Long-term boron-deficiency-responsive genes revealed by cDNA-AFLP differ between Citrus sinensis roots and leaves

Seedlings of Citrus sinensis (L.) Osbeck were supplied with boron (B)-deficient (without H3BO3) or -sufficient (10 μM H3BO3) nutrient solution for 15 weeks. We identified 54 (38) and 38 (45) up (down)-regulated cDNA-AFLP bands (transcript-derived fragments, TDFs) from B-deficient leaves and roots, respectively. These TDFs were mainly involved in protein and amino acid metabolism, carbohydrate and energy metabolism, nucleic acid metabolism, cell transport, signal transduction, and stress response and defense. The majority of the differentially expressed TDFs were isolated only from B-deficient roots or leaves, only seven TDFs with the same GenBank ID were isolated from the both. In addition, ATP biosynthesis-related TDFs were induced in B-deficient roots, but unaffected in B-deficient leaves. Most of the differentially expressed TDFs associated with signal transduction and stress defense were down-regulated in roots, but up-regulated in leaves. TDFs related to protein ubiquitination and proteolysis were induced in B-deficient leaves except for one TDF, while only two down-regulated TDFs associated with ubiquitination were detected in B-deficient roots. Thus, many differences existed in long-term B-deficiency-responsive genes between roots and leaves. In conclusion, our findings provided a global picture of the differential responses occurring in B-deficient roots and leaves and revealed new insight into the different adaptive mechanisms of C. sinensis roots and leaves to B-deficiency at the transcriptional level.


Introduction
Boron (B), as an essential micronutrient required for higher plants, is absorbed from soil solution by plant roots mainly in the form of boric acid. As boric acid in soils is easily leached under high rainfall conditions, B-deficient symptoms are often observed in many important agricultural crops, including citrus (Chen et al., 2012). According to investigation, a fair number of cultivated soils in southern and eastern China had very low concentration of hot water soluble B (less than 0.25 mg kg −1 DW) (Liu et al., 1989). Up to 45.5 and 9.0% of "Guanximiyou" pummelo (Citrus grandis) orchards in Pinghe, Zhangzhou, China were deficient in soil water-soluble B and leaf B, respectively (Huang et al., 2001).
There are several comparative studies showing that the effects of B-deficiency on gas exchange, carbohydrates and OAs and related metabolic enzymes, nitrogen and phenolic metabolisms differ between roots and leaves (Stavrianakou et al., 2006;Camacho-Cristóbal et al., 2008a;Hajiboland et al., 2013;Lu et al., 2014a). Thus, B-deficiency-induced changes in transcriptomics should be different between roots and leaves. To our knowledge, such data are not yet available in woody plants.
In this study, we first compared B-deficiency-responsive genes in Citrus sinensis roots and leaves using cDNA-amplified fragment length polymorphism (cDNA-AFLP) in order to (a) determine the mechanisms of plants to deal with B-deficiency at the transcriptional level and (b) understand the differences in Bdeficiency-induced alterations in gene expression between roots and leaves.

Plant Materials
This study was conducted at Fujian Agriculture and Forestry University, Fuzhou, China (26 • 5 ′ N, 119 • 14 ′ E). "Xuegan" (Citrus sinensis) seedlings were used in this study, because C. sinensis is polyembryonic seed development, an apomictic process in which many embryos are initiated directly from the maternal nucellar cells surrounding the embryo sac containing a developing zygotic embryo (Aleza et al., 2010).

Experimental Design
Plant culture and B-treatments were performed according to Yang et al. (2013b). In late-May (5 weeks after germination), uniform seedlings were transplanted into 6 L pots (two per pot) containing river sand and grown in a greenhouse under natural photoperiod. Ten weeks after transplanting, each pot were irrigated every other day until dripping with B-deficient (without H 3 BO 3 ) or -sufficient (10 µM H 3 BO 3 ) nutrient solution for 15 weeks. There were 20 pots per treatment in a completely randomized design. At the end of the experiment, fully expanded (about 7-week-old) leaves (midribs and petioles removed) and ca. 5-mm-long root apices were collected at noon under full sun from different replicates and treatments and immediately frozen in liquid N 2 . Both leaf and root samples were stored at −80 • C until they were used for cDNA-AFLP and qRT-PCR analysis. The remaining seedlings that were not sampled were used to measure root, stem, and leaf dry weights (DWs) and root and leaf concentration of B.

