ABSTRACT
In concert with other phytohormones, auxin regulates plant growth and development. However, how auxin and other phytohormones coordinately regulate distinct processes is not fully understood. In this work, we uncover an auxin-abscisic acid (ABA) interaction module in Arabidopsis that is specific to coordinating activities of these hormones in the hypocotyl. From our forward genetics screen, we determine that ABA biosynthesis is required for the full effects of auxin on hypocotyl elongation. Our data also suggest that ABA biosynthesis is not required for the inhibitory effects of auxin treatment on root elongation. Our transcriptome analysis identified distinct auxin-responsive genes in root and shoot tissues, which is consistent with differential regulation of growth in these tissues. Further, our data suggest that many gene targets repressed upon auxin treatment require an intact ABA pathway for full repression. Our results support a model in which auxin stimulates ABA biosynthesis to fully regulate hypocotyl elongation.
INTRODUCTION
Seedling establishment is a particularly tenuous time in the life of a plant. After a seed germinates, typically below the soil line, the nascent seedling must quickly burn through seed reserves to fuel the growth necessary to emerge from the soil and capture sunlight. In Arabidopsis, this growth of the hypocotyl is driven solely by cell expansion (Gendreau et al., 1997). Multiple pathways converge on regulating Arabidopsis hypocotyl elongation; however, the plant hormone auxin is a key controller of the cell expansion driving this growth.
Plant hormones work together to coordinate plant growth and development (reviewed in Vanstraelen and Benková, 2012). The plant hormone auxin is a central regulator of many plant growth and developmental processes through regulation of cell division and expansion (reviewed in Perrot-Rechenmann, 2010), sitting at the nexus of multiple signaling pathways. How auxin coordinates with other signaling pathways to regulate specific growth processes is not fully understood.
The plant hormone abscisic acid (ABA) is often thought of as a ‘stress hormone’; however, it also plays roles under non-stress conditions (Yoshida et al., 2019). Indeed, roles for ABA as a regulator of plant growth have become increasingly apparent (Brookbank et al., 2021). Mounting evidence suggests that auxin and ABA work together to regulate various processes (reviewed in Emenecker and Strader, 2020), including seed germination (Belin et al., 2009; Liu et al., 2013; Monroe-Augustus et al., 2003; Thole et al., 2014), hypocotyl elongation (Lorrai et al., 2018a), root elongation (Monroe-Augustus et al., 2003; Strader et al., 2008; Thole et al., 2014), lateral root formation (Lu et al., 2019; Shin et al., 2007; Shkolnik-Inbar and Bar-Zvi, 2010; Xing et al., 2016) and cotyledon expansion (Rinaldi et al., 2012). Notably, in all cases where it has been investigated, including seed germination, root elongation and lateral root formation, ABA requires functional auxin biosynthesis and signaling to regulate these processes (Emenecker and Strader, 2020).
Because many mutants that exhibit resistance to auxin in root growth assays are not markedly resistant to the inhibitory effects of auxin on dark-grown hypocotyl elongation (Strader et al., 2008, 2011), we hypothesized that our understanding of how auxin regulates this crucial process was incomplete. Here, we describe the identification of an Arabidopsis mutant that is resistant to the suppressive effects of supplied auxin on dark-grown hypocotyl elongation. The ABSCISIC ALDEHYDE OXIDASE 3 (AAO3) gene is required for full responsiveness to auxin in the regulation of hypocotyl elongation. AAO3 is the enzyme responsible for catalyzing the final step in ABA biosynthesis (Seo et al., 2000). Examination of various ABA biosynthetic and signaling mutants revealed that auxin requires functional ABA biosynthesis and signaling to fully exert its regulatory effects on hypocotyl elongation. We found that auxin treatment upregulated expression of NCED5, which encodes an enzyme that performs a rate-limiting step in ABA biosynthesis, specifically in shoot tissues. Further, seedling shoots displayed increased ABA levels in response to auxin treatment, raising the possibility that auxin directly regulates ABA levels in this tissue. Finally, we found that the root and the shoot have distinct transcriptional responses to auxin and that many auxin-responsive transcripts in the hypocotyl and root require intact ABA biosynthesis. Our data support a model in which auxin relies on intact ABA biosynthesis to regulate some aspects of hypocotyl elongation. Our findings further support that some tissues can display hormone interactions distinct from those in other tissues and raise the possibility that ABA alters auxin-regulated hypocotyl and root growth under stressful conditions.
RESULTS
Mutations in ABSCISIC ALDEHYDE OXIDASE 3 confer auxin resistance in dark-grown hypocotyl elongation assays
Depending on context, auxin can either promote or inhibit cell expansion (reviewed in Perrot-Rechenmann, 2010). It is thought that there exists an optimal range of auxin concentrations to promote cell expansion, and that supraoptimal auxin concentrations result in reduced cell expansion. Supraoptimal concentrations of exogenous auxin inhibit hypocotyl elongation in dark-grown seedlings (Strader et al., 2011); however, many mutants strongly resistant to the inhibitory effects of auxin in root elongation assays are only mildly resistant to the inhibitory effects of auxin on dark-grown hypocotyl elongation, which is consistent with the possibility that distinct mechanisms drive auxin responses in hypocotyl and root tissues. We therefore expanded from our previous Arabidopsis hypocotyl resistance (HR) screen (Strader et al., 2011) to identify ethyl methanesulfonate (EMS)-mutagenized, dark-grown M2-generation seedlings that display resistance to the inhibitory effects of the auxin indole-3-butyric acid (IBA) on hypocotyl elongation. Hypocotyl resistant 12 (HR12) displayed resistance to the inhibitory effects of the natural active auxin indole-3-acetic acid (IAA), the natural long-chain auxin precursor IBA, the synthetic active auxin 2,4- dichlorophenoxyacetic acid (2,4-D) and the synthetic long-chain auxin precursor 2,4-dichlorophenoxybutyric acid (2,4-DB) on dark-grown hypocotyl elongation (Fig. 1A). This suggests that the defect in HR12 generally affects auxin responsiveness in the hypocotyl.
