Nitrogen (N) is an essential macronutrient element for plant growth. Plants take up N from the soil mainly in the form of nitrate (NO₃^–) and/or ammonium (NH₄⁺) (Wang et al., 2012; Fan et al., 2016). In crops, the demand for nitrogen is in accordance with its importance as the basic constituent of nucleotides, amino acids, and chlorophyll (Zhang M. et al., 2021). Nitrogen utilization involves multiple limiting factors, mainly including the varieties and the amount of nitrogen fertilizer, the method of nitrogen fertilizer application, and the expression of nitrogen genes (Ju et al., 2009; Yang et al., 2015). However, some elite rice germplasm can be harvested and produce the same biomass and grain yield under low N (LN) fertilizers, as compared with high N levels. This phenomenon is called low nitrogen-use efficiency (NUE), and thus, it mitigates detrimental effects on ecosystems (Good et al., 2004; Zhang X. et al., 2015). It is critically important to understand the molecular mechanisms of low-N-induced growth response (LNGR) and to develop crop varieties with improved NUE (Fan et al., 2017; Du et al., 2021; Yu et al., 2021).

The NUE is a complex trait controlled by multiple essential genes and quantitative trait loci (QTLs). Nitrogen-use efficiency is closely related to agronomic traits, such as plant height, tiller number, chlorophyll content, and yield, which can be markedly affected by N application in rice. In recent years, several NUE-related genes and loci have been identified by mutants and QTL mapping, targeting various components of nitrogen metabolism and potential regulators of NUE. DENSE AND ERECT PANICLES 1 (DEP1) is a major typical QTL that controls rice yield traits, and its allele, DEP1-1, affects rice NUE. The dominant allele is a nitrogen-insensitive vegetative growth mutation that increases nitrogen uptake and assimilation, resulting in improved harvest index and grain yield at moderate levels of nitrogen fertilization (Sun et al., 2014, 2018).

Allelic variants for the nitrate-transporter gene NRT1.1B display distinctive divergence between two subspecies of rice, namely, indica and japonica, which show the difference in NO₃^– utilization (Hu et al., 2015). Notably, ARE1 (for abc1 PEPRESSOR1) is prominent due to its enhanced NUE under nitrogen-limiting conditions (Wang et al., 2018a). Subsequently, grain number, plant height, and heading date7 (Ghd7) confer tolerance to LN, in order to positively regulate nitrogen utilization (Good et al., 2004; Saito et al., 2019; Wang et al., 2021). Additionally, the allelic variation in OsNR2 (Gao Z. et al., 2019) caused differences in the NO₃^– assimilation capacity and NUE of the indica and japonica subspecies. Moreover, it has been reported that other transcription factors, such as OsGRF4 (Li et al., 2018), OsTCP19 (Liu Y. et al., 2021), OsNLP4 (Yu et al., 2021), OsCCA1 (Zhang S. et al., 2021), MYB61 (Gao Y. et al., 2020), and OsNAC42 (Tang et al., 2019), participate in nitrogen utilization regulation. The molecular mechanisms of NUE between diverse rice varieties are still worthy of intensive exploration.

Increasing evidence has revealed that N supply has a significant impact on plant growth and development (McAllister et al., 2012; Chen et al., 2016). However, the conventional approach of linkage mapping has been used to identify potential genes with phenotypic variation, and even when this approach was limited to a biparental experimental population, it was time-consuming (Liu Z. et al., 2016; Wei et al., 2021).

Genome-wide association study (GWAS) is a high-resolution method for analyzing complex genetic traits in plants (Liu N. et al., 2016; Gao H. et al., 2020). Association mapping simultaneously reveals multiple loci of candidate genes with important agronomic traits. In this study, based on the genomic database and bioinformatics analysis, we carried out GWAS with 225 natural varieties of rice from the Rice 3K Sequencing Project with 4 LNGR agronomic phenotypes and analyzed their differences between normal nitrogen (NN) and LN conditions. We subsequently identified 56 LN-induced loci and 4 candidate genes for further analysis.

Nitrogen (N) is an indispensable element for crop production. Excessive chemical nitrogen fertilizer is applied to obtain high crop yields (Krapp, 2015; Chen et al., 2016). The sequence-based GWAS is a highly efficient approach for analyzing rice agronomic traits, and it can identify the elite genes and can evaluate rice core accessions to assist in specifically improving the usage of nitrogen. In this study, we collected 225 worldwide rice germplasms and performed GWAS on the trait differences (△NN-LN) of NUE-related agronomic traits. More importantly, compared with the hydroponic or seedling stage phenotyping experiment, our study was completed in the field under two different N levels (NN, 180 kg^⋅hm^–2; LN, 90 kg^⋅hm^–2) until maturity. The results obtained exhibited meaning and persuasiveness.

