Ecotype-specific responses to HLC growth for CBFs and CBF-regulated genes
The fact that the strongest induction of CBF1-3 was achieved in both ecotypes under HLC indicates a synergistic effect of cool temperature and high light on CBF1-3 induction. Elevated excitation pressure in the chloroplast may contribute to maintenance of elevated CBF1-3 expression during long-term growth in HLC (Hüner et al., 2012; Hüner et al., 2016). Other possible contributing signals include tetrapyrrole Mg-ProtoIX-mediated retrograde signaling that impacts CBF1-3 expression levels in specific mutant backgrounds (Lee & Thomashow, 2012; Noren et al., 2016). Ecotypic differences presumably also shape how HL-dependent signals are translated toCBF1-3 induction. For instance, the finding that SW in HLC maintained a more oxidized QA state under experimental exposure to high light relative to IT in HLC is consistent with other phenotypic measures that indicate superior adaptation of SW to either high light or cold temperature than IT (Adams et al., 2014, 2016, 2018; Ågren & Schemske, 2012; Agrena, Oakley, McKay, Lovell, & Schemske, 2013; Cohu et al. 2013a,b; Oakley et al., 2014; Park et al., 2018; Stewart et al., 2015, 2016, 2017).
CBF1-3 are essential for long term maintenance of freezing tolerance in both ecotypes
The present finding that CBF1-3 are essential to the full induction of freezing tolerance in the SW and IT ecotypes under HLC confirms their importance in A. thaliana as a key survival trait for overwintering plants. Previous studies showed that CBFs are essential for full induction of freezing tolerance in mature A. thaliana plants grown under warm conditions and transferred in one step to chilling conditions (Jia et al., 2016; Park et al., 2018; Zhao et al., 2016). The present study on plants grown from seed under differing temperature regimes demonstrates that CBFs also have an essential role in the long-term maintenance of elevated freezing tolerance. Moreover, just as was concluded from short-term studies on warm-grown cbfmutants abruptly transferred to cold conditions (Jia et al., 2016; Park et al., 2018; Zhao et al., 2016), both CBF-dependent and CBF-independent pathways also appear to be required for long-term freezing tolerance in HLC plants – as illustrated here by the moderately enhanced freezing tolerance in both it:cbf123 and sw:cbf123 under HLC versus LLW. This finding is also consistent with the observation that gene expression of CBF-target genes under HLC was strongly reduced, but not fully blocked, in cbf123 triple mutants in the present study. Furthermore, the contribution of the CBF-dependent pathway to freezing tolerance after long-term growth under HLC was similar in both ecotypes (reduction of LT50 of freezing tolerance by 3.5 ºC and 3.4ºC under HLC in it:cbf123 and sw:cbf123 versus IT and SW, respectively). These findings indicate that the greater freezing tolerance of SW versus IT in HLC is due to CBF-independent pathways contributing to freezing tolerance (see also Park et al., 2018).
The absence of a clear effect under HLC in the sw:cbf2 mutant is best explained by the impact of paralog compensation (Gilmour et al., 2004; Jia et al., 2016; Park et al., 2018; Zhao et al., 2016), i.e., functionally overlapping components that can partly compensate for each other’s loss. The observation in the present study of stronger induction of many CBF-target genes under HLC in IT with its pre-existingcbf2 mutation relative to SW could also be interpreted in the context of paralog compensation by CBF1 and CBF3 in acbf2 mutant background. Several independent A. thalianalineages evolved loss-of-function mutations in individual CBFgenes without apparent severe adverse effects on survival in regions with mild winters (Kang et al., 2013; Monroe et al., 2016). The full suite of CBF1–3 may thus only be required for tolerance to colder conditions than used here (daytime air temperature of 8ºC and leaf temperature of ~14 ºC).
However, paralog compensation among CBF1–3 does not explain the observed significant induction of CBF-target genes and moderately elevated freezing tolerance of cbf123 mutants in HLC, i.e., completely CBF-independent induction of some level of freezing tolerance. The concept of paralog compensation can also apply to entire gene families that may functionally overlap and compensate for each other’s loss. CBF1–3 belong to the ERF/AP2 A-1 subfamily that includes three additional members located outside the CBF1–3gene locus in A. thaliana (Mizoi, Shinozaki, & Yamaguchi-Shinozaki, 2012). However, these three other ERF/AP2 A-1 subfamily members (DDF1, AT1G12610; DDF2 , AT1G63030;CBF4 , AT5G51990) are not expressed at detectable levels in leaf tissue of IT or SW under any of the four growth regimes. Additionally, Park et al. (2018) observed no induction of these three genes in the it:cbf123 and sw:cbf123 in their various cold treatments. It is thus unlikely that paralog compensation accounts for the CBF-independent induction of freezing observed in cbf123 mutants. Instead, the present findings suggest involvement of unrelated signaling networks.