Characterization of HR12 and confirmation of the causative mutation. (A) Mean±s.d. hypocotyl lengths (n≥78) of wild-type (Wt; Col-0), HR12 and aao3-4 seedlings grown in the dark on medium supplemented with ethanol (mock) or the indicated treatment. Statistically significant (P≤0.01) differences from wild type for each treatment are indicated with an asterisk (one-way ANOVA with Tukey HSD test). (B) Bulk segregant analysis revealed homozygous EMS mutations within the HR12 backcrossed population. Schematic chromosome map shows loci in which mutations were identified. (C) AAO3 (At2g27150) gene schematic depicting the EMS-consistent point mutation identified in aao3-12 (HR12) and the aao3-4 T-DNA insertion allele (SALK-072361c). The aao3-12 line carries a C-to-T mutation at position +3846 (where the A of ATG is +1) that results in a Q1081-to-stop substitution (nt, nucleotide). (D) Non-complementation assay of HR12 (aao3-12; 3-12) and aao3-4 (3-4). Mean±s.d. hypocotyl lengths (n=20) of dark-grown seedlings of the indicated genotype for ethanol (mock) and 30 µM IBA treatments. Significant differences (P≤0.01) between wild type and the indicated genotype are indicated by an asterisk (one-way ANOVA with Tukey HSD test). (E) Photograph of seedlings grown in the dark on medium supplemented with ethanol (mock) or 4 µM 2,4-DB. Examined genotypes include wild type, aao3-12 (3-12/3-12), aao3-4 (3-4/3-4) and the F1 transheterozygote from a cross between aao3-12 and aao3-4 (3-12/3-4). (F) The AAO3 gene complements aao3-12 (HR12) phenotypes. Mean±s.d. hypocotyl lengths (n=40) of wild-type and aao3-12 seedlings carrying either no transgene, 35S:AAO3 or 35S:HA-AAO3, grown in the dark on medium supplemented with ethanol (mock) or the indicated auxin. Significant differences (P≤0.01) in comparison with wild type carrying no transgene are indicated with an asterisk (one-way ANOVA with Tukey HSD test). NS denotes that there was not a significant difference between the indicated genotype-treatment combination and its wild-type equivalent.
Characterization of HR12 and confirmation of the causative mutation. (A) Mean±s.d. hypocotyl lengths (n≥78) of wild-type (Wt; Col-0), HR12 and aao3-4 seedlings grown in the dark on medium supplemented with ethanol (mock) or the indicated treatment. Statistically significant (P≤0.01) differences from wild type for each treatment are indicated with an asterisk (one-way ANOVA with Tukey HSD test). (B) Bulk segregant analysis revealed homozygous EMS mutations within the HR12 backcrossed population. Schematic chromosome map shows loci in which mutations were identified. (C) AAO3 (At2g27150) gene schematic depicting the EMS-consistent point mutation identified in aao3-12 (HR12) and the aao3-4 T-DNA insertion allele (SALK-072361c). The aao3-12 line carries a C-to-T mutation at position +3846 (where the A of ATG is +1) that results in a Q1081-to-stop substitution (nt, nucleotide). (D) Non-complementation assay of HR12 (aao3-12; 3-12) and aao3-4 (3-4). Mean±s.d. hypocotyl lengths (n=20) of dark-grown seedlings of the indicated genotype for ethanol (mock) and 30 µM IBA treatments. Significant differences (P≤0.01) between wild type and the indicated genotype are indicated by an asterisk (one-way ANOVA with Tukey HSD test). (E) Photograph of seedlings grown in the dark on medium supplemented with ethanol (mock) or 4 µM 2,4-DB. Examined genotypes include wild type, aao3-12 (3-12/3-12), aao3-4 (3-4/3-4) and the F1 transheterozygote from a cross between aao3-12 and aao3-4 (3-12/3-4). (F) The AAO3 gene complements aao3-12 (HR12) phenotypes. Mean±s.d. hypocotyl lengths (n=40) of wild-type and aao3-12 seedlings carrying either no transgene, 35S:AAO3 or 35S:HA-AAO3, grown in the dark on medium supplemented with ethanol (mock) or the indicated auxin. Significant differences (P≤0.01) in comparison with wild type carrying no transgene are indicated with an asterisk (one-way ANOVA with Tukey HSD test). NS denotes that there was not a significant difference between the indicated genotype-treatment combination and its wild-type equivalent.
To identify the causative mutation in HR12, we backcrossed the mutant to the wild-type line and isolated auxin-resistant individuals in the F2 generation. We then used a whole-genome sequencing of bulk segregants approach (Thole et al., 2014; Thole and Strader, 2015) on DNA from pooled F3 individuals to identify nine homozygous EMS-induced mutations in the HR12 mutant (Fig. 1B). One of these mutations was in At2g27150 and resulted in a premature stop codon in AAO3, prompting us to further investigate whether this mutation was causative for the auxin resistance seen in HR12. We named the AAO3 mutation in HR12 aao3-12 (Fig. 1C).
To determine whether the aao3-12 mutation was causative for the auxin resistance observed in HR12, we examined auxin responsiveness of a T-DNA insertional allele, Salk_072361C (hereafter referred to as aao3-4; Fig. 1C). Like HR12, the aao3-4 mutant displayed resistance to auxin in dark-grown hypocotyl elongation assays (Fig. 1A) suggesting that the mutation in AAO3 is causative for the observed auxin resistance in hypocotyl elongation assays in HR12. We note that aao3-4 and HR12 exhibited slight but statistically significant shorter hypocotyl lengths than wild-type individuals under mock conditions (Fig. 1A); however, we did not find this consistently across additional hypocotyl elongation assays (Fig. 1D,E), and this effect might be due to extraneous factors such as seed batch variation.
To further determine whether the AAO3 mutation was causative in HR12, we crossed aao3-4 to HR12 and examined auxin sensitivity of the F1 progeny, finding that the aao3-12/aao3-4 transheterozygote displayed resistance to the inhibitory effects of IBA (Fig. 1D) and 2,4-DB (Fig. 1E) on hypocotyl elongation. Thus, aao3-4 and aao3-12 fail to complement one another.
As a final means of verifying that the aao3-12 lesion caused the observed resistance to the inhibitory effects of auxin in dark-grown hypocotyl elongation, we transformed aao3-12 with a wild-type genomic copy of AAO3 under the control of the cauliflower mosaic virus 35S promoter. Expression of a wild-type copy of AAO3 was sufficient to restore auxin sensitivity to aao3-12 (Fig. 1F). We note that lines overexpressing AAO3 frequently displayed increased sensitivity to the inhibitory effects of auxin on dark-grown hypocotyl elongation (Fig. 1F). Although statistically significant, these differences were small and not consistent across all auxins tested. However, this might present indirect evidence of the influence that ABA has on auxin regulation of hypocotyl elongation.
Altogether, our results demonstrate that multiple alleles of aao3 result in reduced auxin sensitivity in dark-grown hypocotyl elongation assays, that these alleles fail to complement one another, and that provision of a wild-type copy of AAO3 rescues mutant phenotypes, supporting the conclusion that the auxin response defects in HR12 (aao3-12) result from reduced AAO3 function.
Auxin requires intact ABA biosynthesis and signaling for full regulation of hypocotyl elongation
AAO3 catalyzes the final step in ABA biosynthesis (Fig. 2A) (Seo et al., 2000). We reasoned that if disruption of ABA biosynthesis, rather than an unrelated role for AAO3, causes the auxin resistance seen in aao3 mutants, then mutants with defects in other steps in the ABA biosynthetic pathway (Fig. 2A) should also display auxin resistance in dark-grown hypocotyl elongation assays. We therefore obtained T-DNA insertional alleles for ABA DEFICIENT 2 (ABA2) and ABA DEFICIENT 3 (ABA3). We found that disruption of either ABA2 or ABA3 resulted in auxin resistance in dark-grown hypocotyl elongation assays (Fig. 2B), supporting a model wherein ABA biosynthesis is necessary for full auxin-mediated inhibition of dark-grown hypocotyl elongation. Further, mutants with defects in ABA signal transduction displayed partial resistance to the inhibitory effects of auxin on dark-grown hypocotyl elongation (Fig. S1). Thus, ABA biosynthesis and response are required for full sensitivity to the effects of exogenous auxin in dark-grown hypocotyls.
