Iron uptake, signaling, and sensing in plants.

Iron (Fe) is an essential micronutrient that affects the growth and development of plants because it participates as a cofactor in numerous physiological and biochemical reactions. As a transition metal, Fe is redox active. Fe often exists in soil in the form of insoluble ferric hydroxides that are not bioavailable to plants. Plants have developed sophisticated mechanisms to ensure an adequate supply of Fe in a fluctuating environment. Plants can sense Fe status and modulate the transcription of Fe uptake-associated genes, finally controlling Fe uptake from soil to root. There is a critical need to understand the molecular mechanisms by which plants maintain Fe homeostasis in response to Fe fluctuations. This review focuses on recent advances in elucidating the functions of Fe signaling components. Taking Arabidopsis thaliana and Oryza sativa as examples, this review begins by discussing the Fe acquisition systems that control Fe uptake from soil, the major components that regulate Fe uptake systems, and the perception of Fe status. Future explorations of Fe signal transduction will pave the way for understanding the regulatory mechanisms that underlie the maintenance of plant Fe homeostasis.

Introduction

Iron (Fe) is one of the nutrients that are indispensable for plant growth and development. Fe takes part in many chemical reactions, including photosynthesis and respiration, as it is redox active owing to its ability to easily accept and donate electrons. Fe deficiency usually leads to a reduction in plant viability. Plants mainly absorb Fe from the soil (Zelazny and Vert, 2014). Because of its high redox activity, Fe is readily oxidized in aerobic or alkaline soils. The oxidized Fe is insoluble and hence unavailable to plants. Approximately one-third of the world’s cultivated lands are calcareous (Mori, 1999), making Fe deficiency a limiting factor for crop yield and quality. However, excess Fe is toxic to cells, as Fe can cause the formation of highly reactive hydroxyl radicals through the Fenton reaction (Dixon and Stockwell, 2014). Plants have evolved sophisticated mechanisms for tightly controlling the influx of Fe to maintain Fe homeostasis. Thus, elucidating the mechanisms by which plants control Fe uptake and transport is essential for breeding of crops that are rich in Fe or tolerant to Fe-deficient soils.

This review discusses how plants integrate different components into the Fe deficiency response, with a focus on research in Arabidopsis and rice, and highlights the most recently characterized molecular components involved in Fe sensing, Fe signaling, and downstream Fe uptake processes. Fe translocation from roots to shoots through vascular tissues and Fe allocation from pools to sinks are not the aims of this review (for review, see Curie et al., 2009; Kobayashi and Nishizawa, 2012; and Jeong et al., 2017). There are many excellent reviews about the Fe-deficiency response that may also interest readers (Grillet and Schmidt, 2019; Gao and Dubos, 2021; Riaz and Guerinot, 2021; Vélez-Bermúdez and Schmidt, 2022).

Fe acquisition systems

Plants use two different strategies for Fe uptake to adapt to Fe-deficient environments, “strategy I” and “strategy II” (Römheld and Marschner, 1986).

Strategy I is a reduction-based mechanism that involves acidification of the rhizosphere by root-released protons, reduction of Fe3+ to Fe2+ by a membrane-localized reductase, and root uptake of Fe2+ by a membrane-localized high-affinity Fe transporter (Figure 1). In Arabidopsis thaliana, AHA2 (Arabidopsis H+-ATPase 2), AtFRO2 (ferric reduction oxidase 2) and AtIRT1 (iron-regulated transporter 1) are involved in these three steps (Robinson et al., 1999; Henriques et al., 2002; Varotto et al., 2002; Vert et al., 2002; Santi and Schmidt, 2009), and they interact with one another to form an iron-acquisition protein complex that optimizes Fe uptake at the cell surface of Arabidopsis root epidermal cells (Figure 1; Martín-Barranco et al., 2020). Notably, Fe3+ bio-availability decreases 1000-fold with an increase of one pH unit between 4 and 9 (Lindsay, 1997). The low efficiency of AtFRO2 for reduction of Fe3+ in alkaline soil results in the failure of strategy I. Coumarins were found to be crucial for the adaptation of Arabidopsis plants to alkaline soils (Rodríguez-Celma et al., 2013; Schmid et al., 2014; Siwinska et al., 2018; Tsai et al., 2018). To cope with inadequate soluble Fe in alkaline soil, Arabidopsis plants secrete Fe-mobilizing coumarins into the rhizosphere (Robe et al., 2021; Figure 1). In Arabidopsis, AtF6′H1 (feruloyl-coenzyme A [CoA] 6′-hydroxylase 1), a 2-oxoglutarate-dependent dioxygenase enzyme, is responsible for the conversion of feruloyl CoA to 6′-hydroxyferuloyl CoA, which is the common precursor for the synthesis of coumarins (Kai et al., 2006; Rodríguez-Celma et al., 2013; Schmid et al., 2014). The conversion of 6′-hydroxyferuloyl-CoA into scopoletin can occur spontaneously, but coumarin synthase (COSY) improves the reaction efficiency (Vanholme et al., 2019). With the addition of a catechol group, coumarins possess the ability to bind Fe (Rajniak et al., 2018). Fraxetin and sideretin are two main catechol coumarins that are biosynthesized by AtS8H (scopoletin 8-hydroxylase), a 2-oxoglutarate-dependent dioxygenase, and AtCYP82C4, a P450-dependent monooxygenase, respectively (Siwinska et al., 2018; Tsai et al., 2018). The expression of AtCYP82C4 is reduced at alkaline pH but increased at acidic pH under Fe-deficient conditions, explaining why alkaline pH and acidic pH favor the biosynthesis of fraxetin and sideretin, respectively (Figure 1; Tsai and Schmidt, 2020; Gautam et al., 2021). Fraxetin is a major type of coumarin that dissolves and binds Fe to form stable complexes at neutral to alkaline pH. It has been established that the ABCG transporter AtPDR9 (pleiotropic drug resistance 9) mediates the secretion of coumarins from root to rhizosphere (Fourcroy et al., 2014; Figure 1). It is hypothesized that Fe-mobilizing coumarins promote the reduction of Fe3+ to Fe2+, contributing to reduction-based Fe uptake (Rodríguez-Celma et al., 2013; Rajniak et al., 2018). Similar to coumarins, flavins also facilitate reduction-based Fe uptake (Rodríguez-Celma et al., 2013; Sisó-Terraza et al., 2016). Some species synthesize and secrete large quantities of flavins in response to Fe deficiency (Rodríguez-Celma et al., 2011). Moreover, the induction of coumarins and flavins under Fe-deficient conditions is mutually exclusive in Arabidopsis and Medicago truncatula (Rodríguez-Celma et al., 2013).

