Identification of cis-regulatory elements related to water-deficit and low-temperature stress within the promoter of Citrus sinensis APETALA1

Citrus flowering is promoted by water-deficit and low-temperature stress and inhibited by gibberellins, providing a unique system to investigate regulation of floral development. In silico analysis, using plant cis-acting regulatory element databases, PLACE and PlantCARE, was conducted to compare transcriptional regulation of LEAFY (LFY), APETALA1 (AP1), and TERMINAL FLOWER (TFL) by water deficit, low temperature, abscisic acid and gibberellins in C. sinensis, Arabidopsis thaliana and Populus trichocarpa. A striking enrichment of response elements upregulated by water deficit was found in the C. sinensis AP1 (CsAP1) promoter, but neither the A. thaliana AP1 (AtAP1), nor P. trichocarpa AP1-1 (PtAP1-1) promoter. Notably, a tandem array of three response elements, each containing a LFY binding site, coupling element 3 site (CE3), and dehydration-responsive element (DRE), was found within a 100-bp region of the CsAP1 promoter. The CE3 and DRE sites are associated, respectively, with abscisic acid (ABA)-dependent and ABA-independent signaling pathways induced by water deficit. The CE3 sequence is present in the AtAP1, but not PtAP1-1, promoter. The DRE site is lacking in both AtAP1 and PtAP1-1. Three LFY binding sites are located in the CsAP1 and AtAP1 promoters, with only one in PtAP1-1. Multiple C-repeat binding factor (CBF) response elements associated with low-temperature responses through an ABA-independent signaling pathway are present in the CsAP1 and AtAP1 promoters, but not PtAP1-1. The unique 100-bp regulon of the CsAP1 promoter suggests that flower formation in C. sinensis in response to water-deficit and low-temperature stress is mediated at AP1 through ABA-dependent and ABA-independent pathways.

Identification of cis-regulatory elements… of pathways regulating flowering, but also identified features distinct from A. thaliana.For example for Citrus spp., overexpression studies using the floral meristem identity genes LFY and AP1 from A. thaliana were able to promote early flowering when transformed into juvenile citrange rootstocks (C.sinensis L. Osbeck x Poncirus trifoliata L. Raf.) (Peña et al., 2001).In addition, CsLFY,CsAP1,CsTFL,C. unshiu FT (CiFT), and CsSOC1-like (CsSL1 and CsSL2) successfully complemented their respective A. thaliana mutants, thus demonstrating functional equivalence (Endo et al., 2005;Pillitteri et al., 2004aPillitteri et al., , 2004b;;Tan & Swain, 2007).Taken together, these results strongly suggest that key components of the Citrus floral development pathway are the same as those of A. thaliana.
However, distinct differences in the regulation of flowering between the two species have been found.In Washington navel orange (WNO) (C.sinensis), flowering is promoted by water-deficit (< -2.1 to -3.0 MPa for > 10 days) and low-temperature stress (day 15-18 ºC, night 10-13 ºC), with inflorescence number positively correlated with the increasing severity or duration of either stress up to 8 weeks (Chica & Albrigo, 2013a, 2013b;Lovatt et al., 1988;Pillitteri et al., 2004a;Tang & Lovatt, in press).In contrast, in A. thaliana, flowering time is either advanced or delayed by water deficit as part of the plant's drought escape response (Riboni et al., 2016;Schmalenbach et al., 2014).Recently published floral gene transcription data for WNO provided evidence that FT, SOC1, LFY and AP1 transcripts were present in buds that subsequently produced inflorescences, as well as in buds that continued vegetative shoot growth, rendering the two bud populations

INTRODUCTION
Floral development is the essential first step in fruit production.To be able to regulate floral intensity, it is imperative to understand the underlying processes governing floral induction and bud determinacy (irreversible commitment of the shoot apical meristem [SAM] to floral development).Knowledge of the molecular basis of these two events in floral development in Citrus spp. is largely built on results derived from the model plant Arabidopsis thaliana.Floral induction begins with endogenous or environmental cues that act on several genetic pathways to upregulate genes that promote flowering (FLOWERING LOCUS T [FT],