ABA biosynthesis modulates auxin effects on dark-grown hypocotyl elongation. (A) Schematic of the ABA biosynthetic pathway with intermediates and enzymatic steps indicated. Unless otherwise indicated, the number of genes encoding enzymes involved in each step in the biosynthetic process is one. The asterisk indicating three genes for AAO denotes that, whereas single mutants of AAO3 have dramatic reductions in ABA concentration, mutations of AAO1 and AAO4 only cause reduced ABA levels if they are in a double mutant that also has loss-of-function mutation of AAO3 (Seo et al., 2004). (B) Mean±s.d. hypocotyl lengths (n=40) of wild-type (Wt; Col-0), aao3-4, aao3-12, aba2-1, aba2-3, aba2-4, aba3-1 and gin1-3 seedlings grown in the dark on medium supplemented with ethanol (mock) or the indicated hormone. The gin1-3 mutant is defective in ABA2. Significant differences (P≤0.01) in comparison with wild type for each treatment are marked by an asterisk (one-way ANOVA with Tukey HSD test). (C) Quantification of ABA levels in dark-grown seedling shoots in response to auxin treatment (10 µM IAA; FW, fresh weight). Bars represent the distribution of the data with the exception of outliers (any data outside of the range of the 25th percentile of all data minus 1.5x interquartile range to the 75th percentile of all data plus 1.5x interquartile range). The line in the box represents the median, and each box represents the interquartile range. Individual data points from each replicate are denoted by the circles. Six replicates were carried out for each treatment, and for each treatment at least thirty seedlings were used. (D) Mean±s.d. (n≥19) dark-grown hypocotyl elongation following 5 µM IAA treatment, shown as a percentage of the elongation in the corresponding mock-treated control. Seedlings were transferred to the indicated concentration of ABA or ethanol (mock) after germination. Significant differences (P≤0.01) in comparison with wild type for each treatment are marked by an asterisk (NS, no significant difference; one-way ANOVA with Tukey HSD test).
ABA biosynthesis modulates auxin effects on dark-grown hypocotyl elongation. (A) Schematic of the ABA biosynthetic pathway with intermediates and enzymatic steps indicated. Unless otherwise indicated, the number of genes encoding enzymes involved in each step in the biosynthetic process is one. The asterisk indicating three genes for AAO denotes that, whereas single mutants of AAO3 have dramatic reductions in ABA concentration, mutations of AAO1 and AAO4 only cause reduced ABA levels if they are in a double mutant that also has loss-of-function mutation of AAO3 (Seo et al., 2004). (B) Mean±s.d. hypocotyl lengths (n=40) of wild-type (Wt; Col-0), aao3-4, aao3-12, aba2-1, aba2-3, aba2-4, aba3-1 and gin1-3 seedlings grown in the dark on medium supplemented with ethanol (mock) or the indicated hormone. The gin1-3 mutant is defective in ABA2. Significant differences (P≤0.01) in comparison with wild type for each treatment are marked by an asterisk (one-way ANOVA with Tukey HSD test). (C) Quantification of ABA levels in dark-grown seedling shoots in response to auxin treatment (10 µM IAA; FW, fresh weight). Bars represent the distribution of the data with the exception of outliers (any data outside of the range of the 25th percentile of all data minus 1.5x interquartile range to the 75th percentile of all data plus 1.5x interquartile range). The line in the box represents the median, and each box represents the interquartile range. Individual data points from each replicate are denoted by the circles. Six replicates were carried out for each treatment, and for each treatment at least thirty seedlings were used. (D) Mean±s.d. (n≥19) dark-grown hypocotyl elongation following 5 µM IAA treatment, shown as a percentage of the elongation in the corresponding mock-treated control. Seedlings were transferred to the indicated concentration of ABA or ethanol (mock) after germination. Significant differences (P≤0.01) in comparison with wild type for each treatment are marked by an asterisk (NS, no significant difference; one-way ANOVA with Tukey HSD test).
Although the ABA biosynthesis mutants we examined showed resistance to auxin in the inhibition of dark-grown hypocotyl elongation, it is important to note that the ABA biosynthesis mutants are capable of responding to auxin treatment in these assays, albeit to a lesser extent than the wild-type response. This suggests that there exists ABA-dependent and ABA-independent mechanisms by which auxin can regulate hypocotyl elongation.
Because all examined ABA biosynthesis mutants displayed resistance to the inhibitory effects of auxin in dark-grown hypocotyl elongation assays, we suspected that auxin might be regulating hypocotyl elongation by directly altering endogenous ABA levels. We therefore quantified ABA levels following auxin treatment of dark-grown seedling shoots. We treated 4-day-old, dark-grown Arabidopsis seedlings with auxin or mock treatment for 2 h or 4 h. Following treatment, we bisected the seedlings at the root-shoot junction and quantified ABA levels in the shoot using mass spectrometry. We found that both the 2 h and 4 h auxin treatments resulted in significant increases in endogenous ABA levels in comparison to ABA levels following mock treatment (Fig. 2C), suggesting that auxin treatment rapidly alters endogenous ABA levels. We next tested whether ABA supplementation could restore auxin responsiveness to an ABA biosynthesis mutant. We found that the inhibitory effects of auxin on dark-grown hypocotyl elongation could be restored to wild-type levels in aba2-3 seedlings transferred to 25 nM ABA after germination (Fig. 2D). Altogether, these data suggest that ABA levels modulate auxin responses in dark-grown hypocotyls.
ABA biosynthesis mutants exhibit reduced promotion of hypocotyl elongation in response to auxin during growth in the light
During growth in the light, exogenous auxin promotes hypocotyl elongation (Chapman et al., 2012), and growth under elevated temperature causes a natural increase in auxin levels that results in elongated hypocotyls (Gray et al., 1998). To further interrogate the conditions under which auxin relies on intact ABA biosynthesis to regulate hypocotyl elongation, we examined the response of ABA biosynthesis mutants to auxin under these conditions. Growing wild-type seedings in the presence of the synthetic auxin picloram resulted in promotion of hypocotyl elongation in light-grown seedlings and inhibition of hypocotyl elongation in dark-grown seedlings (Fig. 3A). In wild-type seedlings, growth under continuous illumination in the presence of picloram resulted in a 120% increase in hypocotyl length compared to growth in the absence of auxin (Fig. 3A,B). Whereas the nonsense allele aao3-12 displayed an increase in hypocotyl elongation in response to picloram similar to that observed in the wild type, the strong ABA biosynthesis mutants aao3-4, aba2-3, aba3-1 and gin1-3 (a mutant of ABA2) displayed a 60% increase in hypocotyl elongation in response to picloram, less than that observed in wild type (Fig. 3B). In addition, we found that aba2-3 failed to display hypocotyl elongation when grown under elevated temperatures (Fig. 3C). Thus, ABA-dependent aspects of auxin-mediated regulation of hypocotyl elongation are independent of whether the seedlings are grown in the dark or in the light. Alternatively, it is possible that the ABA biosynthesis mutants achieve near-maximal hypocotyl lengths in the absence of auxin and thus only appear to be auxin resistant because they are unable to grow further. This alternative seems unlikely, as hypocotyls are much longer when grown in the dark or under elevated temperatures (Gray et al., 1998), suggesting that these seedlings have not reached a maximum of growth.