Figure 1

Fe acquisition systems in Arabidopsis and rice.

Arabidopsis plants absorb Fe from soil using strategy I, which involves a protein complex consisting of AHA2, FRO2, and IRT1. AHA2 secretes protons to acidify rhizosphere soil and mobilize Fe3+. After reduction by FRO2, Fe2+ enters roots with the assistance of IRT1. In Arabidopsis, coumarins also contribute to the reduction strategy, probably by converting Fe3+ to Fe2+. Sideretin is the major coumarin at acidic pH, whereas fraxetin is the major coumarin at neutral or alkaline pH. PDR9 mediates the efflux of coumarins from roots. Rice plants use 2 different strategies. In strategy I, PEZ2 may be responsible for the secretion of PCA/CA, which can reduce Fe3+ to Fe2+, and IRT1 directly transports Fe2+, which is abundant in paddy fields. In strategy II, DMA is excreted by TOM1 to chelate Fe3+, and Fe3+-DMA is transferred into roots by YSL15. Names in red indicate that their transcripts are inducible in response to Fe deficiency. COSY is not inducible in response to Fe deficiency. The dotted line indicates that it has not been validated experimentally.

Strategy II is a chelation-based mechanism, well known in graminaceous plants (Figure 1). This strategy involves release of the mugineic acid (MA) family of phytosiderophores from roots to solubilize and chelate Fe3+ in soil (Takagi et al., 1984). The synthesis of MA derives from methionine. Methionine is activated to SAM (S-adenosyl methionine) by SAM, and three SAM molecules are combined by NAS (nicotianamine synthase) to form NA (nicotianamine) (Mori and Nishizawa, 1987). NA is the direct precursor for MA synthesis and is converted into a 3″-keto intermediate by OsNAAT1 (NA amino transferase 1) (Cheng et al., 2007). The reduction of the keto intermediate by OsDMAS1 (deoxymugineic acid synthase 1) (Bashir et al., 2017) results in the production of deoxymugineic acid (DMA), which is excreted into the rhizosphere via OsTOM1 (transporter of mugineic acid 1) (Nozoye et al., 2011). DMA chelates Fe3+ to form the Fe3+-DMA complex, which is transported into roots by OsYSL15 (yellow stripe-like 15) (Inoue et al., 2009; Lee et al., 2009).

Although the genes that encode IRT1 and FRO2 exist in barley and maize, the Fe-deficiency-inducible Fe2+ uptake system is absent (Zaharieva and Römheld, 2000). The ferric-chelate reductase activity of rice roots is very low and not induced by Fe deficiency (Ishimaru et al., 2006). However, rice plants possess a partial strategy I (Figure 1). Rice as a specific graminaceous species grows better under anaerobic and Fe2+-rich conditions such as paddy fields. As a result of the pressures of surviving under submerged conditions with abundant Fe2+, rice plants have retained their ability to directly absorb Fe2+ by OsIRT1 (Ishimaru et al., 2006). Similar to Arabidopsis plants, which secrete various phenolic compounds, rice plants also secret protocatechuic acid (PCA) and caffeic acid (CA) into the rhizosphere under low Fe availability conditions. OsPEZ2 (phenolics efflux zero 2), a phenolic efflux transporter, is expressed in the plasma membrane of the root epidermis and exodermis and may account for the secretion of PCA and CA (Bashir et al., 2011). PCA and CA possess the chemical properties to chelate and reduce Fe3+ in vitro (Yoshino and Murakami, 1998), hence improving Fe uptake by strategy I. However, the expression of OsPEZ2 is not inducible in response to Fe deficiency (Kobayashi et al., 2014), implying that this effect is limited.

Regulation of Fe acquisition systems

Plants have evolved an intricate network for the regulation of their Fe acquisition systems. The basic helix-loop-helix (bHLH) transcription factor LeFER was the first Fe signaling regulator identified in plants. LeFER is crucial for Fe acquisition in tomato (Lycopersicon esculentum), as the fer mutant has no ability to activate strategy I under Fe-deficient conditions (Ling et al., 2002). Later, its Arabidopsis ortholog AtFIT (Fer-like Fe deficiency-induced transcription factor)/AtbHLH29 was found to positively regulate the expression of AtIRT1 and AtFRO2 (Colangelo and Guerinot, 2004; Jakoby et al., 2004; Yuan et al., 2005). Loss-of-function fit mutants display extreme Fe-deficiency symptoms and cannot survive in soil without Fe supplementation (Colangelo and Guerinot, 2004). Correspondingly, rhizosphere acidification ability, Fe reductase activity, and Fe content are considerably lower in the fit mutants. However, the overexpression of AtFIT alone is insufficient to constitutively stimulate the expression of AtIRT1 and AtFRO2 (Colangelo and Guerinot, 2004). It has been established that AtFIT interacts with each of the four members of the bHLH Ib clade (AtbHLH38, AtbHLH39, AtbHLH100, and AtbHLH101) to form heterodimers. The overexpression of both AtFIT and AtbHLH Ib is sufficient for constitutive activation of AtIRT1 and AtFRO2, irrespective of Fe status (Yuan et al., 2008; Wang et al., 2013). Recently, Cai et al. (2021) revealed that the quadruple mutant (bhlh4x) of bHLH Ib displays Fe-deficiency symptoms similar to those of fit. However, it is still unclear how AtFIT and AtbHLH Ib interdependently regulate the expression of AtIRT1 and AtFRO2. Comparative analysis of phenotypes and transcriptomes between fit and bhlh4x will provide insights into the molecular functions of AtFIT and AtbHLH Ib.

