To examine the circadian rhythm in CssPLA₂α and CsPLA₂β gene expression in citrus, 17-year-old Citrus sinensis cv. ‘Valencia’ citrus trees on ‘Swingle’ rootstock were used. Field samples of leaf blades (LB), leaf abscission zones (LAZ), fruit flavedo (FF), and fruit abscission zones (FAZ) were collected as described by Malladi and Burns (2008). Four biological replicates of each tissue were harvested from random canopy locations of four trees at 4 h intervals over 48 h from 19 June to 21 June in 2007. To test diurnal changes under controlled conditions, leaf blades were collected from light-entrained Citrus sinensis cv. ‘Valencia’ potted citrus trees as described by Malladi and Burns (2008). Briefly, 7-year-old trees were entrained to test conditions by transferring to growth rooms set at 25/19 °C in light/dark (LD) for 12:12 h photoperiod length for 7 d prior to sampling. The plants were then transferred into either constant light (LL), constant dark (DD) or kept in LD. Metal halide lamps (1000 W Clear BT37, Philips, Somerset, NJ) with a fluence rate of 245 μmol m⁻² s⁻¹ were used as the light source in the growth chambers. Temperature was maintained at 24 °C during the experiments. Mature leaf blades were collected at 4 h or 8 h intervals for 48 h. At least four biological replications from eight trees were collected for each time period.

To determine the effects of different spectra of light on CssPLA₂α, and CsPLA₂β gene expression, and CsPLA₂ enzyme activity, 1-year-old seedlings were transferred to light boxes (see below) set to LD photoperiods with various light spectra and intensity for at least 1 week prior to sampling. Leaf blades were harvested at 4 h intervals over 60 h. Temperature and humidity in all light boxes ranged from 23.5 °C to 24.5 °C and 60–65% RH during the sampling period. At least four biological replications from 16 seedlings were collected for each time period.

Light boxes were constructed with blue, green, red, and red/far-red spectra LED lights as described by Folta et al. (2005). The light boxes containing blue, green, and red lights were supplied with 36 Luxdrive 7007 Endor Star LED light modules whereas the red/far-red lights were supplied by Snap-Lite l (Quantum Devices, Barneveld, WI). Light emission range and fluence rate were measured using a spectroradiometer (SpectraWiz PS-100R, Apogee, Roseville, CA). Blue, green, red, and red/far-red LED lights emitted light from 410–540, 470–620, 580–670, and 600–780 nm, respectively, with emission peaks of blue, green, and red light of 456, 530, and 630 nm, respectively. Two emission peaks of 655 nm and 725 nm were measured with red/far-red light. Light intensity (μmol m⁻² s⁻¹) in each box was adjusted to the appropriate illuminating rate as indicated below for experiments.

For CssPLA₂ promoter analysis and gene structure characterization, total DNA was extracted from leaf blades of Citrus sinensis cv. ‘Valencia’ using the Wizard Genomic DNA Purification kit (Promega, Madison, WI). For gene cloning and expression analysis using quantitative RT- PCR, LB, LAZ, FF, and FAZ were frozen in liquid N₂, and stored at –80 °C as described by Malladi and Burns (2008). Total RNA was extracted and purified using the RNeasy Mini Kit (Qiagen, Valencia, CA), and DNA contamination was removed using DNase treatment (Qiagen). Total RNA (1 μg) was reverse transcribed using ‘Superscript III’ reverse transcriptase (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. The cDNA was diluted 1.5-fold and stored at –20 °C until quantitative RT-PCR analysis. Gene expression analysis was performed using quantitative RT-PCR. Citrus glyceraldehyde-3-phosphate-dehydrogenase (CsGAPDH) was used as the constitutively expressed internal calibrator. Optimal primer concentration used in quantitative RT-PCR analysis for CssPLA₂α (PLA2af1: 5′-GTGGGCTGCTGTACAGTGGAT-3′; PLA2ar1: 5′-CAAGCATCAAGGCCATCACA-3′) was 0.25 μM, and was 0.125 μM for CsPLA₂β (PLA2bf1: 5′-TCGGATTTTCTCGGGAATGT-3′; PLA2br1: 5′-CATGTCCATACCCTGTACCATAGTG-3′). Gene expression analysis was performed on each biological replicate in duplicate and results averaged for each replicate. The lowest expression value was used to compare all others within the experiment. To compare expression levels between CssPLA₂α and CsPLA₂β, relative expressions were compared within an experiment to the lowest expression sample of CssPLA₂α.

