To analyze the effects of Fe, P, or S deficiency on the induction of Fe-, P-, and S-related responses, we used WT Arabidopsis (Arabidopsis thaliana (L.) Heynh ecotype Columbia) plants. In addition, we used the Arabidopsis ctr1 mutant that shows constitutive upregulation of ET responses (Guo and Ecker, 2003; Huang et al., 2003) and the ein2-1 and ein3eil1 mutants that show insensitivity to ET (Shakeel et al., 2013; Wang et al., 2013; Dubois et al., 2018; Binder, 2020; all mutants were obtained from the European Arabidopsis Stock Center). Arabidopsis plants were grown under controlled conditions as previously described (Lucena et al., 2006, 2007). Briefly, seeds were germinated in black peat and, when appropriate, seedlings were transferred to individual containers (of 70 ml volume) with complete nutrient solution continuously aerated. The nutrient solution had the following composition: macronutrients: 2 mM Ca(NO₃)₂, 0.75 mM K₂SO₄, 0.65 mM MgSO₄, 0.5 mM KH₂PO₄, and micronutrients: 50 μM KCl, 10 μM H₃BO₃, 1 μM MnSO₄, 0.5 μM CuSO₄, 0.5 μM ZnSO4, 0.05 μM (NH₄)₆Mo₇O₂₄, and 20 μM Fe-EDDHA. Plants were grown in a growth chamber at 22°C day/20°C night temperatures, with relative humidity between 50 and 70%, and an 8-h photoperiod (to postpone flowering) at a photosynthetic irradiance of 300 μmol m⁻² s⁻¹ provided by fluorescent tubes (Sylvania Cool White VHO).

When plants were ~45 days old, they were directly transferred, without washing, from this complete nutrient solution to the different treatments. The treatments imposed were: control: complete nutrient solution with 40 μM Fe-EDDHA; –Fe: nutrient solution without Fe; –P: nutrient solution without P (0.5 mM KH₂PO₄ was not added to the nutrient solution; instead, 0.5 mM KCl was added); –S: nutrient solution without S (only 2 μM S from the micronutrients added; 0.75 mM K₂SO₄ and 0.65 mM MgSO₄ were not added to the nutrient solution; instead, 1.5 mM KCl and 0.65 mM Mg(NO₃)₂ were added). After 2, 4, 6, and 8 days of the treatments, root ferric reductase activity (FRA) and acid phosphatase activity (PA) were determined in plants from the four treatments, as previously described (Lucena et al., 2006, 2019; see also below). After FRA determination, roots were collected and kept at −80°C for subsequent analysis of mRNA levels. Each experiment was repeated at least twice, and representative results are presented.

It was determined as previously described (Zakhleniuk et al., 2001). Briefly, roots of intact plants were placed in Petri dishes containing a solution with 5-bromo-4-chloro-3′-indolyphosphate p-toluidine salt (BCIP) 0.01% (w/v) for 4 h. Blue color of roots is higher with increased PA. After 4 h, photographs of roots were taken with a stereoscopic microscope.

It was determined as previously described (Lucena et al., 2006, 2007). Briefly, intact plants were placed in a Fe(III) reduction assay solution containing Fe(III)-EDTA and ferrozine for 60 min. The FRA was determined spectrophotometrically by measuring the absorbance (562 nm) of the Fe(II)-ferrozine complex and using an extinction coefficient of 29,800 M⁻¹ cm⁻¹. After the reduction assay, roots were excised and weighed, and the results were expressed on a root fresh weight basis. Data are given as means of six replicates.

Roots were ground to a fine powder with a mortar and pestle in liquid nitrogen. Total RNA was extracted using the Tri Reagent solution (Molecular Research Center, Inc. Cincinnati, OH, USA) according to the manufacturer's instructions. M-MLV reverse transcriptase (Promega, Madison, WI, USA) was used to generate cDNA from 3 μg of DNase-treated root RNA as the template and random hexamers as the primers.

