Transcriptome Profiling of Cu Stressed Petunia Petals Reveals Candidate Genes Involved in Fe and Cu Crosstalk

Copper (Cu) is an essential element for most living plants, but it is toxic for plants when present in excess. To better understand the response mechanism under excess Cu in plants, especially in flowers, transcriptome sequencing on petunia buds and opened flowers under excess Cu was performed. Interestingly, the transcript level of FIT-independent Fe deficiency response genes was significantly affected in Cu stressed petals, probably regulated by basic-helix-loop-helix 121 (bHLH121), while no difference was found in Fe content. Notably, the expression level of bHLH121 was significantly down-regulated in petals under excess Cu. In addition, the expression level of genes related to photosystem II (PSII), photosystem I (PSI), cytochrome b6/f complex, the light-harvesting chlorophyll II complex and electron carriers showed disordered expression profiles in petals under excess Cu, thus photosynthesis parameters, including the maximum PSII efficiency (FV/FM), nonphotochemical quenching (NPQ), quantum yield of the PSII (ΦPS(II)) and photochemical quenching coefficient (qP), were reduced in Cu stressed petals. Moreover, the chlorophyll a content was significantly reduced, while the chlorophyll b content was not affected, probably caused by the increased expression of chlorophyllide a oxygenase (CAO). Together, we provide new insight into excess Cu response and the Cu–Fe crosstalk in flowers.


Introduction
Copper (Cu) is an essential micronutrient for most living plants as a cofactor in metalloproteins involved in many biological processes, such as ethylene signaling, reactive oxygen species (ROS) response, photosynthesis and respiration [1]. About 30% of leaf Cu is found in chloroplasts [2,3]. The most abundant Cu-containing protein in high plants is plastocyanin (PC), which localizes in chloroplast thylakoid lumen and functions as the electron carrier transporting electrons from the cytochrome b 6 /f complex (Cyt b 6 /f ) to photosystem I (PSI) [4,5]. Under Cu limitation, chlorophyll biosynthesis and photosynthesis are inhibited causing chlorosis of young leaves and reproductive organs [6,7]. As a redox metal, Cu also can be toxic to plants by generating reactive oxygen species (ROS) under excess conditions. Therefore, the content of Cu in plant cells must be tightly regulated. Several proteins have been reported involved in Cu detoxification, such as metallothioneins (MTs), heavy metal P-type ATPases (HMA5I and HMA5II) and Copper chaperons (CCH) [8][9][10]. However, the transcription factors (TFs) involved in Cu detoxification have not been identified yet. The Cu deficiency response in Arabidopsis roots is controlled by squamosa promoter binding protein-Like7 (SPL7), which is constitutively expressed in roots [11]. Under Cu deficiency, SPL7 activates expression of copper transporter2 (COPT2), ferric reductase oxidase 4 (FRO4) and FRO5 which promotes Cu uptake into the roots [11,12].

Transcriptome Profiling Analysis
In order to study the regulatory mechanism in petunia petal under excess Cu, RNAseq was performed using stage 4 and stage 7 petals from petunia treated with 40 μM Cu. The comparison of these transcriptomes identified 292 genes (106 up-regulated and 186 down-regulated) and 824 genes (303 up-regulated and 521 down-regulated) with significant differential expression in Cu treated stage 4 petals and stage 7 petals, respectively, using fold-change > 2 and false discovery rate correction adjusted (FDR) <0.05 as a cutoff (Figure 2a). Among those differentially expressed genes (DEGs), 213 genes were found in both stage 4 and stage 7 transcriptomes. Except for one gene, encoding a Glycine-rich protein, all of them had the same expression profile under excess Cu treatment, of which 74 were up-regulated and 138 were down-regulated (Figure 2b).
To verify the RNAseq results, petunia genes were selected from stage 4 and stage 7 transcriptomes for real-time RT-PCR (qRT-PCR) analysis. The expression patterns of these selected genes were in line with the RNA-seq results, and there was a strong positive correlation between RNAseq and qRT-PCR results in both stage 4 and stage 7 (Figures S1 and S2), suggesting that the RNA-seq data were reliable.

KEGG and GO Enrichment Analysis
To gain more information about the response mechanism in petunia petals under excess Cu treatment, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed. "Photosynthesis-antenna proteins" and "Photosynthesis" were the top2 enriched KEGG pathways in both stage 4 and stage 7 petals. Beside these two pathways, "carbon fixation in photosynthetic organisms", "carbon metabolism", "porphyrin and chlorophyll metabolism", all of them being photosynthesis-related, were all shared by stage 4 and stage 7 petals (Figure 3a).

