To generate transgenic rice lines expressing AtCHLH, we amplified the complete coding sequence of Arabidopsis CHLH (At5g13630; U21294 from the Arabidopsis Biological Resource Center) using the primers 5′-ACCAACTAGTAAAATGTCGACCGCCGCTCCCA-3′ (SpeI site underlined) and 5′-TTTTGGTACCTTATCGATCGATCCCTTCGATCTTGTC-3′ (KpnI site underlined). The resulting PCR product was digested with SpeI and KpnI, gel-purified, and ligated into pBluescript-SK (Stratagene, Cedar Creek, TX, USA) linearized with the same restriction sites. After verifying the DNA sequence by sequencing, the SpeI/KpnI fragment of AtCHLH was ligated into the same restriction sites of the pGA1611 vector (modified by the insertion of a SpeI restriction site) between the maize Ubiquitin promoter and the Nos 3′ terminator sequences. The resulting pGA1611:AtCHLH clone was transformed into Agrobacterium tumefaciens strain LBA4404. For rice transformation, Agrobacterium harboring the pGA1611:AtCHLH construct was co-cultured with scutellum-derived rice (Oryza sativa cv. Dongjin) calli as previously described (Lee et al., 2000). The independent transgenic rice lines expressing AtCHLH were selected based on hygromycin resistance.

To analyze transgene expression and porphyrin metabolism, rice seedlings of WT and homozygous transgenic lines (generations T₂ to T₅) were grown on soil for 4 weeks in a greenhouse at 28–30°C. For testing plant responses to nutrient deficiency, seeds of WT and two representative transgenic lines (AtCHLH-OE7 and AtCHLH-OE10) were germinated in water under dark conditions for 4 days and then hydroponically grown in different nutrient solutions under a 14-h-light/10-h-dark photoperiod with a 200 μmol m^–2 s^–1 photosynthetic photon flux density (PPFD) for 10 days. Growth conditions included half-strength Hoagland solution as control, water only, and water containing 50 µM Fe-EDTA, 0.5 mM MgSO₄, or 1.25 mM NH₄NO₃. In other experiments to test tolerance to Fe deficiency, 1-cm emerged seedlings (to achieve uniform emergence) of WT and transgenic lines were transferred into half-strength Hoagland solution without (–Fe) or with 50 µM Fe-EDTA (+Fe) under a 14-h-light/10-h-dark photoperiod with a 200 μmol m^–2 s^–1 PPFD for 7 days. Samples were collected for physiological analysis and measurement of growth parameters 7 days after exposure to Fe-deficient conditions. The lengths of shoots were measured from 30–40 hydroponically grown plants per treatment using a Vernier caliper. To measure dry biomass, the shoot and root parts of plants were dried at 80°C for 48 h and then weighed.

For ABA treatment, 1-cm emerged seedlings of WT and transgenic lines were transferred to half-strength Hoagland solution under a 14-h-light/10-h-dark photoperiod with 200 μmol m^–2 s^–1 PPFD for 3 days. Fe deficiency was applied by transferring the seedlings to half-strength Hoagland solution without (–Fe) or with 50 µM Fe-EDTA (+Fe) for 3 days. Then, Fe-sufficient WT and transgenic seedlings were transferred to solutions without or with 0.5 µM ABA (Sigma-Aldrich, St. Louis, MO, USA) for 7 days: +Fe or +Fe+ABA. Fe-deficient WT and transgenic seedlings were also transferred to solutions without or with 0.5 µM ABA for 7 days: –Fe or –Fe+ABA.

Total RNA was prepared from leaf and root tissues using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) in accordance with the manufacturer’s instructions, and 5 µg of RNA from each sample was used for the reverse transcription reaction (ImProm-II™ Reverse Transcription System, Promega, WI, USA). Subsequently, cDNA was used for qPCR analysis, which was carried out with a StepOnePlus™ Real-Time PCR system (Applied Biosystems, Waltham, MA, USA) using Power SYBR™ Green PCR Master Mix (Applied Biosystems) and gene-specific primers ( Supplementary Table 1 ). The RT-qPCR program consisted of 2 min at 50°C, 10 min at 95°C, and 40 cycles of 15 s at 95°C and 1 min at 60°C. All reactions were set up in triplicates. Actin was used as an internal control. The WT control sample was used as a calibrator, with the expression level of the sample set to 1.

