A combined physiological and biophysical approach to understand the ligand‐dependent efficiency of 3‐hydroxy‐4‐pyridinone Fe‐chelates

Abstract Ligands of the 3‐hydroxy‐4‐pyridinone (3,4‐HPO) class were considered eligible to formulate new Fe fertilizers for Iron Deficiency Chlorosis (IDC). Soybean (Glycine max L.) plants grown in hydroponic conditions and supplemented with Fe‐chelate [Fe(mpp)3] were significantly greener, had increased biomass, and were able to translocate more iron from the roots to the shoots than those supplemented with an equal amount of the commercially available chelate [FeEDDHA]. To understand the influence of the structure of 3,4‐HPO ligand on the role of the Fe‐chelate to improve Fe‐uptake, we investigated and report here the effect of Fe‐chelates ([Fe(mpp)3], [Fe(dmpp)3], and [Fe(etpp)3]) in addressing IDC. Chlorosis development was assessed by measurement of morphological parameters, quantification of chlorophyll and Fe, and other micronutrient contents, as well as measurement of enzymatic activity (FCR) and gene expression (FRO2, IRT1, and Ferritin). All [Fe(3,4‐HPO)3] chelates were able to provide Fe to plants and prevent IDC but with a different efficiency depending on the ligand. We hypothesize that this may be related with the distinct physicochemical characteristics of ligands and complexes, namely, the diverse hydrophilic–lipophilic balance of the three chelates. To test the hypothesis, we performed an EPR biophysical study using liposomes prepared from a soybean (Glycine3 max L.) lipid extract and spin probes. The results showed that the most effective chelate [Fe(mpp)3] shows a preferential location close to the surface while the others prefer the hydrophobic region inside the bilayer. Significance statement The 3‐hydroxy‐4‐pyridinone Fe‐chelates, [Fe(mpp)3], [Fe(dmpp)3], and [Fe(etpp)3], were all able to provide Fe to plants and prevent IDC. Efficacy is dependent on the structure of the ligand. From an EPR biophysical study using spin probes and liposomes, prepared from a soybean lipid extract, we hypothesize that this may be related with the distinct preferential location close to the surface or on the hydrophobic region of the lipid bilayer. [Fe(mpp)3] provide higher amounts of Fe in the leaves.