Transcriptome Profiling Analysis
In order to study the regulatory mechanism in petunia petal under excess Cu, RNAseq was performed using stage 4 and stage 7 petals from petunia treated with 40 µM Cu. The comparison of these transcriptomes identified 292 genes (106 up-regulated and 186 down-regulated) and 824 genes (303 up-regulated and 521 down-regulated) with significant differential expression in Cu treated stage 4 petals and stage 7 petals, respectively, using fold-change > 2 and false discovery rate correction adjusted (FDR) <0.05 as a cutoff (Figure 2a). Among those differentially expressed genes (DEGs), 213 genes were found in both stage 4 and stage 7 transcriptomes. Except for one gene, encoding a Glycine-rich protein, all of them had the same expression profile under excess Cu treatment, of which 74 were up-regulated and 138 were down-regulated ( Figure 2b).

Figure 1.
Dry weight and element concentrations in stage 4 and stage 7 petunia petals in plants that were unexp (Control), or exposed to 40 μM Cu (+Cu). (a) Dry weight of stage 4 and stage 7 petal upon exposure to excess Cu. Bar mean ± SD (n = 6). (b) Cu, (c) Fe, (d) Zn and (e) Mn content in petunia petals under excess Cu. Bars represent mean (n = 4×3 petals). * indicates significant differences by Tukey's HSD test (p < 0.05).

Transcriptome Profiling Analysis
In order to study the regulatory mechanism in petunia petal under excess C seq was performed using stage 4 and stage 7 petals from petunia treated with 40 The comparison of these transcriptomes identified 292 genes (106 up-regulated down-regulated) and 824 genes (303 up-regulated and 521 down-regulated) wit cant differential expression in Cu treated stage 4 petals and stage 7 petals, resp using fold-change > 2 and false discovery rate correction adjusted (FDR) <0.05 a ( Figure 2a). Among those differentially expressed genes (DEGs), 213 genes were both stage 4 and stage 7 transcriptomes. Except for one gene, encoding a Glycinetein, all of them had the same expression profile under excess Cu treatment, of were up-regulated and 138 were down-regulated ( Figure 2b).
To verify the RNAseq results, petunia genes were selected from stage 4 an transcriptomes for real-time RT-PCR (qRT-PCR) analysis. The expression pattern selected genes were in line with the RNA-seq results, and there was a strong pos relation between RNAseq and qRT-PCR results in both stage 4 and stage 7 (Figur S2), suggesting that the RNA-seq data were reliable.

KEGG and GO Enrichment Analysis
To gain more information about the response mechanism in petunia peta excess Cu treatment, Kyoto Encyclopedia of Genes and Genomes (KEGG) path richment analysis was performed. "Photosynthesis-antenna proteins" and "Phot sis" were the top2 enriched KEGG pathways in both stage 4 and stage 7 petal these two pathways, "carbon fixation in photosynthetic organisms", "carbon lism", "porphyrin and chlorophyll metabolism", all of them being photosynthesis were all shared by stage 4 and stage 7 petals (Figure 3a). To verify the RNAseq results, petunia genes were selected from stage 4 and stage 7 transcriptomes for real-time RT-PCR (qRT-PCR) analysis. The expression patterns of these selected genes were in line with the RNA-seq results, and there was a strong positive correlation between RNAseq and qRT-PCR results in both stage 4 and stage 7 (Figures S1 and S2), suggesting that the RNA-seq data were reliable.

KEGG and GO Enrichment Analysis
To gain more information about the response mechanism in petunia petals under excess Cu treatment, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed. "Photosynthesis-antenna proteins" and "Photosynthesis" were the top2 enriched KEGG pathways in both stage 4 and stage 7 petals. Beside these two pathways, "carbon fixation in photosynthetic organisms", "carbon metabolism", "por-phyrin and chlorophyll metabolism", all of them being photosynthesis-related, were all shared by stage 4 and stage 7 petals (Figure 3a).
Gene ontology (GO) annotation enrichment was also performed with p-value < 0.05 as a cut-off ( Figure 3b). Both stage 4 and stage 7 petals appeared to be enriched in "oxidoreductase activity", "electron transfer activity", "metal cluster binding", "transmembrane transporter activity", "chlorophyll binding" and "lyase activity", within in the molecular function category. Notably, most of these categories are closely related to photosynthesis and heavy metal homeostasis. .

