Adaptive Responses of Hormones to Nitrogen Deficiency in Citrus sinensis Leaves and Roots

Some citrus orchards in China often experience nitrogen (N) deficiency. For the first time, targeted metabolomics was used to examine N-deficient effects on hormones in sweet orange (Citrus sinensis (L.) Osbeck cv. Xuegan) leaves and roots. The purpose was to validate the hypothesis that hormones play a role in N deficiency tolerance by regulating root/shoot dry weight ratio (R/S), root system architecture (RSA), and leaf and root senescence. N deficiency-induced decreases in gibberellins and indole-3-acetic acid (IAA) levels and increases in cis(+)-12-oxophytodienoic acid (OPDA) levels, ethylene production, and salicylic acid (SA) biosynthesis might contribute to reduced growth and accelerated senescence in leaves. The increased ethylene formation in N-deficient leaves might be caused by increased 1-aminocyclopropanecarboxylic acid and OPDA and decreased abscisic acid (ABA). N deficiency increased R/S, altered RSA, and delayed root senescence by lowering cytokinins, jasmonic acid, OPDA, and ABA levels and ethylene and SA biosynthesis, increasing 5-deoxystrigol levels, and maintaining IAA and gibberellin homeostasis. The unchanged IAA concentration in N-deficient roots involved increased leaf-to-root IAA transport. The different responses of leaf and root hormones to N deficiency might be involved in the regulation of R/S, RSA, and leaf and root senescence, thus improving N use efficiency, N remobilization efficiency, and the ability to acquire N, and hence conferring N deficiency tolerance.


Introduction
Nitrogen (N) is one of the key factors affecting crop growth and yield [1,2].Some citrus orchards often experience N deficiency.In the main citrus producing areas of China, 59.8% of the orchards were deficient in soil-available N [3].
Evidence shows that N deficiency-induced increases in root/shoot dry weight (DW) ratio (R/S), alterations in root system architecture (RSA), and delaying in root senescence play a role in plant N deficiency tolerance [4][5][6].Hormones are involved in diverse aspects of plant growth and development [7].Therefore, these changes caused by N deficiency should be related to the changes in hormones in roots and shoots (leaves) [8][9][10].In maize, Tian et al. [11] observed that N supply decreased indole-3-acetic acid (IAA) concentration in roots and that the inhibitory impact of high N on primary root growth could be restored by the application of IAA.Further study indicated that low N increased shoot-to-root IAA transport, thereby enhancing IAA concentration in root tips which promoted root elongation through auxin (AUX)-induced acid growth and AUX-regulated targeting of the rapamycin pathway in maize, and that when shoot-to-root IAA transport was hindered, the increase in root elongation induced by low N was greatly repressed [12].
Plants 2024, 13,1925 2 of 25 The application of ethylene (ETH) can reduce root growth.N deprivation lowers ETH formation in maize roots [9].Cytokinins (CKs) repress root growth and promote shoot growth [13].High nitrate may enhance CK accumulation in roots, which can increase ETH biosynthesis and hinder AUX action, thereby lowering root elongation [9].However, in pumpkins, Mardanov et al. [14] observed that N-starved roots had increased levels of CKs, and reduced levels of IAA and abscisic acid (ABA), and N-starved shoots had increased concentrations of ABA and reduced concentrations of CKs and IAA, concluding that N starvation induced an increase in root growth and a decrease in shoot growth was related to a high CK/ABA ratio in roots and a low CK/ABA ratio in shoots.
The first characteristic physiological impact of jasmonic acid (JA) is to inhibit plant growth [15].ABA is regarded as a growth inhibitor [16].High concentrations of salicylic acid (SA) have an inhibitory effect on plant growth [17].In tomatoes, N deficiency increased the levels of ABA, IAA, SA, jasmonoyl-L-isoleucine (JA-ILE), and JA in the xylem sap, and decreased leaf area and plant biomass [18].
Camut et al. [1] showed that nitrate promoted leaf sheath elongation in wheat, and stem elongation and shoot branching in Arabidopsis through improving gibberellin (GA) biosynthesis and accumulation.N deficiency lowered the concentrations of GA 1 , GA 4 , and GA 3 in maize roots [19], and increased the concentration of GA 3 in wheat roots [20].
Arabidopsis strigolactone (SL)-deficient mutants had reduced length of the primary roots and number of cells in the meristem zone than the wild-type plants.A low concentration of the SL analog GR24 increased the two parameters in the SL-deficient mutants and wild-type plants [9].Growing evidence shows that the elevated accumulation and exudation of SLs in N-deficient roots is a key adaption, because SLs may enhance N uptake through altering RSA [21,22].
Leaf senescence plays a key role in the economy and recycling of N in plants [23].Studies suggested that N deficiency-induced leaf senescence was an adaptive response to N deficiency by enhancing N remobilization from senescing to N-demanding tissues [24,25].ABA, SA, ETH, JA, and SLs can promote leaf senescence, while CKs can delay leaf senescence [26].The roles of AUXs and GAs in leaf senescence are elusive [27,28].Zakari et al. [29] indicated that the elevated levels of ABA and reactive oxygen species in N-deficient rice leaves was the main cause for accelerating leaf senescence and that the application of ABA accelerated the senescence of rice leaves.However, Oka et al. [23] demonstrated that ABA accumulated in N-limited cucumber shoots and that the application of ABA inhibited leaf senescence of cucumber plants grown under N limitation.N starvation reduced ABA levels in tomato leaves [30] and did not alter ABA levels in Arabidopsis shoots [5].Application of transzeatin (tZ) repressed N deficiency-induced senescence in wheat leaves [31].N deficiency increased ETH evolution in mustard leaves [32], SA levels in Arabidopsis shoots [5], (+)-5-deoxystrigol (5DS) levels in grape leaves [33], and wound-induced accumulation of JA in maize leaves [34].However, N starvation decreased JA concentrations in Arabidopsis shoots [5] and SA and OPDA levels in tomato leaves [30].
Although some researchers have investigated plant hormone responses to N deficiency, most of the work has focused on model plants and herbs, and the results are inconsistent.To date, very little is known about the impacts of N deficiency on hormones in woody horticultural plants [9].Additionally, most researchers only examined the effects of N deficiency on one or several hormones in roots or leaves [8,32,35,36], which may miss a comprehensive investigation on the contribution of hormones in the N deficiency tolerance of plants.To our knowledge, data on the effects of N deficiency on hormone metabolomics in leaves and roots are rare.
Recent work from our laboratory showed that N deficiency increased R/S and rootsurface-per-unit volume, delayed root senescence, and accelerated leaf senescence in sweet orange (Citrus sinensis (L.) Osbeck cv.Xuegan) seedlings, thereby improving N remobilization efficiency (NRE), N use efficiency (NUE), and the ability to acquire N, and hence conferring N deficiency tolerance [2,4,6,25,37].Based on the previous studies, we used targeted metabolomics to examine N-deficient effects on hormones in 'Xuegan' leaves and Plants 2024, 13,1925 3 of 25 roots.Also, we used RNA-Seq to investigate N-deficiency-responsive genes related to hormone metabolism in leaves and roots.The aims were (a) to validate the hypothesis that hormones are involved in citrus N deficiency tolerance by regulating R/S, RSA, and leaf and root senescence and (b) to reveal the differences in hormone changes caused by N deficiency in leaves and roots.