The synthesis of coumarins is also controlled by the AtFIT-AtbHLH Ib module, as induction of AtS8H and AtCYP82C4 is dependent on AtFIT (Schwarz and Bauer, 2020). The transcription factor AtMYB72 is crucial for Arabidopsis growth in alkaline soil (Palmer et al., 2013) and is a key regulator of coumarin synthesis (Stringlis et al., 2018; Robe et al., 2021). AtMYB72 tightly controls the expression of AtF6′H1, and the loss-of-function myb72 mutant can produce almost no scopoletin (Stringlis et al., 2018). By contrast, scopoletin can still accumulate in the fit mutant (Chutia et al., 2019), in line with the fact that expression of AtMYB72 is partially regulated by AtFIT (Colangelo and Guerinot, 2004; Zamioudis et al., 2015). AtMYB72 is a typical Fe-deficiency-inducible gene. Interestingly, the expression of AtMYB72 is also regulated by the AtFIT-AtbHLH Ib module because the transcript abundance of AtMYB72 is higher in AtFIT/AtbHLH Ib dual overexpression plants than in the wild type (Zamioudis et al., 2015). Notably, AtMYB72 is still induced in fit mutants in response to Fe deficiency, implying that other regulators mediate the upregulation of AtMYB72 in an AtFIT-independent manner. Recent studies reveal that AtbHLH121/AtURI (upstream regulator of IRT1) directly regulates AtbHLH Ib and AtMYB72 (Kim et al., 2019; Gao et al., 2020; Lei et al., 2020). Consistent with this finding, mutation of AtbHLH121 results in very low levels of two main catechol coumarins, fraxetin glycoside and sideretin glycoside (Gao et al., 2020). Unlike AtMYB72 and AtS8H, both of which are induced under high-pH conditions, AtCYP82C4 is not induced under these conditions (Gautam et al., 2021), implying that the expression of AtCYP82C4 is independent of AtMYB72. Taken together, AtFIT-bHLH Ib and AtMYB72 coordinate the synthesis of coumarins according to changes in Fe availability and environmental pH.

A regulatory module similar to AtFIT-bHLH Ib also exists in rice and governs its Fe acquisition systems. OsFIT/OsbHLH156 is an ortholog of AtFIT (Wang et al., 2020a; Liang et al., 2020), and the rice bHLH Ib member OsIRO2 (iron-related bHLH transcription factor 2/OsbHLH056) corresponds to the four AtbHLH Ib members (Supplemental Figure 1; Ogo et al., 2007). OsFIT and OsIRO2 interact with each other to form a functional transcription complex (Wang et al., 2020a; Liang et al., 2020). Owing to functional redundancy among the four AtbHLH Ib members, the Fe-deficiency symptoms of their double or triple mutants are not as severe as those of the fit mutants (Wang et al., 2013; Maurer et al., 2014), whereas their quadruple mutants phenocopy the fit mutants (Cai et al., 2021). In rice, OsIRO2 is the only member corresponding to AtbHLH Ib, and this conclusion is also supported by genetic evidence that the iro2 mutant is similar to the fit mutants in terms of Fe-deficiency symptoms and expression of Fe-deficiency-inducible genes (Liang et al., 2020). The strategy II genes (OsNAS1, OsNAS2, OsNAAT1, OsDMAS1, OsTOM1, and OsYSL15) are under the control of the OsFIT-OsIRO2 module because their induction is almost completely blocked in roots of iro2 and fit mutants under Fe-deficient conditions (Wang et al., 2020a; Liang et al., 2020). By contrast, the strategy I gene OsIRT1 is only weakly regulated by OsFIT-OsIRO2 (Wang et al., 2020a; Liang et al., 2020) or not regulated by OsIRO2 (Ogo et al., 2007). There are no counterparts of AtMYB72 and the coumarin synthesis-associated genes among the Fe-deficiency-inducible genes in rice, suggesting that the coumarin-based Fe acquisition strategy is not employed in rice.

In Arabidopsis, AtbHLH39 is mainly localized in the cytoplasm, but it accumulates in the nucleus in the presence of AtFIT (Trofimov et al., 2019). Similarly, the cytoplasmic fraction of OsIRO2 decreases and its nuclear fraction increases in the presence of OsFIT (Wang et al., 2020a; Liang et al., 2020). Although Arabidopsis and rice have evolved different Fe acquisition systems, they have retained the same regulatory mechanism to control them. Close homologs of FIT and bHLH Ib have also been identified in plant species beyond Arabidopsis and rice (Du et al., 2015; Li et al., 2018), suggesting that the Fe acquisition systems of plants are controlled by the conserved FIT/bHLH Ib module. Monocot plants (e.g., rice, maize) have only one bHLH Ib gene, whereas dicot plants (e.g., Arabidopsis, tomato, soybean, Populus) have more than three (Huang and Dai, 2015; Kurt and Filiz, 2018; Li et al., 2018), implying that the regulatory module evolved further after the monocot-dicot divergence.