Expression Pattern of Cu Homeostasis Genes
Cu detoxication genes in roots and shoots have been studied in Arabidopsis, rice and Silene vulgaris [9, 10,28]. We investigated the expression of known Cu homeostasis genes in petunia petal under excess Cu treatment. Surprisingly, only COPT1, encoding COPTfamily transporter, CCH, one Cu chaperone, and HMA1, encoding chloroplast copper transporter, were differently expressed in petunia under excess Cu ( Figure 4, Table 1). Gene ontology (GO) annotation enrichment was also performed with p-value < 0.05 as a cut-off ( Figure 3b). Both stage 4 and stage 7 petals appeared to be enriched in "oxidoreductase activity", "electron transfer activity", "metal cluster binding", "transmembrane transporter activity", "chlorophyll binding" and "lyase activity", within in the molecular function category. Notably, most of these categories are closely related to photosynthesis and heavy metal homeostasis.

Expression Pattern of Cu Homeostasis Genes
Cu detoxication genes in roots and shoots have been studied in Arabidopsis, rice and Silene vulgaris [9, 10,28]. We investigated the expression of known Cu homeostasis genes in petunia petal under excess Cu treatment. Surprisingly, only COPT1, encoding COPT-family

Expression Profile of Fe Homeostasis Genes
In contrast to Cu homeostasis related genes, many more genes involved in Fe deficiency response were differently expressed in petunia petal transcriptomes (Table 1, Figure 5).
Unlike that of Arabidopsis, the Petunia axillaris genome encodes four proteins belonging to the bHLH IVb subfamily, including two Pypeye (PYE) orthologous proteins and two proteins closed related to bHLH121 ( Figure S3). The expression levels of these two PYEs, named PYE1 and PYE2, were both elevated under excess Cu in petunia petals (Table 1). In contrast with these PYEs, the ortholog of bHLH121 was repressed under excess Cu treatment in both stage 4 and stage 7 petals (Table 1). Two bHLH38/39/100/101 orthologs were found in the DEG lists, and both were more highly expressed in petals after excess Cu treatment (Table 1, Figure S4).
Members of the Brutus (BTS) hemerythrin E3 ligase family, e.g., Brutus-like1 (BTL1) and BTL2 in Arabidopsis thaliana, also play an important role in the Fe deficiency response [29,30]. Unlike in Arabidopsis, three BTS and one BTL homologs were identified in the Petunia axillaris genome ( Figure S5). Two, one BTS and one BTL, were overexpressed in stage 4 and stage 7 petunia petals under Cu treatment (Table 1, Figure 5).
The ferrous transporter IRT1 and ferric chelate reductase2 (FRO2) are two major Fe acquisition genes, which are strongly upregulated under Fe deficiency in Arabidopsis [33][34][35]. Petunia axillaris encodes three AtIRT orthologous proteins, all of them not expressed in flowers (Table 1 and Figure S7). Three genes encoding FRO proteins were found in stage 4 and stage 7 transcriptomes, two belonging to the AtFRO1/2/3 subgroup and one to the FRO8 subgroup ( Figure S8). The expression of all of these three FROs was induced The ferrous transporter IRT1 and ferric chelate reductase2 (FRO2) are two major Fe acquisition genes, which are strongly upregulated under Fe deficiency in Arabidopsis [33][34][35]. Petunia axillaris encodes three AtIRT orthologous proteins, all of them not expressed in flowers (Table 1 and Figure S7). Three genes encoding FRO proteins were found in stage 4 and stage 7 transcriptomes, two belonging to the AtFRO1/2/3 subgroup and one to the FRO8 subgroup ( Figure S8). The expression of all of these three FROs was induced by excess Cu in petals (Table 1). One oligopeptide transporter OPT3 is also involved in Fe transport [36]. Petunia OPT3 was more expressed upon exposure to Cu excess in petals ( Table 1, Figures 5 and S9). Yellow stripe-like proteins (YSLs) mediate metal-nicotinamine (NA) transport [37][38][39][40]. Two YSL genes, all encoding proteins belonging to YSL1/2/3 subfamily, were up-regulated in petals under excess Cu treatment (Table 1, Figures 5 and S10). The Petunia axillaris genome encodes two NAS proteins, one of them showing lower ex-pression under excess Cu treatment, while the other was barely expressed, in S4 and S7 petals (Table 1, Figure 5).
Two genes encoding Natural resistance-associated macrophage proteins (NRAMPs) were found in our transcriptomes. According to the phylogenetic analysis, they were related to NRAMP2/5 and NRAMP3/4 in Arabidopsis, respectively. NRAMP2, belonging to AtNRAMP2/5 subgroup, only showed lower expression in stage 7 petunia petals under Cu stress, while no difference was found in stage 4 petals. NRAMP3 belonging to AtNRAMP3/4 subgroup showed enhanced expression in petunia petals under Cu stress. (Table 1, Figures 5 and S11).
Ferritins are the major proteins involved in Fe storage. The Petunia axillaris genome encodes three Ferritin (Fer) proteins; notably, the expression of all of these three genes was decreased in stage 4 and stage 7 upon exposure to Cu excess ( Table 1, Figures 5 and S12).
Together, all the results indicated that the network controlling the Fe deficiency response was significantly affected in petunia petals under excess Cu.