ABA biosynthesis mutants are resistant to promotion of hypocotyl elongation by auxin in the light. (A) Photograph of wild-type (Wt; Col-0) and aba2-3 seedlings grown under continuous illumination (top) or in the dark (bottom) on medium supplemented with DMSO (mock) or 2.5 µM picloram (a synthetic auxin). (B) Hypocotyl lengths (n=30) of wild-type, aao3-12, aao3-4, aba2-3, aba3-1 and gin1-3 seedlings grown under continuous illumination on medium supplemented with the synthetic auxin picloram (2.5 µM), shown as a percentage of the hypocotyl length of seedlings of the same genotype grown in the presence of DMSO (mock). Significant differences (P≤0.01) in comparison with wild type for each treatment are marked by an asterisk (one-way ANOVA with Tukey HSD test). (C) Hypocotyl lengths (n≥15) of wild-type and aba2-3 seedlings grown under continuous illumination at 22°C or 28°C, shown as a percentage of hypocotyl length of seedlings of the same genotype at 22°C. Box plots in B and C show the median (line), interquartile range (box) and range excluding outliers (bars). Outliers (any data outside of the range of the 25th percentile of all data minus 1.5x interquartile range to the 75th percentile of all data plus 1.5x interquartile range) are denoted by circles.
ABA biosynthesis mutants are resistant to promotion of hypocotyl elongation by auxin in the light. (A) Photograph of wild-type (Wt; Col-0) and aba2-3 seedlings grown under continuous illumination (top) or in the dark (bottom) on medium supplemented with DMSO (mock) or 2.5 µM picloram (a synthetic auxin). (B) Hypocotyl lengths (n=30) of wild-type, aao3-12, aao3-4, aba2-3, aba3-1 and gin1-3 seedlings grown under continuous illumination on medium supplemented with the synthetic auxin picloram (2.5 µM), shown as a percentage of the hypocotyl length of seedlings of the same genotype grown in the presence of DMSO (mock). Significant differences (P≤0.01) in comparison with wild type for each treatment are marked by an asterisk (one-way ANOVA with Tukey HSD test). (C) Hypocotyl lengths (n≥15) of wild-type and aba2-3 seedlings grown under continuous illumination at 22°C or 28°C, shown as a percentage of hypocotyl length of seedlings of the same genotype at 22°C. Box plots in B and C show the median (line), interquartile range (box) and range excluding outliers (bars). Outliers (any data outside of the range of the 25th percentile of all data minus 1.5x interquartile range to the 75th percentile of all data plus 1.5x interquartile range) are denoted by circles.
Overall, our data that ABA biosynthesis mutants do not respond as robustly to auxin or increased temperatures in the context of hypocotyl elongation are consistent with the possibility that ABA biosynthesis is required for full auxin response in the promotion of light-grown hypocotyl elongation.
ABA biosynthesis mutants display wild-type responsiveness to auxin in root elongation assays
Previous assays to examine interactions between auxin and ABA have focused on the effects of these hormones on root elongation (Thole et al., 2014) or seed germination (Liu et al., 2013; Strader et al., 2008). Indeed, our previous studies have demonstrated that auxin signaling is required for ABA effects on each of these processes (Rinaldi et al., 2012; Strader et al., 2008; Thole et al., 2014). Thus, we were surprised to identify an ABA biosynthesis mutant in our screen for resistance to the inhibitory effects of auxin on dark-grown hypocotyl elongation. We therefore examined root auxin responses of the ABA biosynthesis mutants more closely. Auxin-responsive root elongation is typically studied in light-grown seedlings. Consistent with previous studies (Thole et al., 2014), we found that mutants defective in ABA biosynthesis displayed wild-type sensitivity to auxin in light-grown root elongation assays (Fig. 4A). Because we examined hypocotyl elongation under both light and dark conditions, we also assessed sensitivity of the ABA biosynthesis mutants to the inhibitory effects of auxin on dark-grown root elongation, finding that they displayed wild-type sensitivity (Fig. 4B). Thus, altering ABA biosynthesis appears to dampen the response to exogenous auxin in hypocotyl, but not root, elongation.
ABA biosynthesis mutants show wild-type sensitivity to the inhibitory effects of auxin on root elongation. (A,B) Mean±s.d. root lengths of wild-type (Wt; Col-0), aao3-4, aao3-12, aba2-1, aba2-3, aba2-4, aba3-1 and gin1-3 seedlings grown on medium supplemented with ethanol (mock) or the indicated hormone either (A) under continuous illumination (n≥35) or (B) in the dark (n≥25).
ABA biosynthesis mutants show wild-type sensitivity to the inhibitory effects of auxin on root elongation. (A,B) Mean±s.d. root lengths of wild-type (Wt; Col-0), aao3-4, aao3-12, aba2-1, aba2-3, aba2-4, aba3-1 and gin1-3 seedlings grown on medium supplemented with ethanol (mock) or the indicated hormone either (A) under continuous illumination (n≥35) or (B) in the dark (n≥25).
Auxin and ABA transcriptional responses in dark-grown roots and shoots
Because our data suggested specific requirements for ABA biosynthesis in shoot tissues, we hypothesized that auxin transcriptional responses might be distinct in root and shoot tissues. To this end, we examined auxin- and ABA-regulated transcriptional programs in dark-grown roots (everything below the root-hypocotyl junction) and dark-grown shoots (everything above the root-hypocotyl junction) (Fig. 5A). Working under green safe lights, we exposed 4-day-old dark-grown seedlings to mock (ethanol), auxin or ABA treatment for 2 h prior to bisecting the root and the shoot at the root-hypocotyl junction and examining the differential effects of these hormone treatments in these two tissues (Fig. 5B).
The shoot and the root show different responses to auxin and ABA treatment. (A) Dark-grown seedlings were treated with ethanol (mock), 10 µM IAA or 10 µM ABA prior to dissection of the root and the shoot by cutting at the root-hypocotyl junction. RNA was extracted from the separated tissues, which was then used for RNA-seq analysis. (B) Volcano plots displaying pairwise transcript accumulation differences between auxin (IAA)- or ABA-treated and mock-treated wild-type (Col-0) shoots (top) and roots (bottom). Dashed lines indicate significantly differentially expressed genes defined as follows: y-axis (horizontal dashed line), −log10(P-value) where the P-value is the adjusted P-value (which is what we used for all P-values for the RNA-seq data) with a cutoff of less than or equal to 0.05; x-axis (vertical dashed lines), log2(FoldChange) less than or equal to −1 or greater than or equal to 1. The data points are colored in orange if they are significantly downregulated, in magenta if they are significantly unregulated or in gray if they are not significnatly differentially regulated. Datapoints are slightly transparent to help visualize overlapping datapoints. (C) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin treatment in the root and the shoot of wild-type seedlings. Gene sets are broken down into whether auxin treatment resulted in upregulation of the gene (left) or downregulation of the gene (right). (D) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to ABA treatment in the root and the shoot of wild-type seedlings. Gene sets are broken down into whether ABA treatment resulted in upregulation of the gene (left) or downregulation of the gene (right). (E) NCED family transcript accumulation differences in auxin-treated wild-type root and shoot tissues. Numbers are shown as auxin-induced fold change in transcript accumulation in comparison to mock treatment. ND, not detected in our RNA-seq data; NS, not significant (|log2 fold change|≤ 1.0; FDR≥0.05).