Transcriptional regulation of the FIT-bHLH Ib module

Although AtbHLH Ib genes are positive regulators of Fe uptake genes, they themselves are also dramatically induced by Fe deficiency. It has been established that the upregulation of bHLH Ib genes is dependent on the AtbHLH IVc subgroup, which consists of AtbHLH34/IDT1 (iron deficiency tolerant 1), AtbHLH104, AtbHLH105/ILR3 (IAA-leucine-resistant 3), and AtbHLH115 (Supplemental Figure 1). A yeast one-hybrid screen with the AtbHLH101 promoter as a bait against an Arabidopsis cDNA library identified AtbHLH34 (Li et al., 2016). As a close homolog of bHLH34, AtbHLH104 also binds to the AtbHLH101 promoter in yeast one-hybrid assays. bhlh34 and bhlh104 mutants are sensitive to Fe-deficient conditions and have low levels of AtbHLH Ib transcripts (Li et al., 2016). Zhang et al. (2015) found that BHLH104 and ILR3 are required for upregulation of AtbHLH Ib genes and activation of the Fe-deficiency response. Similarly, bhlh115 loss-of-function mutants also develop Fe-deficiency symptoms similar to those of bhlh34, bhlh104, and ilr3-2 (Liang et al., 2017). Correspondingly, the overexpression of each of the AtbHLH IVc members enhances the expression of AtbHLH Ib genes and rescues the bhlh115-1 mutant (Zhang et al., 2015; Li et al., 2016; Liang et al., 2017). This suggests that these four members share similar molecular functions. Although the four members of the AtbHLH IVc subgroup have different expression patterns, their expression overlaps in some regions (Liang et al., 2017). Thus, AtbHLH IVc proteins have not only their own independent functions but also interdependent functions, a conclusion that is further supported by the fact that AtbHLH IVc members interact with one another to form homodimers or heterodimers, and their high order mutants develop more severe Fe-deficiency symptoms than their single mutants (Li et al., 2016; Liang et al., 2017). The expression of both AtbHLH Ib and AtFIT decreases in single and high-order mutants of AtbHLH IVc. However, unlike AtbHLH Ib members, which are the direct target genes of AtbHLH IVc, AtFIT is not a direct target gene of AtbHLH IVc, as its promoter is not associated with AtbHLH IVc (Zhang et al., 2015; Li et al., 2016; Liang et al., 2017; Tissot et al., 2019).

The AtbHLH IVb subgroup contains three members, AtbHLH121, AtbHLH11, and AtPYE (POPEYE/AtbHLH47) (Supplemental Figure 1; Heim et al., 2003). Three different groups have reported that AtbHLH121 interacts with AtbHLH IVc in the nucleus and is required for the induction of AtbHLH Ib and AtFIT under Fe-deficient conditions (Kim et al., 2019; Gao et al., 2020; Lei et al., 2020). All three groups established that bHLH121 associates with the promoters of AtbHLH Ib genes. It is still controversial whether AtbHLH121 directly regulates AtFIT. Lei et al. (2020) proposed that AtbHLH121 directly targets the AtFIT promoter, but the other two groups did not support this conclusion. It is noteworthy that AtFIT is expressed mainly in the root and is slightly induced by Fe deficiency. Lei et al. (2020) used roots in ChIP (chromatin immunoprecipitation) assays, whereas the other groups used whole seedlings, which may have produced different results. AtbHLH11 is the closest homolog of AtbHLH121, and its overexpression causes a decrease in AtFIT but an increase in bHLH Ib (Tanabe et al., 2019). By analyzing bhlh11 mutants and transgenic plants conditionally overexpressing AtbHLH11, Li et al. (2022) confirmed that AtbHLH11 negatively regulates the expression of AtbHLH Ib. They revealed that AtbHLH11 interacts with AtbHLH IVc members and inhibits the transactivation activity of the latter toward AtbHLH Ib genes by its two ethylene response factor-associated amphiphilic repression (EAR) motifs (LxLxL), which recruit the co-repressors topless/topless-related (AtTPL/AtTPRs). They also proposed that the severe Fe-deficient status caused by AtbHLH11 overexpression indirectly results in the upregulation of AtbHLH Ib genes. AtPYE is a negative regulator of Fe homeostasis. Although the pye-1 loss-of-function mutant displays severe Fe-deficiency symptoms during Fe limitation, expression of bHLH Ib genes (bHLH39 and bHLH101) is increased in pye-1 (Long et al., 2010). AtPYE interacts with AtbHLH105, and both of them negatively regulate several structural genes involved in the control of Fe homeostasis, such as AtFERs (Ferritins) and AtNAS4 (Tissot et al., 2019). AtPYE contains another type of EAR motif (DLNxxP) that was also predicted to interact with the co-repressors AtTPL/AtTPRs (Kagale et al., 2010; Causier et al., 2012). However, this protein interaction has not been validated experimentally. It is noteworthy that the loss of function of AtPYE results in upregulation of AtbHLH Ib genes but causes a reduction in both Fe chelate reductase activity and rhizosphere acidification ability (Long et al., 2010). Extensive exploration is required to clarify how AtPYE negatively regulates Fe acquisition systems.

Expression of AtbHLH Ib genes is significantly higher in fit mutants than in the wild type, suggesting that AtFIT is not required for upregulation of AtbHLH Ib genes (Wang et al., 2007). By contrast, expression of AtFIT is higher in AtbHLH39-overexpressing plants than in the wild type (Naranjo-Arcos et al., 2017), suggesting that AtbHLH Ib proteins positively regulate the expression of AtFIT. When AtFIT is absent, β-glucuronidase activity decreases in transgenic plants expressing an AtFIT promoter-driven β-glucuronidase, implying that AtFIT is crucial for its own upregulation under Fe-deficient conditions (Wang et al., 2007). It is possible that the expression of AtFIT is also controlled by the AtFIT-AtbHLH Ib complex (Figure 2).

Figure 2

Fe perception and the regulation of Fe acquisition systems.

Hexagon (Fe sensor); oval (protein); and rectangle (transcript). IVc (bHLH34, bHLH104, bHLH105, and bHLH115) and Ib (bHLH38, bHLH39, bHLH100, and bHLH101). The arrow indicates a positive effect, the blunt arrow indicates a negative effect, the black dotted line indicates an indirect regulation, and the red dotted line indicates a hypothesis. The letter U in the black circle indicates ubiquitin.