The study of gene expression by qRT-PCR was performed by using a qRT-PCR Bio-Rad CFX connect thermal cycler and the following amplification profile: initial denaturation and polymerase activation (95°C for 3 min), amplification, and quantification repeated 40 times (90°C for 10 s, 57°C for 15 s, and 72°C for 30 s), and a final melting curve stage of 65 to 95°C with increment of 0.5°C for 5 s, to ensure the absence of primer dimer or nonspecific amplification products. PCR reactions were set up in 20 μl of SYBR Green Bio-RAD PCR Master Mix, following the manufacturer's instructions. Controls containing water instead of cDNA were included to check for contamination in the reaction components. Gene-specific primers used are listed in Table 1. Oligonucleotides in García et al. (2018) were used to amplify FRO2, IRT1, and FIT cDNA, and those in Lucena et al. (2019) and Maruyama-Nakashita et al. (2006) to amplify PAP17 and SULTR1;1, respectively. Oligonucleotides used to amplify PHT1;5, PHR1 and SLIM1/EIL3 were designed by using the Primer-BLAST software from the NCBI site. Standard dilution curves were performed for each primer pair to confirm appropriate efficiency of amplification (E = 100 ± 10%). Constitutively expressed SAND1 and YLS8 genes, which do not respond to changes in the Fe conditions (Han et al., 2013), were used as reference genes to normalize qRT-PCR results. The relative expression levels were calculated from the threshold cycles (Ct) values and the primer efficiencies by the Pfaffl method (Pfaffl, 2001). Each PCR analysis was conducted on three biological replicates and each PCR reaction repeated twice.

All experiments were repeated at least twice and representative results are presented. The values of qRT-PCR represent the mean of three independent biological replicates. The values of FRA represent the mean of six replicates. Within each day and genotype, ^*P < 0.05 or ^**P < 0.01 indicate significant differences in relation to the control treatment using one-way analysis of variance (ANOVA) followed by a Dunnett's test.

In WT Columbia plants, as expected, FRO2, IRT1, and FIT expression, and ferric reductase activity (FRA; associated with FRO2), were greatly induced after 2 days of Fe deficiency. Later on, the induction decayed and tended to recover some days after the deficiency (Figures 1–3). In relation to the other deficiencies, most of the Fe-related genes (FRO2, IRT1, and FIT) were also induced in WT Columbia plants under both P and S deficiency but to a much lower extent than under the Fe deficiency itself (Figures 1–3). However, FRA was not appreciably induced in WT Columbia plants under either P or S deficiency (Figure 1E).

In the ctr1 mutant, FRA was also greatly induced after 2 days of Fe deficiency while the Fe-related genes (FRO2, IRT1, and FIT) were also induced but after 4 days of the Fe deficiency and, in general, to lower levels than in the WT Columbia (Figures 1–3). In relation to the other deficiencies, neither FRA nor FIT were induced under either P or S deficiency while FRO2 and IRT1 were slightly induced under S deficiency (Figures 1–3).

In the ein2-1 mutant, FRO2, FRA, and IRT1 induction under Fe deficiency were delayed in relation to the WT Columbia and, in general, the maximum values achieved were lower than in the WT (Figures 1, 2). Surprisingly, despite FRO2 and IRT1 induction under Fe deficiency, FIT expression was not upregulated in this mutant at any time studied (Figures 1–3). In relation to the other deficiencies, FRA was not induced in this mutant under either P or S deficiency (Figure 1G). Similarly to the results obtained under Fe deficiency, neither FIT nor the Fe acquisition genes FRO2 and IRT1 were clearly upregulated in the ein2-1 mutant under either P or S deficiency (only FRO2 was slightly induced after 2 days of S deficiency; Figures 1–3).

In the ein3eil1 mutant, as occurred in the WT Columbia, FRO2, FRA, IRT1, and FIT were greatly induced after 2 days of Fe deficiency. It should be noted that the maximum values of FRO2, IRT1, and FIT expression were attained in this mutant, where some values were several times those obtained in the WT Columbia (Figures 1–3). In relation to the other deficiencies, neither FRA nor the Fe acquisition genes FRO2 and IRT1 were significantly induced under either P or S deficiency (Figures 1, 2). However, FIT was upregulated under both P and S deficiency in this mutant (Figure 3D).

Collectively, the results show that Fe deficiency responses can also be induced under P or S deficiency but to lower intensities and more transiently. The induction of the Fe deficiency responses under Fe deficiency itself or under P or S deficiency differs depending on the ET signaling mutants. However, FRA was only induced by Fe deficiency itself but neither by P deficiency nor by S deficiency in any of the genotypes studied.

In WT Columbia plants, PHT1;5, PAP17, and PHR1 expression, and phosphatase activity (PA; associated with PAP genes), were induced after 2–4 days of P deficiency (Figures 4A, 5, 9A). In relation to the other deficiencies, PHT1;5 was also induced under Fe deficiency (Figure 4A). PAP17 was also induced under both Fe and S deficiency while PA was induced by Fe deficiency but not by S deficiency (Figure 5). PHR1 was also upregulated by both Fe and S deficiency (Figure 9A).