Measurements of Root, Stem, and Leaf DWs, and Root and Leaf B Concentration
Ten plants per treatment from different replicates were harvested and divided into roots, stems and leaves. After being dried 70 • C for 48 h, their DWs were weighted.
For the determination of B concentration, about 7-weekold leaves (midribs and petioles removed) and fibrous roots were collected and dried at 70 • C for 48 h. Dried samples were ground in a mortar to pass a 40-mesh sieve, then ashed at 500 • C for 5 h, finally dissolved in 0.1 M HCl. B concentration in the solution was assayed by the modified curcumin method (Kowalenko and Lavkulich, 1976). There were six replicates per treatment.

RNA Preparation, cDNA Synthesis and CDNA-AFLP Analysis
Equal amounts of frozen leaf (root) samples collected from five plants (one per pot) were mixed as a biological replicate. There were three biological replicates for each treatment. Total RNA was independently extracted three times from the frozen samples using Recalcitrant Plant Total RNA Extraction Kit (Centrifugal column type, Bioteke Corporation, China) according to manufacturer's instructions. cDNA synthesis and cDNA-AFLP analysis were performed according to Zhou et al. (2013). After the integrity and quantity of total RNA being checked, first-strand cDNA was synthesized. The resulting double-stranded cDNA was purified using equal volume of phenol: chloroform: isoamyl alcohol (25: 24: 1). Double-stranded cDNA (600 ng) was digested with restriction enzymes: 5 U each of EcoR I (Thermo Scientific, Massachusetts, USA; 3 h at 37 • C) and Mse I (Tru1I, Thermo Scientific, Massachusetts, USA; 3 h at 65 • C). The resulting restricted fragments were ligated to adaptors (EcoR I, 0.2 µM forward primer: 5 ′ -CTC GTAGACTGCGTACC-3 ′ and reverse primer: 3 ′ -CATCTGACG CATGGTTAAP -5 ′ ; Mse I, 2 µM forward primer: 5 ′ -GACGAT GAGTCCTGAG-3 ′ and reverse primer: 3 ′ -TACTCAGGACTC ATP-5 ′ ) with T4-DNA ligase (Thermo Scientific, Massachusetts, USA) for 10-16 h at 16 • C. The resulting ligated products were pre-amplified with the corresponding pre-amplification primers: EcoR I, 5 ′ -GACTGCGATCCAATTC-3 ′ and Mse I, 5 ′ -GATGAGTCCTGAGTAA-3 ′ . From a 100-fold dilution of the pre-amplified samples, a 5 µL diluted sample was used for the selective amplification using 256 combinations of the following primers: 16 derivatives of EcoR I primers 5 ′ -GACTGCGAT CCAATTCEE-3 ′ and 16 derivatives of Mse I primers 5 ′ -GAT GAGTCCTGAGTAAMM-3 ′ ; where EE and MM represented AA, AT, AC, AG, TA, TC, TT, TG, CA, CT, CG, CC, GA, GC, GT, and GG. The selective amplification products were separated on a 6% (w/v) polyacrylamide gel run at 50 W for 2.5 h. The gels were silver stained to visualize the cDNA bands. Samples for cDNA-AFLP analysis were run in three replicates at least.
The up-or down-regulated cDNA-AFLP bands (transcriptderived fragments, TDFs) were selected based on their presence, absence, or differential intensity and cut out with a scalpel, and incubated in 50 µL of dd H 2 O for 30 min in a boiling water bath, then centrifuged at 10000 rpm (Eppendorf 5418R, Hamburg, Germany). The supernatant (eluted DNA) was re-amplified by PCR using the same primer combinations. The re-amplified products were checked on 1% (w/v) agarose gels, each band was isolated and eluted using DNA Agarose Gel Recovery Kit (Solarbio, China). Before being sequenced by BGI Technology Corporation (Shenzhen, China), these TDFs fragments were ligated to pGEM-T EASY vector according to usage information of pGEM R -T Easy Vector System I (Promega, USA), then transduced into Escherichia coli (DH5α) competent cells using ampicillin as the selecting agent. All sequences were input into the VecScreen (http://www.ncbi.nlm.nih.gov/ VecScreen/VecScreen) to identify and remove all of the vector sequence. Homology of TDFs' sequences was analyzed using the BLASTX and BLASTN searching engines (http://www. blast.ncbi.nlm.nih.gov/Blast). Their functional categories were assigned based on the analysis of information reported for each sequence by The Gene Ontology (http://amigo.geneontology. org/cgi-bin/amigo/blast) and Uniprot (http://www.uniprot. org/).