| INTRODUC TI ON
is an essential nutrient for most living organisms including plants. Despite being the fourth most abundant element in Earth's crust, Fe is only available in the environment in the form of very insoluble oxides and hydroxides, which are inappropriate for an adequate Fe uptake, in particular, in alkaline soils. Fe has a key role in fundamental biological processes, such as photosynthesis, chlorophyll synthesis, respiration, nitrogen fixation, enzyme activation, and electron transfer. When this micronutrient is unavailable to plants, they frequently develop yellowing of the younger leaves, exhibit reduced leaf areas and shoot and root dry weight (Santos et al., 2015). Iron Deficiency Chlorosis (IDC) is a major constraint for successful cultivation of crops in calcareous or alkaline soils around the world. Considering that ca 30 % of the world's arable land lies in alkaline soils, farmers must rely on supplementing their crops with Fe to avoid severe growth deficiencies and disorders.
Plants can sense Fe deficiency and respond to the induced stress by triggering mechanisms in order to improve Fe uptake.
A reduction-based strategy (strategy I) and chelation-based strategy (strategy II) have been identified as plants' mechanisms for improving Fe uptake. Soybean, used as model in the present study, utilizes strategy I type mechanisms in which root Hþ-ATPases acidify the rhizosphere so that Fe(III) solubility is increased, al- Soybean is very susceptible to IDC and it has been used to study physiological and molecular mechanisms related to Fe uptake, transport, and accumulation (Roriz et al., 2014, Vasconcelos andGrusak, 2014). Various management strategies to correct Fe chlorosis are implemented in agriculture to increase yields (Wiersma, 2005, Liesch et al., 2011. The application of Fe fertilizers is effective in counteracting IDC of plants grown on calcareous soils and is the most commonly applied technique in agriculture (Lucena, 2006). The use of Fe salts is limited to low reactive media such as hydroponics or foliar applications due to their rapid precipitation under neutral-alkaline pH, conditions that occur in calcareous soils. Traditionally, products based on synthetic Fe-chelates, prepared from polyaminocarboxylate ligands such as EDTA (ethylenediamine tetraacetic acid) and EDDHA (ethylenediamine-N, N0-bis(o-hydroxyphenylacetic), have been used to control and solve the problem of IDC (Rodríguez- Lucena et al., 2010). However, although the use of polyaminocarboxylate synthetic Fe-chelates in organic farming is legally permitted in the case of a severe deficiency of micronutrients, they do present some drawbacks, including environmental risks due to the persistence of the synthetic ligands in the environment (Nowack, 2002, Lucena, 2003 and only recently one biodegradable compound was reported (López-Rayo et al., 2019). Therefore, an urgent need to test new Fe-chelates with less impact on the environment and with properties that allow more efficient pathways for root uptake, root-to-shoot translocation, and maintenance of metal homeostasis is obvious.
In order to improve Fe uptake in Strategy I plants, an Fe-chelate must be stable and it is known that its performance is largely determined by: (a) the ability of the ligand to maintain large amounts of Fe in solution and (b) the ability of the ligand to, once its original Fe has been delivered to the plant, take more Fe and supply it again to the developing tissues (López-Rayo et al., 2009, Nadal, 2012. Ligands of the 3-hydroxy-4-pyridinone (3,4-HPO) class are well known for their biological and analytical applications (Burgess and Rangel, 2008, Rangel et al., 2009, Moniz et al., 2011. The possibility of using the ligands in such a variety of fields is mainly due to their high affinity towards M(III) and M(II) metal ions and their versatility in synthesis, which allows preparation of chelators of variable denticity and distinct physicochemical properties , Moniz, Nunes et al. 2013. Since the molecules contain, in their chemical structure, both hydrophilic and hydrophobic parts, 3,4-HPO ligands are considered amphiphilic molecules. For that reason, the concept of hydrophiliclipophilic balance (HLB), originally defined for surfactants (Griffin, 1949), may also be applied to 3,4-HPO ligands. Most ligands are non-toxic and have been utilized in biomedical applications, namely, in the treatment of iron overloaded patients suffering from β-thalassemia (Galanello, 2007). The structural features of the ligand, in particular, size and HLB, have proved to be of relevance for the efficiency of both the ligand and the complexes to achieve a particular biological effect (Galanello, 2007, Rangel et al., 2009, Moniz et al., 2011. The distinct interaction of structurally different ligands and complexes with biological membrane models has allowed rationalization of the dependence of the biological effect on the nature of the ligand (Galanello, 2007, Rangel et al., 2009, Moniz et al., 2011.  (Santos, Carvalho et al., 2016).
To comprehend the ligand-dependent efficiency of [Fe(3,4-HPO) 3 ] chelates in addressing IDC, we compared the effect of three structurally different complexes whose formulae and structure are shown in Figure 1  Information about the interaction of biologically active molecules with biological membranes can be useful not only to understand their mechanism of action but also to infer about structure-activity relationships. Biophysical studies performed using liposomes as membrane models have been extensively used taking advantage of a set of spectroscopic techniques that provide information about the affinity of a molecule towards lipid bilayers and its preferential location within the hydrophilic and hydrophobic regions (Alves et al., 2016, Moniz et al., 2017. Electron Paramagnetic Resonance (EPR) spectroscopy is particularly valuable in this area of research since it allows the use of liposomes marked with spin probes located at the surface and deep inside the lipid bilayer. The ESR spectrum of each spin probe is sensitive to alterations in its molecular environment, thus, reporting the presence of molecules that are not part of the original bilayer. The analysis of the spectral features and EPR parameters of the probes in the absence and presence of the external molecule permits to get insight about its preferential location and permeation properties (Melnyk et al., 2016).
In the present work, the effect of the three Fe-chelates ([Fe(mpp) 3 ], [Fe(dmpp) 3 ], and [Fe(etpp) 3 ]) in addressing IDC was inspected by assessing the chlorosis development in hydroponically grown soybean plants. Also, an EPR biophysical study was performed using spin probes and liposome membrane models prepared from a soybean lipid extract, to get information on the distinct interaction of the [Fe(3,4-HPO) 3 ] chelates with membranes and understand the preferential location of the chelates in hydrophilic or lipophilic environments. were prepared in our laboratory following previously described procedures (Schlindwein et al., 2006, Queiros et al., 2011

| Plant material, growth conditions, and treatments
Seeds of G. max cultivar "Williams 82" were germinated for 7 days in the dark at 25°C in moist paper. Germinated seedlings were trans-

| Experiment 3 -Examination of plant supplementation with [Fe(mpp) 3 ] in alkaline conditions
Plants were grown with 20 µM [Fe(mpp) 3 ] supplementation or with no added Fe (n = 5) with the hydroponic solution described above (two vessels) or buffered with the addition of bicarbonate buffer at pH 8.8 (two vessels).