In the ctr1 mutant, PHT1;5 expression under P deficiency achieved the highest values of its induction (Figure 4). PA, PAP17, and PHR1 were also induced in this mutant under P deficiency (Figures 6, 9). In relation to the other deficiencies, PHT1;5 was also induced under S deficiency (Figure 4B). PA was also induced under both Fe and S deficiencies while PAP17 was not (Figure 6). PHR1 was not induced by Fe deficiency, but it was slightly upregulated by S deficiency (Figure 9B).

In the ein2-1 mutant, PHT1;5 was induced under P deficiency (Figure 4C). In this mutant, PAP17 expression was induced under P deficiency but PA was not (Figure 7). PHR1 was also upregulated under P deficiency but after 6 days of P deficiency and to lower levels than in the other genotypes (Figure 9). In relation to the other deficiencies, neither PHT1;5 nor PHR1 were induced under either Fe or S deficiency in the ein2-1 mutant (Figures 4C, 9C). However, both PA and PAP17 were induced under Fe deficiency in the ein2-1 mutant (Figure 7).

In the ein3eil1 mutant, PHT1;5 expression was delayed, starting after 6 days of the P deficiency, attaining the lowest values within the different genotypes, while PHR1 expression reached its highest level of induction (Figures 4, 9). Both PA and PAP17 were also induced under P deficiency (Figure 8). In relation to the other deficiencies, PHT1;5 was induced after 8 days of S deficiency (Figure 4D). PA, PAP17, and PHR1 were induced under both Fe and S deficiencies (Figures 8, 9D).

Collectively, the results show that P deficiency responses can also be induced under Fe or S deficiency. In this case, some P deficiency responses attained similar or higher intensities under Fe deficiency, and in some cases under S deficiency, than under P deficiency itself: see, for example, PAP17 expression and PA under Fe deficiency (Figures 5, 8). The induction of the P deficiency responses under P deficiency itself or under Fe or S deficiency differs depending on the ET signaling mutants.

In WT Columbia plants, SULTR1;1 expression was upregulated after 2 days of S deficiency. By contrast, SLIM1/EIL3 was not upregulated under S deficiency at any time. In relation to the other deficiencies, both SULTR1;1 and SLIM1/EIL3 were upregulated under both P and Fe deficiency (Figures 10A, 11A).

In the ctr1 mutant, SULTR1;1 reached the highest values of its expression under S deficiency while SLIM1/EIL3 was not upregulated by S deficiency. In relation to the other deficiencies, neither SULTR1;1 nor SLIM1/EIL3 were upregulated by either Fe or P deficiency (Figures 10B, 11B)

In the ein2-1 mutant, SULTR1;1 was upregulated after 2 days of S deficiency while SLIM1/EIL3 was slightly upregulated but only after 8 days of the deficiency. In relation to the other deficiencies, SULTR1;1 was not upregulated by either Fe or P deficiency while SLIM1/EIL3 was upregulated under Fe deficiency (Figures 10C, 11C).

In the ein3eil1 mutant, SULTR1;1 was upregulated after 2 days of S deficiency while SLIM1/EIL3 was not upregulated under S deficiency at any time. In relation to the other deficiencies, SULTR1;1 was upregulated by both Fe and P deficiency while SLIM1/EIL3 was not (Figures 10D, 11D).

Collectively, the results show that SULTR1;1 expression (a S deficiency response) can also be upregulated under Fe or P deficiency in the WT Columbia and in the ein3eil1 mutant (Figure 10). In the case of SLIM1/EIL3, the results obtained in this work show its lack of induction upon S deficiency in the WT Columbia. However, it can be upregulated under S, Fe, or P deficiency depending on the ET-related genotype of the plants (Figure 11).

The ctr1 mutant displays the known “triple-response” morphology in the absence of exogenously added ET (Huang et al., 2003). In relation to CTR1, the results presented in this work (Figures 1–4, 6, 9–11) show that the ctr1 mutant does not present constitutive upregulation of any of the Fe-, P-, and S-related genes studied, all of them are upregulated upon the imposition of the deficiencies. In the same way, this mutant does not show constitutive activation of either FRA or PA (Figures 1F, 6B). These results agree with previous ones showing that PHT1;4 (PT2) is not constitutively upregulated in the hsp2 (ctr1) mutant (Lei et al., 2011) and that Fe-related genes, like FRO2, IRT1, and FIT, and FRA, are not constitutively induced in the ctr1 mutant (García et al., 2014). Taken together, all these results suggest that, for the activation of physiological responses by ET, a defective CTR1 is not enough and an internal decrease of some nutrient-specific repressive signals would be necessary. This decrease would not be required for some morphological responses, like the development of subapical root hairs, since the ctr1 mutant does present them even when grown in complete nutrient solution (Romera and Alcántara, 2004).