Excess Cu Affects Photosynthesis and Fe-Dependent Proteins in Petal
The KEGG and GO enrichment results led us to investigate the effects of excess Cu on photosynthesis-related symptoms. The chlorophyll a content, chlorophyll a+b content and chl a/b ratio was significantly decreased, while the chlorophyll b content was not affected by excess Cu in petals (Figure 6a-d). Accordingly, the expression level of HEMA2, encoding glutamyl-tRNA reductase, was down-regulated in petals under excess Cu, while chlorophyllide a oxygenase (CAO), encoding enzyme for chlorophyll b biosynthesis, was induced ( Figure 6e and Table S1). petals under excess Cu (Figure 6e). Accordingly, the CAT activity was significantly lower in petals under excess Cu (Figure 6j).

Expression Levels of DEGs in Petunia Petals under Fe-Deficiency Treatment
To gain more information about the underlying mechanisms in petunia petal response to Cu stress, the expression pattern of Fe homeostasis related genes in petunia petals under Fe deficiency was studied using real-time PCR (qRT-PCR). Petunias were grown in -Fe half Hoagland's nutrient solution and the total RNA was extracted from stage 7 petals for the qRT-PCR analysis (Figure 7).
The expression level of BTSL, NAS1, OPT3, FER2, FER3 and NRAMP3 was affected by Fe deficiency (fold change > 2), and the fold change was not different from that in excess In addition, the F V /F M , nonphotochemical quenching (NPQ), ΦPS(II) and qP values were significantly decreased in the petals under excess Cu (Figure 6f-i). All these results agreed with the RNA-seq result that genes involved in electron transport, including plas-tocyanin (PC), ferredoxin-NADP reductase (FNR) and two ferredoxin (FD) genes, were significantly down-regulated in petals under excess Cu (Figure 6e). Moreover, the photosynthesis changes also were in agreement with the KEGG pathway enrichment results that the expression of genes related with photosystem II (such as PsbO, PsbP, PsbQ, PsbW and PsbY), photosystem I (such as PsaD, PsaF, PsaH, PsaK, PsaI and PsaN), the cytochrome b 6 /f complex (PetC) and the light-harvesting chlorophyll II complex were also markedly down-regulated by excess Cu treatment in petunia petals ( Figure S13).
Iron-sulfur (Fe-S) clusters are the most abundant iron cofactors in the plant, and the Fe-S cluster biosynthesis and mobilization inside the cell is tightly controlled, to avoid toxicity of free or loosely bound iron [41]. According to the GO enrichment analysis, Fe-S cluster binding is significantly enriched in stage 4 and stage 7 petals (Figure 3b). The expression profile of genes involved in Fe-S cluster metabolism was investigated in more detail (Figure 6e). Except for CAO, all the genes involved in Fe-S cluster binding were down-regulated in Cu-stressed petals, including those encoding two subunits of the Fe-S scaffold complex (SurB and SurD) and NEET (Figure 6e).
Fe metabolism is tightly linked with ROS homeostasis in plants, especially with regard to Fe-containing proteins, such as catalase (CAT), ascorbate peroxidase (APX) and Fe superoxide dismutase (FeSOD). CAT, APX and FeSOD were all down-regulated in petunia petals under Cu excess (Figure 6e). Notably, the Petunia axillaris genome encodes four CAT proteins and all the four CAT genes were significant down-regulated in petunia petals under excess Cu (Figure 6e). Accordingly, the CAT activity was significantly lower in petals under excess Cu (Figure 6j).