Effects of N Supply on Hormones and Related Metabolites (HRMs) in Leaves and Roots
We tested 88 HRMs in leaves and roots (Table S1), 80 of which were detected in leaves and/or roots, including 48, 53, 55, and 65 HRMs in the leaves of 0 mM N-treated seedlings (LN0), leaves of 15 mM N-treated seedlings (LN15), roots of 0 mM N-treated seedlings (RN0), and roots of 15 mM N-treated seedlings (RN15), respectively.Among the 80 HRMs, three, one, two, and six HRMs were detected only in LN0, LN15, RN0, and RN15, respectively.Thirty HRMs were simultaneously detected among the four samples (Table 1 and Figure 1A).Different letters behind the values in the same row for the same tissue (leaves or tissue) represent significant differences at p < 0.05.ND, not detected.An HRM was considered downregulated or upregulated when it had both a relative change of less or more than 1 and a p < 0.05.Also, an HRM was considered upregulated or downregulated when it was detected only in RN0 (LN0) or RN15 (LN15), respectively, in RN0 vs. RN15 (LN0 vs. LN15).
We detected 59 HRMs in LN0 and/or LN15, including 2 ABA and its metabolic products (hereafter referred to as abscisates (ABAs)), 17 AUXs, 22 CKs, ACC, 4 GAs, 9 JAs, 2 SLs, and 2 SAs.Isoprenoid and aromatic CKs are present in plants, with the former being more common and having higher concentrations than the latter [38].Among these 22 CKs, 14 belong to isoprenoid CKs and 8 belong to aromatic CKs.A total of 6 and 11 HRMs were detected only in LN0 and LN15, respectively.There were 42 common HRMs between the two (Table 1 and Figure 1A).
In leaves, N0 increased the tZOG, 2MeScZR, and iP7G concentrations relative to N15, decreased the BAP and BAP7G concentrations, and did not significantly affect the tZR, cZROG, cZ9G, DHZROG, iP9G, mT, oTR, and pT9G concentrations.cZR, DHZR, DHZ7G, K, and mTR were detected only in LN0, while iP, iPR, and 2MeSiP were detected only in LN15 (Table 1).Different letters behind the values in the same row for the same tissue (leaves or tissue) represent significant differences at p < 0.05.ND, not detected.An HRM was considered downregulated or upregulated when it had both a relative change of less or more than 1 and a p < 0.05.Also, an HRM was considered upregulated or downregulated when it was detected only in RN0 (LN0) or RN15 (LN15), respectively, in RN0 vs. RN15 (LN0 vs. LN15).We detected 59 HRMs in LN0 and/or LN15, including 2 ABA and its metabolic products (hereafter referred to as abscisates (ABAs)), 17 AUXs, 22 CKs, ACC, 4 GAs, 9 JAs, 2 SLs, and 2 SAs.Isoprenoid and aromatic CKs are present in plants, with the former being more common and having higher concentrations than the latter [38].Among these 22 CKs, 14 belong to isoprenoid CKs and 8 belong to aromatic CKs.A total of 6 and 11 HRMs were detected only in LN0 and LN15, respectively.There were 42 common HRMs between the two (Table 1 and Figure 1A).
In leaves, N0 decreased the ABA concentration relative to N15, increased the ACC concentration, and did not significantly affect the ABA-GE, SA, SAG, 5DS, and ST concentrations (Table 1).In leaves, GA 1 , GA 3 , GA 9 , and GA 53 were detected in LN0.N0 increased the OPDA concentration relative to N15, decreased the H2JA and JA-Val concentrations, and did not significantly alter the JA, MEJA, OPC-6, OPC-4, JA-ILE, and JA-Phe concentrations (Table 1).
In leaves, N0 decreased the ABA concentration relative to N15, increased the ACC concentration, and did not significantly affect the ABA-GE, SA, SAG, 5DS, and ST concentrations (Table 1).
In roots, N0 increased the GA 15 and GA 24 concentrations relative to N15, and had no significant impacts on the GA 1 , GA 4 , and GA 19 concentrations.GA 53 and GA 3 were detected only in RN15.N0 decreased the JA, OPDA, H2JA, and JA-Val concentrations relative to N15, and did not significantly alter the MEJA, OPC-6, OPC-4, and JA-ILE concentrations (Table 1).
In roots, N0 decreased the ABA and SAG concentrations relative to N15, and increased the 5DS and ACC concentrations, and did not significantly affect the SA and ST concentrations.ABA-GE was detected only in N0 (Table 1).