In silico identification of response elements and comparative analysis
In silico analyses were performed to identify cis-responsive elements in the promoters of genes encoding homologues of LFY, AP1, and TFL.Promoters were queried against two databases of plant cis-acting regulatory elements, PLACE and PlantCARE (Higo et al., 1999;Rombauts et al., 1999).Motifs > 6 bp were used in the analysis.Spatial patterning of putative response elements was visualized on respective promoters using Vector NTI Advance 10 software (Invitrogen, Carlsbad, CA).Unless otherwise noted, analyses used a threshold level of 100% maximum homology.In addition, MatInspector, which used position-weighted matrices (PWM), was used to identify transcription factor binding sites (Cartharius et al., 2005).When binding sites identified by MatInspector were redundant within PLACE and PlantCARE at thresholds > 75% core motif similarity indistinguishable at the level of gene transcription during early development (Chica & Albrigo, 2013a;Tang & Lovatt, in press).However, under low-temperature and water-deficit conditions that resulted in intense flowering in WNO, bud determinacy correlated with increasing AP1 and AP2 transcript levels in response to the stress treatments, but only after the stress was alleviated (Tang & Lovatt, in press).This observation is consistent with the critical role of the class A genes, AP1 and AP2, in sepal formation in the ABC model of floral organ specification in A. thaliana (Coen & Meyerowitz, 1991).In WNO, sepal formation is the developmental marker indicating bud determinacy (Lord & Eckard, 1987).In fact, in WNO, profuse flowering only occurred under conditions that increased AP1 and AP2 transcript accumulation to a level sufficient to result in full SEP, PI and AG expression (Tang & Lovatt, in press).
The role of TFL in Arabidopsis and Citrus may be similar.In A. thaliana TFL, maintains shoot indeterminacy by downregulating LFY and AP1 (Liljegren et al., 1999).In WNO, TFL was highly expressed in buds of seedling (juvenile) trees, which did not express LFY or AP1 and did not flower under low temperature; whereas TFL was expressed only at extremely low levels in buds of adult trees before, during and after a low-temperature treatment that increased LFY and AP1 expression and resulted in intense flowering (Pillitteri et al., 2004a).
Exposure of subtropical evergreen species, such as Citrus spp., to water deficit and low temperature constitutes a stress that elicits a range of biochemical, physiological, and molecular responses (Nakashima & Yamaguchi-Shinozaki, 2006).In A. thaliana, differential gene expression induced by water deficit occurs through the binding of upstream transcription regulators to response elements in the promoters of downstream target genes.This action occurs through ABA-dependent and ABA-independent signaling pathways (Yamaguchi-Shinozaki & Shinozaki, 1994;Yamaguchi-Shinozaki & Shinozaki, 2005).In order to identify potential molecular mechanisms by which Citrus Identification of cis-regulatory elements… approximately 500 bp upstream from the translation start codon of CsAP1, a unique arrangement of three repeating tandem arrays of stress-responsive elements was identified (Figures 1 and 2).Each tandem repeat contains a LFY binding site (CCANTG) (Benllock et al., 2011) and a dehydration-responsive element (DRECRTCOREAT, or simply DRE/CRT; RCCGACA) (Dubouzet et al., 2003) on the forward strand and a coupling element 3 site (CE3) (GCGTGTC) on the reverse-strand (Shen et al., 1996).
The DRE/CRT and CE3 elements are essential for gene induction in response to water-deficit and low-temperature stress through ABA-independent and ABA-dependent pathways, respectively (Liu et al., 1998;Yamaguchi-Shinozaki & Shinozaki, 2005).The DRE/CRT element (RCCGACA) (Figure 1) is essential in the ABA-independent pathway response to dehydration, high salinity and low temperature in A. thaliana (Sakuma et al., 2006;Yamaguchi-Shinozaki & Shinozaki, 1994).In addition, the C-repeat (CRT) and low temperature-responsive element (LTR) found within the CsAP1 promoter each contain the A/GCCGAC motif that conforms to the core of DRE but are specific to low temperature-inducible gene expression (Figure 1) (Sakuma et al., 2006).It is significant that use of a second method of motif discovery analysis, MEME and MAST, also uncovered the 28-bp region that coincides with the tandem repeats found in the promoter of CsAP1 at positions -382, -351 and -320 (Figures 1 and 3).The three DRE sequences identified are located in close proximity to the LFY binding site within each tandem repeat in the CsAP1 promoter (Figures 1 and 3).Of interest, a single sequence similar to the DRE core sequence was found near a LFY binding site within the upstream regulatory and > 75% matrix similarity, the IUPAC sequence listed in the MatBase transcription factor database was reported (http://www.genomatix.de/products/portfolio.html).Promoter locations are reported as the distance upstream from ATG start codon.
To discover additional highly conserved regions within homologous promoters, sequences were subjected to Multiple EM for Motif Elicitation (MEME, v3.5.7) analysis with motif width setting > 6 bp and < 50 bp (Bailey & Elkan, 1994).Promoters were then searched for motifs found by MEME using the Motif Alignment and Search Tool (MAST v3.5.7).Multiple sequence alignment was performed with AlignX software bundled with the VectorNTI Advance 10 software package.