Expression Levels of DEGs in Petunia Petals under Fe-Deficiency Treatment
To gain more information about the underlying mechanisms in petunia petal response to Cu stress, the expression pattern of Fe homeostasis related genes in petunia petals under Fe deficiency was studied using real-time PCR (qRT-PCR). Petunias were grown in -Fe half Hoagland's nutrient solution and the total RNA was extracted from stage 7 petals for the qRT-PCR analysis (Figure 7).
The expression level of BTSL, NAS1, OPT3, FER2, FER3 and NRAMP3 was affected by Fe deficiency (fold change > 2), and the fold change was not different from that in excess Cu treated petals. The expression level of bHLH121-like, PYE1, BTS1, YSL1, YSL2, NRAMP2, FRO8 and CCH was not significantly affected by Fe deficiency treatment (fold change < 2), indicating these genes are responsive to excess Cu, rather than Fe deficiency. The expression level of COPT1, PYE2, bHLH038, bHLH039, FRO1 and NEET was differently expressed in both Fe deficiency treated and excess Cu treated petals, while the fold change in petals under excess Cu treatment was higher than that under Fe deficiency, indicating these genes might be involved in both Fe deficiency and Cu stress response pathways (Figure 7).
Heavy metal-induced Fe deficiency responses in above-ground tissues have also been observed in cadmium-stressed Noccaea caerulescens shoots [42,43]. The expression profiles of Fe homeostasis-related genes in both transcriptomes were also compared ( Figure 8). The transcript levels of bHLH039, BTSL, OPT3, NEET, FER2 and FER3 were significantly affected by excess Cd treatment in N. caerulescens shoots and Fe deficiency treatment in Arabidopsis shoots (fold change > 2 and FDR < 0.05). The transcript levels of bHLH121-like, PYE1, FRO8 and NRAMP2 were not affected by excess Cd treatment in N. caerulescens shoots and by Fe deficiency treatment in Arabidopsis shoots, confirming that these genes may specifically respond to Cu stress in petunia petals. On the other hand, the transcript fold-change of NEET in petunia petals under Fe deficiency is similar to that in excess Cdtreated N. caerulescens shoots and Fe deficiency-treated Arabidopsis shoots, and markedly lower than that in excess Cu-treated petunia petals. Heavy metal-induced Fe deficiency responses in above-ground tissues have also been observed in cadmium-stressed Noccaea caerulescens shoots [42,43]. The expression profiles of Fe homeostasis-related genes in both transcriptomes were also compared (Figure 8). The transcript levels of bHLH039, BTSL, OPT3, NEET, FER2 and FER3 were significantly affected by excess Cd treatment in N. caerulescens shoots and Fe deficiency treatment in Arabidopsis shoots (fold change > 2 and FDR < 0.05). The transcript levels of bHLH121-like, PYE1, FRO8 and NRAMP2 were not affected by excess Cd treatment in N. caerulescens shoots and by Fe deficiency treatment in Arabidopsis shoots, confirming that these genes may specifically respond to Cu stress in petunia petals. On the other hand, the transcript fold-change of NEET in petunia petals under Fe deficiency is similar to that in excess Cd-treated N. caerulescens shoots and Fe deficiency-treated Arabidopsis shoots, and markedly lower than that in excess Cu-treated petunia petals. The value of petunia is log2 fold change of qRT-PCR. The log2 fold change of genes with fold change < 2 or FDR (p-value for Col) >0.05, all been marked as 0. * data was download from Halimaa et al. [42]. ** data was download from Schuler et al. [44]. *** data was download from Kastoori Ramamurthy et al. [21].

Discussion
In this study, we investigated the transcriptional response to excess Cu in petunia petals, and provided new insights into the crosstalk between iron and copper homeostasis. Interestingly, the primary effects of excess Cu in petunia petals are related to Fe deficiency, such as disordered expression profiles of Fe deficiency response genes, the blocking of Fe-S cluster metabolism, and a dysfunctioning photosynthesis. .05, all been marked as 0. * data was download from Halimaa et al. [42]. ** data was download from Schuler et al. [44]. *** data was download from Kastoori Ramamurthy et al. [21].

Discussion
In this study, we investigated the transcriptional response to excess Cu in petunia petals, and provided new insights into the crosstalk between iron and copper homeostasis. Interestingly, the primary effects of excess Cu in petunia petals are related to Fe deficiency, such as disordered expression profiles of Fe deficiency response genes, the blocking of Fe-S cluster metabolism, and a dysfunctioning photosynthesis.