Total protein was extracted with a buffer containing 12% (w/v) sucrose, 56 mM Na₂CO₃, 2% (w / v) SDS, 2 mM EDTA (pH 8.0), and 56 mM DTT. The extract was centrifuged at 12,000 × g and 4°C for 20 min, and the resulting supernatants were collected to obtain total soluble proteins. Soluble proteins were separated on a 12% SDS-PAGE for AtCHLH and electroblotted onto PVDF membranes. Immunodetection was performed according to standard procedures (Roche, Basel, Switzerland). The polyclonal antibodies against AtCHLH and α - tubulin were produced, by minimizing risk of cross-reactivity with OsCHLH (Agrisera antibody production service; Supplementary Figure 5 ), and purchased, respectively from Agrisera (Agrisera, Vännäs, Sweden).

To measurement porphyrin contents, leaf tissue was ground in a methanol:acetone:0.1 N NaOH mixture (9:10:1, v/v/v), and the homogenate was centrifuged at 10,000 × g and 4°C for 10 min to remove cell debris (Lermontova and Grimm, 2006). Porphyrin was separated by high-performance liquid chromatography (HPLC) using a Novapak C₁₈ column (4-µm particle size, 4.6 × 250 mm, Waters, Milford, MA, USA). Porphyrins were eluted with a gradient solvent system from 0.1 M ammonium phosphate (pH 5.8) and methanol (20:80, v/v) to 100% methanol at a flow rate of 1 mL min^–1. The eluates were detected by a fluorescence detector (2474, Waters) at excitation and emission wavelengths of 400 nm and 630 nm, respectively, for Proto IX and 415 and 595 nm for Mg-Proto IX and Mg-Proto IX ME. For heme determination, heme was extracted and quantified as previously described (Schneegurt and Beale, 1986). Protoheme was separated by HPLC on a Novapak C₁₈ column (Waters) with a solvent system of ethanol:acetic acid:water (66.5:17:16.5, v/v) and detected by a detector (SPD-M20A, Shimadzu) at 402 nm. Chlorophyll contents were determined spectrophotometrically according to the method of Lichtenthaler (1987).

For measurement of ALA-synthesizing capacity, leaf squares were incubated in 20 mM phosphate buffer containing 40 mM levulinic acid in the light (Papenbrock et al., 1999). Samples were homogenized, resuspended in 1 mL of 20 mM potassium phosphate buffer, pH 6.9, and centrifuged at 10,000 × g and 4°C for 5 min. The 500-μL supernatant was mixed with 100 μL ethylacetoacetate, boiled for 10 min, and cooled for 5 min. An equal volume of modified Ehrlich’s reagent (Mauzerall and Granick, 1956) was added, and the absorption was measured by a spectrophotometer (UV-2550, Shimadzu, Kyoto, Japan) at 553 nm.

Mg-chelatase (MgCh) was assayed as described in Lee et al. (1992) , with modifications. Leaf tissue was homogenized in homogenization buffer (50 mM Tricine (pH 7.8), 0.5 M sorbitol, 1 mM MgCl₂, 0.1% (w/v) bovine serum albumin (BSA), and 1 mM dithiothreitol (DTT)), and centrifuged at 5,000 × g and 4°C for 10 min. The resulting chloroplast extracts were incubated in homogenization buffer (–BSA) containing 4 mM MgATP in a regenerating system (60 mM phosphocreatine/creatine phosphokinase, 10 units mL⁻¹) and 10 m MgCl₂. Enzyme reactions were started by adding Proto IX (in DMSO) to a final concentration of 100 µM and stopped after 1 h at 30°C. Mg-Proto IX in hexane-washed water-acetone extracts was measured by a fluorescence detector (2474, Waters) at excitation and emission wavelengths of 415 and 595 nm.

Chlorophyll a fluorescence was measured in vivo using a pulse amplitude modulation fluorometer (JUNIOR-PAM, Walz, Effeltrich, Germany) after dark adaptation for 20 min. The minimum fluorescence (F _o) at open PSII centers was determined by measuring light, while the maximum fluorescence (F _m) at closed PSII centers was examined after an application of a 0.8 s pulse of saturating light (3500 μmol m⁻² s⁻¹). The F _v/F _m ratio, which is the ratio of variable fluorescence (F _v) to F _m after dark-adaptation and represents the maximum quantum yield of PSII, was used to assess the impairment of photosynthetic activity. Electron transport rate (ETR) was derived from photochemical quantum yield of PSII. Non-photochemical quenching (NPQ) was also quantified, as previously done by Bilger and Björkman (1990) according to the Stern–Volmer equation, NPQ = (F _m – F _m΄)/F _m΄.