In silico identification and comparative analyses of stress-responsive elements
Several classes of putative response elements related to water deficit, low temperature, ABA-dependent, and ABA-independent gene regulation pathways were identified in the promoters of CsLFY, CsAP1 and CsTFL.There were more putative stress-responsive elements in the promoter of CsAP1 than in CsLFY or CsTFL.In addition, numerous water-deficit and low-temperature stress-responsive elements in the CsAP1 promoter were common to the promoter regions of one or both AtAP1 and PtAP1-1.Thus, the CsAP1 promoter region was the primary focus of this investigation.Within a 100-bp region results suggest that the DRE site for the ABA-independent pathway present in the CsAP1 is absent from the promoter region of both AtAP1 and PtAP1-1 (Figures 2 and 3).
LFY binding sites (CCANTG) (Benllock et al., 2011) were found in the promoter region of AP1 for all three species (Figure 3) at similar relative positions from the translational start site.The promoters of both CsAP1 and AtAP1 have three LFY binding sites, whereas PtAP1-1 has only one (Figure 2).The sequences of the three CsAP1 LFY bindings sites are identical (Figure 3), whereas the sequences of three LFY binding sites of AtAP1 are uniquely different from each other (Benllock et al., 2011).The three CsAP1 LFY binding sites are identical regions of both AtAP1 and PtAP1-1 at positions -296 and -335, respectively.However, in the promoters of AtAP1 and PtAP1-1, the putative DRE sequences in close proximity to the LFY binding site do not fully conform to the A/GCCGAC core sequence, being GTCGACA and TTCGACA, respectively (Figure 3).Consequently, A. thaliana dehydration-responsive element-binding (DREB) proteins may fail to bind (Dubouzet et al., 2003).Consistent with this interpretation, promoter sequence analysis by MatInspector, which takes into account adjacent nucleotide similarity to minimize false positives (Cartharius et al., 2005), failed to detect DRE in the promoter of either AtAP1 or PtAP1-1.Taken together, these

Identification of cis-regulatory elements…
The presence of the CE3 core sequence, a functionally equivalent ABRE, within the upstream region of CsAP1 suggests a role for ABA in regulating floral development in C. sinensis.This is reinforced by the isolation of a gene encoding a C-repeat binding factor (CBF) isolated from C. sinensis that was demonstrated to be upregulated by low temperature (< 20 °C), high salinity and ABA (He et al., 2016).Regulation of the C. sinensis CBF by ABA is unusual.To date CBF has been reported to be exclusively involved in ABA-independent pathway stress responses in other plants (He et al., 2012).The core motif for the CBF (CBFHV) response element (RYCGAC) (Gu & Cheng, 2014) was found in multiple positions within the promoter region of CsAP1, as well as the AtAP1 promoter, but was not detected in PtAP1-1 (Figure 2).