qRT-PCR Analysis
Both sample collecting and total RNA extraction were performed as described above. qRT-PCR analysis was performed according to Zhou et al. (2013). Specific primers were designed from the sequences of 51 differentially expressed TDFs using Primer Premier Version 5.0 (PREMIER Biosoft International, CA, USA). The sequences of the F and R primers used were listed in Table S1. Samples for qRT-PCR were run in three biological replicates with two technical replicates. Relative gene expression was calculated using ddCt algorithm. For the normalization of gene, citrus actin (GU911361.1) was used as an internal standard and the sample from B-sufficient treated plants was used as reference sample, which was set to 1.

Statistical Analysis
Results represented the mean ± SD. Statistical analyses of data were carried out by unpaired t-test at P < 0.05 level.

Plant Growth and B Concentration in Roots and Leaves
Seedlings treated without H 3 BO 3 had slower growth and less leaf and root level of B than those treated with 10 µM H 3 BO 3 ( Figure  S1). B concentration in leaves from seedlings treated without H 3 BO 3 was lower than the sufficient range of 30-100 µg g −1 DW (Chapman, 1968). Also, a typical B-deficient symptom (i.e., corky split veins) was observed in leaves from seedlings treated without H 3 BO 3 (Han et al., 2008). Thus, seedlings treated without H 3 BO 3 are considered B-deficient, and those treated with 10 µM H 3 BO 3 are considered B-sufficient.

Differentially Expressed Genes in B-deficient Roots and Leaves
A total of 256 selective primer combinations were used to isolate the differentially expressed TDFs from B-deficient roots and leaves. Figure S2 displayed the typical picture of a silver-stained cDNA-AFLP gel. We amplified a total of 5247 (5579) reproducible, clear, and unambiguous cDNA-AFLP bands (TDFs) from B-deficient roots (leaves), with an average of 20.5 (21.8) TDFs in roots (leaves) for each primer combination. A TDF with both a P-value of less than 0.05 and an average fold change of more than 1.5 was considered differentially expressed. Here, 131 and 165 differentially expressed and reproducible TDFs were obtained from B-deficient roots and leaves, respectively. After all these TDFs were reamplified, cloned, and sequenced, 114 and 129 TDFs from roots and leaves produced useable sequence data. All these data were blasted against the sequence data available in GenBank. Eighty-three root TDFs and 92 leaf TDFs showed significant homology to genes encoding known, putative uncharacterized, hypothetical and unknown proteins, and the remaining 31 root TDFs and 37 leaf TDFs did not share homologous with any nucleotide or AA sequence in the public databases.

Functions of the Differentially Expressed Genes in Roots and Leaves
Among the 83 matched root TDFs, 38 TDFs were up-regulated and 45 TDFs were down-regulated by B-deficiency. According to the biological properties, these TDFs were associated with carbohydrate and energy metabolism (11), nucleic acid metabolism (13), protein and AA metabolism (10), cell transport (9), signal transduction (7), stress response and defense (8), lipid metabolism (4), cell wall modification (2), and others (19) (Table 1; Figure S3). For the 92 matched leaf TDFs, 54 TDFs were increased and 38 TDFs were decreased by B-deficiency. These TDFs were classified into the following categories: carbohydrate and energy metabolism (12), nucleic acid metabolism (11), protein and AA metabolism (19), cell transport (10), signal transduction (5), stress response and defense (8), lipid metabolism (6), cell wall modification (1), and   others (20) (Table 2; Figure S3). As shown in Figure S3, the majority of the differentially expressed TDFs were isolated only from B-deficient roots or leaves, only seven TDFs with the same GenBank ID were isolated from the both.

Validation of cDNA-AFLP Data
Twenty-five TDFs from roots and 26 TDFs from leaves were selected for qRT-PCR to check their expression patterns obtained by cDNA-AFLP. The expression profiles of all these TDFs produced by qRT-PCR well-matched with the expression patterns revealed by cDNA-AFLP except for three TDFs (i.e., TDFs #R67-2b, R190-1a and L199-1a; Figure 1). Thus, this technique was validated in 94% of cases.