For Perls staining, seeds from WT and transgenic plants grown under the same conditions were soaked in water overnight. Dehusked seeds were then soaked in solution containing 2% (w/v) HCl and 2% (w/v) potassium hexacyanoferrate II trihydrate for 1.5 h (Ishimaru et al., 2010). The stained samples were washed with water and photographed.

For Fe content analysis, seeds, shoots, or roots were digested completely with 1 mL of 13 M HNO₃ and 1 mL of 8.8 M H₂O₂ in an UltraWAVE instrument (Milestone Srl., Sorisole, Italy) according to the method of Ishimaru et al. (2010). Digests were analyzed for Fe content by inductively coupled plasma optical emission spectroscopy (ICP-OES) using an Optima 7300 DV instrument (Perkin Elmer, Waltham, MA, USA).

All data are shown as means ± standard error (SE). Statistically significant differences were determined by the Duncan’s multiple range test at P < 0.05. The analyses were performed using SPSS software (SPSS Inc., Chicago, IL, USA).

To increase MgCh activity in rice, we generated transgenic lines expressing AtCHLH under the control of the Ubiquitin promoter via Agrobacterium-mediated transformation ( Figures 1A, B ). We examined the expression of the transgene by PCR ( Supplementary Figure 1A ) and RT-qPCR in independent homozygous lines (AtCHLH-OE1–18). Strong AtCHLH expression was detected in all transgenic lines examined but not in nontransgenic WT plants ( Figure 1C ). We also examined the accumulation of a band with a predicted molecular weight of 140 kDa in transgenic plants by immunoblot analysis with an antibody specific to AtCHLH protein, which localizes to chloroplasts in Arabidopsis (Gibson et al., 1996; Wu et al., 2009). We observed the 140-kDa band in immunoblots of proteins extracted from leaves, stems, and roots ( Figure 1D ; Supplementary Figure 1B ) of all transgenic lines but not the WT, confirming the successful accumulation of AtCHLH in transgenic rice. The band appearing at 90 kDa seems nonspecific. As CHLH is the H subunit of MgCh complex, we investigated the effects of AtCHLH expression on MgCh activity in the leaves of 4-week-old soil-grown transgenic plants. MgCh activity was five to six times greater in the transgenic lines AtCHLH-OE6, 7, and 10 than in WT plants, indicating that expression of AtCHLH substantially increases MgCh activity ( Figure 1E ). The transgenic lines were indistinguishable from the nontransgenic WT in their growth phenotype when grown under optimal conditions ( Supplementary Figure 1C ).

We next assessed the effect of heterologous AtCHLH expression on metabolic flux within the Mg and Fe branches for chlorophyll and heme biosynthesis, respectively. Transgenic plants exhibited a higher (58–102%) synthesizing capacity for ALA, a Proto IX precursor along the common branch, than WT plants ( Table 1 ). The levels of Proto IX, a common precursor for the chlorophyll and heme branches, were not significantly different in WT and transgenic plants, but Mg-Proto IX and Mg-Proto IX ME accumulated to higher levels in transgenic plants than in WT plants ( Table 1 ). These results demonstrate that the expression of AtCHLH in transgenic plants increases not only MgCh activity and abundance of its enzymatic product, Mg-Proto IX, but also its downstream metabolite Mg-Proto IX ME. Levels of chlorophylls and heme were not significantly different between WT and transgenic plants ( Table 1 ). We selected two representative transgenic lines, AtCHLH-OE7 and AtCHLH-OE10, for further physiological experiments.