Principal Coordinate Analysis (PCoA) of HRMs in Leaves and Roots
To reveal the response patterns of HRMs in leaves and roots to N deficiency, a PCoA was conducted using the 80 HRMs detected in RN0, RN15, LN0, and/or LN15 (Table 1).The PCoA indicated that the four treatments, namely RN0, RN15, LN0, and L10 were separated, while the three replications per treatment were clustered.It was found that LN0 and LN15 (RN0 and RN15) were clustered on the right (left) side, suggesting that the responses of leaf and root HRMs to N deficiency were different.PCo2, which contributed 25.07% of the total variation, could separate N0 from N15 in leaves but not in roots (Figure 2).
Plants 2024, 13, 1925 10 of 25 Also, we identified upregulated methylesterase (MES) 17 (MES17; Cs7g29470) and short-chain dehydrogenase/reductase SDRA (Cs2g02410) in RN0 vs. RN15 (Table S2).MES hydrolyzes MEIAA to IAA [50].SDRA is involved in the β-oxidation of IBA to form IAA [51].N deprivation promoted shoot-to-root IAA transport [12].Taken together, the N deficiency-induced reduction of IAA levels in leaves might be due to elevated leaf-to-root IAA transport, increased formation of the IAA analog (ILE), and reduced biosynthesis caused by reduced precursors (IAN), while the unchanged IAA level in RN0 might be due to increased leaf-to-root IAA transport, β-oxidation of IBA, biosynthesis of IAA via hydrolysis of IAA conjugates (IAA-Leu and IAA-Ala) and precursor/storage (MEIAA), oxidation of IAAld, and reduced formation of the IAA analog (ILE) (Figure 3).Tian et al. [11] reported that low nitrate increased IAA levels in maize roots and that high nitrate-induced inhibition of root growth could be restored by the application of IAA.Sun et al. [12] demonstrated that N deprivation improved shoot-to-root AUX transport, thus enhancing root elongation in maize seedlings.Rampey et al. [49] observed that both IAA-Leu and IAA-Ala could inhibit the root elongation of A. thaliana seedlings.These results indicated that N deficiency increased leaf-to-root IAA transport and maintained IAA homeostasis in roots, and increased the hydrolysis of IAA-Leu and IAA-Ala and decreased their levels in roots, thus increasing R/S [37].

The Responses of CKs in Leaves and Roots to N Deficiency
Free CK bases are biologically active in plants, including tZ, cZ, iP, dihydrozeatin (DZ), BAP, K, mT, oT, and pT, but once they are ribosylated, they are converted to transport forms with weaker CK activities [38,52].As shown in Table 1 and Figure 4, we obtained downregulated iP and BAP in LN0 vs. LN15, and K was detected only in LN0; downregulated iP in RN0 vs. RN15 and tZ and pT (oT) were detected only in RN15 (RN0).Generally viewed, N0 decreased the accumulation of free CK bases in leaves and roots relative to N15.To become bioactive, CK nucleotides are converted to nucleobase forms by dephosphorylation and deribosylation or are directly catalyzed to the bioactive forms by the CK riboside 5 ′ -monophosphate phosphoribohydrolase called LOG [38,53].CK dehydrogenase (CKX) is involved in the degradation of CKs by cleavage of the side chain.tRNA isopentenyltransferase (tRNA-IPT) is involved in cZ biosynthesis.CK hydroxylase catalyzes the conversion of iP nucleotides to tZ nucleotides.Zeatin O-glucosyltransferase (ZOGT) catalyzes the conversion of CKs from bioactive to inactive forms [38].We obtained 1 upregulated LOG1 (Cs3g25930) involved in the CK biosynthetic process and 1 upregulated (LOG1) and 2 downregulated [CK dehydrogenase 7 (CKX7; Cs4g06150) and CKX1 (orange1.1t00627)]genes involved in the CK metabolic process in LN0 vs. LN15, as well as 2 upregulated genes [tRNA dimethylallyltransferase 2 (IPT2; Cs3g07650) and CK hydroxylase (CYP735A1; Cs9g17460)] involved in the CK biosynthetic process and 5 upregulated genes [IPT2, CYP735A1, CKX3 (Cs4g14450), and 2 ZOGT (Cs4g16640 and orange1.1t02338)]involved in the CK metabolic process in RN0 vs. RN15 (Table S2).The downregulation of iP in LN0 vs. LN15 might be caused by the decreased biosynthesis via the dephosphorylation and deribosylation of iPRMP, as indicated by the downregulation of iPR, and the increased formation of iP-conjugates (iP7G) (Figure 4).The downregulation of iP (tZ) in RN0 vs. RN15 might be caused by the decreased biosynthesis via the dephosphorylation and deribosylation of iPRMP (tZPRM), as indicated by the downregulation of iPR (tZR), and the increased degradation of iP (tZ) catalyzed by CKX, as indicated by the upregulated expression of CKX3 (Figure 4).Also, the increased formation of tZ conjugates (tZOG) catalyzed by ZOGT might contribute to the downregulation of tZ in RN0 vs. RN15 (Figure 4).However, the elevated accumulation of tZOG in LN0 vs. LN15 could not be explained in this way.CKX-overexpressing transgenic tobacco and Arabidopsis plants had increased breakdown of CKs and reduced accumulation of CKs (iP and zeatin), exhibiting reduced shoot growth and increased root growth [13].The vegetative characteristics of transgenic maize overexpressing a zeatin O-glucosyltransferase gene (ZOG1) from Phaseolus lunatus resembled those of CK-deficient plants, including decreased shoot growth, and increased root mass and branching [54].Wang et al. [55] suggested that an N starvation-induced decrease in CK (zeatin) levels in rice roots improved seminal root growth by enhancing meristem cell proliferation and cell elongation.Tian et al. [56] reported that low nitrate reduced the level of zeatin + zeatin nucleotide in the roots of maize inbred line 478, but not in the roots of maize inbred line Wu312, that nitrate with a concentration of 5 mM or higher inhibited root elongation in 478, whereas root elongation in Wu312 was repressed only at 20 mM nitrate, and that exogenous 6-benzylaminopurine (6-BA) completely counteracted the stimulating action of low nitrate on root elongation.Gan and Amasino [57] showed that overexpressing IPT in A. thaliana inhibited leaf senescence by increasing CK (iP) levels.These results suggested that N0 decreased the levels of CKs in leaves (iP and BAP) and roots (iP, tZ, and pT), thus inhibiting leaf growth, accelerating leaf senescence, and increasing R/S [37].
Unlike in leaves, we detected 9 upregulated and 2 downregulated genes involved in the GA biosynthetic process in RN0 vs. RN15 (Table S2).Achard et al. [62] found that the Arabidopsis serine/threonine-protein kinase ctr1-1 mutant contained decreased concentrations of bioactive GAs (GA 1 and GA 4 ) and increased concentrations of some intermediate GAs (GA 24 and GA 53 ) accompanied by smaller vegetative rosette sizes and shorter petiole lengths.Overexpression of AtCPS and AtKS in Arabidopsis led to an increase in ent kaurene production but did not result in an increase in bioactive GA production [63].Overexpression of GA20 ox conferred a GA-overproduction phenotype in Arabidopsis accompanied by increased concentrations of GA 1 , GA 9 , and GA 20 and unaltered concentrations of GA 4 [64].Radi et al. [65] found that overexpression of a pumpkin GA3 ox1 in Arabidopsis led to increased GA 4 concentrations.Overexpression of a pea GA3 ox1 in tobacco plants resulted in an increase in GA 1 levels [66].Vidal et al. [67] indicated that overexpression of a citrus GA20 ox1 in tobacco enhanced the biosynthesis of GAs through the non-13-hydroxylation pathway but decreased the biosynthesis through the early-13-hydroxylation pathway, thus causing GA 4 to become the main bioactive GA in the transgenic tobacco plants.We found that the GA 4 and GA 1 concentrations displayed an upward trend in RN0 vs. RN15, while the GA 3 level was significantly decreased in RN0 vs. RN15 and the level of GA 1 (GA 4 ) in RN0 was much higher than that of GA 3 .These results suggested that the upregulation of CTR1 (Cs4g05990), CPS, chloroplastic (Cs5g31210), KS (orange1.1t03278),KO1 (orange1.1t01909and orange1.1t01910),GA20 ox1 (Cs9g16520), GA20 ox2 (orange1.1t00272),and GA3 ox1 (Cs4g20350) in RN0 vs. RN15 might lead to a redirection of GA biosynthesis to GA 1 and GA 4 , thereby preventing their decline and maintaining them at higher levels.Lv et al. [20] reported that N limitation increased GA 3 concentrations in wheat roots.However, Wang et al. [19] reported that N deprivation reduced the levels of GA 1 , GA 3 , and GA 4 in maize roots.GA deficiency can reduce crop NUE [68].Therefore, the relatively stable GA levels in RN0 vs. RN15 might be beneficial for improving R/S, NUE, and the ability to acquire N.Evidence shows that IAA can promote GA biosynthesis and accumulation [69].The different responses of leaf and root GAs to N deficiency agreed with our results that LN0 reduced IAA concentration in leaves but did not significantly alter its concentration in roots (Table 1).

