Differential expression of genes in olive leaves and buds of ON- versus OFF-crop trees

Alternate bearing (AB) refers to the tendency of trees to have an irregular crop load from 1 year (ON) to the next year (OFF). Despite its economic importance, it is not fully understood how gene networks and their related metabolic pathways may influence the irregular bearing in olive trees. To unravel molecular mechanisms of this phenomenon in olive (cv. Conservalia), the whole transcriptome of leaves and buds from ON and OFF-trees was sequenced using Illumina next generation sequencing approach. The results indicated that expressed transcripts were involved in metabolism of carbohydrates, polyamins, phytohormones and polyphenol oxidase (POD) related to antioxidant system. Expression of POD was increased in leaf samples of ON- versus OFF-trees. The expression pattern of the greater number of genes was changed more in buds than in leaves. Up-regulation of gene homologues to the majority of enzymes that were involved in photorespiration metabolism pathway in buds of ON-trees was remarkable that may support the hypotheses of an increase in photorespiratory metabolism in these samples. The results indicated changes in expression pattern of homologous to those taking part of abscisic acid and cytokinin synthesis which are connected to photorespiration. Our data did not confirm expression of homologue (s) to those of chlorogenic acid metabolism, which has been addressed earlier that have a probable role in biennial bearing in olive. Current findings provide new candidate genes for further functional analysis, gene cloning and exploring of molecular basses of AB in olive.


Results
A total of 325,125,070 clean reads were obtained from eight libraries after filtering. In every sample, an average of 20,320,317 paired-end 150 bp reads was obtained with average Q20 and GC of 97.36% and 45.3%, respectively (Table 1). In total, 512,658 contigs with N50 lengths of 1348 base, average length of 806.45 bases and mean of % GC 41.21 was obtained from de novo assembly analysis. On average about 93% of clean reads of each library were mapped with these contigs.

Comparing of ON-and OFF-transcriptome profiles.
Towards comparing ON-and OFF-trees transcriptome profiles, a set of fragments with differential expression (DE) pattern were studied in leaves and buds.

Comparison of leaf transcriptome in ON-vs. OFF-trees.
Out of that 454 significantly differentially expressed genes were identified in leaves, 277 were found to be up-regulated and 177 were found to be down- www.nature.com/scientificreports/ regulated genes in ON-vs. OFF-trees (see Supplementary Table S1). GO analysis of DEGs represented at least one GO ID at database for 368 of them (e-Value ≤ 0.05) (see Supplementary Table S2). Allocation of transcripts to three main groups of biological process, cellular component, and molecular function were shown in Fig. 1A. KEGG enrichment analysis revealed that only three pathways (out of 80 pathways covered) were significantly enriched (corrected P-Value ≤ 0.05) (see Supplementary Table S3). These pathways included photosynthesis, metabolic and fatty acid degradation pathways. Using the results of GO and KEGG, and based on the previous studies on AB in olive and other fruit trees, homologues genes from the photosynthesis, carbohydrates metabolism, phenolic compounds, antioxidant enzymes, polyamines, and phytohormones biosynthesis pathways were studied in more detail ( Table 2).