B-deficiency-responsive-genes Differed between Roots and Leaves
We isolated less up-regulated TDFs from B-deficient roots than from B-deficient leaves, and more down-regulated TDFs from the former than from the latter (Tables 1, 2; Figure S3). This agrees with our report that mitochondrial respiration, OA metabolism and AA biosynthesis were increased in B-deficient C. sinensis leaves with more accumulation of carbohydrates, but decreased in B-deficient C. sinensis roots with less accumulation of carbohydrates (Lu et al., 2014a Figure S3). To conclude, B-deficiency-induced changes in gene expression differed between roots and leaves.

Genes Involved in Carbohydrate and Energy Metabolism
The expression levels of many carbohydrate and energy metabolism-related TDFs were altered in B-deficient roots and leaves (Tables 1, 2; Figure S3). Plant UDP-glycosyltransferases (UGTs) play important parts in enhancing the tolerance of plants to environmental stresses (Bowles et al., 2006). Over-expression of UGT85A5 conferred salt tolerance in tobacco (Sun et al., 2013). However, over-expression of UGT73B2 lowered oxidative stress tolerance in Arabidopsis (Kim et al., 2010). Our results showed that the expression levels of UGTs were decreased in B-deficient roots (i.e., TDFs #R255-1b, R256-2b, R242-1b, R244-2b, and 63-3b) and increased in B-deficient leaves (i.e., TDFs #L201-1a, L3-2a and L29-2a) (Tables 1, 2), which might be related with the less and more accumulation of carbohydrates in B-deficient C. sinensis roots and leaves, respectively (Lu et al., 2014a). Thus, we proposed that the adaptive responses of UGPs differed between B-deficient roots and leaves. The expression of cytochrome P450s (CytoP450s) was increased in Vicia sativa seedlings when exposed to plant hormone methyl jasmonate  and in Btoxic C. grandis leaves (Guo et al., 2014). Transgenic tobacco and potato plants over-expressing CytoP450 displayed increased monooxygenase activity and enhanced tolerance of oxidative stress after herbicide treatment (Gorinova et al., 2005). Therefore, the up-regulation of CytoP450 (TDF #R132-6a) in B-deficient roots ( Table 1) might contribute to the tolerance of plants to B-deficiency.
that B-deficient C. sinensis leaves had higher mitochondrial respiration and activities of enzymes in glycolysis and TCA cycle (Lu et al., 2014a). Hamilton et al. (2001) observed that both root vacuolar ATPase and mitochondrial ATP synthase were induced by aluminum (Al) in an Al-resistant wheat cultivar and suggested that increased ATP synthase activity was required for supporting V-ATPase induction and other energy-dependent processes associated with Al-resistance. Over-expression of a mitochondrial ATP synthase small subunit gene enhanced the salt-tolerance of transgenic tobacco plants (Zhang et al., 2006). Thus, the up-regulation of ATP synthase subunit β (i.e., TDFs #R3-1a and R16-2a) in B-deficient roots (Table 1) might be advantage to maintaining energy balance by enhancing ATP biosynthesis, when ATP synthesis was reduced due to decreased root respiration (Lu et al., 2014a).
Heat shock transcription factors (HSFs) play a role in various stresses, including oxidative stress. Davletova et al. (2005) demonstrated that HSFs were indispensable in the early sensing of H 2 O 2 stress in Arabidopsis. Mukhopadhyay et al. (2004) FIGURE 1 | Effects of B-deficiency on gene expression of Citrus sinensis roots (A,B) and leaves (C,D). qRT-PCR was run in three biological replicates with two technical replicates. For the normalization of gene, citrus actin (GU911361.1) was used as an internal standard and the sample from B-sufficient plants was used as reference sample, which was set to 1. Bars represent means ± SD. Different letters above the bars indicate a significant difference at P < 0.05.
Frontiers in Plant Science | www.frontiersin.org reported that transgenic tobacco plants over-expressing a zinc finger protein (ZFP) gene from rice had enhanced tolerance to cold, dehydration, and salt stress. Our results showed that HSF24like (TDF #R23-2a) and ZFP4 (TDF #R15-1a) were induced in B-deficient roots (Table 1), indicating a possible role of the two genes in B-deficiency-tolerance.
Plant methyl-CpG-binding domain (MBD) proteins, which control chromatin structure mediated by CpG methylation, play crucial roles in plant development (Grafi et al., 2007). Peng et al. (2006) observed that the mbd9 mutants had more shoot branches by increasing the outgrowth of axillary buds than wild-type Arabidopsis. Our results showed that B-deficiency down-regulated the expression of MBD-containing protein 9like isoform X2 (TDF#L175-1b) in leaves (Table 2), indicating that DNA methylation and plant development were probably impaired, thus increasing shoot branching. This agrees with the report that B-deficient plants displayed a relatively weak apical dominance, and a subsequent sprouting of lateral buds (Wang et al., 2006). Jacobs and Kück (2011) demonstrated that RNA-binding proteins (RBPs) could modulate chloroplast RNA stability and chloroplast splicing, facilitate RNA editing. Lezhneva and Meurer (2004) showed that photosystem I (PSI) function was specifically affected in the high Chl fluorescence (HCF) 145 mutant of Arabidopsis due to the lack of the two PSI core proteins. The defect seemed to be a result of increased instability of the tricistronic psaA, psaB, and rps14 transcript (Jacobs and Kück, 2011). We observed that leaf expression of RBPHCF152 (TDF#L253-3b, Table 2), which functions in the processing of polycistronic chloroplast psbB-psbT-psbH-petB-petD transcript, was repressed by B-deficiency. Thus, it is very likely that PSI was impaired in B-deficient leaves, as observed on B-deficient C. grandis leaves (Han et al., 2009).