The Responses of JAs in Leaves and Roots to N Deficiency
We detected two upregulated (4-coumarate--CoA ligase-like 5 (4CLL5; Cs7g21790) and 4CLL9 (Cs5g06990)) and six downregulated (four allene oxide cyclase (AOC), chloroplastic (Cs3g06080, Cs6g18900, Cs6g18910, and novel.2517),lipoxygenasem (LOX) 2, chloroplastic (LOX2; novel.2136),and Protein GRIM REAPER (GRI; novel.1617))genes involved in the JA biosynthetic process in RN0 vs. RN15, and upregulated 4CLL5 (Cs3g27010) and 4CLL9 (Cs5g06980) in LN0 vs. LN15 (Table S2).Transgenic tomato plants silencing loxD had lower LOX activity and JA concentration [70].Hazman et al. [71] reported that two aoc rice mutants (hebiba and cpm2) had decreased levels of JA, OPDA, and JA-ILE than those of their wild-type plants under control conditions and 6 h of salinity stress with the exception that the OPDA level did not differ between the two under the control conditions.Silencing of Gh4CL30 in cotton inhibited JA biosynthesis in the uninfected plants but not in the Verticillium dahlia-infected plants [72].Therefore, JA biosynthesis might be downregulated in RN0 vs. RN15.This was supported by our results that RN0 had lower concentrations of JA, OPDA, H2JA, and JA-Val than RN15 (Table 1 and Figure 6).Although the expression levels of 4CLL5 and 4CLL9 were upregulated in LN0 vs. LN15, N0 only increased the concentration of OPDA and decreased or did not alter the concentrations of the other JAs.This suggested that the biosynthesis of JAs was not upregulated in LN0 vs. LN15, but it might be redirected to OPDA biosynthesis, and that the 4CLL5 and 4CLL9 were not the rate limiting enzymes for the biosynthesis of JAs in leaves.It is known that JA can inhibit plant growth [15].The downregulated JA in RN0 vs. RN15 and the unaltered JA in LN0 vs. LN15 suggested that N deficiency promoted root growth (relative to shoot growth), thus increasing R/S [37].The current finding that LN0 had increased OPDA levels and unaltered JA and JA-ILE levels relative to LN15 agreed with the findings that drought increased the accumulation of OPDA in Arabidopsis seedlings rather than JA [74], and that salt-stress increased OPDA levels but did not affect JA and JA-ILE levels in rice shoots [71].A study indicated that OPDA negatively regulated Arabidopsis stomatal opening independently and synergistically with ABA, and that the opr3 mutant lines with increased accumulation of OPDA  The current finding that LN0 had increased OPDA levels and unaltered JA and JA-ILE levels relative to LN15 agreed with the findings that drought increased the accumulation of OPDA in Arabidopsis seedlings rather than JA [74], and that salt-stress increased OPDA levels but did not affect JA and JA-ILE levels in rice shoots [71].A study indicated that OPDA negatively regulated Arabidopsis stomatal opening independently and synergistically with ABA, and that the opr3 mutant lines with increased accumulation of OPDA displayed reduced stomatal aperture and enhanced drought tolerance [74].Thus, the increased accumulation of OPDA in LN0 might cause stomatal closure [37], thus lowering transpiration water loss and enhancing N deficiency tolerance.Zhang and Turner [75] reported that exogenous OPDA reduced the leaf area of the wild-type Arabidopsis plants, but not of the opr3 plants, concluding that the OPDA-induced reduction in the leaf area of the wild-type plants might be due to its conversion to JA or one of its metabolites.However, Mueller et al. [76] reported that OPDA reduced the overall growth of wild-type Arabidopsis seedlings, particularly root growth.Root growth inhibition by OPDA was less in the opr3 than in the wild-type.Wu et al. [77] indicated that in rice seedlings, exogenous applications of MeJA lowered the uptake of N by the roots and the root-to-shoot translocation of recently-absorbed 15 N.
To conclude, N deficiency reduced the biosynthesis and accumulation of JAs, while it might redirect to the OPDA biosynthesis.The different operations of the JA biosynthesis pathway in N-deficient leaves and roots might confer citrus N deficiency tolerance by improving R/S, NUE, N distribution in roots, and the ability to acquire N [2,37].