In rice, the expression of OsIRO2 is regulated by the OsbHLH IVc and OsbHLH IVb proteins. The OsbHLH IVc subgroup contains four members, positive regulator of iron homeostasis1 (OsPRI1)/OsbHLH060, OsPRI2/OsbHLH058, OsPRI3/OsbHLH059, and OsbHLH057 (Supplemental Figure 1). Loss of function or gene silencing of any of OsPRI1/2/3 causes downregulation of OsIRO2 and enhanced sensitivity to Fe deficiency (Zhang et al., 2017, 2020; Kobayashi et al., 2019). OsPRI1/2/3 can bind to and activate the promoter of OsIRO2 (Zhang et al., 2017, 2020). Despite its uncharacterized function in the Fe-deficiency response, OsbHLH057 may also regulate the expression of OsIRO2 based on its sequence similarity to the other members (Zhang et al., 2020). There are four members in the OsbHLH IVb subgroup: OsbHLH061, OsbHLH062, OsIRO3/OsbHLH063, and OsbHLH064 (Supplemental Figure 1). OsIRO3 is a negative regulator of the Fe-deficiency response. Overexpression of OsIRO3 causes reduced expression of Fe-deficiency-inducible genes, including OsIRO2, and severe Fe-deficiency symptoms (Zheng et al., 2010). By contrast, loss of function of OsIRO3 results in elevated expression of Fe-deficiency-inducible genes and necrotic leaf spots under Fe-deficient conditions (Wang et al., 2020b, 2020c; Carey-Fung et al., 2022). Similar to Arabidopsis AtbHLH11, OsbHLH061 is inducible in response to high Fe, and bhlh061 mutants are sensitive to excess Fe (Wang et al., 2022). Wang et al. (2022) revealed that OsbHLH061 also has two EAR motifs, one of which recruits the co-repressors OsTPL/OsTPRs, and OsbHLH061 interacts with and inhibits the transcription activation of OsPRI1 toward OsIRO2 (Figure 2). Thus, OsbHLH061 is also a negative regulator of Fe homeostasis. The transcript abundance of OsFIT increases under Fe-deficient conditions (Wang et al., 2020a; Liang et al., 2020). Liang et al. (2020) found that expression of OsFIT is reduced in the pri1-1, pri2-1, and pri3-1 mutants, suggesting that OsPRI1/2/3 positively regulate OsFIT (Figure 2). It is still unclear whether OsbHLH062 and OsbHLH064 are involved in the regulation of the Fe-deficiency response.

Interaction between bHLH IVb and bHLH IVc

Among the three AtbHLH IVb members, only AtbHLH11 is repressed by Fe deficiency (Tanabe et al., 2019; Li et al., 2022). However, the upstream regulators of AtbHLH11 have not yet been identified. AtbHLH121 is slightly responsive or unresponsive to Fe deficiency (Kim et al., 2019; Gao et al., 2020; Lei et al., 2020). AtPYE is a typical Fe-deficiency-inducible gene. AtbHLH IVc members have been reported to mediate the upregulation of AtPYE by directly binding to its promoter (Zhang et al., 2015; Liang et al., 2017; Tissot et al., 2019). AtPYE protein can directly bind to its own promoter, suggesting that AtPYE transcription is also under the control of its own protein (Tissot et al., 2019). Among the OsbHLH IVb members, only OsIRO3 has been reported to be induced by Fe deficiency (Zheng et al., 2010). Three of the OsbHLH IVc members, OsPRI1/2/3, have been confirmed to directly regulate the expression of OsIRO3 (Zhang et al., 2017, 2020). It is possible that the fourth member, OsbHLH057, also directly regulates OsIRO3 as do OsPRI1/2/3.

Unlike AtbHLH IVc proteins, which are predominantly localized in the nucleus, AtbHLH IVb proteins are preferentially localized in both the nucleus and cytoplasm. AtbHLH11 and AtbHLH121 physically interact with all four AtbHLH IVc proteins, and AtPYE also interacts with all except AtbHLH34 (Long et al., 2010; Kim et al., 2019; Gao et al., 2020; Lei et al., 2020; Li et al., 2022). Interestingly, the nuclear fractions of AtbHLH11 and AtbHLH121 proteins dramatically increase in the presence of their interaction partners, the AtbHLH IVc members (Lei et al., 2020; Li et al., 2022). By contrast, OsIRO3 is constitutively expressed in the nucleus (Zheng et al., 2010). It will be interesting to investigate whether the other OsbHLH IVb members are preferentially localized in the cytoplasm. AtbHLH11 and OsbHLH061 inhibit the transcription activation of bHLH IVc (Li et al., 2022; Wang et al., 2022). By contrast, AtbHLH121 promotes the transcription activation of AtbHLH IVc (Lei et al., 2020). OsbHLH064 is a counterpart of AtHLH121 in rice, but it remains an open question whether it plays a similar role to AtbHLH121.

Stability of bHLH proteins in response to Fe deficiency

In addition to the induction of AtFIT transcripts under Fe-deficient conditions, AtFIT protein levels also increase (Meiser et al., 2011; Sivitz et al., 2011; Cui et al., 2018; Lei et al., 2020). Two groups found that AtFIT proteins undergo 26S proteasome-mediated degradation under Fe-deficient conditions (Meiser et al., 2011; Sivitz et al., 2011). Native AtFIT proteins increase in response to Fe deficiency, whereas AtFIT proteins derived from a transgene decrease. The authors proposed that a small pool of active AtFIT is sufficient for the induction of Fe uptake genes, and degradation of inactive AtFIT from the target gene promoters allows fresh AtFIT to reinitiate further transcription cycles. AtBTSL1 (BTS like1) and AtBTSL2 are two Fe sensors (see the next section) that negatively regulate Fe homeostasis (Hindt et al., 2017). Rodríguez-Celma et al. (2019) found that AtFIT protein abundance is higher in the btsl1 btsl2 double mutant than in the wild type. They used AtFIT as a bait to conduct pull-down assays and found AtBTSL2 among the prey. Further investigation suggested that both AtBTSL1 and AtBTSL2 physically interact with AtFIT and mediate its ubiquitination (Rodríguez-Celma et al., 2019).