Comparison of bud transcriptome in ON-vs. OFF-trees.
Out of the total 1809 significantly differentially expressed sequences were identified in buds, 1125 sequences were found to be up-regulated genes, and 684 were found to be down-regulated genes in ON-vs. OFF-trees (see Supplementary Table S4). GO analysis of 1809 DE sequences represented at least one GO ID at database for 1414 transcript (e-Value ≤ 0.05) (see Supplementary  Table S5). Transcripts were allocated into three main groups of molecular function, cellular component, and biological process shown in Fig. 1B. Out of that 1809 DEGs were mapped to canonical pathways of KEGG, 106 were assigned to different KEGG pathways, and 33 pathways were significantly enriched in ON-vs. OFF-trees (corrected P-Value ≤ 0.05) (see Supplementary Table S6). The first five significant pathways were included metabolic pathways, carbon metabolism, and CO 2 fixation in photosynthetic species, protein processing and photosynthesis. Merging the results of GO and KEGG along with previous reports, homologues genes from the photosynthesis, carbohydrates metabolism, phenolic compounds, antioxidant enzymes, polyamines, phytohormones biosynthesis pathways and flowering control genes were more account (see Supplementary Table S7). We found that the gene homologues of the  www.nature.com/scientificreports/ majority of enzymes that are involved in photorespiration metabolism pathway were up-regulated in buds of ON-trees (Fig. 2).
Validation of RNA-Seq results using Q-RT-PCR. The relative expression of four selected sequences of RNA-Seq data, were confirmed by Q-RT-PCR matched to those of the RNA-Seq (Fig. 3). In leaves, POD homologue was up-regulated while SOD homologue was down-regulated. In buds, POD and rubisco genes homologous both were up-regulated in ON vs. OFF-trees. comparative conditions. Quantification of nine housekeeping genes homologous including cyclophilin, 18 s rRNA, GAPDH, 25 s rRNA, ubiquitin (UBQ), beta-actin, RuBP, alpha-tubulin and beta-tubulin showed, that none of them had significant DE in any of compared groups. Prior to DE analysis, the Principal Component Analysis (PCA) was done to explore the similarity of biological replicates. The result indicated the accuracy of sample selection, and that the genetic distance between replications was not significant (Fig. 4).
Biochemical parameters. No significant difference was seen between ON-versus OFF-trees in term of inhibition of DPPH activity and total flavonoid content in leaves (Fig. 5A,B). However, a significant difference www.nature.com/scientificreports/ was revealed in total phenol content (Fig. 5C). The activity of SOD and POD antioxidant enzymes was significantly different between trees. SOD, showed high amount of activity in OFF-trees, while the POD had its highest activity in ON-trees (Fig. 5D,E).

Discussion
We sequenced and studied the eight cDNA libraries obtained from leaves and buds of ON and OFF olive trees during flower induction in July, 2018. The quantity of DEGs in the bud-derived libraries was nearly four times more than that of leaf-derived one. These results indicate that a large number of differentially expressed mRNAs are involved in buds metabolism than in the leaves under AB condition. No wonder, because the bud is the only organ in which a phase transition to the reproductive phase takes place.

Figure 2.
The schematic of the photorespiration metabolism pathway; Genes homologues to the ten enzymes through this pathway, corresponding to number 1-10, showed up-regulation pattern in olive's bud samples of ON-tree. The homologue sequence identification code of each enzyme resulted from current research, has been noticed. Adapted from Wingler et al. 25 and Buchanan et al. 26 .   www.nature.com/scientificreports/ OFF-trees in both leaf and bud samples. In leaves, for instance, the homologue of galactosyl transferase 6 (XP_022863390.1) showed near 25-fold up-regulation in ON-trees. There are some contradictory reports on the involvement of photosynthesis, and carbohydrate contents on the flower induction events in olive trees 6,11 . While some researchers suggested that photosynthesis was not changed between leaves of ON-and OFF-trees 7,27 , others showed an increase rate of CO 2 assimilation in high fruit-bearing trees 28,29 . Shalom et al. 1 showed that induction of gene expression was related to photosynthesis in the citrus OFF-buds, which is in agreement with our finding in olive's buds. Andreini et al. 30 revealed that the anatomical and histochemical analysis of starch and total carbohydrate accumulation changes at the cellular level, did not allow discrimination of the buds sampled from ON-and OFF-shoots in the olive trees during flower induction stage. These all may suggest the oxygenase activity of rubisco (i.e., photorespiration) in olive's OFF-buds. Gene homologues related to the majority of enzymes involved in photorespiration pathway were up-regulated in buds of ON-trees. Buds of ON-trees have been suggested to withstand a stressful conditions resulting from heavy fruit load 31 . Photorespiratory metabolism is prerequisite for rapid physiological acclimation against different unusual plant growth conditions. In this way, photorespiration in connection with phytohormones signaling contributes for mitigating stressor effects 32 .
In Pisum sativum, for instance, Fedina et al. 33 reported that the salinity stress resulted in increase of photorespiration when, treated with ABA eliminate the adverse effect of NaCl. Furthermore, the precursors for other metabolic processes, e.g., glycine for the synthesis of glutathione, are provided through the photorespiration pathway, which is also involved in stress protection 25 . Our data showed that a gene homologue to glutathione reductase (GR) (XP_022860602.1) in olive's bud samples of ON-trees was up-regulated 6.85-fold in comparison to those of OFF-trees. Endogenous plant hormones are involved in AB phenomenon. Hormones such as ABA, GAs and auxins 11,14 , and miRNAs that targeted genes involved in hormone-mediated signaling 2 are expressed differentially in olive ON-and OFF-trees. We found one sequence homologue to zeaxanthin epoxidase (ZEP) (XP_012839200.1) in the leaves and four sequences (XP_022863205.1) in the buds which were up-regulated in samples of ON-trees. The conversion of zeaxanthin to violaxanthin is catalyzed by the ZEP enzyme. This reaction contributes to the biosynthesis of ABA 34 which regulates the expression of stress-responsive genes. The high expression of ZEP might be related to the presence of increased amount of ABA in olive leaves, and particularly in buds, as a result of stressful conditions imposed by fruit load in bearing trees. An increasing level of ABA in citrus' buds from ON-trees has previously been reported 31 . ABA is considered as an inhibitor of flower induction in A. thaliana, where mutants with high ABA levels showed late flowering 35 . It is likely that ABA plays the same role in olive buds of ON-trees, and up-regulation of sequence homologues to the biosynthesis of ABA may reflect its alleviating role for increasing photorespiration metabolisms in olive buds on ON-trees. However, relation between photorespiration metabolism and ABA-mediated signaling pathway remains unclear.
Two gene homologues to cytokininriboside 5′-monophosphate phosphoribohydrolase LOG1-and LOG3-like (XP_022894450.1 and XP_022887256.1) were up-regulated in olive's bud of ON-trees. These enzymes mediate production of cytokinin nucleobases and contribute in regulating cytokinin activity during plant growth 36 . Cytokinins are a class of phytohormones which regulate leaf senescence, lateral root formation, cell division and stress tolerance 37