The Responses of ETH in Leaves and Roots to N Deficiency
It was reported that the overexpression of the citrus ACC synthase (ACS) gene (CiACS4) led to a dwarfing phenotype, increased ETH release, and decreased IAA, GA 3 , and GA 20 concentrations in transgenic lemon (Citrus limon) plants, while the inhibition of CiACS4 expression resulted in exactly the opposite [78].Wi and Park [79] reported that the antisense expression of carnation ACS or ACC oxidase (ACO) attenuated stress-induced senescence by lowering ETH formation in transgenic tobacco plants.Therefore, N0 might reduce ETH biosynthesis in roots, as indicated by the downregulated expression of ACS8 (Cs3g16400), ACO (Cs1g15700), and ACO5 (Cs3g20140; Table S2).As well as this, the upregulation of CTR1 (a negative regulator of ETH response) in RN0 vs. RN15 (Table S2) might contribute to the lower ETH formation.This agreed with our results that N0 increased or did not affect the concentrations of 5 out of 7 GAs in roots (Table 1), and with the reports that ETH formation was lowered in N-starved maize roots [8].Tian et al. [36] reported that in Arabidopsis, the high N-induced downregulation of the high affinity nitrate transporter NRT2.1 and the upregulation of the low affinity nitrate transporter NRT1.1 in roots could be inhibited by the ETH synthesis antagonist, aminoethoxyvinylglycine.In Arabidopsis roots, NRT2.1 expression was upregulated by aminoethoxyvinylglycine and N starvation and downregulated by ETH [80].Zhang et al. [81] indicated that the application of ethephon reduced maize N uptake.Léran et al. [82] showed that Arabidopsis NRT1.1 was a bidirectional transporter and was involved in root-to-shoot nitrate transport.Recently, it was found in our laboratory that N0 upregulated the expression of NRT2.4 (Cs8g16010 and orange1.1t02415)and NRT2.5 (Cs7g09040) and downregulated the expression of NRT1.1 (Cs5g09050) in roots [6].These results suggested that N0 decreased ETH biosynthesis, thereby delaying root senescence and improving the R/S, the capacity of roots to acquire N and the distribution of N in roots [37].The increased accumulation of ACC in RN0 vs. RN15 (Table 1) might be mainly caused by its decreased utilization for ETH biosynthesis, as indicated by the downregulated expression of ACO and ACO5 and the upregulated expression of CTR1.This suggested that ACS8 was not the rate limiting enzyme for ACC biosynthesis in RN0.
Rauf et al. [83] found that in A. thaliana, waterlogging triggered the NAC transcription factor Speedy Hyponastic Growth (SHYG), which then activated the expression of ACO5 and ASC, thus increasing ETH biosynthesis.Here, N0 upregulated the expression of ACS2 (orange1.1t00414),ACO (Cs2g20590), and SHYG (Cs5g26130) and increased the levels of ACC in leaves (Table S2), thereby enhancing ETH formation and release.Varsani et al. [84] indicated that OPDA enhanced the expression of ETH biosynthesis genes (ACS2, ACS6, and ACO15) in maize plants.Thus, the upregulation of OPDA in LN0 vs. LN15 might contribute to the increased ETH biosynthesis.The increase in ETH biosynthesis in LN0 agreed with our results that N0 decreased the levels of GA 1 , GA 3 , GA 9 , and GA 53 in leaves (Table 1), and with the reports that N deficiency increased ACS activity and ETH evolution in mustard leaves [32], and N deficiency accelerated C. sinensis leaf senescence [25].However, Iqbal et al. [32] reported that the application of ethephon increased leaf stomatal conductance, photosynthesis, area, and plant growth in mustard.Thus, the increase in ETH formation in LN0 might be an adaptive response to N deficiency.