| Evaluation and analysis of the potential to prevent IDC
After 14 days of growth the plants were collected and the morphological and physiological parameters were measured. Samples for the several analyses were prepared according to the procedure detailed below for each parameter. Chlorosis development was assessed by measurement of morphological parameters, quantification of chlorophyll (SPAD) and Fe, and other micronutrients concentration (ICP-OES). Measurements of enzymatic activity (FCR) and gene expression (FRO2, IRT1, and Leaf Ferritin) were also performed.

| Morphological parameters
Sampled roots, stems, and leaves of the five biological replicates were separated, measured, and weighed.

| Physiological parameters
Leaf chlorosis was assessed with Soil and Plant Analyzer Development (SPAD) readings, measured with a portable chlorophyll meter (Konica Minolta SPAD-502Plus; Minolta, Osaka, Japan), using the youngest trifoliate leaf of five independent biological replicates.

| Root iron reductase activity measurements
Root iron reductase was quantified as described before (Vasconcelos et al., 2006). The measurements were carried out in intact roots of five plants via the spectrophotometric determination of Fe 2+ chelated to BPDS (bathophenanthroline disulfonic acid

| Determination of Fe contents and ionome study
The plant material was dried at 70°C until constant weight and 100 mg of dried plant tissue (root, stem, cotyledon, unifoliate, and tri-

| Gene expression analysis
Plants grown for 'Experiment 1' were individually pulverized thoroughly with a mortar and pestle, until a fine powder was obtained, and total RNA was extracted using Qiagen RNeasy Mini Kit (#74904) according to the manufacturer's instructions. RNA quality and quantity were checked by UV-spectrophotometry, using a nanophotometer (Implen, Isaza, Portugal). Single-stranded cDNA was then synthesized   (Neves et al., 2007) were determined from the experimental spectra.

| Electron paramagnetic resonance spectroscopy
The acquisition conditions used for the probe 16-DSA were as follows: magnetic field window of 150G (3285G to 3435G) and microwave power of 20 mW, modulation frequency of 100 kHz, modulation amplitude of 2G, gain of 60 dB, acquisition time of 100 ms, and 2 scans. The values of the rotational correlation time were calculated from the experimental spectra.

| Statistical analysis
Data were analyzed with GraphPad Prism version 6.00 for Windows (GraphPad Software, www.graph pad.com). For 'Experiment 1' and 'Experiment 3', differences between treatments were tested with ANOVA corrected for multiple comparisons using Holm-Sidak method; for 'Experiment 2' differences were tested using Pearson correlation test. Statistical significance was considered at p < .05.

| RE SULTS AND D ISCUSS I ON
The  Figure 1).

| Experiment 1 -Comparative evaluation of supplementation with [Fe(mpp) 3 ], [Fe(dmpp) 3 ], and [Fe(etpp) 3 ]
Soybean plants were grown in hydroponic controlled conditions as described in the experimental section and the three Fe-chelates were used for Fe supplementation, and in agreement with previous work (Santos, Carvalho et al., 2016). After 14 days of growth, the plants were collected and the chlorosis development was analyzed through determination of key morphophysiological, enzymatic, and molecular parameters (Roriz et al., 2014, Santos et al., 2015. In Figure 2 (Figure 3b).
Due to the essential role of Fe in the chlorophyll biosynthesis (Santos, Serrão, et al., 2016), the yellowing of the leaves is one of the main visual symptoms of Fe deficiency. The relative chlorophyll content of the leaves was measured in SPAD values, which are shown in Figure 4.  (Santos, Carvalho et al., 2016), and seem to suggest that Fe delivered by this Fe-chelate seems to be more bioavailable to the plants.
To confirm this observation, the enzymatic activity of root reductase enzyme, responsible for the reduction of Fe(III) to Fe(II) under stress conditions (Santos, Serrão, et al., 2016), was examined and the results are depicted in Figure 5.
We recognize that the lower value of root reductase activity obtained for plants grown without Fe supplementation can be misleading since plants in this condition are under severe stress and is not in agreement with studies is most plant species (Qiu et al., 2017).
However, identical result has already been shown to happen in bean (Blair et al., 2010) and soybean (Santos et al., 2013), possibly due to the fact that the enzyme needs Fe for its functioning (Krishnan, 2005). For this reason, comparison of the activity of the enzyme is made between plant supplemented with the three Fe-chelates.