Phenolic compounds, antioxidant enzymes-related genes and polyamines. Homologues of
genes involved in few secondary metabolism pathways including caffeic acid and cinnamic showed DE pattern in leaves and mostly in buds. A gene homologue to chalcone synthase-like (XP_022852825.1), a key enzyme of the flavonoid/isoflavonoid biosynthesis pathway, was up-regulated more than seven-fold in buds of ON-trees compare with OFF-trees. One gene homologue to caffeic acid (XP_022880295.1), was in common with leaves and buds, and another homolog (XP_022880297.1) was down-regulated just in buds in ON-trees. The phenolic compounds are secondary plant metabolites that change in response to environmental stimuli, and even these changes may vary from one tissue to another 39 . In an attempt to establish a correlation between AB and total phenol content and its constituents, Mert et al. 40 addressed lower accumulation of caffeic acid and chlorogenic acid in the leaves in ON-compared with OFF-years. The data from present study revealed that the gene homologue to caffeic acid metabolism pathway in bud samples of ON-trees was down-regulated. Total phenolic compounds in leaf samples, showed significant differences between ON-and OFF-trees parallel to the data from Mert et al. 40 . However, the analysis here could not detect gene homologue(s) to those of chlorogenic acid metabolism, which has been considered to have a probable role in olive flower induction and AB [41][42][43]  www.nature.com/scientificreports/ POD activity was consistent with the results of RNA-Seq in terms of its increasing expression in leaf samples from ON-versus OFF-trees. The electron transport processes in photosynthesis and respiration are the main production source of the reactive oxygen species (ROS) which in turn activates signal transduction processes. The imbalance between ROS detoxification and generation imposed oxidative stress to the plant cell 44 . From the point view of a stressful condition of the bearing tree and a high metabolism of oxidation-reduction process, one can explain the up-regulation of gene homologues to POD. Therefore, the accumulation of the antioxidant enzymes might dissipate the excess ROS resulting from the stressful condition in ON-trees. Unlike POD, the expression of a gene homologue to SOD (XP_022888361.1) showed DE just in leaves and decreased in samples of ON-trees. This result was in agreement with that of SOD activity determination in leaf samples carried out here. The involvement of copper-regulated microRNAs in down-regulation of copper/zinc SOD expression in Arabidopsis thaliana and response to low copper has been previously reported 45 . Since there is a high competition between vegetative organs and growing fruits for nutrients, one may assume that such micronutrient-regulated microRNA system might be involved in down-regulation of SOD in olive's leaves in response to probably lower micronutrients such as Mn, Fe, Cu, and Zn under fruit bearing conditions. Polyamines (PAs) such as spermine, spermidine, and putrescine (their obligate precursor) are a group of phytohormone-like aliphatic amine natural compounds, which contribute in the regulation of physiological processes and plant developmental 46 . Our results showed DE of seven gene homologues which involved in polyamines metabolism of which, up-regulation of a gene homologue to S-adenosyl methionine decarboxylase (SAMDC) (XP_022844785.1) in leaf and bud samples of ON-trees was in common. SAMDC is an enzyme which is involved in the synthesis of polyamines in plants, mammals, and other species 47 . PAs are well known for their anti-stress and anti-senescence effects due to their antioxidant properties and acid neutralizing, as well as for their cell wall and membrane stabilizing abilities 46 . The present study suggests that the up-regulation of the gene homologue to PAs biosynthesis in leaves and bud samples may represent a biochemical response of the tree to cope with high fruit load stress. Supporting this, it has been shown that the transformation of a plant with genes encoding SAMDC, spermidine synthase and/or arginine decarboxylase improved tolerance to environmental stress 47 . A gene homologue to thermospermine synthase (PIN17113.1) in bud samples of ON-trees showed more than eight-fold up-regulation in comparison to OFF-trees. In Arabidopsis, it has been assumed that thermospermine acts as a regulator of stem elongation, as sever dwarf plants were obtained in mutants that encoded thermospermine synthase 48  Three gene homologues to flowering locus K (FLK) homology domain showed DE in bud samples; two of them (XP_022876625.1; XP_022889081.1) were down-regulated and another one (XP_022876624.1) was up-regulated in ON-trees. The FLK, a protein containing three RNA-binding domains, acts in the autonomous flowering-promotive pathway 53 . Regardless of its role, simultaneous up-and down-regulation of different homologues of FLK is in the debate. Probably, different isoforms of these genes present in olive. This is probable because of relatively huge (2n = 46), complex (heterozygosity and different alleles at gene locus) and not well-known genome in olive to compare with model plant species 54 . A gene homologue to Upstream of Flowering Locus C (UFC) isoform X1 (XP_022850620.1) showed a down-regulation trend in buds of ON-trees. The UFC is component of a cluster containing three genes; Flowering Locus C (FLC), UFC and Downstream of FLC (DFC), which are coordinately regulated in response to environmental stimulus 55 . The UFC is considered as the flowering inhibitor. Thus, we expected to find its up-regulation in buds of ON-trees (OFF-buds for return blooming).