Both WT and AtCHLH transgenic seedlings grown hydroponically in water for 10 days developed chlorosis and stunted growth of young leaves ( Figure 2A ). Interestingly, the chlorosis phenotype exhibited by the AtCHLH transgenic plants was more moderate than that seen in WT seedlings. To identify which missing element caused the differential chlorosis between WT and transgenic lines, we grew seedlings hydroponically in half-strength Hoagland solution (control) or in water containing a single element whose deficiency is known to cause chlorosis in seedlings: Fe, Mg, or nitrogen (N) (Uchida et al., 2000). When grown with 0.5 mM MgSO₄ or 1.25 mM NH₄NO₃, both WT and transgenic seedlings still developed chlorosis and had lower chlorophyll contents, with a more severe chlorosis observed in WT seedlings than in AtCHLH transgenic seedlings ( Figures 2A, B ). By contrast, WT and transgenic seedlings grown in water supplemented with 50 µM Fe-EDTA resulted in no visible leaf chlorosis, although shoot growth was still impaired. AtCHLH transgenic seedlings had greater chlorophyll contents under these conditions than did the WT seedlings, even those grown in half-strength and Fe-replete Hoagland solution. The chlorophyll a/b ratio was lower in transgenic seedlings than that of WT seedlings under Fe-deficient conditions; 3.89, 3.23, and 3.03 in WT, AtCHLH-OE7, and AtCHLH-OE10, respectively.

Next, we grew WT and AtCHLH transgenic seedlings hydroponically in half-strength Hoagland solution without or with 50 µM Fe-EDTA to measure the extent of leaf chlorosis, the typical visible symptom of Fe deficiency (Terry, 1980). Although the leaves of both WT and transgenic seedlings exhibited chlorosis after transfer to Fe-free Hoagland solution for 7 days, the transgenic lines AtCHLH-OE7 and AtCHLH-OE10 showed milder chlorosis symptoms than WT seedlings ( Figure 2C ). Leaf chlorosis was accompanied by a decrease in chlorophyll contents under Fe-deficient conditions, but to a lesser extent in AtCHLH transgenic seedlings than in WT seedlings ( Figure 2D ). Fe deficiency also caused lower photosynthetic performance, as determined by F _v/F _m values and ETR, which indicate photochemical efficiency of photosystem II (PSII) ( Figures 2E, F ). Under Fe-deficient conditions, F _v/F _m values and ETR dropped more in WT seedlings than in transgenic seedlings, indicating that impairment of photosynthesis is less pronounced in transgenic seedlings. To evaluate the occurrence of protective mechanism of PSII, we measured NPQ in leaves of seedlings treated with Fe deficiency. In controls, NPQ levels were the same in WT and transgenic seedlings ( Figure 2G ). Compared to the controls, non-radiative energy dissipation through NPQ greatly decreased by 70% in WT, whereas the transgenic lines AtCHLH-OE7 and AtCHLH-OE10 did not show a noticeable decline in NPQ.

To examine the effect of AtCHLH expression on plant growth, we measured morphological characteristics in WT and transgenic seedlings. Under Fe-sufficient conditions, shoot length and biomass of the transgenic seedlings were comparable to those of the WT ( Figures 2H, J ). However, after 7 days of exposure to Fe deficiency, shoot length of the WT only reached about 46% that seen under Fe-sufficient conditions. Although transgenic seedlings also showed impaired shoot growth, the effect was not as pronounced, as shoot length was 24% and 15% longer in transgenic lines AtCHLH-OE7 and AtCHLH-OE10, respectively, than in WT seedlings under Fe deficiency ( Figure 2H ). Shoot biomass decreased by 50% in WT seedlings and by 32% and 31% in AtCHLH-OE7 and AtCHLH-OE10 seedlings, respectively, under Fe deficiency ( Figure 2I ). By contrast, transgenic seedlings had a lower root biomass than did WT seedlings under both Fe-sufficient and Fe-deficient conditions ( Figure 2J ). The Fe deficiency-induced decreases in total plant biomass were 42.8%, 27.8%, and 28.3% in WT, AtCHLH-OE7, and AtCHLH-OE10, respectively, resulting in a higher total biomass in transgenic seedlings (data not shown). Our results demonstrate that heterologous expression of AtCHLH in transgenic rice plants results in alleviation of chlorosis as well as smaller declines in chlorophyll contents, photosynthesis, and shoot growth compared to those seen in WT plants under a limited Fe pool, successfully ameliorating Fe deficiency-induced stress in rice seedlings.

