and potentially reduce N and P demands, promoting sustainable agriculture and increasing food security (An et al., 2018; Wu et al., 2019). Nutrient utilization efficiencies have been associated with harvest index, a key trait determining grain yield (Meng et al., 2022). Genetic variations in nutrient utilization efficiencies have been reported for various crops, including wheat (Ortiz-Monasterio et al., 1997; Sadras and Lawson, 2013), maize (Ciampitti and Vyn, 2012), barley (Muurinen et al., 2006; Bingham et al., 2012), and cotton (Rochester and Constable, 2015). However, the effects of breeding on nutrient use efficiencies vary among different species. For example, P and N utilization efficiencies significantly increased in cotton cultivars released in Australia from 1973 to 2006 (Rochester and Constable, 2015), while N utilization efficiency did not change with year of release in barley cultivars (Muurinen et al., 2006). These inconsistent results suggest that changes in nutrient utilization efficiency during cultivar improvement may be species-specific.

Understanding changes in nutrient partitioning among plant organs could help reduce N and P demands by directing limited nutrients to essential organs, such as reproductive organs. For example, increased seed biomass accumulation changes the N partitioning between other plant organs (Sinclair, 1998). Nutrient partitioning is associated with biomass partitioning (Donald and Hamblin, 1976; Tamagno et al., 2017). While soybean breeding has increased plant biomass (Yang et al., 2022), it remains unclear if dry weights and nutrient accumulation have improved in all organs, how P and N partitioning among organs has changed, and how P and N utilization efficiencies have been affected.

This study investigated changes in the agronomic efficiency of P fertilizer (AEp), P and N accumulation and utilization efficiencies, and their partitioning to different organs in a historic set of 12 soybean cultivars bred for high seed yield. The study was condcuted under two P rates [0 (P0) and 35 (P35) kg ha⁻¹ P] at two field sites [Dafang (DF) and Shiqian (SQ)] during the 2019 growing season. The hypotheses tested were: (1) soybean breeding has increased P and N utilization efficiencies with seed yield; (2) enhanced P and N utilization efficiencies are associated with high P and N partitioning to pods.

This study evaluated a historic set of 12 soybean cultivars (released from 1995 to 2016, Supplementary Table S1) grown by local farmers (past and present) at two field sites [Shiqian (SQ) and Dafang (DF)] in Guizhou Province, China, in 2019. The cultivars were collected from three provinces in southwest China (Sichuan, Yunnan, and Guizhou), where soybean breeding focused on increasing seed yield. All cultivars can grow in Guizhou province, with maturity times ranging from 115 to 121 days after sowing (DAS) for SQ and 124 to 130 DAS for DF. The soil pH, total P, and plant available soil P were 6.8, 0.77 g kg⁻¹ and 33 mg kg⁻¹ for SQ, and 6.7, 0.92 g kg⁻¹ and 33 mg kg⁻¹ for DF, respectively. The mean temperature and precipitation were 849 mm and 20.2°C for SQ and 573 mm and 23.6°C for DF, respectively (Supplementary Figure S1). The split-plot design had two P levels [zero P (P0) and 35 kg ha⁻¹ (P35) applied as calcium superphosphate] as the main plots, with cultivars as the sub-plots. Each cultivar in each main plot had three replicates, for a total of 72 plots at each site. Each plot was 12.8 m² (3.2 m wide × 4 m long), with rows spaced 0.4 m apart. The straight line between the two main plots was 2 m. Two days before sowing, N and K fertilizers were applied to all plots as urea (75 kg N ha⁻¹) and K₂SO₄ (40 kg K ha⁻¹) according to our previous study (Zhang et al., 2022), with the same N and K rates used at both experimental sites. The fertilizers were broadcast and mixed into the soil using a rotary cultivator. The seeds were sown (April 2019) at about 5 cm depth with a 40 cm row spacing. After germination, each plot was thinned to 18 seedlings per m². No irrigation was applied. Weeds were removed by hand, with pesticides used as needed. The upper 20 cm of soil was collected to analyze the basic nutrient status before applying the fertilizer. Weather data were collected from weather stations near the field sites (straight-line distance ranged from 0.5–19.2 km).