Conclusions
This study clarified, to some extent, the transcriptional alterations of endogenous metabolic pathways in olive leaves and buds under AB conditions during flower induction using Illumina HiSeq 2000 system. De novo assembly strategy was used for creating the reference genome and further analysis and mapped the majority of clean reads in every library with constructed contigs and blasted over 69% of the obtained transcripts with available databases. Additionally, some transcripts could not be associated with a specific function. This may be mainly due to the lack of sufficient reference database of olive in GenBank, it's relatively large genome size, and divergent gene functions in such species. DE of 57 gene homologues to carbohydrate metabolism were revealed in leaves' and buds' samples of ON-versus OFF-trees, supporting an important role for carbohydrate metabolism especially in sugar-mediating signaling pathway(s) under AB condition. The increase of oxygenase activity of rubisco (i.e., photorespiration) in olive's buds from ON-trees was hypothesized as a contribution to multiple signaling pathways, particularly those that govern plant hormonal and defense responses. In connection with that, four homologous sequences of ABA biosynthesis pathway in buds and another homologue in leaves of ON-samples showed up-regulation. The results could not detect DE of gene homologue(s) to those involved in Scientific Reports | (2020) 10:15762 | https://doi.org/10.1038/s41598-020-72895-7 www.nature.com/scientificreports/ chlorogenic acid metabolism, which was considered to have a role in olive flower induction and AB. However, the down-regulation of gene homologous to those involved in caffeic acid metabolism, as a probable metabolic precursor of chlorogenic acid metabolism pathway, was shown in buds samples of ON-trees. It is suggested that a few number of genes corresponding to flower development are expressed differentially during flower induction period in olive's buds under AB. More probably, the majority of these genes would be differentially expressed during flower initiation and developmental phases in late summer onwards. Indeed, genes such as MADS-box transcription factor 27-like and/or FCA-like are likely a component of a posttranscriptional cascade involved in the control of flowering. Proteomic analysis of olive buds and leaves is our ongoing project to shed much more light to the pathway for the transition to flowering in the olive tree under AB condition. From practical and horticultural view of points, we approved that the change in expression pattern of many genes from different metabolism pathway, especially those direct or indirect related to stressful condition, have been started in July, correspond to flower induction period in olive tree. Hence, general encouraging of tree tolerance to abiotic stress at this time or even sooner, improving nutritional status (especially those of micronutrients) and spraying with different antioxidant compounds for instances, could modulated to some extent the AB behavior in this fruit tree species. Of course, continuous treatment of tree and orchard later on, during flower initiation and development, should also be considered for getting satisfied results in this respect.