In silico identification and comparative analyses of GA-responsive elements
Whereas it is well document that GA 3 inhibits flowering in Citrus spp., the mechanism remains unclear.A GA-responsive sequence associated with blocking gene transcription was not identified in the promoters of CsAP1, CsLFY or CsTFL.In contrast, promoters of the LFY homologues of C. sinensis, A. thaliana, and P. trichocarpa contain a conserved region having a motif similar to a known GA response element (GARE), which was originally found in the AtLFY promoter through comparative analyses with the PtLF promoter (Blázquez & Weigel, 2000).This motif was not observed in the CsAP1 or CsTFL promoter.The GARE motif (CAACTGTC) in PtLF and AtLFY differs from the sequence found in the CsLFY promoter by 1 bp, (CAAATGTC).Mutation in this region has been shown to abolish GA-responsiveness under long days in A. thaliana (Blázquez & Weigel, 2000).In WNO, repeated GA 3 applications during water-deficit and low-temperature floral-promoting treatments dramatically reduced floral intensity.Bud LFY expression was not affected.However, AP1 and AP2 expression were dramatically reduced and activity of the downstream floral organ identity genes was totally repressed.Mediation of the floral inhibitory effect of GA 3 through the activity of CsAP1 and CsAP2, which are essential for sepal formation, is consistent with the fact that GA 3 can no longer prevent C. sinensis flowering once the SAM has initiated sepal formation (Lord & Eckard, 1987).
It must be emphasized that a GARE that results in the repression of floral gene activity has not been identified.
to LFY binding site 2 of the AtAP1 promoter (Figure 3).The sequence of the PtAP1-1 LFY binding site, although not identical to LFY binding sites 1, 2 or 3 of AtAP1, conforms to the core sequence CCANTG (Figure 3).The presence of a LFY binding site located within the promoters of AP1 for all three species is indicative of the positive regulation of AP1 by LFY originally described in A. thaliana (Wagner et al., 1999).Research has determined that mutations in the AtAP1 LFY binding site 1 prevent AP1 activation and flowering under long day conditions, whereas mutations in LFY binding site 2 do not (Benllock et al., 2011).The results suggest that AtLFY binding site 2 may trigger flower development using a cue other than photoperiod.Given that all three CsAP1 LFY binding sites are identical to LFY binding site 2 in A. thaliana, it raises the question of whether day-neutral plants like WNO broadly use this specific consensus sequence over others.