In AtbHLH11-overexpressing plants, protein abundance of AtbHLH11 increases under Fe-excessive conditions and decreases under Fe-deficient conditions, implying that AtbHLH11 proteins may be degraded when Fe availability is low (Li et al., 2022). Fe-deficient plants benefit from the degradation of AtbHLH11 because it is a repressor of Fe uptake. Kim et al. (2019) found that phosphorylated AtbHLH121 undergoes 26S proteasome-mediated degradation in an AtBTS (BRUTUS, an Fe sensor; see the next section)-dependent manner. However, yeast two-hybrid assays showed that AtBTS does not interact with AtbHLH121 (Gao et al., 2020). There is no report about whether AtPYE proteins undergo degradation in response to Fe deficiency. Further identification of the E3 ligases that interact with and ubiquitinate the AtbHLH IVb proteins will enhance our understanding of the Fe signaling pathway.

The transcription of AtbHLH IVc members is not inducible or slightly inducible in response to Fe deficiency (Zhang et al., 2015; Li et al., 2016; Liang et al., 2017). Tissot et al. (2019) found that protein abundance of AtbHLH105 increases in response to Fe deficiency. Protein interaction assays suggest that both AtbHLH105 and AtbHLH115 interact with AtBTS, which possesses E3 ligase activity (Long et al., 2010; Selote et al., 2015). The protein abundance of AtbHLH105 and AtbHLH115 is increased in the bts-1 mutant, and AtBTS facilitates 26S proteasome-mediated degradation of AtbHLH105 and AtbHLH115 (Selote et al., 2015; Xing et al., 2021). Both AtbHLH34 and AtbHLH104 are required for the full activation of Fe uptake genes under Fe-deficient conditions (Zhang et al., 2015; Li et al., 2016), implying that their protein activity or abundance may increase under these conditions. Indeed, analysis of GFP-IDT1-overexpressing plants suggests that GFP-IDT1 protein abundance is higher under Fe-deficient conditions than under Fe-sufficient conditions (Sharma and Yeh, 2020). When a mutated version of GFP-IDT1 with the last amino acid changed from A to V is overexpressed, GFP intensity no longer responds to Fe deficiency (Sharma and Yeh, 2020). These data suggest that AtbHLH34 protein abundance responds to Fe deficiency. AtbHLH104 is also an interaction partner of AtBTS; however, unlike AtbHLH105 and AtbHLH115, which are degraded by AtBTS, AtbHLH104 is insensitive to AtBTS (Selote et al., 2015; Xing et al., 2021). Although the abundance of AtbHLH104 protein is expected to increase in response to Fe deficiency, experimental evidence is still lacking. AtBTSL1/2 are homologs of AtBTS, raising the possibility that protein stability of AtbHLH104 is under the control of AtBTSL1/2. Further investigation is required to reveal how plants modulate the protein abundance of AtbHLH34 and AtbHLH104 according to Fe status. Although AtbHLH105 and AtbHLH115 are degraded by AtBTS, they directly promote the transcription of AtBTS under Fe-deficient conditions. This feedback loop may be crucial for the maintenance of Fe homeostasis, as reduced expression of AtBTS in bts-1 or overexpression of AtbHLH105/115 causes Fe overaccumulation (Long et al., 2010; Zhang et al., 2015; Liang et al., 2017). In addition, AtBTS was reported to facilitate the degradation of VOZ1 (vascular plant one-zinc finger 1) and VOZ2, which negatively regulate drought and cold stress responses in plants (Selote et al., 2018).

In rice, OsHRZ1 (hemerythrin motif-containing really interesting new gene (RING)- and zinc-finger protein 1) and OsHRZ2 are two Fe sensors (see the next section). Yeast two-hybrid, pull-down, and co-immunoprecipitation assays indicated that OsHRZ1 interacts with the three bHLH IVc members OsPRI1/2/3 (Zhang et al., 2017, 2020). By contrast, Kobayashi et al. (2019) showed that OsHRZs interact with OsPRI1/2, but not with OsPRI3, in yeast two-hybrid assays. Both OsHRZ1 and OsHRZ2 have E3 ligase activity (Kobayashi et al., 2013). Zhang et al. (2017) confirmed that OsHRZ1 mediates ubiquitination of OsPRI1. They also found that OsHRZ1 promotes the degradation of OsPRI1/2/3 (Zhang et al., 2017, 2020). However, Kobayashi et al. (2019) showed that OsHRZs do not ubiquitinate OsPRI1/2/3. Recently, Kobayashi et al. (2022) found that the basic leucine zipper transcription factor OsbZIP83 and the glutaredoxins OsGRX6 and OsGRX9 facilitate rice Fe utilization under the control of OsHRZ1/2. In addition, OsHRZ2 has been reported to ubiquitinate the Gγ protein GS3 and OsPHR2 (phosphate starvation response 2) (Yang et al., 2021; Guo et al., 2022).

Fe sensors

Because expression of Fe uptake genes is tightly controlled in an Fe-dependent manner, cells must be able to sense changes in Fe bioavailability. Kobayashi and Nishizawa (2014) proposed three criteria for an Fe sensor: (1) binds Fe or an intimately related molecule(s); (2) changes function in response to Fe status; and (3) regulates Fe homeostasis.

In mammals, Fe homeostasis is maintained by IRP1 (iron-regulatory protein 1) and IRP2, both of which coordinate the post-transcriptional regulation of Fe homeostasis genes by binding to IREs (iron-response elements) within their transcripts (Anderson et al., 2012). When Fe is sufficient, IRP1 binds a 4Fe-4S cluster (criterion 1), loses its affinity for IREs (criterion 2), and functions as a cytosolic ACO (aconitase), whereas IRP2 is degraded by the proteasome. IRP1 complies with the three criteria and is thus an Fe sensor in mammals. FBXL5 (F-box and leucine-rich repeat protein 5) was characterized as an Fe sensor in mammals and also conforms to all three criteria. FBXL5 possesses an iron-responsive hemerythrin domain capable of binding Fe (criterion 1) and an F-box domain that mediates the ubiquitination of IRP2 (criterion 3) (Salahudeen et al., 2009; Vashisht et al., 2009). FBXL5 protein abundance increases under Fe-sufficient conditions and decreases under Fe-deficient conditions (criterion 2). When the hemerythrin domain is mutated or removed, FBXL5 protein is no longer responsive to changes in Fe concentration. Thus, FBXL5 senses Fe status by its hemerythrin domain and accordingly modulates its own protein stability.