Since Fe deficiency-induced chlorosis reflects a decrease in chlorophyll levels, we examined the consequences of heterologous AtCHLH expression on the metabolic activities associated with porphyrin biosynthesis in seedlings grown hydroponically under Fe-sufficient and Fe-deficient conditions. In Fe-sufficient controls, AtCHLH transgenic seedlings accumulated more Mg-Proto IX than did WT seedlings ( Figure 3A ), in agreement with our earlier result with soil-grown plants ( Table 1 ). Fe deficiency resulted in decreases in Proto IX and Mg-Proto IX levels, in addition to lower ALA-synthesizing capacity, in both WT and transgenic seedlings, with a stronger decrease in Mg-Proto IX levels in WT seedlings ( Figure 3A, B ; Supplementary Figures 2 , 3 ). By contrast, Mg-Proto IX ME accumulated to a lesser extent in Fe-deficient transgenic seedlings compared to WT seedlings. Although Fe deficiency was accompanied by lower MgCh activity in both WT and transgenic seedlings, MgCh activity remained high in transgenic seedlings relative to WT seedlings ( Figure 3C ). In the Fe-porphyrin branch, Fe deficiency lowered the heme contents in WT and transgenic seedlings, with the level of heme being greater in transgenic seedlings ( Figure 3D ; Supplementary Figure 4 ). The differences in porphyrin metabolite levels between soil-grown plants ( Table 1 ) and hydroponically grown plants ( Figure 3 ) may result from different growth stages (i.e., 4-week-old plants versus young seedlings) and conditions (i.e., light level and soil versus hydroponics).

To explore the molecular mechanisms underlying these changes in porphyrin metabolism induced by Fe deficiency, we determined relative transcript levels for key genes in porphyrin biosynthesis by RT-qPCR. Under Fe-sufficient conditions, we observed no significant difference in expression for most porphyrin biosynthetic genes between WT and transgenic seedlings ( Figure 4 ). In the common pathway, we analyzed two genes encoding enzymes critical for ALA-synthesizing activity, HEMA1 and glutamate 1-semialdehyde aminotransferase (GSA) (Beale and Weinstein, 1990). Relative HEMA1 expression decreased by approximately 80% upon Fe deficiency in WT and transgenic seedlings compared to Fe-sufficient conditions ( Figure 4A ). By contrast, relative transcript levels of GSA, ALAD (encoding ALA dehydratase), and PPO1 increased by 50–219% under Fe-deficient conditions in WT and transgenic seedlings relative to Fe-sufficient conditions, with a greater increase for GSA and PPO1 in WT seedlings.

We next assessed genes belonging to the Mg-porphyrin or Fe-porphyrin branch as a possible explanation for the observed changes in Mg-porphyrins and heme under Fe-deficient conditions. OsCHLH and CHLI, which encode the H and I subunits of MgCh, respectively, were downregulated in WT and transgenic seedlings experiencing Fe-deficient conditions, with a smaller reduction in OsCHLH transcript levels in transgenic seedlings; CHLD remained constant in all lines and in both growth conditions ( Figure 4B ). Fe deficiency downregulated protochlorophyllide oxidoreductase B (PORB), with transcript levels in WT seedlings reaching about 26% of those seen in Fe-sufficient conditions; relative PORB transcript levels dropped only 40–47% in Fe-deficient transgenic seedlings compared to Fe-sufficient seedlings. In the Fe-porphyrin branch, Fe-chelatase 2 (FC2), encoding the plastidic isoform of Fe-chelatase (Beale and Weinstein, 1990), was downregulated by about 26% in response to Fe deficiency in WT seedlings but remained constant in transgenic seedlings regardless of growth condition ( Figure 4C ). Under Fe-deficient conditions, HO1 encoding heme oxygenase, which catalyzes the formation of biliverdin-IXα, carbon monoxide (CO), and Fe²⁺ through the oxidation of heme (Tanaka and Tanaka, 2007), remained constant in transgenic seedlings but downregulated by 28% in WT seedlings. HO2 exhibited a 10-fold increase in transcript levels in Fe-deficient WT seedlings but only reached 4-fold higher levels in transgenic seedlings under the same conditions. In transgenic seedlings, AtCHLH transcript levels increased up to 3-fold upon Fe deficiency ( Figure 4D ), showing that AtCHLH is a Fe deficiency-responsive gene. Overall, porphyrin metabolism was differentially modulated by Fe deficiency stress between WT and transgenic seedlings. These results indicate that heterologous expression of AtCHLH reprograms porphyrin metabolism in transgenic seedlings under Fe-deficient conditions.