Roots of plants treated with [Fe(etpp) 3 ] registered a significant in-
crease when compared to those treated with [Fe(mpp) 3 ] (17 %) and to [Fe(dmpp) 3 ] (12 %) in root reductase activity, putatively demonstrating higher nutritional stress levels, a result which is coherent with the biomass results presented in Figure 3.
The bioavailability of Fe, in the form of the three distinct Fechelates, was also analyzed by considering the Fe distribution profiles ( Figure 6). The results of total Fe content ( Figure 6a)  Alkyl-3-hydroxy-4-pyridinones show the expected decrease in water solubility as the sizes of the alkyl groups in positions 1 and 2 increase, thus, pointing out the influence of this property (Burgess and Rangel, 2008).
The values of the Fe content measured separately in roots and leaves are shown in Figure 6b and the results are very interesting.  After analysis of morphophysiological parameters, the relative expression of three genes (FRO2-like, IRT1-like, and ferritin) was evaluated, according to their importance in IDC response (Ivanov et al., 2012, Santos, Carvalho et al., 2016, Santos, Serrão, et al., 2016. FRO2 and IRT1 gene transcripts are usually accumulated in response to Fe stress conditions in order to increase Fe uptake (Jeong and Connolly, 2009, Xiong et al., 2014, Santos, Serrão, et al., 2016 and, as long as the plant keeps sensing this nutritional stress, the transcripts continue to be accumulated (Vert et al., 2003, Fuentes et al., 2018. Consistent with the general results obtained in this experiment, among the three Fe complexes, [Fe(mpp) 3 ] induced the lowest transcript levels of FRO2-like and IRT1-like genes at the root level (Figure 7a,b). Also, as seen for reductase activity (presented in Figure 5), plants supplemented with [Fe(etpp) 3 ] have the highest FRO2-like and IRT1-like gene expression levels (Figure 7a, Figure 7a,b) but, as mentioned before, Fe-deficiency response gene expression is induced by Fe (Vert et al., 2003). The leaf expression levels of the ferritin gene ( Figure 7c) followed the patterns registered for SPAD ( Figure 4) and leaf Fe accumulation results (Figure 6b). ] is also promising to formulate a fertilizer and will be further studied.

| Experiment 2 -Examination of plant supplementation with [Fe(mpp) 3 ] in different concentrations
Considering growth, relative chlorophyll content, and Fe accumulation profile (Table 2). No significant differences were found in total dry weight of plants treated with 20 and 10 µM concentrations, but differences were found for treatment with a 5 µM concentration of chelate. In what concerns SPAD results, for all concentrations, values were relatively high and no significant differences were found. Coherently with the SPAD results, no significant differences were registered between Fe chelate concentrations in leaf Fe accumulation but, as the concentration of Fe chelate was lower, a significant decrease in Fe accumulation in the roots was also observed (between 20 and 5 µM). It has been shown that supplementation with an Fe chelate at a moderate concentration might be more beneficial in Fe fertilization (Hasegawa et al., 2012), since it avoids formation of Fe oxides and consequently result in higher concentration of dissolved Fe (Bin et al., 2016). Given that, with 10 µM Fe chelate concentration, plants were able to produce a similar amount of biomass, to maintain similar levels of relative chlorophyll and to accumulate the same amount of Fe in leaves. In conclusion, it is possible to assert that chelate [Fe(mpp) 3 ] may be an efficient Fe fertilizer, even at lower dosages (half a dose in this experiment) than those previously described for the compound (Santos, Carvalho et al., 2016), and for the commer- Data are means ± SE of five biological replicates.
Different letters indicate significant differences (p < .05) within tissue types for each nutrient by ANOVA with Holm-Sidak correction test.