Materials and methods
Plant materials. The experiment was carried out on 12-year-old own-rooted olive trees cv. 'Conservalia' , which is considered as a severely AB cultivar 3 , at Tarom  The leaf and bud samples were collected at flower induction time in July parallel with fruit pit hardening 56 . Samples were collected from the middle part of current year growth (fifth and sixth nodes from the top of the shoot) which is responsible for the fruit production in the following year 57 . Although a little mix of vegetative and generative buds is possible in bud sampling however, the majority of collected bud sample from 'ON' tree were surly vegetative buds for return blooming in the next year and vice versa. According to literature, in mentioned sampling time there is no differences between OFF-and ON-buds from the morphological points of view however, molecular and histochemical evidence have confirmed their differences at flower induction period 30 . In order to minimize the environmental noise on transcriptome profiles, the sampling in all replicates was performed only from one side (southeast) of the trees at shoulder height.
For assessment of some biochemical traits, five branches (10 replications of each 'OFF' and 'ON' status in total) were selected. Collected samples were immediately immersed in liquid nitrogen after a surface cleaning with ethanol 70% and then transferred to the lab and stored in a − 80 °C freezer until use.
RNA preparation and next-generation sequencing. Total RNA was extracted using Trizol Reagent kit (Invitrogen) according to the manufacturer's instructions with two biological replicates. The RNA concentration and purity was examined using both Nanodrop spectrophotometer (Thermo Fisher Scientific) and Bioanalyzer (Agilent technologies). High-quality total RNA (260/280, 1.8-2.0; 260/230 > 2.0; RIN > 7.5) was subjected to cDNA library synthesis using SMART-Seq protocol. The libraries were sequenced on an Illumina HiSeq 2000, 150 base pair-end reads at Beijing Genomics Institute (BGI) (Hong Kong) according to standard procedures. Bioinformatics analysis. Raw reads were filtered for low-quality reads, adaptor sequences and unexpected contamination. The sequencing quality control of clean reads was obtained by FastQC v.0.11.5 software 58 (https:// www. bioin forma tics. babra ham. ac. uk/ proje cts/ fastqc). There was only one released data by "International Olive Genome Consortium" includes 50,684 genes (cds), at the time of analyzing and preparing the current manuscript, and we used this data for GO analysis. However, the results were very poor and unreliable, hence were ignored. The assembly of clean reads has been performed by Trinity v.2.4.0 software 59 . For this aim, first a de novo assembly file was done by the combination of all reads of every eight libraries. The output obtained was used as a reference genome for further analysis. This procedure is reliable and strong strategy which has been used for molecular study of wide plant species such as olive 18,20 , pear 60 and pistachio 61 . By de novo assembly strategy, we mapped the majority of clean reads in every library with made contigs and blasted over 69% of the obtained transcripts with available databases.
Each sample read was mapped by Bowtie2 v.2.3.4.1 software 62 on the contigs derived from the Trinity analysis as the reference genome. Using the Trinity output and by the Kallisto v.0.43.1 software 63 , the counting of the reads was done in the desired samples. The fragments per kilobaseof transcript per million mapped reads (FPKM) derived from the counting of the reads were used to compare the differences among the samples. DE analysis by RNA-Seq reads was performed for two groups of data including "Bud ON vs. OFF" and "Leaf ON vs. OFF". The significant threshold for the detection of genes with DE adjusted P-value ≤ 0.05 was considered and genes with log twofold change ≥ 1 and ≤ − 1 were considered as genes with DE. DE analysis was performed by the R package, DESeq2 v.3.7 64 . Then, BLASTx was performed with the Nr database by Diamond v.0.9.14 software and using blast result, the gene ontology (GO) analysis was done by Blast2GO software 65 (https:// www. blast 2go. com) for differential expressed genes. GO terms with e-Values ≤ 0.05 were considered significant. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed using KOBAS v.3.0 (https:// kobas. cbi. pku. edu. cn/ anno_ iden. php) 66 . Pathways with corrected P-values ≤ 0.05 were considered significant. www.nature.com/scientificreports/ Quantitative reverse transcription PCR (Q-RT-PCR) validation. In order to validate the RNA-Seq data, four cDNA sequences, including two from leaf samples (TRINITY_DN81593_c2_g2_i7 and TRINITY_DN 78486_c0_g1_i14, gene homologous of POD and SOD respectively) and another two from bud samples (TRIN-ITY_DN81593_c2_g2_i7 and TRINITY_DN78486_c0_g1_i14, gene homologous of POD and rubisco respectively) were analyzed using Q-RT-PCR. Primers were designed using Primer3 (https:// prime r3. ut. ee/). The actin gene was used as an internal reference control. The primer sequences used for Q-RT-PCR have been shown in Supplementary Table S8. The Q-RT-PCR reaction was performed by a Rotor-Gene 3000 system (Qiagen). Amplification was done in two technical replicates with a reaction volume of 20 μl containing 0.5 μl (0.1 μM) of each forward and reverse primer, 2 μl (160 ng) of diluted cDNA, 4 μl of 5 × HOT FIREPol EvaGreen qPCR Mix Plus (Solis BioDyne) and 13 μl of sterile distilled water. The cycling parameters were as follows: one cycle at 95 °C for five min to activate the Taq enzyme, followed by 40 cycles of denaturation at 95 °C for 20 s, annealing at 55 °C for 20 s and extension at 72 °C for 12 s. The final followed by a melting step in a 50-99 °C range by an increase of 1 °C every 5 s to evaluate the "melting curve" for confirming the occurrence of a unique PCR product. Relative expression levels were calculated using the 2 −△△CT method 67 .