The shoot and the root show different responses to auxin and ABA treatment. (A) Dark-grown seedlings were treated with ethanol (mock), 10 µM IAA or 10 µM ABA prior to dissection of the root and the shoot by cutting at the root-hypocotyl junction. RNA was extracted from the separated tissues, which was then used for RNA-seq analysis. (B) Volcano plots displaying pairwise transcript accumulation differences between auxin (IAA)- or ABA-treated and mock-treated wild-type (Col-0) shoots (top) and roots (bottom). Dashed lines indicate significantly differentially expressed genes defined as follows: y-axis (horizontal dashed line), −log10(P-value) where the P-value is the adjusted P-value (which is what we used for all P-values for the RNA-seq data) with a cutoff of less than or equal to 0.05; x-axis (vertical dashed lines), log2(FoldChange) less than or equal to −1 or greater than or equal to 1. The data points are colored in orange if they are significantly downregulated, in magenta if they are significantly unregulated or in gray if they are not significnatly differentially regulated. Datapoints are slightly transparent to help visualize overlapping datapoints. (C) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin treatment in the root and the shoot of wild-type seedlings. Gene sets are broken down into whether auxin treatment resulted in upregulation of the gene (left) or downregulation of the gene (right). (D) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to ABA treatment in the root and the shoot of wild-type seedlings. Gene sets are broken down into whether ABA treatment resulted in upregulation of the gene (left) or downregulation of the gene (right). (E) NCED family transcript accumulation differences in auxin-treated wild-type root and shoot tissues. Numbers are shown as auxin-induced fold change in transcript accumulation in comparison to mock treatment. ND, not detected in our RNA-seq data; NS, not significant (|log2 fold change|≤ 1.0; FDR≥0.05).
We found that auxin regulates largely non-overlapping sets of transcripts in dark-grown root and shoot tissues despite over 63% of basal transcripts being shared between these two tissues (Fig. 5C; Table S1), consistent with a model in which auxin has distinct effects in these two tissues. Gene ontology (GO) analysis (Mi et al., 2013) revealed that auxin-regulated root-specific differentially expressed genes (DEGs) included those with annotated roles in cell wall loosening, cell growth, lateral root formation, root system development and gravitropism (Table S2). In contrast, auxin-regulated shoot-specific DEGs included genes with roles in shade avoidance, positive regulation of growth, meristem maintenance and brassinosteroid response. Shared auxin-induced DEGs included genes involved in regulation of hormone levels, auxin homeostasis, response to auxin, auxin transport, regulation of transcription, auxin-activated signaling pathways and hormone metabolic processes (Table S2).
In contrast, ABA-induced DEGs were largely overlapping in root and shoot tissues (Fig. 5D). Root-specific ABA-responsive DEGs included those with roles in response to salicylic acid, jasmonic acid signaling, amino acid export, cell wall organization, response to auxin, cellular response to stress and glutamine-family amino acid catabolic processes. Shoot-specific ABA-responsive DEGs included genes with annotated functions in response to salt stress, response to oxidative stress, defense response to fungus and negative regulation of response to stimulus. Finally, DEGs shared between the root and the shoot in response to ABA treatment had roles including immune response, response to abscisic acid, cellular amide metabolic process and response to stress (Table S2).
Full auxin responses in hypocotyl-based growth assays require intact ABA biosynthesis (Fig. 2B). The 9-CIS-EPOXYCAROTENOID DIOXYGENASE (NCED) family regulates the first committed step of ABA biosynthesis and is often transcriptionally regulated to affect ABA levels (Frey et al., 2012). Examination of NCED transcript accumulation revealed that only NCED5 transcripts were upregulated by auxin treatment in shoot, but not root, tissues (Fig. 5E), consistent with our finding that auxin treatment results in increased endogenous ABA levels in shoot tissue (Fig. 2C).
We found little overlap between the transcriptional effects of auxin and ABA in the shoot (Fig. 6A) or in the root (Fig. 6B). Indeed, ABA and auxin treatment had opposing effects on the levels of many transcripts (Fig. 6C,D). Because our phenotypic data, combined with the observed effect of auxin treatment on ABA levels, suggest that ABA modulates auxin effects on hypocotyl elongation, we examined the shared targets more closely. Downregulated gene targets shared by auxin and ABA in dark-grown shoots included multiple genes in the peroxidase superfamily; various transporters, including a sulfate transporter, a dipeptide transporter and a known cytokinin transporter; as well as genes involved in biotic or abiotic stress responses. These downregulated targets have not previously been implicated in hypocotyl elongation.
Auxin and ABA impact distinct sets of genes in the root and the shoot. (A) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin (IAA) or ABA treatment in shoot tissues. Gene sets are broken down into whether auxin or ABA treatment resulted in upregulation of the genes in the shoot (left) or downregulation of the genes in the shoot (right). (B) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin or ABA treatment in root tissues. Gene sets are broken down into whether auxin or ABA treatment resulted in upregulation of the genes in the root (left) or downregulation of the genes in the root (right). (C) Hierarchical clustering of genes displaying differential expression among hormone-treated shoot samples (FDR≤0.05). (D) Hierarchical clustering of genes displaying differential expression among hormone-treated root samples (FDR≤0.05). Z-score was used for the self-organizing heat map. Scale in C and D shows log2 fold change.
Auxin and ABA impact distinct sets of genes in the root and the shoot. (A) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin (IAA) or ABA treatment in shoot tissues. Gene sets are broken down into whether auxin or ABA treatment resulted in upregulation of the genes in the shoot (left) or downregulation of the genes in the shoot (right). (B) Venn diagrams showing the overlap between the sets of genes differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin or ABA treatment in root tissues. Gene sets are broken down into whether auxin or ABA treatment resulted in upregulation of the genes in the root (left) or downregulation of the genes in the root (right). (C) Hierarchical clustering of genes displaying differential expression among hormone-treated shoot samples (FDR≤0.05). (D) Hierarchical clustering of genes displaying differential expression among hormone-treated root samples (FDR≤0.05). Z-score was used for the self-organizing heat map. Scale in C and D shows log2 fold change.