AtBTS/AtBTSL1/AtBTSL2 and OsHRZ1/OsHRZ2 were thought to be potential Fe sensors in Arabidopsis and rice (Kobayashi and Nishizawa, 2014). They contain hemerythrin domains in the N-terminal region that are capable of binding Fe (criterion 1). Their C-terminal region has a RING domain that contributes to E3 ligase activity for ubiquitination, just like the function of the F-box domain in FBXL5. As mentioned above, AtBTS and OsHRZ1 target bHLH IVc proteins for degradation by their RING domains, thus regulating Fe homeostasis (criterion 3). AtBTS is unstable in the presence of sufficient Fe in vitro, and mutation of its hemerythrin domains makes AtBTS insensitive to Fe fluctuations (Selote et al., 2015), indicating that AtBTS can sense Fe and correspondingly adjust its own stability (criterion 2). OsHRZ2 is also more stable under Fe-deficient conditions (criterion 2) (Guo et al., 2022). Therefore, both AtBTS and OsHRZ2 are Fe sensors. AtBTSL1 and AtBTSL2 are two close homologs of AtBTS in Arabidopsis. Loss of function of AtBTSL1 and AtBTSL2 causes the constitutive activation of Fe-deficiency-responsive genes, just as the loss of AtBTS function does (Hindt et al., 2017; Rodríguez-Celma et al., 2019). OsHRZ1 and OsHRZ2 also share many similarities in protein sequences, domains, and functions (Kobayashi et al., 2013). Evidence therefore suggests that AtBTS/AtBTSL1/AtBTSL2 and OsHRZ1/OsHRZ2 are Fe sensors. AtBTS preferentially localizes to the nucleus (Selote et al., 2015). Similarly, OsHRZ1 is a nuclear-localized protein (Kobayashi et al., 2013; Zhang et al., 2017). However, OsHRZ2 is preferentially localized to the cytoplasm (Kobayashi et al., 2013). Further investigation is needed to clarify whether the cytoplasmic fraction of OsHRZ2 has specific functions that have not been explored.

ACOs, homologs of mammalian IRPs, have been characterized in Arabidopsis and rice (Arnaud et al., 2007; Senoura et al., 2020). Senoura et al. (2020) revealed that OsACO1 is a positive regulator of the Fe-deficiency response, as knockdown of the OsACO1 gene leads to downregulation of many Fe-deficiency-inducible genes. Interestingly, they also found that OsACO1 has RNA-binding activity with affinity to the plant ACO-interacting RNA element (PAIR), similar to mammalian IRE, raising the possibility that OsACO1 senses Fe status and transmits the Fe signal, as does IRP1. Arnaud et al. (2007) analyzed the functions of three Arabidopsis ACOs and concluded that they are not involved in Fe homeostasis. However, they only examined the expression of Fe storage genes, not Fe-deficiency-inducible genes. Further investigation is required to confirm whether ACOs function as Fe sensors in plants.

An IDE1 (iron-deficiency-responsive cis-acting element 1) cis-acting element was identified in the promoter of the barley Fe-deficiency-inducible gene iron-deficiency specific clone no. 2 (Kobayashi et al., 2003). The IDE1 element is similar to the RY element, which is specifically recognized by the ABI3/VP1 family transcription factors. OsIDEF1 (IDE-binding factor 1) from the ABI3/VP1 family can activate the expression of OsIRO2, OsIRO3, and OsIMAs (Kobayashi et al., 2007, 2021; Wang et al., 2020b). Kobayashi et al. (2009) further revealed that OsIDEF1 is a positive regulator of the early Fe-deficiency response. The N-terminal region of OsIDEF1 contains a metal-binding domain that can bind Fe2+ and other metal ions, and the removal of this domain causes impairment of its transcriptional regulation (Kobayashi et al., 2012). However, evidence supporting criterion 2 is still lacking. Therefore, OsIDEF1 is thought to be a potential Fe sensor (Kobayashi and Nishizawa, 2014). Notably, orthologs of OsIDEF1 exist only in graminaceous plants. It is possible that grass plants have developed this specific regulator to modulate the early Fe-deficiency response.

IRON MANs

IRON MAN (IMA)/Fe-uptake-inducing peptide is a ubiquitous family of peptides with an Asp stretch in the C-terminal region (Grillet et al., 2018; Hirayama et al., 2018). Experiments in vitro showed that IMA peptides can bind Fe2+ and that saturation of the binding sites triggers precipitation of the peptides (Grillet et al., 2018). However, it is unclear whether IMA stability responds to Fe excess in plants. Plants overexpressing IMA overaccumulate Fe and display enhanced tolerance to Fe deficiency, whereas loss of function of IMA genes causes the opposite phenotypes (Grillet et al., 2018; Hirayama et al., 2018; Kobayashi et al., 2021). The overexpression of IMA in Arabidopsis and rice results in the upregulation of many Fe-deficiency-responsive genes (Grillet et al., 2018; Li et al., 2021a; Kobayashi et al., 2021), implying that IMA functions upstream of Fe-deficiency response signaling.

Li et al. (2021a) revealed that AtIMA interacts with and inhibits AtBTS to regulate Fe homeostasis. There are eight IMA members in Arabidopsis, and six of them interact with AtBTS and are functional. The C-terminal AtBTS interaction domain of AtIMA accounts for its interaction with AtBTS. AtbHLH105 and AtbHLH115 also contain an AtBTS interaction domain in their C-terminal region. AtIMA1 and AtIMA3 were found to compete with AtbHLH105 and AtbHLH115 for interaction with AtBTS and to alleviate degradation of the latter by AtBTS. Thus, AtIMA functions as an inhibitor of AtBTS. Li et al. also revealed that AtIMA3 peptides are ubiquitinated by AtBTS. It is noteworthy that AtIMA3 does not possess a lysine amino acid that is classically necessary for ubiquitination. In their ubiquitination assays, GFP-tagged AtIMA3 was used, and the lysine amino acids in the GFP protein may account for the ubiquitination. It is still an open question whether AtIMA3 itself can be ubiquitinated by AtBTS.