Fe Deficiency Response in Petunia Petals under Excess Cu
Two sets of Fe-regulated genes have been described in Arabidopsis, i.e., FIT-dependent and FIT-independent [16]. The FIT-dependent pathway is controlled by the FIT and bHLH Ib subfamilies of TFs, including bHLH38/39/100/101. FIT interacts with bHLH Ib TFs to activate the expression of IRT1 and FRO2, which are involved in Fe uptake into the root [45]. Certainly, this is not the case in petunia petals, because the orthologs of FIT and IRT are not expressed in petals (Table 1). Thus, the Fe deficiency response in petunia petals is FIT-independent pathway, like in Arabidopsis shoots [16]. Among the known Fe deficiency response genes, the expression levels of bHLH39, BTSL, OPT3, FRO1, NRAMP3, NEET, FER2, FER3 and NAS1 were significant affected by Fe deficiency or excess Cu in petals. Interestingly, except for PYE2, all the other known Fe-responsive genes that responded to Cu stress in petunia petals, were also significantly affected in Cd-stressed N. caerulescens shoots and/or Arabidopsis shoots upon Fe deficiency, indicating that the Fe deficiency response in the shoot, triggered either directly by Fe deficiency, or by Cd or Cu excess, is conserved among different species.
FER2 and FER3 encode ferritin proteins, which are an iron source for maintaining the function of iron-containing proteins in leaves and seeds [46]. In plants, ferritin proteins are mainly located in the plastids and barely in the cytoplasm [47]. In Arabidopsis seeds, ferritins are proposed to release Fe for the use of Fe-containing proteins after germination [46]. Beside ferritins, the vacuole is another compartment for iron storage, and NRAMP3 and NRAMP4 transport Fe from vacuole to the cytoplasm when needed. NRAMP3 and NRAMP4 are induced under Fe deficiency and functionally redundant in Arabidopsis [48,49]. The increased expression level of NRAMP3 and decreased expression level of FER2 and FER3 in petunia petals under excess Cu or Fe deficiency indicate the redistribution of Fe from the storage compartment to the cytoplasm. Chloroplasts are the largest Fe storage compartment in the plant cell, containing up to 80% of total leaf Fe [2]. A NEET protein recently has been reported to participate in Fe distribution in the plant cell. The Fe content in the chloroplasts of the neet mutant is higher than in wild-type, and the mutant accumulates more Fe in the seedlings compared to the wild-type; however, the neet mutant exhibits an Fe deficiency response in its shoot. In addition, the Fe-S cluster biogenesis and chlorophyll content were also affected by the disruption of NEET function [50,51].
All these results suggest that the redistribution of Fe inside the cell is a primary response to Fe deficiency.