The Responses of ABAs in Leaves and Roots to N Deficiency
We identified 2 downregulated (zeaxanthin epoxidase (ZEP), chloroplastic (Cs1g22620) and NDR1/HIN1-like protein 6 (NHL6; novel.774))and 6 upregulated (3 ZEP, chloroplastic (or-ange1.1t04051,orange1.1t05125and orange1.1t04849), 2 AO1 (Cs8g13760 and Cs8g13770), and molybdenum cofactor sulfurase (ABA3; Cs4g12520)) genes involved in the ABA biosynthetic process, as well as 1 downregulated (Cs3g23530) and 2 upregulated (Cs3g21210 and Cs1g24480) ABA 8 ′ -hydroxylases 3 (CYP707A3) genes involved in the oxidative degradation of ABA in RN0 vs. RN15 (Table S2).Transgenic Nicotiana plumbaginifolia plants overexpressing a ZEP and expressing an antisense ZEP displayed elevated and decreased ABA concentrations in transgenic seeds, respectively [85].The final step in ABA biosynthesis is the oxidation of abscisic aldehyde to ABA.This reaction is catalyzed by abscisic aldehyde oxidase 3 (AAO3), which requires a molybdenum cofactor that is biosynthesized by the ABA3 [86].This reaction can also be catalyzed by AO [87].Overexpression of NHL6 increased the levels of ABA in transgenic Arabidopsis plants, but the levels of ABA did not differ between the nhl6 and the wild-type plants [88].Overexpression of PvCYP707As in Nicotiana sylvestris led to increased phaseic acid (PA) concentrations but decreased ABA concentrations in transgenic plants [89].Therefore, the decreased ABA levels in RN0 vs. RN15 (Table 1) might be caused by the elevated degradation, rather than by the decreased biosynthesis (Figure 7).The decrease in ABA levels in RN0 might also be due to the increased formation of the inactive storage form ABA-GE by the glucosylation of ABA and the decreased release of ABA from ABA-GE, because ABA-GE was detected only in RN0 (Table 1). Lee et al. [90] reported that under low nitrate conditions, the NITROGEN RESPONSE DEFICIENCY 1 (nid1) knockout Arabidopsis mutants displayed longer primary roots and more lateral roots than the Col-0 plants but lower accumulation of ABA.The decrease in ABA levels in RN0 might contribute to the increase in R/S during N deficiency.
Here, we identified 3 downregulated genes (2 ZEP, chloroplastic (Cs4g20590 and Cs4g20560) and 9-cis-epoxycarotenoid dioxygenase NCED1, chloroplastic (Cs2g03270)) involved in the ABA biosynthetic process in LN0 vs. LN15 (Table S2).Tan et al. [93] found that the ABA levels were lower in the maize vp14 embryos than in the wild-type embryos, and that the vp14 phenotype could be rescued by ABA.The gene encodes NCED1.Thus, the decrease in ABA levels in LN0 vs. LN15 might be caused by a decrease in ABA biosynthesis.This agreed with the reports that ABA levels in Catalpa bungei, rice, and pea leaves did not significantly alter under N deficiency but displayed a decreased trend [94][95][96].Studies indicated that ABA addition could indirectly enhance leaf growth via reducing ETH formation [97], or directly repress leaf expansion via lowering cell wall extensibility [98].In cucumbers, Oka et al. [23] found that ABA inhibited leaf senescence under N starvation.Taken together, N deficiency caused a decrease in ABA levels and an increase in ETH biosynthesis in LN0 vs. LN15, thus reducing leaf growth and accelerating leaf senescence.However, Zakari et al. [29] indicated that the N deficiency-induced increase in ABA levels in rice leaves was the important factor for the accelerated leaf senescence under N deprivation, and that the exogenous application of ABA could accelerate rice leaf senescence.
Here, we identified 3 downregulated genes (2 ZEP, chloroplastic (Cs4g20590 and Cs4g20560) and 9-cis-epoxycarotenoid dioxygenase NCED1, chloroplastic (Cs2g03270)) involved in the ABA biosynthetic process in LN0 vs. LN15 (Table S2).Tan et al. [93] found that the ABA levels were lower in the maize vp14 embryos than in the wild-type embryos, and that the vp14 phenotype could be rescued by ABA.The gene encodes NCED1.Thus, the decrease in ABA levels in LN0 vs. LN15 might be caused by a decrease in ABA biosynthesis.This agreed with the reports that ABA levels in Catalpa bungei, rice, and pea leaves did not significantly alter under N deficiency but displayed a decreased trend [94][95][96].Studies indicated that ABA addition could indirectly enhance leaf growth via reducing ETH formation [97], or directly repress leaf expansion via lowering cell wall extensibility [98].In cucumbers, Oka et al. [23] found that ABA inhibited leaf senescence under N starvation.Taken together, N deficiency caused a decrease in ABA levels and an increase in ETH biosynthesis in LN0 vs. LN15, thus reducing leaf growth and accelerating leaf senescence.However, Zakari et al. [29] indicated that the N deficiency-induced in-Figure 7. A diagram displaying the mean relative changes of ABAs in leaves (LN0/LN15; green) and roots (RN0/RN15; magenta).Data from Table 1.An asterisk indicated a significant difference between LN0 (RN0) and LN15 (RN15) at p < 0.05.AAO3, abscisic-aldehyde oxidase; ABA2, xanthoxin dehydrogenase; ABA3, molybdenum cofactor sulfurase; CYP707A, ABA 8 ′ -hydroxylase; NCED, 9-cis-epoxycarotenoid dioxygenase; ZEP, zeaxanthin epoxidase (refer to Chen et al. [91] and Watanabe et al. [92]).

The Responses of SLs in Leaves and Roots to N Deficiency
As shown in Table S2, we identified 1 downregulated (carotenoid cleavage dioxygenase 8 (CCD8) homolog B, chloroplastic (CCD8b; Cs4g19470)) and 3 upregulated (β-carotene isomerase D27, chloroplastic (Cs5g30540), CCD7, chloroplastic (Cs1g25090) and CCD8b (Cs4g19460)) genes involved in SL biosynthesis, and 1 upregulated SL esterase D14 (Cs3g16030) and 1 downregulated probable strigolactone esterase DAD2 (Cs4g16860) gene involved in SL degradation in RN0 vs. RN15.A rice d14 mutant with an increased level of 2 ′ -epi-5deoxystrigol displayed enhanced outgrowth of tillers [99].Therefore, the elevated level of 5DS in RN0 vs. RN15 might be caused by the elevated biosynthesis and the reduced degradation.Additionally, the ST level in RN0 displayed an increased trend, whereas the 5DS and ST levels in LN0 displayed a decreased trend (Table 1).This agreed with the results that in sorghum, N starvation increased 5DS biosynthesis and accumulation in roots and its exudation by roots but it did not alter 5DS concentration in shoots [100], and with our results that the levels of GAs were reduced in LN0 vs. LN15, but generally displayed an increased trend in RN0 vs. RN15 (Table 1), because GA can improve SL biosynthesis by inducing the expression of D27 [101].SLs are biosynthesized mainly by roots [21].Sun et al. [10] demonstrated that N starvation-induced accumulation of SLs improved NUE.The different responses of leaf and root SLs to N deficiency might confer citrus N deficiency tolerance by increasing R/S, NUE, and the ability to acquire N.