Measurement of biochemical parameters in leaves.
A number of biochemical parameters including antioxidant capacity, total flavonoid and phenol contents as well as the activity of peroxidase (POD; EC 1.11.1.7) and superoxide dismutase (SOD; EC 1.15.1.1) enzymes in leaf samples were assessed. Bud samples were excluded from those mentioned evaluations. Due to the small size of buds, it was necessary to ruin large amounts of shoots for gathering enough samples, which could result in irreparable damage to trees as well as non-uniformity in plant materials. Total phenolic compounds were measured by reaction with the Folin-Ciocalteu reagent and gallic acid as standard 68 . The results were expressed in mg gallic acid equivalents (GAE) per 100 g FW. Total flavonoids were assessed by reaction with AlCl 3 , as described by Chang et al. 69 , and quercetin was used as standard. Results were expressed in mg quercetin acid equivalent (QE) per 100 g FW. To evaluate the antioxidant activity, DPPH free radical scavenging activity 70 was used. The results of this analysis were expressed in terms of the 'percentage of DPPH free radical scavenging' .
The enzymatic extract was prepared to measure total soluble protein, SOD and POD enzymes activity using 50 mM potassium phosphate buffer with a pH = 7. The SOD enzyme activity was measured by the method of Giannopolitis and Ries 71 using spectrophotometer at a wavelength of 560 nm. The results of this enzyme activity were expressed in 'units per mg protein' . The POD enzyme activity was measured by the method of Abeles and Biels 72 using spectrophotometer. Absorbance changes were measured in 120 s at 470 nm and enzyme activity was expressed as 'micromole per minute per mg protein' . Statistical analyses of this section were performed using unpaired t-test.

Data availability
All the data analyzed or generated during this study are included in this article. This Sequence Read Archive (SRA) submission has been released on 2020-07-01. The data that support the findings of this study are deposited at https:// trace. ncbi. nlm. nih. gov/ Traces/ study/? acc= PRJNA 549052 and received submission code of SUB5704553.