In relation to the crosstalk between the three deficiencies, CTR1 seems to play a role since Fe deficiency did not induce either the P-related genes PHT1;5, PAP17, and PHR1 or the S-related genes SULTR1;1 and SLIM1/EIL3 in the ctr1 mutant while it did in the WT Columbia (Figures 4–6, 9–11). The reasons for this are not clear and would need further research. In addition, PA was induced by Fe deficiency in the ctr1 mutant despite PAP17 was not (Figure 6). This result could be explained by the induction of other PAP genes, besides PAP17, encoding phosphatases (Sun et al., 2016).

The ein2-1 mutant has been described as insensitive in most responses to ET (Alonso et al., 1999; Shakeel et al., 2013; Wang et al., 2013; Dubois et al., 2018; Binder, 2020). In fact, EIN2 has been considered one of the central players in the linear canonical signaling pathway proposed for ET action (see Introduction; Wang et al., 2013; Dubois et al., 2018; Binder, 2020). In relation to each particular deficiency, FIT (encoding a key TF) was not induced under Fe deficiency (Figure 3C) while PHR1 (also encoding a key TF) was only slightly induced under P deficiency in the ein2-1 mutant (Figure 9C). Despite these results, the genes activated by both TFs, such as FRO2 and IRT1, and PAP17, were upregulated under Fe deficiency or P deficiency, respectively (Figures 1C, 2C, 7A). In spite of PAP17 induction, PA was not induced under P deficiency in this mutant (Figure 7B), which conforms to previous results showing the impairment of PA induction in the ein2-5 mutant under P deficiency (Lei et al., 2011). All these results could be partly explained by taking into account that both FIT and PHR1, and perhaps PAP17, can also be subjected to post-transcriptional and post-translational modifications (Lingam et al., 2011; Jung et al., 2018; Sega and Pacak, 2019; Wu and Ling, 2019). Another possibility could be the existence of additional FIT-and PHR1-independent pathways to control the expression of genes like FRO2, IRT1, and PAP17 (Figure 12).

In relation to the crosstalk between the three deficiencies, EIN2 probably plays a key role on it because the Fe-related gene IRT1, the P-related gene PHT1;5, and the S-related gene SULTR1;1 were not appreciably induced in the ein2-1 mutant under nonspecific deficiencies (Figures 2, 4, 10). In the case of Fe and P deficiencies, the relevant role of EIN2 in the crosstalk between them is further supported by the results showing its great influence on the upregulation of the genes encoding the key TFs FIT (Fe) and PHR1 (P). Neither FIT was induced under P deficiency nor PHR1 was induced under Fe deficiency in the ein2-1 mutant, as occurred in the WT Columbia (Figures 3, 9).

Besides its role in the crosstalk between Fe and P deficiencies through its effects on FIT and PHR1 expression (see above), EIN2 could also affect other genes not activated by these TFs, like PHT1;5, encoding an internal P transporter (Nagarajan et al., 2011). PHT1;5 is not activated by the PHR1 TF but by the WRKY75 TF (Nagarajan et al., 2011), which is also regulated by ET through EIN3/EIL1 (Figure 12; Guo et al., 2017). It seems that EIN2 plays an important role in the PHT1;5 upregulation under nonspecific deficiencies, since PHT1;5 was only induced under P deficiency in the ein2-1 mutant while in the WT Columbia was also induced under Fe deficiency (Figure 4).