The newer soybean cultivars exhibited higher seed yields and greater yield responses to P application than older cultivars. The higher seed yields of newer cultivars are the result of soybean breeding efforts worldwide (Jin et al., 2010; Todeschini et al., 2019; de Felipe et al., 2020; Yang et al., 2020, 2022). The greater yield response to P application in the newer cultivars can be attributed to the significant yield improvements compared to older cultivars under P supply, leading to improved agronomic efficiency of P fertilizer (AEp). In maize, increased yield with N fertilizer supply was associated with increased grain numbers (Liu et al., 2022), indicating the important role of yield components in the yield response to fertilizer supply, such as seed number and seed size (Kumudini et al., 2001; de Felipe et al., 2016; Wang et al., 2016; Qin et al., 2017; Zhang et al., 2022). Understanding the underlying mechanisms responsible for soybean’s high response to P supply, particularly related to yield components, would be valuable for future research.

Soybean breeding simultaneously increased yield and P and N accumulation, consistent with similar studies in Argentina (de Felipe et al., 2020) and United States (Donahue et al., 2020). High N accumulation positively correlated with leaf and pod biomass (Figure 7), indicating that increased soil N uptake sustains leaf and seed development (He et al., 2019; Jin et al., 2022). Similarly, high P accumulation may be associated with root traits associated with P acquisition (He et al., 2017b, 2021; Lynch, 2019). While soybean breeding did not change root dry weights, which did not correlate with P or N accumulation (Figure 7), other root traits, such as shallow root growth angle, could improve P uptake (Lynch, 2019) and contribute to P and N accumulation.

In this study, biomass accumulation increased during soybean breeding, consistent with other studies (Jin et al., 2010; Todeschini et al., 2019; de Felipe et al., 2020; Yang et al., 2022), which could play a role in increasing nutrient accumulation. Soybean breeding did not change P and N concentrations despite the increase in biomass, suggesting that the increase in biomass accumulation primarily drives P and N accumulation, which is influenced by factors such as the duration after flowering (Yang et al., 2022) and/or high photosynthesis rate (Todeschini et al., 2019). However, it is important to consider the role of soil nutrient status in nutrient accumulation. For example, DF with high soil-available P had higher P and N concentrations than SQ with low soil-available P, which were associated with the high leaf and stem P and N accumulation driven by the high stem and leaf biomass. Thus, soil nutrient status can increase nutrient accumulation by increasing biomass accumulation.

P and N accumulation increased with P supply. In addition, P and N accumulation positively correlated with leaf and pod dry weights (Figure 7) without diluting P and N concentrations, indicating that biomass accumulation with P supply primarily drove pod and leaf P accumulation. Pods had significantly higher genetic gains in P and N accumulation (average 26.9 mg m⁻² y⁻¹ for P, 239 mg m⁻² y⁻¹ for N) than leaves (11.9 mg m⁻² y⁻¹ for P, 91 mg m⁻² y⁻¹ for N), demonstrating that pods contributed more to P and N accumulation than leaves. The high genetic gains of pod P and N accumulation were also associated with the high demand for P and N during seed development; furthermore, the partitioning of P and N from leaves to seeds also contributed to pod P and N accumulation (Gaspar et al., 2017). Thus, enhanced pod P and N accumulation was associated with high seed numbers, a key driver for seed yield improvement during soybean breeding (Jin et al., 2010; He et al., 2019; Yang et al., 2022).

This study evaluates whether soybean breeding increased P and N utilization efficiencies in southwestern China. We confirmed that selecting for high seed yield during soybean breeding improved P and N utilization efficiencies, attributed to increased P and N partitioning to pods while reducing their partitioning to stems. The increased P and N accumulation in pods, driven by pod biomass, played a significant role in the overall P and N accumulation during soybean breeding. We conclude that soybean breeding increased P and N utilization efficiencies by regulating P and N accumulation and partitioning.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

JH and YJ conceived the experiments. XL and RD performed the experiments. XL, YJ, and RD analyzed the data. XL, KS, and JH wrote this manuscript. All authors contributed to the article and approved the submitted version.

This research was supported by the National Natural Science Foundation of China (32060427), Guizhou Science and Technology Support Program Project (Qiankehezhicheng (2021) yiban217).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.