Two OsIMA genes have been identified in rice (Grillet et al., 2018; Kobayashi et al., 2021). An IMA gene was also identified in tomato, and fruits of tomato plants overexpressing AtIMA1 accumulated more Fe than those of control plants (Grillet et al., 2018). Both Fe sensors and IMA peptides are conserved across different species (Grillet et al., 2018; Rodríguez-Celma et al., 2019). It is likely that the regulatory mechanism of IMA inhibiting the Fe sensor is conserved across different species. Further investigation is needed to confirm this hypothesis.

IMA genes are inducible in response to Fe deficiency (Grillet et al., 2018; Kobayashi et al., 2021). Li et al. (2021a) revealed that the transcription of AtIMA1–3 is positively regulated by AtbHLH105 and AtbHLH115, both of which directly bind to the promoter of AtIMA3. Thus, there is a positive feedback loop, in which AtbHLH105 and AtbHLH115 promote the expression of AtIMA3, and AtIMA3 then stabilizes AtbHLH105 and AtbHLH115 by inhibiting AtBTS. Meanwhile, AtIMA1 and AtIMA2 are also positively regulated by AtbHLH121, and the latter binds to the promoters of AtIMA1–3 (Gao et al., 2020). Similarly, the transcription induction of OsIMA1 and OsIMA2 is regulated positively by OsPRI2 and OsPRI3 (Kobayashi et al., 2021). However, it remains unknown whether OsIMAs stabilize OsPRI2 and OsPRI3 proteins.

Future perspectives

In the past decade, significant progress has been made in understanding the regulation of Fe uptake in plants. However, our understanding of the mechanisms by which plants sense Fe status and transmit the Fe signal is still incomplete. The first step in Fe signaling is Fe sensing. There is therefore an urgent need for research to clarify how plants sense Fe status and coordinate Fe uptake systems. It will be helpful to analyze the conformation and activity of Fe sensors with or without Fe ions. The protein stability of AtBTS and OsHRZ2 increases in response to Fe deficiency, but it remains unclear whether the protein stability of AtBTSL1/2 and OsHRZ1 is also affected by Fe status. A challenge for the characterization of Fe sensors is the instability of full-length Fe sensor proteins. At least four different domains exist in Fe sensors, and functional analysis of each domain will provide insights into Fe sensing. Moreover, their functions diverge in the Fe-deficiency response. Careful in-depth characterization of each Fe sensor will provide us with a detailed view of how Fe sensors coordinate and integrate Fe uptake.

There has been remarkable progress in our understanding of Fe uptake regulation, particularly the characterization of bHLH IVc, bHLH IVb, bHLH Ib, FIT, and AtMYB72. These proteins are involved in the transcriptional control of Fe acquisition system genes. Although loss of function of each bHLH IVc member results in disruption of Fe homeostasis, their Fe-deficiency symptoms differ from one another. Each member is likely to have evolved its own specific functions. For example, AtbHLH105 and AtPYE specifically mediate photoprotection under Fe-deficient conditions (Akmakjian et al., 2021). Exploring their spatiotemporal expression differences and globally identifying their specific target genes will shed further light on the functional specificity of each member.

Once plants are exposed to Fe deficiency, bHLH IVc proteins initiate the transcription of downstream Fe-deficiency-inducible genes. It is unknown how the bHLH IVc transcription factors are activated. Although IMAs can protect bHLH IVc from degradation, the transcription induction of IMAs depends on bHLH IVc. Thus, bHLH IVc must be activated before the transcription induction of IMAs under Fe-deficient conditions. The bHLH IVc protein activity may be modulated by phosphorylation. An apple bHLH IVc member was reported to promote the expression of Fe-deficiency-responsive genes after phosphorylation by MAP kinase 6 (Li et al., 2021b). Phosphoproteomics research will be helpful for the characterization of many changes in protein phosphorylation that may contribute to addressing this question.

Both AtbHLH11 and AtbHLH121 are responsive to Fe deficiency at the post-translational level. Moreover, AtbHLH121 protein undergoes phosphorylation under Fe-deficient conditions. Further investigation is required to verify their interacting E3 ligases and kinases. It is still an open question whether bHLH IVb proteins play similar roles in rice. The protein modification of AtFIT involves ubiquitination and phosphorylation. Although AtFIT protein can be ubiquitinated by AtBTSL1/2, it still undergoes degradation in the btsl1 btsl2 mutant after Fe supply (Rodríguez-Celma et al., 2019). It is plausible that unknown E3 ligases mediate the degradation of AtFIT. Several phosphorylation sites exist in AtFIT, and S272 is phosphorylated by CBL-interacting protein kinase 11 (Gratz et al., 2019, 2020). Future investigation should aim to identify the kinases responsible for the phosphorylation of these bHLH proteins and clarify the impact of phosphorylation on their protein activity.

The subcellular localization of AtBTS and OsHRZ1 is different from that of OsHRZ2. Work on bHLH IVb and Ib also highlights the relevance of subcellular localization to the Fe-deficiency response. IMAs are also localized in both the cytoplasm and nucleus. Does their subcellular localization change in response to Fe status? Do they undergo protein modification in the cytoplasm? Do the interactions between Fe sensors and bHLH proteins occur in the cytoplasm or nucleus? These questions are important for understanding how Fe sensors transmit subcellular Fe status for downstream transcriptional regulation.

During the past decade, the identification of Fe sensors and many key transcription factors has helped us to build a basic regulatory network of the Fe-deficiency response. Further dissecting the functions of each component will make a substantial contribution to plant biology, agriculture, and human nutrition.

Funding

This work was supported by the (32070278) and the Youth Talent Support Program of Yunnan Province (YNWR-QNBJ-2018-134). No conflict of interest declared.