Genes and promoters were obtained from total RNA and genomic DNA extracted from leaf blades in Citrus sinensis cv. ‘Valencia’. Gene specific primers designed based on the partial sequences of PLA₂α (5′-TCTGATATGAGAATACTCTGGTGC-3′) and β (5′-GCCACGAAAAGTTTAAGAGATGC-3′) obtained from citrus HarvEST database (http://harvest.ucr.edu/) were used for CssPLA₂α and CsPLA₂β cloning. 5′-RACE (Rapid Amplification of cDNA End, Invitrogen), and 3′-RACE were performed for specific gene amplification via Polymerase Chain Reaction (PCR) using Taq DNA polymerase (Qiagen). Inverse PCR was performed to obtain CssPLA₂α and CsPLA₂β promoter sequences by restriction enzyme digestion of genomic DNA, self-ligation, followed by PCR amplification using Elongase enzyme mix (Invitrogen) with inverse primers. The amplified DNA fragments were cloned into pGEM-T easy vector (Promega). Transformation of Escherichia coli JM109 cells (Promega) was performed, and the resulting plasmid DNA was purified using a Wizard DNA purification kit (Promega). Sequencing was performed at the Interdisciplinary Center for Biotechnology Research, University of Florida (Gainesville, FL). Nucleotide sequences of CssPLA₂α (accession no. GU075396), CsPLA₂β (accession no. GU075398), and 1000 bp promoter regions of CssPLA₂α (accession no. GU075397) from Citrus sinensis cv. ‘Valencia’ were used for this study. Alignment of amino acid sequences were performed using ClustalW2 software (http://www.ebi.ac.uk/Tools/clustalw2/index.html). Signal peptide prediction was performed using SIG-Pred: Signal Peptide Prediction web server (http://bmbpcu36.leeds.ac.uk/prot_analysis/Signal.html).

Citrus leaves (1 g) were ground in liquid N₂, dissolved in 2.5 ml of cold extraction buffer (1 mM EDTA, 100 mM TRIS-HCl, 2% (w/v) PVPP, 0.15 M sorbitol, pH 7) with 1 mM PMSF added, and the resulting homogenate centrifuged at 12 000 rpm at 4 °C for 15 min. Supernatant proteins were desalted using a PD-10 column (GE Healthcare, Buckinghamshire, UK) equilibrated and eluted with column buffer (10 mM HEPES, 1 mM MgCl₂, pH 7). Total proteins in the eluate were precipitated overnight at 4 °C using ammonium sulphate to saturation, centrifuged 14 000 rpm at 4 °C, washed with acetone, and resuspended in extraction buffer. Concentrated proteins were quantified and stored at –20 °C for the total PLA₂ enzyme activity assay. Total PLA₂ activity assay (Cayman Chemical Co, Ann Arbor, MI, USA) was performed as described (Alferez et al., 2005). sPLA₂ was fractionated from total protein samples using centrifugal filter (Microcon, Bedford, MA, USA) with a molecular weight cut-off of 30 kDa, and enzyme activity measured.

Gene specific primers designed based on the EST sequences of PLA₂α and β from the citrus HarvEST database were used for CssPLA₂α and CsPLA₂β cloning. 5′- and 3′-RACE was performed to obtain the full-length sequences. Full-length CssPLA₂α and CsPLA₂β cDNA sequences encoded 17.1 and 31.6 kDa proteins of 156 and 279 amino acids, respectively, and shared 31% identity (Fig. 1). CssPLA₂α and CsPLA₂β contained key phospholipase catalytic sites of a Ca²⁺-binding loop, the active site motif with a conserved his/asp dyad (HD), and 12 conserved Cys residues that form six disulphide bonds for conformational integrity. A eukaryotic signal peptide sequence was present in the N-terminus of CssPLA₂α but not CsPLA₂β. The KxEL endoplasmic reticulum (ER) retention sequence (Pagny et al., 1999; Seo et al., 2008) was found in the 3′-terminus of CsPLA₂β but not CssPLA₂α. The presence of catalytic Ca²⁺-binding loop and active site motifs suggests that CssPLA₂α and CsPLA₂β encoded PLA proteins that catalyse hydrolysis of the phospholipase sn-2 bond in a Ca²⁺-dependent manner (White et al., 1990; Yu and Dennis, 1991; Six and Dennis, 2000; Lee et al., 2005; Hsu et al., 2008). The molecular weight of CssPLA₂α suggests that it encoded a secretory phospholipase A₂ (Six and Dennis, 2000).