Genes Involved in Signal Transduction
Protein phosphorylation and dephosphorylation play a key role in plant stress signal transduction pathways. Transgenic Arabidopsis over-expressing TaSnRK2.4 encoding SNF1-type serine/threonine protein kinase had enhanced tolerance to drought, salt, and freezing stresses (Mao et al., 2010). Arabidopsis dual-specificity protein tyrosine phosphatase 2 (AtDsPTP2) plays a role in the tolerance of plants to oxidative stress generated by ozone (Lee and Ellis, 2007). AtPTP1 has been suggested to function in stress responses of higher plants (Xu et al., 1998). Thus, the up-regulation of four TDFs related to phosphorylation (i.e., TDFs #L219-2a and L160-1a) and phosphatases (i.e., TDFs #L123-2a and L98-4a) ( Table 2) in B-deficient leaves indicated that protein phosphorylation and dephosphorylation probably played a role in B-deficiency-tolerance.

Genes Involved in Stress Response and Defense
Senescence is a genetically programmed process governed by the developmental age and induced by (a) biotic stresses. The up-regulation of plant senescence-associated protein (TDF #R100-1a) and putative senescence-associated protein (i.e., TDFs #R148-4a and L177-3a) in B-deficient roots and leaves (Tables 1, 2) indicated that senescence might be accelerated in these tissues. However, the expression of plant senescence-associated protein (TDF # L224-2b) was repressed in B-deficient leaves ( Table 2).
Germin-like proteins (GLPs), which have different enzyme functions, including oxalate oxidase (OXO) and superoxide dismutase (SOD), play a role in plant development and various abiotic stress responses (Dunwell et al., 2008). Transgenic Arabidopsis plants overexpressing Arachis hypogaea GLP2 and 3 displayed enhanced salt-tolerance . Plant heat shock protein 70s (HSP 70s) function in various cellular processes including protein import into organelles (Shi and Theg, 2010) and folding of de novo-synthesized polypeptides (Hartl, 1996). A nuclear-localized HSP70 conferred heat and drought tolerance on tobacco plants (Cho and Choi, 2009). The upregulation of GLP (TDF #L98-3a) and HSP70-binding protein 1-like (TDF #L235-2a) in B-deficient leaves demonstrated the possible involvement of the two genes in B-deficiency-tolerance.
In conclusion, most of the differentially expressed TDFs in stress defense (tolerance) were up-regulated in B-deficient leaves, while the amount of the down-regulated TDFs was more than that of the up-regulated ones in B-deficient roots (Tables 1, 2; Figure S3). Obviously, great difference existed in B-deficiency-responsive genes related to stress defense between roots and leaves.