Cu Detoxification and Crosstalk with Fe in Petunia Petal under Excess Cu
Copper is translocated as Cu(II)-NA in the xylem [52]. Cu(II) needs to be reduced to Cu(I) before being taken up by shoot cells. FRO4/5 are the key proteins involved in Cu(II) reduction in Arabidopsis roots [12,33]. However, the orthologs of FRO4/5 in petunia, are barely expressed and also not induced by excess Cu in petals (Table 1), suggesting there might be another FRO protein(s) involved in Cu(II) reduction in petals, if there is any. FRO2 is the key protein involved in the Fe(III) reduction in Arabidopsis roots, which takes place at the root surface, prior to Fe uptake (as Fe(II)), and is strongly induced under Fe deficiency [33]. Petunia FRO1, an ortholog of Arabidopsis FRO1/2/3, was up-regulated in petals under excess Cu or Fe deficiency, but the fold change in Fe deficiency is much lower than that under excess Cu treatment (Figure 7). This result indicates that petunia FRO1 maybe involved in the response to excess Cu, rather than to Fe deficiency, and might primarily reduce Cu(II), rather than Fe(III).
Its reduction to Cu(I), possibly through FRO1, is expected to increase the availability of Cu for transmembrane (intracellular or cell-to-cell) transport, either as Cu + or Cu(I)-NA. The cuprous ion is mainly taken up by high-affinity copper transport proteins (COPTs), and the Petunia axillaris genome encodes six COPT proteins, the same number as Arabidopsis, but only one, named COPT1, has been found in the DEG lists. The expression level of COPT1 was down-regulated in petunia petals under excess Cu or Fe deficiency (Table 1 and Figure 7), indicating that it may be involved in Cu and/or Fe homeostasis. Similarly, Arabidopsis COPT2 has been reported to be involved in iron and copper crosstalk [18,53]. AtCOPT2 encodes a plasma membrane-located Cu(I) transporter in yeast [54]. The transcripts level of AtCOPT2 is down-regulated by excess Cu [55]. However, unlike petunia COPT1, AtCOPT2 is up-regulated in the roots upon exposure to Fe deficiency [18]. In any case, the downregulation of COPT1 in response to excess Cu in petunia petals may be taken to suggest that it is primarily a Cu transporter, possibly involved in the reloading of Cu, as Cu + from the vascularity.
Beside COPTs, ZIP2 and ZIP4 are also involved in the Cu uptake in Arabidopsis [56,57], but their expression level is not effected by excess Cu in petunia petals.
The subgroup I of YSL proteins, including AtYSL1/2/3 and OsYSL16, have been reported to be involved in metal-NA transport [58]. Arabidopsis YSL1 and YSL3 both encode plasma membrane-located proteins and are abundantly expressed in shoots, roots and flowers. Their expression levels are down-regulated both under Fe deficiency and excess Cu [37,55]. The single mutants ysl1 and ysl3 have no apparent phenotypes, while the ysl1ysl3 double mutant accumulates less Fe in the shoot when grown on MS medium [37]. Recently, it has been reported that YSL1 and YSL3 are the down-stream genes of the SUMO E3 ligase, SIZ1. The siz1 mutant is more sensitive to excess Cu than wild-type. After excess Cu treatment, siz1 mutants accumulate more Cu in the leaves, because the siz1 mutant is unable to repress the expression of YSL1 and YSL3 [55]. YSL2 encodes a plasma membrane-located protein, which is expressed in both shoots and roots, and its expression in the shoots is repressed by excess Cu and Fe deficiency [40]. YSL2 transports both Fe(II)-NA and Cu(I)-NA in yeast, but the ysl2 mutant doesn't show any phenotype. The Fe and Cu concentrations are not significantly different between wild type and the ysl2 mutant, irrespective of Fe and Cu treatment levels [40]. Rice YSL16 is also involved in Cu homeostasis. YSL16 encodes a plasma membrane-located protein, which transports Cu-NA complex in yeast. The rice ysl16 mutant accumulates less Cu in shoots [59]. The ortholog of OsYSL16 in Brachypodium distachyon, BbYSL3, recently has been reported to be localized at the plasma membrane, and to mediate copper transport. The expression of BbYSL3 was induced by Cu deficiency and repressed by Fe deficiency [60,61]. The Petunia axillaris genome encodes three proteins belonging to YSL1/2/3 subgroup. Petunia YSL1 and YSL2 were both down-regulated in Cu-stressed petals, but not affected by Fe deficiency, indicating that these two genes are involved in the homeostasis of Cu, rather than Fe.
Once inside the cell, Cu(I) interacts with Cu chaperones to avoid Cu + toxicity. CCH is the only chaperone affected by excess Cu in petunia petals, in agreement with the observations in Arabidopsis shoots [62]. Heavy metal ATPase 5 (HMA5) plays an important role in Cu detoxification [8][9][10]. The Petunia axillaris genome encodes three HMA5 orthologs, one of HMA5I and two of HMA5II [10]. The HMA5II genes are barely expressed in petunia petals, and the expression of HMA5I is not affected by excess Cu (Table 1), which is not surprising, because its ortholog in rice, OsHMA4, is also not affected by excess Cu in shoots [8]. Another HMA superfamily protein HMA1 is involved in the delivery of Cu to the chloroplast. Under excess Cu, petunia HMA1 was down-regulated in petals (Table 1), probably to reduce the copper uptake into the chloroplast.
Recently, it has been reported that the crosstalk between Cu and Fe in Arabidopsis root under Cu and Fe deficiency is regulated by SPL7, FIT and bHLH Ib proteins. FIT interacts with bHLH Ib proteins to activate the expression of COPT2, FRO4 and FRO5 [12,18].
Again, this is apparently not the regular network underlying the Fe and Cu crosstalk in petunia petals, because FIT is barely expressed in petals. There were several TFs showing significantly altered expression levels in petals under excess Cu and Fe deficiency, such as PYE2, bHLH38 and bHLH39. Although the expression level of these genes was induced by both Cu excess and Fe deficiency, the fold change in petals under excess Cu is higher than that under Fe deficiency. Besides these three TFs, one bHLH121-like protein was down-regulated in petals under excess Cu, but not affected by Fe deficiency. Arabidopsis bHLH121/URI and bHLH11 both belong to the bHLH IVb subfamily, and play important roles in the Fe deficiency response [17,63,64]. Arabidopsis bHLH11 is mainly expressed in roots, and induced under Fe sufficient conditions. It acts as a transcriptional repressor. The expression level of FRO2 and FER1 was down-regulated in bHLH11 overexpression lines, while NRAMP4 and NAS4 were induced in the roots of bHLH11-overexpressing lines under Fe deficiency [63]. In contrast to bHLH11, bHLH121 functions as a direct transcriptional activator of genes involved in the Fe deficiency response. Around 50% of the Fe homeostasis-related genes are regulated by bHLH121, including bHLH39, BTS, BTSL1, BTSL2, YSL3, OPT3, MYB10, FER1, FER3, FER4, NEET, NAS4 and more [17]. Notably, the expression of these genes' orthologs is affected by excess Cu in petunia with the same expression trends, and the expression profile in excess Cu and Fe deficiency treated petunia petals showed a strong correlation ( Figure S14). The expression of bHLH121 is not affected by the Fe status in Arabidopsis. In contrast, Fe availability affects the cellular localization of bHL121 protein in roots. However, the expression of bHLH121 orthologs are down-regulated (about 40-fold) in petunia petals under excess Cu, indicating that it might be involved in Cu homeostasis. Whether bHLH121 plays an important role in Cu and Fe crosstalk still needs to be confirmed by detailed genetic analysis.
In conclusion, excess Cu treatment induced a FIT-independent Fe deficiency response in petunia petals, with strong effects on Fe-S cluster assembling and photosynthesis. In addition, the Fe deficiency response network that controlled by bHLH121 was markedly disordered. The expression level of petinia bHLH121-like protein is significantly reduced under excess Cu, indicating it might also be involved in Cu homeostasis in petals. Further genetic studies are needed to evaluate this possibility.