| Experiment 3 -Examination of plant supplementation with [Fe(mpp)3] in alkaline conditions
To be considered as an effective Fe chelate, [Fe(mpp) 3 ] must be able to provide Fe to plants grown under neutral or alkaline conditions, which is the case of [FeEDDHA], commonly used in agricultural context (Lucena, 2006). In 'Experiment 3', plants were supplemented with [Fe(mpp) 3 ] and grown hydroponically at both pH 5.5 and 8.8. Plants' growth, relative chlorophyll content, root reductase activity, and Fe accumulation profiles were evaluated to understand the Fe chelate effectiveness at two pH values (Table 3) (Table 3). As mentioned before, under alkaline conditions, root reductase activity is expected to increase, due to the hindering of the Fe reduction and uptake system (Blair et al., 2010), but in a recent study with sugar beet, the maximum reaction rate of a chloroplast ferric chelate reductase was in a pH range 6.5-7.0, decreasing to lower activity levels between 7.5 and 8.5 (Solti et al., 2014). Here, root reductase activity did not vary significantly with the pH, possibly because in plants grown at pH 8.8, the enzyme was not at its maximum reaction rate, maintaining about the same levels that those of plants grown at pH 5.5. Regarding the effect of pH on the Fe accumulation profile, total Fe content greatly increased in plants grown under alkaline conditions, accumulating mainly in the root tissue (Table 3). This Fe pool is most likely fixed in the root apoplast given the harsh pH conditions, as seen in maize in response to Fe deficiency (Shi et al., 2018

| Interaction of Fe-chelates with model membranes
As previously stated, the chemical nature of the substituents on the heterocyclic ring of a 3,4-HPO ligand (Figure 1) does not significantly influence its affinity for Fe or the redox potential of the corresponding Fe-chelate, but it does modify properties like size and the HLB of both ligand and Fe-chelate as demonstrated by studies regarding their solvation properties and n-octanol-water partitions coefficients Rangel, 2008, Santos et al., 2012). The HLB has been considered relevant to select 3,4-HPO ligands for biomedical applications (Moniz et al., 2011, Moniz et al., 2017 and interestingly it is also crucial for enhancement of Fe uptake in plants. In fact, the HLB has an important influence on the interaction of the molecules with biological membranes and can determine not only the uptake´ pathway but also the ability to crossmembrane barriers within a cell (Moniz et al., 2017.
To study the effect of different substituents on the interaction of several ligands and complexes with biological membranes, our group has been using a biophysical approach performed with liposomes as biological membrane models and spectroscopic methods (Fluorescence, NMR, and EPR; Moniz et al., 2017. The information provided is, in our opinion, more realistic not only because it is possible to prepare liposomes with the appropriate type of lipids according to the target cell but also because the use of spectroscopic techniques permits to get detailed structural and topographical information concerning the preferential location of the molecule of interest. Considering the nature of the compounds and the application, we chose to prepare liposomes from a soybean lipid extract and incorporating nitroxide spin probes for EPR spectroscopic studies. As spin labels, we used the nitroxide probes, 5-DSA and 16-DSA (Figure 8  and [Fe(dmpp) 3 ], the results concerning their distinct total Fe content ( Figure 7a) and Fe distribution in roots and in shoots ( Figure 7b) we hypothesize that the two chelates may follow different pathways within the plant. We believe that in further experiments in which plants are grown for longer periods of time, namely, in the future experiments performed in soil, we will be able to investigate fluid composition and corroborate this preliminary hypothesis.
The higher affinity of the Fe-chelate [Fe(mpp)3] for the surface of the membrane seems to favor it's the rate of reduction and improve both uptake and translocation of Fe from the root to the leaves. an explanation for the differences in Fe distribution. Also, preferential location at the surface may favor the uptake by the reduction strategy.

| CON CLUS IONS
The studies performed for chelate [Fe(mpp) 3 ] at variable concentration and pH values of the hydroponic culture medium demonstrate that the compound may be an efficient Fe fertilizer, even at lower dosages than those previously described (Santos, Carvalho et al., 2016) and used for the commercially available fertilizers. Chelate [Fe (

CO N FLI C T S O F I NTE R E S T
None declared.