In contrast, multiple upregulated shared auxin and ABA targets identified in dark-grown hypocotyls have previously annotated roles in regulating hypocotyl elongation. Examples that are thought to repress hypocotyl elongation include OVATE FAMILY PROTEIN 1 (AT5G01840, OFP1) (Zhang et al., 2020), SHI-RELATED SEQUENCE 5 (AT1G75520, SRS5) (Yuan et al., 2018) and IAA METHYLTRANSFERASE 1 (AT5G55250, IAMT1) (Qin et al., 2005), whereas SHORT HYPOCOTYL 1 (AT1G52830, IAA6) (Kim et al., 1996) and CYCLING DOF FACTOR 5 (AT1G69570, CDF5) (Martin et al., 2020) are thought to promote hypocotyl elongation. These shared targets might be mediators of the shared effects of short-term treatment with ABA and auxin on dark-grown hypocotyl elongation.
An aba2 mutant reveals ABA-dependent auxin-regulated gene transcription
Our physiological assays suggested that ABA is required for full auxin effects on hypocotyl elongation, but our transcript analysis revealed minimal overlap in the DEGs for these two hormones; we therefore sought to understand the effects of loss of ABA biosynthesis on auxin-regulated transcript accumulation.
Because aba2 mutants are defective in ABA biosynthesis (Laby et al., 2000), we reasoned that comparing the auxin-induced transcriptional response of aba2 (specifically aba2-3) to the wild-type transcriptional response in the hypocotyl would allow us to identify the effects of loss of endogenous ABA on transcriptional changes elicited by auxin application. We found striking differences in auxin-regulated transcripts between wild-type and aba2 dark-grown shoot tissues (Fig. 7A,B). Transcripts downregulated by auxin were more strongly impacted by loss of endogenous ABA than those upregulated by auxin (Fig. 7B). Of those transcripts with decreased accumulation in response to auxin treatment of dark-grown wild-type shoots, 71% failed to display auxin-responsive decreases in dark-grown aba2 shoots. Conversely, only 22% of transcripts that displayed increased accumulation in response to auxin treatment in wild-type shoots had a disrupted auxin response in aba2 shoots. Thus, ABA biosynthesis largely affects downregulated auxin targets in the hypocotyl.
Disruption of ABA biosynthesis impacts the transcriptional response to auxin in the shoot. (A) Volcano plots displaying the transcript changes in response to auxin (IAA) treatment compared to mock treatment for wild-type (Col-0; left) or aba2-3 (right) shoots. Dashed lines indicate the significance cutoff values abs(LFC)>1, −log10 P-value>1. Transcripts below these cutoffs are in gray, transcripts displaying LFC>1 are in purple and transcripts displaying LFC<1 are in yellow. (B) Venn diagrams showing overlap between the sets of genes that are differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin treatment in the shoots of aba2-3 and wild-type (Wt) seedlings.
Disruption of ABA biosynthesis impacts the transcriptional response to auxin in the shoot. (A) Volcano plots displaying the transcript changes in response to auxin (IAA) treatment compared to mock treatment for wild-type (Col-0; left) or aba2-3 (right) shoots. Dashed lines indicate the significance cutoff values abs(LFC)>1, −log10 P-value>1. Transcripts below these cutoffs are in gray, transcripts displaying LFC>1 are in purple and transcripts displaying LFC<1 are in yellow. (B) Venn diagrams showing overlap between the sets of genes that are differentially expressed (|log2 fold change|≥1.0; FDR≤0.05) in response to auxin treatment in the shoots of aba2-3 and wild-type (Wt) seedlings.
We found 110 transcripts that displayed auxin-regulated differential accumulation specifically in aba2, but not wild-type, dark-grown shoot tissue (Fig. 7B). Thus, in the absence of intact ABA biosynthesis, auxin regulates accumulation of a distinct set of transcripts, which might reflect differences in the physiology of ABA-replete and ABA-deficient seedlings, even though we did not note any gross morphological differences between wild-type and aba2 seedlings. Auxin-induced DEGs that were specific to aba2 (Table S2) have annotated functions in cell wall modification, regulation of growth, cell growth, response to hormones and regulation of biosynthetic processes.
We identified 96 genes that displayed altered accumulation in response to auxin in wild-type shoots but not in aba2 shoots (Fig. 7B). We also identified 110 genes that displayed altered accumulation in response to auxin in aba2 shoots but not in wild-type shoots (Fig. 7B). Considering the complexity of regulatory mechanisms governing hypocotyl elongation, which exhibits sensitivity to numerous environmental cues, we were unsurprised to find the absence of a singular obvious change in gene expression between aba2 shoots and wild-type shoots that neatly explains the aba2 partial resistance to auxin. Nonetheless, in examining the literature associated with the genes that either depended on ABA biosynthesis to respond to auxin or were only auxin responsive when ABA biosynthesis was disrupted, we found numerous examples of genes with annotated functions associated with regulation of hypocotyl elongation (Table 1).
Genes implicated in hypocotyl elongation that are auxin-responsive in only wild-type shoots or aba2-3 shoots

In addition to examining ABA2-dependent genes with annotated roles in hypocotyl elongation, we examined whether genes associated with additional plant hormones might be differentially expressed in auxin-treated hypocotyls in an ABA-dependent manner. Unsurprisingly, we identified genes associated with brassinosteroids, cytokinins, gibberellins, jasmonates, ethylene and salicylic acid that specifically displayed altered accumulation in response to auxin in wild-type shoots or in aba2 shoots (Table 2). This observation suggests that the orchestration of plant hormone interactions governing hypocotyl elongation likely incorporates many hormones, underscoring the significant complexity in the regulation of this relatively simple growth process. The mechanisms by which these additional plant hormones interface with auxin and ABA to regulate hypocotyl elongation will be an exciting area of investigation in the future.
Phytohormone-related genes with auxin-responsive transcript accumulation specific to either wild-type shoots or aba2-3 shoots

Despite these differences, most auxin-upregulated transcripts were unaltered in the aba2 mutant (Fig. 7B). This finding is consistent with our phenotypic data demonstrating that whereas ABA biosynthesis is necessary for full auxin responsiveness in hypocotyls, ABA biosynthesis mutants are only mildly resistant to the effects of auxin in hypocotyl elongation assays (Figs 1–3). Therefore, our phenotypic data and our RNA-seq data support a model whereby some aspects of auxin regulation of hypocotyl elongation are affected by ABA, but clearly auxin regulation of hypocotyl elongation is not fully dependent on ABA.
DISCUSSION
Hormone interactions in plants are complex. Because hypocotyl elongation is driven solely by cell expansion, dissecting hormone signaling interactions in this tissue is simpler than understanding hormone interactions in developmental events requiring a combination of cell division, cell fate changes and cell expansion.