The Responses of SAs in Leaves and Roots to N Deficiency
Xia et al. [102] showed that an Arabidopsis aspartic proteinase CDR1 (CDR1-D) dominant mutant had a dwarf stature and greatly elevated levels of SA and its glucoside than the wild-type plants.Lipase-like PAD4, serine/threonine-protein kinase PCRK1, and Protein SAR DEFICIENT 1 (SARD1) can upregulate isochorismate synthase 1 (ICS1), one key gene involved in SA biosynthesis [103].Arabidopsis plants overexpressing both PAD4 and ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1) were stunted compared to wild-type plants accompanied by increased accumulation of free SA [104].Vogelmann et al. [105] observed early senescence and cell death, as well as increased concentrations of SA + SA glucoside (total SAs) in Arabidopsis senescence-associated ubiquitin ligase1 (saul1) mutants but not in saul1-1/pad4 double.They suggested that the PAD4-dependent SA pathway was necessary for saul1 senescence and cell death.The upregulation of PAD4 (Cs1g08280), PCRK1 (Cs6g12050), SARD1 (Cs7g27120), and CDR1 (Cs5g18300) in LN0 vs. LN15 (Table S2) implied that the biosynthesis of SA might be upregulated in LN0.As shown in Table 1 and Table S2, we obtained 3 upregulated (2 Protein DMR6-LIKE OXYGENASE 2 (DLO2; Cs9g14480 and orange1.1t01963)and 1 Protein DOWNY MILDEW RESISTANCE 6 (DMR6; Cs5g16310)) and 1 downregulated (DLO1; Cs5g28730) gene involved in the SA catabolic process, 2 upregulated UDP-glycosyltransferase 74F2 (UGT74F2; Cs5g21200 and Cs5g21220) genes involved in the conversion (glucosylation) of SA to SAG and SA glucose ester (SGE) [106], and an unchanged SA level and an increasing trend in SAG level in LN0 vs. LN15.Zhang et al. [107] found that Arabidopsis salicylate 3-hydroxylase DLO1 (s3h) knockout mutants failed to convert SA to 2,3-dihydroxybenzoic acid (2,3-DHBA) sugar conjugates, contained high concentrations of SA and its sugar conjugates, and displayed accelerated leaf senescence, while the reverse was the case for the gain-of-function lines.Arabidopsis plants overexpressing UGT74F2 showed decreased accumulation of free SA and elevated susceptibility to Pseudomonas syringae [108].The upregulation of UGT74F2 (Cs5g21200 and Cs5g21220) and the downregulation of DLO1 (Cs5g28730) in LN0 vs. LN15 implied that the glucosylation and catabolism of SA might be upregulated and downregulated, respectively, in LN0.These results suggested that the increase in SA glycosylation induced by N0 offsets the increase in SA biosynthesis and decrease in SA catabolism induced by N0, thereby keeping the SA concentration in LN0 unchanged, and that the increased SA biosynthesis and glucosylation, reduced catabolism, and an increased trend in the level of SA + SAG (Table 1) might be responsible for the reduced growth and accelerated senescence in these leaves.Unlike the leaves, the biosynthesis of SA might be downregulated in RN0 vs. RN15, as indicated by the downregulated GRI involved in the SA biosynthetic process, as well as the 2 downregulated CDR1 (Cs5g18300 and Cs5g18330) genes (Table S2).Also, we obtained 3 upregulated (2 DLO1 (Cs5g28720 and Cs5g28750) and 1 DLO2 (Cs9g14500)) and 3 downregulated (2 DLO1 (Cs5g28710 and Cs1g12310) and 1 DLO2 (Cs5g28780)) genes involved in the SA catabolic process, as well as 2 upregulated (Cs2g18300 and novel.306)and 3 downregulated (Cs2g18240, Cs5g21200, and Cs5g21220) UGT74F2 genes in RN0 vs. RN15 (Table S2).These results suggested that the biosynthesis and glucosylation of SA might be reduced in RN0 vs. RN15, thereby keeping the SA concentration in RN0 unchanged while decreasing its SAG concentration (Table 1).It is known that moderate levels of SA promote root and shoot growth, while higher levels of SA have the opposite impact [17].Chen et al. [35] showed that low N increased the concentrations of SA in cotton main and lateral roots, that the main root average diameter (lateral root length) was negatively correlated with the SA concentration, and that moderate N levels improved root growth by lowering SA concentrations in the main and lateral roots.Conesa et al. [5] indicated that 50 µM of SA inhibited Arabidopsis root growth.Therefore, the N deficiency-induced downregulation of SA biosynthesis and glucosylation and reduction of SAG concentration in roots might be beneficial for root growth, thereby improving R/S.

Plant Materials and Treatments
Plant materials and treatments referred to Lai et al. [6].Six weeks after germination of sweet orange (Citrus sinensis (L.) Osbeck cv.Xuegan) seeds, the uniform seedlings were transplanted into 6 L flowerpots (2 plants per flowerpot) containing sand.The seedlings were grown in an unheated greenhouse at the Fujian Agriculture and Forestry University, Fuzhou, China (26 • 5 ′ N, 119 • 14 ′ E) under a natural light environment.The annual average temperature, relative humidity, and sunshine hours were ~20 • C, 76%, and 1600 h, respectively [109].One week after transplantation, the seedlings were supplied thrice weekly with quarter-, half-, and full-strength nutrient solutions for 1, 2, and 3 weeks, respectively, until dripping (~500 mL pot −1 ).The full-strength nutrient solution refers to the nutrient solution used for the 15 mM N treatment.Seven weeks after transplantation, they were provided thrice weekly with nutrient solutions containing 0 (N0) or 15 (N15 or control) mM N [i.e., macronutrients (Table 2) and micronutrients: 10 µM H 3 BO 3 , 20 µM Fe-EDTA, 2 µM MnCl 2 , 2 µM ZnSO 4 , 0.5 µM CuSO 4 , and 0.065 µM (NH 4 ) 6 Mo 7 O 24 ] until dripping.There were 10 flowerpots of seedlings per treatment arranged in a randomized design.After 10 weeks of N treatments, the recently fully expanded (~7-week-old) leaves and ~5 mm long white root apices were used for all determinations except for N in roots.Leaves (petioles, winged leaves, and midribs removed) and ~5 mm long white root apices were harvested at sunny noon and immediately frozen in liquid N 2 , then stored at −80 • C until extraction of hormones and RNA.