Plant NUE is an inherently complex trait (Masclaux-Daubresse et al., 2010; Gao Z. et al., 2019; Gao H. et al., 2020). At present, some studies exist that focus on NUE in rice (Hu et al., 2015; Wang et al., 2018b; Zhang S. et al., 2021). However, the studies on NUE by the GWAS analysis are limited, and to the best of our knowledge, few studies have been previously performed. In rice, a NIN-like protein, OsNLP4, was associated with NUE and was identified through GWAS in rice core populations. The previous studies showed that OsNLP4 imparts nitrogen assimilation efficiency and promotes the number of tillers and yield of rice (Yu et al., 2021). Additionally, OsNPF6.1^HapB is an elite haplotype derived from the GWAS analysis of extreme nitrogen-related phenotypes, such as PH, effective panicle number (EPN), and yield per plant (YPP) (Tang et al., 2019).

We measured the LN-related agronomic traits of SV, LL, PH, and TN of the 225 rice varieties under a NN level and an LN level. There was a correlation between different LN-related traits under NN and LN levels, suggesting that there were close physiological bases among different traits (Xu et al., 2012). We, therefore, focused on the difference (△NN-LN) between the phenotypic value of NN and LN and divided the varieties into hyposensitive and sensitive varieties. The critical first step before initiating association mapping is to collect a wide range of germplasm resources with different genetic backgrounds (Zheng M. et al., 2017). Our selection of 225 rice germplasms covered 5 continents around the world, which indicated that the results for genetic diversity were reliable for a feasible GWAS. Furthermore, we selected high-quality SNPs from the Rice 3K Sequencing Project and performed the association analysis on the phenotypic value difference (△NN-LN) covering the entire genome.

Numerous studies have supported the idea that N regulation is involved in plant development (Zhang S. et al., 2020; Zhang M. et al., 2021). Based on the previously reported genes and published studies, we conducted candidate gene identification. More importantly, the reliability of our GWAS for association mapping of NUE was previously confirmed by Gao (Gao Y. et al., 2020), who found that the indica allele of MYB61 exhibited superiority in governing nitrogen use and biomass production. Encouragingly, we also identified the NUE-associated genes OsOAT and MOC2 from our GWAS, and both of them play a critical role in modulating N use. Additionally, we also identified several other genes encoding N-related proteins worthy of mining in our study.

As expected, the gene LOC_Os12g41090, which encodes calcium/calmodulin-dependent protein kinases, displays a signal close to OsCPK12. OsCPK12 itself can participate in the signal transduction pathway in the LN stress response and can potentially improve the NUE of rice. One concern with studying this candidate gene is that the gene function must be verified. We, therefore, have constructed various vectors for transgenic verification. There are multiple haplotypes of genes in different varieties, which cause differences in phenotypic value. MYB61 was classified into five haplotypes based on three missense SNPs in its coding region. MYB61^Hap⋅5 was found to be an elite haplotype for its lower phenotypic difference (△NN-LN). MOC2 regulates the N-related function of rice, and further haplotype analysis revealed that MOC2 ^Hap⋅2 is the elite haplotype that enhances NUE. A previous study suggested that OsOAT has an important role in regulating NUE, and OsOAT^Hap⋅3 was significantly associated with the NUE of rice compared with other haplotypes. Moreover, Menjiagao_1 and Xiangaizao_10hao with MYB61^Hap⋅5 were identified as promising candidate rice accessions in NUE under LN-induced phenotypic data, which have shown an insignificant phenotypic difference (△NN-LN).

Different varieties responded differently to N levels, and the physiological differences may be caused by genotype and N interactions (Watanabe and Inubushl, 1986; He et al., 2020). The analysis of agronomic traits for two different N levels suggests a critical role for NUE. We successfully deduced that four candidate genes are involved in N regulation, and these were consistent with what currently appears in the available literature. Importantly, our research provides a theoretical basis for the selection and breeding of NUE genes and materials.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

YL, LG, QQ, and LS conceived and designed the study. LS and QW performed GWAS. YW and XL provided advice for the experimental design. YL, JM, HH, and YW performed the phenotyping measurement. YL, JM, and YW performed the data analysis. YL wrote the manuscript. All authors have read the manuscript and approved it for publication.

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.