Our phenotypic analysis (Fig. 2B) and our finding that endogenous levels of ABA in the shoot increase in response to auxin treatment (Fig. 2C) support a model in which aspects of ABA biosynthesis and signaling are necessary for auxin to exert its full regulatory capacity on hypocotyl elongation. This auxin-ABA interaction is converse to those previously discovered, in which auxin signaling is required for full response to ABA. Meanwhile in the root, diverse ABA biosynthesis mutants display wild-type sensitivity to auxin (Fig. 4), highlighting the tissue specificity of hormone interactions. From a developmental perspective, the downstream influence of ABA on auxin responses in the regulation of hypocotyl elongation might be a mechanism that confers to a seedling the ability to rapidly respond to external stressors during the tenuous stage of seedling establishment. During initial stages of dark-grown seedling growth, seedlings prioritize rapid hypocotyl elongation to emerge from the soil and to reach sunlight. ABA action downstream of auxin likely ostensibly provides a check upon this growth, which if unabated could lead to a seedling increasing its exposure to an external stressor such as drought. Hence, stress signals through ABA could be used to curtail growth, in a similar manner as we observe when treating with supraoptimal levels of exogenous auxin, to pause growth until permissive conditions exist for seedling establishment.
By comparing auxin-regulated genes in wild-type and aba2 seedling shoots, we found striking differences in the genes regulated by auxin when ABA biosynthesis was disrupted (Fig. 7B). This shows that, although the total number of genes that are auxin responsive in the shoot of wild-type and aba2 seedlings is largely equivalent (Fig. 7B), the set of genes that are auxin responsive is very different in aba2 seedlings. Interestingly, we found less overlap between genes downregulated by auxin in aba2 and wild-type shoots than we did between genes that were upregulated by auxin (Fig. 7B). In light of the complex transcriptional interactions between auxin and ABA, and the lack of an obvious ‘core’ set of genes associated with hypocotyl elongation in our data, it is possible that the additive impact of many genes may explain the auxin resistance of ABA biosynthesis mutants in hypocotyl elongation.
In contrast to our understanding of auxin-regulated activation of gene transcription, auxin-mediated repression is not well understood. Our data provide some clues to how auxin might induce the repression, specifically that in the case of dark-grown hypocotyls, ABA biosynthesis is required for repression of a subset of targets. This finding raises the possibility that, more generally, auxin-induced downregulation of genes in other tissues might be through secondary pathways, such as ABA signaling.
Although our phenotypic analyses and RNA-seq experiments support a model whereby aspects of auxin-mediated regulation of hypocotyl elongation require intact ABA biosynthesis, we acknowledge that the reality of the interactions between auxin and ABA in the regulation of this process are likely far more complex. Indeed, examining the genes that we identified as being dependent on ABA biosynthesis to respond to auxin in the shoot, we identified many genes known to be involved in aspects of other phytohormone responses, including genes involved in their biosynthesis, catabolism, modification and signaling (2). Within this small subset, we found genes associated with gibberellin, brassinosteroid, salicylic acid, jasmonate, cytokinin and ethylene phytohormones. These findings reflect that the regulation of something as simple as hypocotyl elongation is extraordinarily complex and unlikely to be teased apart from a single study. Furthermore, recent research on the importance of the asymmetric distribution of auxin across the hypocotyl to regulate elongation (Du et al., 2022) highlights an additional interesting question for future studies – is there a similar phenomenon with respect to spatial regulation of ABA concentration to regulate hypocotyl elongation? Nonetheless, our finding that auxin relies on intact ABA biosynthesis to exert its full regulatory capacity on hypocotyl elongation provides one more piece of information that brings us closer to disentangling the immense complexities in the regulation of this process by phytohormones.
MATERIALS AND METHODS
Plant growth
All Arabidopsis lines, with the exception of abi1-1, were in the Col-0 background, which was used as the wild type. The abi1-1 line is in the Ler background, and Ler was used as the wild type for assays using the abi1-1 line. Before plating, seeds were surface sterilized as described previously (Last and Fink, 1988) and then resuspended in 0.1% agar. The seeds were then stratified for 2 days at 4°C. After stratification, the seeds were plated on plant nutrient (PN) medium (Haughn and Somerville, 1986) containing 0.6% agar and supplemented with 0.5% sucrose (PNS). Seedlings were grown in continuous light at 22°C.
All seeds were obtained from the Arabidopsis Biological Resource Center: aba2-1 (Rook et al., 2021); and aba2-3 and aba2-4 (Laby et al., 2000).
Phenotypic assays
For dark-grown hypocotyl elongation assays, stratified seeds were plated on PNS supplemented with hormones or the equivalent amount of ethanol for mock treatment and then exposed to continuous light for 24 h filtered through a thin yellow long-pass plexiglass filter (Piedmont Plastics) to help reduce breakdown of the added hormones (Stasinopoulos and Hangarter, 1990). Next, the plates were wrapped in three layers of aluminium foil to ensure complete blocking of light. The seedlings were then allowed to grow for 5 days before hypocotyls were measured.
For light-grown hypocotyl elongation assays, stratified seeds were plated on PNS supplemented with picloram or the equivalent amount of DMSO for mock treatment. The seedlings then grew for 6 days under continuous light filtered through a thin yellow long-pass plexiglass filter to help reduce breakdown of the added hormones (Stasinopoulos and Hangarter, 1990). The seedlings were then imaged, and the hypocotyls were measured using ImageJ software (National Institutes of Health, Bethesda, MD).
For root elongation assays, stratified seeds were plated on PNS supplemented with hormones or the equivalent amount of ethanol for mock treatment. The seedlings were grown under continuous light filtered through a thin yellow long-pass plexiglass filter to help reduce breakdown of added hormones (Stasinopoulos and Hangarter, 1990). After 8 days of growth, roots were measured.
For statistical analysis of data acquired for the various phenotypic assays, first all groups were compared using a one-way ANOVA test to test for significant differences between the groups. In addition, pairwise comparisons between groups using the Tukey HSD test were carried out to identify statistically significant differences between specific groups. A P-value of less than 0.01 was used as the cutoff for statistical significance.
EMS mutagenesis and HR mutant isolation
Ethyl methanesulfonate (EMS) Col-0 seeds were mutagenized as described previously (Normanly et al., 1997). The M2-generation seeds were then surface sterilized as described previously (Last and Fink, 1988), stratified for 2 days and then plated on PNS containing 30 µM IBA. The seeds were then incubated for 24 h under continuous light filtered through a thin yellow long-pass plexiglass filter to help reduce breakdown of the added hormones (Stasinopoulos and Hangarter, 1990). The plates were then wrapped three times in aluminium foil and incubated in darkness for 4 days. At this time, the seedlings were screened for those exhibiting resistance to the inhibitory effects of auxin on dark-grown hypocotyl elongation. Resistant seedlings were transferred to PNS plates using sterile technique. Once large enough, the seedlings were transferred to soil. The M3 progeny from the auxin-resistant isolates were then re-tested for auxin resistance.
Whole-genome sequencing
EMS-induced mutations were identified using a whole-genome sequencing of bulk segregants approach as described previously (Thole and Strader, 2015). Briefly, seedlings isolated from the hypocotyl resistance screen were backcrossed to Col-0. The F2 generation from this cross was then used to re-isolate seedlings exhibiting resistance to auxin in dark-grown hypocotyl elongation assays. The resultant F3 seed from the re-isolated seedlings was then grown, and DNA isolated from the tissue of six F3 lines was pooled together and used for whole-genome sequencing. The sequencing data was then analyzed to identify EMS-generated mutations.