Extraction and Determination of HRMs in Leaves and Roots
Six frozen leaf samples (2 treatments × 3 biological replicates) and 6 frozen root samples (2 treatments × 3 biological replicates) were sent to Wuhan MetWare Biotechnology Co., Ltd.(Wuhan, China) (https://www.metware.cn/,accessed on 1 June 2023) for extraction and determination of HRMs.Equal amounts of leaves (roots) from 3 plants (1 per pot) were mixed as a biological replicate.After the frozen samples were ground to a fine powder in liquid N 2 using a mixer mill (30 Hz, 1 min; MM400, Retsch GmbH, Haan, Germany), 50 mg of the powder was accurately weighed into a 2 mL tube containing 1 mL of methanol/water/formic acid (15:4:1, v:v:v) and 10 µL of internal standard mixed solution (100 ng/mL).The mixture was vortexed for 10 min.After centrifugation at 4 • C and 16,000× g for 5 min, the supernatant was collected, evaporated with a CentriVap centrifuge concentrator (Labconco, Kansas City, MO, USA) to dryness, redissolved in 100 µL of 80% methanol (v/v), and filtered through a 0.22 µm filter.The yielding filtrate was used for the determination of HRMs [110].
The ESI source operation parameters were as follows: ion source, ESI+/−; source temperature, 550 • C; ion spray voltage (IS), −4500 V for negative ion mode and 5500 V for positive ion mode; and curtain gas (CUR), 35 psi.HRMs were analyzed using scheduled multiple reaction monitoring (MRM).Data acquisitions were performed using the Analyst 1.6.3software (Sciex).The Multiquant 3.0.3software (Sciex) was used to quantify all metabolites.Mass spectrometer parameters including the declustering potentials (DP) and collision energies (CE) for individual MRM transitions were obtained with further DP and CE optimization.A specific set of MRM transitions were monitored for each period according to the metabolites eluted within this period [112][113][114].

RNA-Seq and Analysis
A total of 6 frozen leaf samples (2 treatments × 3 biological replicates) and 6 frozen root samples (2 treatments × 3 biological replicates) were sent to Novogene Corporation (Beijing, China) for RNA extraction and RNA-Seq.The samples were sequenced on an Illumina (Illumina Inc., San Diego, CA, USA) Novaseq platform and 150 bp paired-end reads were generated [25].RNA was extracted using Recalcirtant Plant Total RNA Extraction Kit (Bioteke Corporation, Beijing, China) following the manufacturer's instructions [115].Highquality clean reads were mapped to the C. sinensis v2.0 genome (http://citrus.hzau.edu.cn/download.php,accessed on 1 June 2023) using HISAT2.DESeq2 1.22.2 by default was used for differential transcriptional analysis between two groups.The yielding p-values were adjusted using the Benjamini and Hochberg's method to control the false discovery rate (FDR).Genes with both FDR < 0.05 and |log 2 (fold change)| > 1 were considered differentially transcribed.The reliability of the RNA-Seq data was validated by qRT-PCR, which was run in 2 technical replicates and 3 biological replicates using U4/U6 small nuclear ribonucleoprotein PRP31 (Cs7g08440) and actin (Cs1g05000) as the internal standards [6,25].

Statistical Analysis
Results were the means ± SE of three replicates.The unpaired t-test was performed for comparison between means.The PCoA was implemented by ChiPlot (https://www.chiplot.online,accessed on 1 June 2023).

Conclusions
Our results indicated that N deficiency reduced leaf growth and accelerated leaf senescence by lowering the concentrations of GAs (GA 9 , GA 3 , GA 53 , and GA 1 ) and IAA, and increasing the concentrations of OPDA and the biosynthesis of ETH and SA.The increased formation of ETH in LN0 might be caused by decreased levels of ABA, and increased levels of ACC and OPDA.The increased level of OPDA in LN0 might lead to stomatal closure, thereby lowering transpiration water loss and conferring N deficiency tolerance.LN0 altered RSA, increased R/S, and delayed root senescence by lowering the levels of bioactive CKs, JA, OPDA, and ABA and the biosynthesis of ETH and SA, increasing the level of 5DS, and maintaining the homeostasis of IAA and GAs.The unchanged concentration of IAA in RN0 involved increased leaf-to-root IAA transport.Additionally, we obtained more DAHs and DTGs related to hormone metabolism in RN0 vs. RN15 (13 upregulated and 26 downregulated HRMs and 40 upregulated and 36 downregulated genes) than in LN0 vs. LN15 (16 upregulated and 16 downregulated HRMs and 23 upregulated and 10 downregulated genes).The different responses of leaf and root HRMs to N deficiency might be responsible for the increase in R/S and the modulations of RSA, thus improving NUE, NRE, and the ability to acquire N, and hence conferring citrus N deficiency tolerance (Figure 8).Therefore, our study provided the most comprehensive analysis of HRMs in response to N deficiency in citrus leaves and roots, and a foundation for the related study.

Plants 2024 , 24 Figure 2 .
Figure 2. A PCoA plot of HRMs detected in leaves and roots from N0-and N15-treated Citrus sinensis seedlings.

Figure 2 .
Figure 2. A PCoA plot of HRMs detected in leaves and roots from N0-and N15-treated Citrus sinensis seedlings.

Figure 5 .
Figure5.A diagram displaying the mean relative changes of GAs in leaves (LN0/LN15; green) and roots (RN0/RN15; magenta).Data from Table1.An asterisk indicated a significant difference between

Table 1 .
I Effects of N supply on mean (±SE, n = 3) concentrations (ng g −1 FW) of HRMs detected in Citrus sinensis leaves and roots.