Sequences (1 kb) upstream from the putative ATG translational start cordon of CssPLA₂α and β were examined for promoter regulatory elements. Core regulatory TATA box (TATAAA) and CAAT/CCAAT elements, light regulatory elements, and some of clock regulated elements, such as G-box and CCA1 binding sites, were found in CssPLA₂α and CsPLA₂β promoter sequences (Fig. 2; Table 1). Clock-regulated elements associated with blue light and evening element were present in CssPLA₂α but absent in CsPLA₂β, whereas a phytochrome A element was present in CsPLA₂β but not CssPLA₂α. The abundance of light/clock and evening regulatory elements in the promoter region of CssPLA₂α suggests CssPLA₂α may function as a light-activated gene diurnally governed with an evening-specific expression pattern. CsPLA₂β may be regulated by light but less responsive.

CssPLA₂α transcript accumulation displayed a distinct diurnal rhythm in leaf blades (LB), leaf abscission zones (LAZ), fruit flavedo (FF), and fruit abscission zones (FAZ) in field samples (Fig. 3). Minimum and maximum CssPLA₂α transcript accumulation in all tissues occurred at approximately 10.00 h and 18.00 h, respectively. The order of maximum amplitude of CssPLA₂α relative expression was LB>LAZ>FAZ>FF (Fig. 3A). At its maximum, relative expression in LB was 1337-fold and 643-fold higher than its minimum over two diurnal cycles. Accumulation of CsPLA₂β RNA oscillated in LB and much less in LAZ. CsPLA₂β transcript accumulation maxima and minima occurred at 18.00 h and 10.00 h, respectively, however, the decline in expression during the dark period was gradual compared with CssPLA₂α. Relative CsPLA₂β expression in LB was 7.7-fold and 5.0-fold higher at its maxima than its minima over two diurnal cycles. Closer inspection of LAZ, FF, and FAZ expression revealed the order of maximum rhythmic amplitude of CsPLA₂β was LB>LAZ, but no oscillations occurred in FAZ or FF (Fig. 3B). The basal expression of CsPLA₂β was over 50-fold higher in LB compared to LB CssPLA₂α expression.

CssPLA₂α and CsPLA₂β expression in LB from potted citrus plants held in light/dark (LD) and constant temperature in the growth room demonstrated diurnal oscillation (Fig. 4). Average maximum amplitude was similar to that from field samples. Transfer of plants from LD to constant dark (DD) resulted in the cessation of CssPLA₂α expression oscillation within 24 h, whereas transfer of plants from LD to constant light (LL) markedly dampened expression. The period of CsPLA₂β expression shifted when transferred from LD to LL or DD conditions. Thus, light was a major factor regulating CssPLA₂α and CsPLA₂β expression.

CssPLA₂α and CsPLA₂β gene expression in seedling LB oscillated in blue and green light (Fig. 5). The amplitude of CssPLA₂α expression was much higher than CsPLA₂β under all light spectra examined. Rhythmicity of CsPLA₂β gene expression in blue and green light was erratic, but oscillations were more pronounced with blue light treatment. CssPLA₂α expression was lowest 1 h before the end of dark period through 3 h after the light period was initiated. Expression increased 7–11 h after the light treatment commenced and peaked 11 h after illumination. Expressions of both genes in blue light were greater at 24 μmol m⁻² s⁻¹ intensity compared with 3 μmol m⁻² s⁻¹. CssPLA₂α gene expression in seedling LB also oscillated in red, red/far-red, and the combination of blue, green, and red light held in LD (see Supplementary Fig. S1 at JXB online). The amplitudes of CssPLA₂α expression were 18-fold and 8-fold lower in red and red/far-red compared to blue light at 24 μmol m⁻² s⁻¹ intensity, respectively. Blue light was as effective as the combination light treatment in inducing maximum CssPLA₂α gene expression. Rhythmicity of CsPLA₂β gene expression in red/far-red light was also erratic and the amplitude was much lower than blue light. Increasing green, red or red/far-red light intensity from 3 to 24 μmol m⁻² s⁻¹ had minor effects on expression of both genes (see Supplementary Fig. S1 at JXB online).

CssPLA₂α relative gene expression in LB of seedlings treated with blue light under LD conditions and total CsPLA₂ enzyme activity showed similar rhythmicity, with the peak occurring 11 h after the light period commenced (Fig. 6A, B). Like CssPLA₂α expression, total PLA₂ activity remained low and constant under DD conditions. The peak activity of the low-molecular-weight PLA₂ fraction corresponded with the maximum amplitude of CssPLA₂α expression.