When subjected to Fe deficiency, plants need to manipulate Fe uptake and translocation efficiently. To determine whether AtCHLH influences Fe uptake and/or homeostasis in transgenic rice, we compared the expression of representative Fe transporter genes between WT and transgenic seedlings under Fe-deficient conditions. In Fe-deficient leaves, the expression of Fe transporter gene IRT1 showed a 3-fold increase in WT seedlings and about 5-fold increase in transgenic seedlings, whereas IRT2 exhibited the opposite pattern, with a higher rise upon Fe deficiency in WT relative to transgenic seedlings ( Figure 5 ). Transcript levels of the other Fe transporter genes, YSL2 and YSL15, markedly increased in WT and transgenic leaves under Fe-deficient conditions; a greater increase of YSL2 and YSL15 in transgenic and WT leaves, respectively. The genes that encode key enzymes for MA biosynthesis including NAS1, NAS2, and NAAT1 greatly upregulated in Fe-deficient leaves of WT and transgenic seedlings, with greater upregulation in transgenic leaves. Transcript levels of IRO2 encoding the iron-deficiency-inducible bHLH transcription factor 2 greatly increased in WT and transgenic leaves in response to Fe deficiency, with a greater increase in transgenic leaves. The genes IDEF1 and IDEF2 belong to ABI3 transcription factor family and NAC-family transcription factor, respectively (Kobayashi et al., 2007; Ogo et al., 2008), showed smaller increases in leaves of WT and transgenic seedlings, compared to those of other Fe-related genes ( Figure 5 ). In Fe-deficient roots, IRT1 and IRT2 were strongly upregulated in WT and transgenic seedlings, with a slightly greater induction for IRT1 observed in WT seedlings ( Figure 6 ). Under Fe-deficient conditions, increased levels of NAAT1, YSL2, and IRO2 were greater in WT roots than in transgenic roots, whereas increased level of NAS1 was greater in transgenic roots. Induction of IRT2, NAS2, and YSL15 upon Fe deficiency was similar between WT and transgenic roots. In response to Fe deficiency, transcript level of IDEF1 increased in WT roots but remained constant in transgenic roots, while IDEF2 slightly increased in WT and transgenic roots ( Figure 6 ).

As Fe is taken up by the roots, we explored whether altered expression of Fe homeostasis-related genes in transgenic seedlings influenced Fe translocation from roots to shoots and seeds. We employed the Fe stain Perls Prussian blue on seeds, which confirmed that Fe contents were higher in transgenic seeds compared to WT seeds, as indicated by the stronger blue color on the surface of transgenic seeds ( Figure 7A ). We also conducted ICP-OES to measure the contents of Fe in seeds. Fe contents were higher by 9–33% in transgenic seeds compared to WT seeds ( Figure 7B ). The contents of Fe were obtained in seedlings grown hydroponically under Fe-sufficient or Fe-deficient conditions. Shoots and roots of WT and transgenic seedlings accumulated comparable levels of Fe when grown under Fe-sufficient conditions ( Figure 7C ). In response to Fe deficiency, Fe levels similarly dropped in shoots and roots of both WT and transgenic seedlings.

To explore a possible role for ABA in the Fe deficiency responses, Fe-sufficient and Fe-deficient rice seedlings were transferred to different solutions without or with 0.5 µM ABA for 7 days. This exogenous ABA treatment did not induce noticeable changes in leaf phenotypes or chlorophyll contents in either WT or transgenic seedlings under any growth condition ( Figures 8A, B ). We also compared the effects of ABA treatment on expression of AtCHLH and OsCHLH under Fe-sufficient and Fe-deficient conditions. AtCHLH transcript level and protein abundance in transgenic leaves did not significantly change following ABA treatment under Fe-sufficient or Fe-deficient conditions ( Figures 8C ). In transgenic roots, ABA did not induce any changes in AtCHLH transcript levels in both conditions ( Figure 8C ). By contrast, exogenous ABA treatment upregulated endogenous OsCHLH transcript levels in leaves of WT and transgenic seedlings about 2-fold under Fe-sufficient conditions ( Figure 8D ). While WT leaves experienced a 56% drop in OsCHLH transcript levels upon Fe deficiency, exogenous application of ABA to Fe-deficient seedlings returned OsCHLH transcripts to levels comparable to those of Fe-sufficient WT leaves. However, ABA treatment had no effect on OsCHLH expression in Fe-deficient transgenic leaves. Similarly, both Fe deficiency and exogenous ABA additively repressed OsCHLH expression in WT roots, but not in transgenic roots, in which OsCHLH expression remained constant.