Plant Material and Growth Conditions
Petunia seeds (cv Mitchell; W115) were sterilized with 1:10 (v/v) diluted household bleach for 10 min. Sterile seeds were stored in water in the fridge for 48 h, then seeds were sown on solid Murashige & Skoog (MS) medium. Two weeks later, the seedlings were transferred to modified half-strength Hoagland's nutrient solution, renewed once per week [65]

RNA-Seq
The experiments were performed with three independent biological replicates, and each replicate was a pooled sample of three individual petals. Total RNA was isolated using RNAprep pure Plant plus kit (Tiangen Biotech, China) and 1 µg total RNA from those samples was used for RNA sequencing library constructed. RNA sequencing was performed using an Illumina Novaseq 6000 instrument at GENEWIZ (Suzhou, China). In order to remove technical sequences, including adapters, polymerase chain reaction (PCR) primers, or fragments thereof, and quality of bases lower than 20, pass filter data of fastq format were processed by Cutadapt V1.9.1 (phred cutoff: 20, error rate: 0.1, adapter overlap: 1 bp, minimum length: 75, proportion of N: 0.1) to be high quality clean data. Reads were aligned to the Petunia axillaris genome sequence [66] using Hista2 v2.0.1. The Hisat2 parameters set were: -x Ref -dta-cufflinks -1 clean_R1.fq -2 clean_R2.fq -knownsplicesite-infile splicesites.txt -p threads -S sam. Gene expression level were estimated as fragments per kilobase of transcript per million fragments (FPKM). Differential expression analysis was performed using EdgeR with adjusted p-value < 0.05 as a cut-off setting.

Real-Time RT-PCR
First strand cDNA was synthesized from 2 µg total RNA using PrimeScriptTM RT reagent Kit using gDNA Eraser (Perfect Real Time) (Takara). Real-time RT-PCR was performed with Green ® Premix Ex Taq™ II (Tli RNaseH Plus) (TaKaRa) on a Bio-Rad CFX Connect. The relative expression level was normalized to EF1α by the 2 −∆∆CT method [69,70]. Primers are listed in Table S2.

Chlorophyll Content and Photochemical Activityp Measurement
Fresh petals were collected from each treatment, and chlorophyll were extracted with Acetone: Ethanol (1:1, v/v) at room temperature in the dark for 24 h. Chlorophyll content was determined photometrically by measuring absorption at 663 and 645 nm using a microplate reader (Synergy2, BioTek, Winooski, VT, USA), and was then calculated using the following formulae.
Dual portable fluorescence (Dual-PAM-100, WALZ, Effeltrich, Germany) was used to determine the chlorophyll fluorescence. Before measurements, all plants were kept in the dark for 30 min. A saturating pulse (20,000 µmol photons m −2 s −1 , 300 ms) was applied to measure the maximum fluorescence, and then petals were illuminated at a saturating light of 358 µmol photons m −2 s −1 for 5 min to activate the electron sink in photosynthesis, and saturation pulses were generated every 20 s. PSII parameters were calculated as follows [72]:

Malondialdehyde (MDA) and Catalase (CAT) Measurement
MDA was measured using the Malondialdehyde (MDA) Content Assay Kit (Solarbio, Beijing, China), based on the thio-barbituric acid assay [73]. MDA was analyzed spec-trophotometrically at 450, 532 and 600 nm. The experiments were performed with three independent biological repetitions, one petunia petal used for each measurement. CAT activity was measured using a Catalase (CAT) Activity Assay Kit (Solarbio, Beijing, China), based on the protocol of Aebi [74]. One stage 4 petal was used for each measurement, with three independent biological repetitions. The absorbance was determined at 240 nm.

Phylogenetic Analysis
The putative orthologs search of interesting proteins was carried out by blast search in BLAST+ v2.2.31, through TBtools v1.098661 software. Full length protein sequences were aligned using MAFFT v7.475 with default setting [75]. The alignment curation was performed using BMGE, with settings as described in Morel et al. [76]. The substitution model for the tree building was calculated with ModelFinder using the Akaike information criterion in IQ-tree v1.6.8 [77], through PhyloSuite v1.2.2 [78]. Then the phylogeny was generated using IQ-tree v1.6.8, with the parameter settings included in the descriptions of individual trees.

Statistical Analysis
Tukey HSD analysis was performed using the "agricolae" package in R v4.1.0. Correlation analysis between RNA-seq and qRT-PCR was also performed using R v4.1.0.