Vector construction and Arabidopsis transformation
AAO3 was amplified from Arabidopsis thaliana (Col-0) genomic DNA using Pfx platinum (Life Technologies) polymerase and the primers 5′-caccATGGATTTGGAGTTTGCAGTTAATGG-3′ and 5′-GTTGCTTACTTGCTTTGCCTTTATTGTC-3′ (lowercase characters indicate the noncoding initial sequence used for dTOPO cloning). The generated PCR product was then cloned into pENTR/D-TOPO (Life Technologies) resulting in pENTR-AAO3. The pENTR-AAO3 vector was sequenced to confirm accurate cloning of AAO3. The sequenced pENTR-AAO3 vector was then recombined into the plant expression vectors pEG100 and pEG201 (Earley et al., 2006), both of which were confirmed by sequencing. pEG100 is a Gateway destination vector that will drive the insert (i.e. AAO3) behind the 35S promoter. pEG201 is a Gateway destination vector that results in an HA-AAO3 fusion driven behind the 35S promoter. The pEG100 and pEG201 vectors were then transformed into Agrobacterium tumefaciens strain GV3101 (Koncz and Schell, 1986). The A. tumefaciens strains carrying the pEG100 and pEG201 vectors were then used to transform Col-0 and HR12 plants via the floral dip method (Clough and Bent, 1998). From the dipped plants, T1 seed was harvested, surface sterilized and then plated on PN medium supplemented with 10 µg/ml BASTA (Haughn and Somerville, 1986). Subsequent generations were tested to identify lines homozygous for the transgene, which were then used for downstream experiments.
ABA quantification
Procedures for extraction, purification and measurement of ABA were as described previously (Yan et al., 2016). Briefly, freeze-dried samples were homogenized, and d6-ABA (Icon Isotopes) was added as an internal standard, followed by extraction using methanol containing 1% (v/v) acetic acid. Extracts were purified with the Oasis cartridge columns (Waters) HLB, MCX and WAX. Purified samples were subjected to liquid chromatography (LC)-electrospray ionisation (ESI)-tandem mass spectrometry (MS/MS) using the Agilent 6410 TripleQuad LC/MS system. LC and MS/MS settings were described previously (Yan et al., 2016).
RNA isolation and RNA-seq
For the RNA-seq experiment, Col-0 (wild-type) and aba2-3 seeds were surface sterilized (Last and Fink, 1988), resuspended in agar and then incubated for 2 days at 4°C. The seeds were then plated on PNS plates and incubated for 24 h under continuous light. The plates were then wrapped in three sheets of aluminium foil and incubated for 2 days. Next, the dark-grown plants were transferred to liquid PNS containing either 10 µM IAA, 10 µM ABA or the equivalent amount of ethanol for the mock treatment in a dark room with only a green safe light for illumination. Each treatment was repeated in triplicate for both lines. After incubation for 2 h, the seedlings were removed from the liquid PNS, cut and separated at the root-hypocotyl junction, and the isolated tissue was then immediately frozen in liquid nitrogen. Total RNA was isolated from each tissue for each respective treatment using the RNeasy Plant Mini Kit from Qiagen according to the manufacturer's instructions. Samples were then prepared for sequencing using the SMARTer cDNA synthesis kit (Clontech) according to the manufacturer's instructions. The samples were sequenced across three (1×50 bp) lanes using an Illumina HiSeq 3000. The reads were then demultiplexed and aligned to the TAIR 10 Ensemble Release 23 assembly using STAR (Dobin et al., 2013). Gene counts were derived from the number of uniquely aligned unambiguous reads by Subread:featureCount.
Gene counts were normalized using the R/Bioconductor package EdgeR (Robinson et al., 2010), and trimmed mean of M-values (TMM) normalization size factors were calculated to adjust values due to differences in library size. Ribosomal genes and genes not expressed in any sample greater than 2 counts per million were not considered for further analysis. The calculated TMM size factors and the matrix of counts were then imported into the R/Bioconductor Limma package (Ritchie et al., 2015), and these values were used to make a Spearman correlation matrix (Fig. S2A) and multi-dimensional scaling plot (Fig. S2B). The weighted likelihoods based on observed mean-variance relationship of every gene for every sample were calculated using the voomWithQualityWeights function. The residual standard deviation of every gene to the gene's average log count was plotted to assess gene performance, and these values were used to fit a trendline of the residuals (Fig. S2C). A generalized linear model was then created to test for gene level differential expression, and results for genes with a Benjamini–Hochberg false discovery rate (FDR)-adjusted P-value of greater than or equal to 0.05 were filtered out. For all RNA-seq analyses, we used FDR-adjusted P-values.
GO analysis was carried out using panther.db (Mi et al., 2013). Genes were identified by searching for statistically overrepresented genes within the different lists of DEGs for the various comparisons. The specific analysis used was the statistical overrepresentation test using the GO biological process complete annotation set.
Heat maps for hierarchical clustering were generated using Seaborn in Python (Waskom, 2021). Clustering used the scipy.cluster.hierarchy.linkage() function (Virtanen et al., 2020). The specific Seaborn function sns.clustermap() was used with row_cluster and col_cluster set to True, metric was ‘euclidean’ and linkage clustering method was ‘complete’. A custom matplotlib (Hunter, 2007) coloring scheme was used for the heatmaps.
Acknowledgements
We acknowledge Ayako Nambara for her technical assistance. We thank Nicholas Morffy, Hongwei Jing and Sunita Pathak for helpful comments. We thank the Genome Technology Access Center in the Department of Genetics at Washington University School of Medicine for help with RNA-seq analyses.
Footnotes
Author contributions
Conceptualization: R.J.E., L.C.S.; Methodology: R.J.E., E.N.; Investigation: R.J.E., J.C., I.Y., S.E.N., E.N., C.H.; Resources: H.S.R., E.N., L.C.S.; Writing - original draft: R.J.E.; Writing - review & editing: R.J.E., L.C.S.; Visualization: R.J.E.; Supervision: H.S.R., L.C.S.; Project administration: L.C.S.; Funding acquisition: L.C.S.
Funding
This research was supported by the William H. Danforth Plant Science Fellowship Program (to R.J.E), the National Science Foundation Center for Engineering Mechanobiology (CMMI-1548571 to L.C.S.), a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (RGPIN-2019-04244 to E.N.), the Plant Phenotyping and Imaging Research Centre (P2IRC) project established by Canada First Research Excellence Fund (to E.N.), the European Regional Development Fund ‘SINGING PLANT’ project (CZ.02.1.01/0.0/0.0/16_026/0008446 to S.E.N. and H.S.R.) and the National Institutes of Health (R35 GM136338 to L.C.S.). Deposited in PMC for release after 12 months.
Data availability
The RNA-seq data discussed in this publication have been deposited in the NCBI Gene Expression Omnibus and are accessible through GEO Series accession number GSE169302.
References
Competing interests
L.C.S. is on the Scientific Advisory Board of Prose Foods. All other authors declare no competing or financial interests.