Genes Involved in Cell Wall Modification
In cell wall, proline-rich proteins (PRPs) cross-link each other or link to other components (i.e., saccharides and lignin) to form effective protection layer after pathogen infection or wounding (Brisson et al., 1994). Zhou et al. (2015) reported that the expression levels of two PRP2 genes (i.e., JK817586 and JK817604) in B-deficiency-sensitive Poncirus trifoliate roots were down-regulated by 24 and/or 6 h B-deficient treatments, while Bdeficiency did not alter their expression in B-deficiency-tolerant Carrizo citrange (C. sinensis × P. trifoliata) roots. Thus, the upregulation of 14 kDa PRP DC2.15-like in B-deficient roots might contribute to the B-deficiency-tolerance by reinforcing cell walls.
Expansins, which enable the growing cell wall to extend, are considered to be crucial regulators of wall extension during growth (Link and Cosgrove, 1998). The down-regulation of expansin-like B1-like (TDF #R55-1b) in B-deficient roots (Table 1) indicated that root cell elongation might be impaired (Lee and Kende, 2002), thus inhibiting root growth ( Figure S1). Similar results have been obtained in B-deficient Arabidopsis roots (Camacho-Cristóbal et al., 2008b) and citrus roots (Zhou et al., 2015).
Laccase, which catalyzes the oxidation of phenolic substrates using oxygen as the electron acceptor, is required for lignin polymerization during vascular development in Arabidopsis (Zhao et al., 2013). Ranocha et al. (2002) reported that neither lignin level nor composition was affected due to a repression of laccase expression in poplar. However, the antisense transgenic population, lac3AS, had a two-to three-fold increase in total soluble phenolic level. In addition, lac3 suppression caused a dramatic alteration of xylem fiber cell walls. We observed that B-deficiency down-regulated the expression level of laccase-3like (TDF #L251-2b) in leaves (Table 2), thus increasing leaf concentration of total soluble phenolics and altering cell wall structure. This agrees with the reports that B-deficient C. sinensis leaves had higher level of total phenolics (Lu et al., 2014a) and that B-deficiency altered cell wall structure, thus causing growth defects of C. sinensis leaves (Liu et al., 2014).
Gibberellin (GA) 2-β-dioxygenase 2-like (GA2OX2; i.e., TDFs #R64-4a and L64-2a) functioned in the catabolism of GAs was induced in B-deficient roots and leaves (Tables 1, 2). This implied that the degradation of GAs might be enhanced in these tissues, thus reducing their concentrations. This agrees with the report that B-deficient plants had lower levels of GAs (Shi and Liu, 2002).

Conclusions
This is the first comparative investigation of B-deficiencyinduced alterations in gene expression profiles in C. sinensis roots and leaves using cDNA-AFLP. We isolated more up-regulated TDFs from B-deficient leaves than from B-deficient roots, and less down-regulated TDFs from the former than from the latter, which agrees with our report that mitochondrial respiration, OA metabolism, and AA biosynthesis are enhanced in B-deficient leaves with more accumulation of carbohydrates, but reduced in B-deficient roots with less accumulation of carbohydrates. The majority of B-deficiency-responsive TDFs were isolated only from roots or leaves, only seven TDFs with the same GenBank ID were isolated from the both. Furthermore, only three differentially expressed TDFs shared by the both displayed similar expression trend in response to B-deficiency. Besides, UGTs were repressed in B-deficient roots, but were induced in B-deficient leaves; however, TDFs related to ATP biosynthesis were up-regulated in the former, but were unaffected in the latter. Most of the differentially expressed TDFs associated with signal transduction and stress defense were down-regulated in B-deficient roots, but up-regulated in B-deficient leaves. Eight (one) TDFs related to protein ubiquitination and proteolysis were induced (inhibited) in B-deficient leaves, while only two down-regulated TDFs involved in ubiquitination were detected in B-deficient roots. Through integration of the present results and the previous data available in the literatures, we presented FIGURE 2 | A proposed model for the responses of C. sinensis leaves (A) and roots (B) to B-deficiency. 40SRP, 40S ribosomal protein; 60SRP, 60S ribosomal protein; ACY, Aminoacylase; ANH, Adenine nucleotide α hydrolases-like superfamily protein; ARP6, Auxin-resistance protein 6; ATS, Asx tRNA synthetase (AspRS/AsnRS) class II core domain-contating protein; BEH, Bifunctional epoxide hydrolase; BIG5, Brefeldin A-inhibited guanine nucleotide-exchange protein 5-like; CAB, Chlorophyll a-b binding protein; a possible model for the responses of C. sinensis roots and leaves to B-deficiency (Figure 2). Obviously, many differences existed in long-term B-deficiency-responsive genes between roots and leaves. These findings presented an integrated view of the differential responses occurring in B-deficient roots and leaves and revealed new insight into the different adaptive mechanisms of C. sinensis roots and leaves to B-deficiency at the transcriptional level. Our results are also useful for obtaining the key genes responsible for citrus B-deficiency-tolerance and for improving citrus productivity and quality. Therefore, these results are of great importance from the citrus breeding and production point of view.