Similar to the ein2-1 mutant, the ein3eil1 double mutant has also been considered insensitive in most responses to ET (Shakeel et al., 2013; Wang et al., 2013; Dubois et al., 2018; Binder, 2020), although it does not present complete insensitivity (Harkey et al., 2018). The results obtained in this work show that most genes considered in this study (except SLIM1) were upregulated in the ein3eil1 double mutant under their specific deficiencies, as occurred in the WT Columbia (Figures 1–5, 8–11). These results coincide with previous ones showing upregulation of FIT under Fe deficiency (Lingam et al., 2011), and of PAP17 (ACP5) under P deficiency (Liu et al., 2017), in the ein3eil1 mutant. Moreover, several genes, like FRO2, IRT1, FIT, PHR1, and SULTR1;1 attained their highest expression, and also the lowest one, in the ein3eil1 mutant (Figures 1–3, 9, 10). PA was also greatly induced in this mutant under P deficiency (Figure 8B). In the case of FIT and PHR1, their highest expression was somewhat surprising since EIN3/EIL1 have been involved in the activation of FIT (Yang et al., 2014) and PHR1 (Liu et al., 2017). A possible explanation for the above results could be the existence of additional EIN3/EIL1-independent pathways for the control of FIT and PHR1 expression, and/or for the Fe and P acquisition genes controlled by them (Figure 12). The idea of an EIN3/FIT-independent pathway has already been proposed by Balparda et al. (2020), showing that FRO2 and IRT1 expression, besides its control by FIT, could also be directly controlled by the ERF1 TF (also associated with ET; see “Introduction”). The existence of EIN3/EIL1-independent pathways for the control of PHR1, and consequently for P acquisition genes, is also possible because several transcriptomic analyses have shown altered expression of ERF genes, like ERF1, ERF2, and ERF5, in Pi-starved Arabidopsis roots (Song and Liu, 2015 and references therein). The possibility exists that the EIN3/EIL1-independent pathways could be potentiated when the EIN3/EIL1-dependent pathway is impaired. This would explain why some nutrient-deficiency responses are greatly induced in the ein3eil1 mutant (see above).

In relation to the crosstalk between the three deficiencies, the EIN3/EIL1 mutation, by contrast to the EIN2 mutation, did not impair the upregulation of most of the genes under nonspecific deficiencies (Figures 1–4, 8–11). This again suggests the existence of EIN3/EIL1-independent pathways for the control of Fe- and P-related genes (Figure 12). In the case of SULTR1;1, its higher upregulation under S deficiency, and also under P deficiency, in the ein3eil1 mutant (Figure 10) could be partly explained by considering that EIN3 can negatively interact with SLIM1/EIL3 for the upregulation of S acquisition genes (Wawrzynska and Sirko, 2016). The results agree with those of Wawrzynska and Sirko (2016) showing higher upregulation of SULTR1;1 in the ein3-1 mutant than in the WT Columbia.

EIN3/EIL1 could also play a role in the crosstalk between Fe and P deficiencies related to PHT1;5 expression. This gene, activated by the WRKY75 TF (Nagarajan et al., 2011), was not upregulated under Fe deficiency in the ein3eil1 mutant while it was in the WT Columbia (Figure 4). It should be noted that WRKY75 expression can be activated by ET through the EIN3/EIL1 TFs (Figure 12; Guo et al., 2017).

In conclusion, the results obtained in this work further support the existence of crosstalk between Fe, P, and S deficiency responses. In general, the responses are induced more intensively and less transiently under the specific deficiency. However, there are some exceptions, like PA induction under Fe deficiency, and SLIM1/EIL3 upregulation under P or Fe deficiency. The results also support a relevant role for ET, through its signaling components, in the regulation of the physiological responses to the three deficiencies and in the crosstalk between them. At first, the ET constitutive ctr1 mutant does not present constitutive activation of any of the responses while several responses are impaired in the ET insensitive ein2-1 mutant, either under specific or nonspecific deficiencies. This suggests an important role for EIN2 in the activation and crosstalk of several nutrient deficiency responses. However, in the ET-insensitive ein3eil1 mutant, several P and Fe deficiency responses attained their highest values of induction. These results are somewhat surprising since EIN3/EIL1 have been implicated in the activation of the key TFs FIT and PHR1 controlling Fe and P acquisition genes. At first, it would suggest the existence of additional EIN3/EIL1-independent pathways for the control of FIT and PHR1 expression.

The results presented in this work along with previous published results unravel the great complexity of ET signaling in the control of nutrient deficiency responses. This complexity becomes even greater if we consider that some ET- and nutrient-related TFs, like EIN3, SLIM1/EIL3, FIT, or PHR1, could promote ET biosynthesis in an autocatalytic manner, as occurs during the ripening of climacteric fruits (Lü et al., 2018). The above TFs have been implicated in the activation of ET synthesis genes, like MTK, SAM, ACS, and ACO (Nagarajan and Smith, 2012; Lucena et al., 2015; Song and Liu, 2015; Liu et al., 2019), and consequently could promote ET synthesis. This perhaps is necessary to keep a consistent ET production along the time of the deficiency.