Discussion
Plants cope with changing environments by adjusting their phenotype through phenotypic plasticity. The four adaptive phenotypes studied here, namely early and final rosette size, flowering time and seed yield, are known to respond to varying N availability (de Jong et al., 2019; Ikram et al., 2012; Masclaux-Daubresse & Chardon, 2011; Meyer et al., 2019; North et al., 2009). However, the genetic basis of their phenotypic plasticity in response to N availability is unknown. In this study, we quantified plasticity of the four phenotypes in response to N by using both CV and FC and used GWA analysis to identify candidate genes. None of the significantly associated loci contained genes known to be involved in N response, indicating that the plasticity to N availability has independent genetic basis in comparison to the focal N-related traits. This was further supported with our observation that the loci associated to the mean trait value of the accessions in different N condition were different from the loci associated with the trait plasticities to N. By using T-DNA mutant lines for the candidate genes, we verified that a regulator of chromosome condensation (RCC1) family protein, named here as PROTON1, influenced the plasticity of ERD in response to N availability. The T-DNA mutant line for PROTON1(SALK_117261) showed significantly larger size under limiting N and was significantly smaller under optimal N (Figure 3c), resulting in reduced plasticity of ERD across the conditions (Figure 3b). This suggests that RCC1-mediated stability in growth is attained by improved performance in the limiting conditions, but with reduced performance in optimal conditions.
RCC1-family proteins were first identified to function in regulating cell cycle, but since then different members of this family are known to be involved in diverse functions (Hadjebi et al., 2008). Arabidopsis contains 24 RCC1-family genes, but only four of them have been characterized so far. PROTON1 is the most highly expressed in roots and young leaves (BAR eFP Browser, Arabidopsis.org), and we found that it was induced by limiting N in young leaves (Figure 4b). Co-expression analysis of PROTON1 showed that it is co-expressed with eight genes with function in the spliceosome machinery (Figure S5) that mediates alternative splicing plants. The four genes in primary co-expression network of PROTON1 include two RNA-binding family protein members. One of these is SR34, a known splicing factors in Arabidopsis (Stankovic et al., 2016). The role of alternative splicing controlling PROTON1-mediated plasticity of ERD and FT in response to N availability remains to be investigated. Alternative splicing regulates gene expression and protein diversity by producing multiple mRNA isoforms from a single gene. In Arabidopsis , alternative splicing is known to regulate transitioning to flowering, but less is known how the alternative splicing is regulated by environmental cues. An intriguing question for future studies would be to investigate if variability in spliceosome, similar to the HSP90 system(Queitsch, Sangster, & Lindquist, 2002; Salathia & Queitsch, 2007; Sangster & Queitsch, 2005; Sangster et al., 2008; Zabinsky, Mason, Queitsch, & Jarosz, 2019), plays a central role in regulating plasticity in plants and animals.
In addition to the genetic basis of plasticity, we used correlation analysis to investigate whether plasticity of specific primary metabolites was associated with the plasticity of any of the four phenotypes. We identified that plasticities of the fumarate and glycerate levels showed significant negative correlation with the plasticity of FRD. The mean levels of fumarate and glycerate were previously found to negatively correlate with the biomass of Arabidopsis plants grown under low N conditions (Sulpice et al., 2013). In our experiments, the mean levels of fumarate showed negative correlation with the mean ERD only under optimal N conditions (Figure S4c). This suggests, similar to the four phenotypes, that the mean metabolites levels associated with N responses are different from the plasticity of metabolite levels in response to N.
To rapidly respond to changes in surroundings, plasticity in an important trait for survival of homozygous organisms, such as Arabidopsis. We found that plasticities of ERD and FT and the plasticity of FRD correlated in across the studied accessions (Figure 2c). In addition, proton1 mutant line was associated with both altered ERD and FT plasticity in response to N availability. Further, inproton1, the ERD and FT plasticities were due to faster development and earlier FT in limiting N conditions (Figure 4c, d). These findings indicate that these traits respond simultaneously to the changing N. Furthermore, we found that the different environments had different impact on the focal traits. This enhances our understanding of the complexity of possible constraints and consequences of selection when acting on plasticity in response to climate change. Altogether, these results highlight the importance to investigate plasticity of different phenotypes in multiple environments alone and in combination, when understanding the past, present and future relationships between the plants and environment.