In silico identification and comparative analyses of ABA-responsive elements
Cis-acting ABA response elements (ABREs) mediate ABA-induced transcription.In a promoter, an ABRE functions with a coupling element in an ABA responsive cis-element complex (ABRC) (Shen & Ho, 1995;Shen et al., 1996).Two distinct coupling elements have been identified, coupling element 1 (CE1), having the core sequence CACC (Shen & Ho, 1995), and CE3 with the core sequence GCGTGTC (Shen et al., 1996).The most common ABREs have an ACGT core, but non-ACGT ABREs, including CE3, have been demonstrated to function in ABA-dependent pathways (Hobo et al., 1999;Liu et al., 1998;Yamaguchi-Shinozaki & Shinozaki, 2005).The CsAP1 promoter contained a putative non-ACGT ABRE, the CE1 core sequence CACC, and notably, the CE3 core sequence GCGTGTC in the reverse strand within each tandem repeat (Figure 1), suggesting regulation by ABA.The CE3 sequence was also found in the reverse strand of the AtAP1 promoter, but the sequence was not present in the promoter of PtAP1-1 (Figure 2).It is only recently that ABA was demonstrated to have a positive role in flowering in A. thaliana, specifically under water-deficit stress as part of the drought-escape response (Riboni et al., 2016).However, as of yet, only ABA-dependent activation of AtFT under water-deficit has been documented (Riboni et al., 2016).
In silico comparison analysis of the CsLFY, CsAP1 and CsTFL promoters with their A. thaliana and P. trichocarpa homologues has a clear benefit for elucidating potential mechanisms regulating Citrus floral development at the molecular level.However, in silico analysis alone is insufficient to define precise regulatory models.To reinforce the results of the in silico analyses presented herein, they were evaluated in light of published data to promote a greater understanding of floral development in Citrus spp.and to help define the objectives of future research.The heuristic nature of this effort will hopefully lead to further research and to the development of new methods for regulating Citrus floral development in order to facilitate optimal floral intensity and improve yield on an annual basis.
Thus, an alternative possibility is that the inhibitory effect of GA on Citrus flowering is not through the direct repression of LFY or AP1, but indirectly through the upregulation of TFL, the antagonist to LFY and AP1 that inhibits flowering and confers meristem indeterminacy in A. thaliana (Liljegren et al., 1999) and C. sinensis (Pillitteri et al., 2004a).In support of this proposed mechanism, motif discovery analysis using MEME and MAST revealed several GARE sequences in the promoter of CsTFL.In particular, the TAACAAA box, first identified in the promoter of a barley (Hordeum vulgare) α-amylase gene, controls the upregulation of gene expression by GA and downregulation by ABA (Gubler & Jacobsen, 1992).It is tempting to speculate that the TAACAAA box in the CsTFL promoter may be under similar regulation.However, in light of the fact that juvenile buds have high levels of CsTFL mRNA, whereas adult (competent) buds have low expression, upregulation of CsTFL by GA may be restricted to juvenile buds in WNO (Pillitteri et al., 2004a).Nevertheless, the effect of GA 3 on bud CsTFL expression for trees in both developmental phases warrants further investigation.

CONCLUSION
Research results published over the years have established that AP1 is an essential factor in the network of genes conferring floral meristem identity, floral organ specification, and flower development (Benllock et al., 2011).In WNO, expression of AP1, with subsequent expression of AP2, the two genes necessary for sepal formation, appears to confer bud determinacy, leading to the upregulation of the downstream floral organ identity genes and flower formation.Given this critical function, it is not surprising that the CsAP1 promoter contains many different regulatory elements in order to provide for fine-tuning of its expression in response to different developmental and environmental cues.The multiple and combinatorial regulatory elements associated with CsAP1 may impart a unique failsafe system to citrus floral buds.Whereas transcript accumulation of AP1 and AP2 increases in a manner paralleling the duration of the water-deficit and low-temperature stress period, this increase in expression occurs only after the stress has been alleviated (Tang & Lovatt, in press).As a result, downstream floral organ identity gene activity increases and flower formation occurs under conditions of adequate water and warm temperature.

Figure 1 .
Figure 1.Motif location in the 100-bp region of the CsAP1 promoter with tandem repeats of water-deficit stress-related response elements and LFY binding sites (black boxes).

Figure 2 .
Figure 2. Comparison of motif locations within the promoters of CsAP1, AtAP1 and PtAP1-1 indicating the enrichment of cis-regulatory elements associated with water-deficit and low-temperature stress in relationship to LFY binding sites in the CsAP1 promoter in comparison to the promoters of AtAP1 and PtAP1-1.

Figure 3 .
Figure 3. Alignment of the conserved region within the promoters of CsAP1, AtAP1, and PtAP1-1.AtAP1 #2 refers to LFY binding site 2 (Benllock et al., 2011); CsAP1 #1-3 refers to the three tandem repeats within the CsAP1 promoter.Black boxes indicate 100% conservation; gray boxes indicate moderate conservation (50-99%).Positions of the LFY binding site and DRE are underlined.Two asterisks denote the first two nucleotides of the DRE motif.Note that AtAP1 and PtAP1-1 do not conform to the consensus RCCGACA.