Olive (Olea europaea L.) trees (cv. Picual and Arbequina) were cultivated in the experimental orchard of Instituto de la Grasa, Seville (Spain), using drip irrigation and fertirrigation (irrigation with suitable fertilizers in the solution) from the time of full bloom to fruit ripening. Young drupes, developing seeds, and mesocarp tissue were collected from at least three different olive trees at different weeks after flowering (WAF) corresponding to distinct developmental stages of the olive fruit: green (9, 12, 16, and 19 WAF); yellowish (23 WAF); turning or veraison (28 and 31 WAF); and mature or fully ripen (35 WAF). Young leaves were harvested as well.

Olive trees (cv. Klon-14, Abou Kanani, Dokkar, and Piñonera) from the Worldwide Olive Germplasm Bank of Córdoba located at IFAPA (Alameda del Obispo, Córdoba, Spain) were grown by using standard culture practices. Mesocarp tissue from these cultivars was collected at distinct developmental stages of the olive fruit: green (20 WAF); yellowish (24 WAF); turning or veraison (27 WAF) and mature or fully ripen (31 WAF).

To study the effect of water deficit, mesocarp tissue at different WAF was collected at the Sanabria orchard, a commercial super high-density olive (cv. Arbequina) orchard near Seville, as previously described (Hernández et al., 2018). Full irrigation (FI) and two regulated deficit irrigation (RDI) treatments (60RDI and 30RDI) were applied.

Stress treatments were conducted as described by Hernández et al. (2019) using olive branches collected from different olive trees (cv. Picual and Arbequina) with around 100 olive fruit at the turning stage (28 WAF) and incubated in a growth chamber with a 12 h light/12 h dark cycle (light intensity of 300 µmol m^-2 s^-1.) at 25°C. These incubation parameters were considered the standard conditions since they mimic the physiological conditions of the tree. For stress treatments, standard conditions were changed according to the effect studied. To assess the influence of low and high temperatures, the branches were incubated at 15 or 35°C, respectively. To study the impact of the darkness, the light was turned off. For wounding experiments, the entire surface of the olive fruit was mechanically damaged affecting mesocarp tissue, with pressure at zero time using forceps with serrated tips. The zero time of each treatment was chosen 2 h after the beginning of the light period to preserve the natural photoperiod day/night of the olive fruit. The duration of stress exposure was 24 h. This short-term exposition was selected because longer times can cause tissue damage in the olive fruits, since they are not in the olive tree but in branches incubated in chambers.

In all cases, olive tissues were frozen in liquid nitrogen immediately after harvest and stored at -80°C.

Candidate olive HPT and MPBQ MT sequences were identified in the olive transcriptome (Muñoz-Mérida et al., 2013) and the wild olive (var. sylvestris) genome (Unver et al., 2017) using the tblastn algorithm together with the amino acid sequences of Arabidopsis HPT and MPBQ MT genes (At2g18950 and At3g63410, respectively). Based on these two new sequences, specific pairs of primers for each gene were designed and used for PCR amplification with VELOCITY DNA polymerase (Bioline, Spain), which has proofreading activity. An aliquot of an olive Uni-ZAP XR cDNA library constructed with mRNA isolated from 13 WAF olive (cv. Picual) fruit (Haralampidis et al., 1998) was utilized as a DNA template. One fragment with the expected size was obtained in each reaction, subcloned into the vector pSpark^® I (Canvax, Spain), and sequenced in both directions. DNA sequencing was carried out by GATC Biotech (Germany).

The DNA sequence data were compiled and analyzed with the LASERGENE software package (DNAStar, Madison, WI). The multiple sequence alignments of olive HPT and MPBQ MT amino acid sequences were calculated using the ClustalX program and displayed with GeneDoc. Phylogenetic tree analysis was performed using the neighbor-joining method implemented in the Phylip package using Kimura’s correction for multiple substitutions and a 1000 bootstrap data set. TreeView was used to display the tree. The conserved domains in the deduced amino acid sequences were analyzed using the NCBI Conserved Domain Search (http://www.ncbi.nlm.nih.gov/structure/cdd/wrpsb.cgi) and Pfam software (https://pfam.xfam.org/). TMHMM analysis was carried out (http://www.cbs.dtu.dk/services/TMHMM/) and subcellular localization was predicted using three different programs: ProtComp 9.0 (http://www.softberry.com), WoLF PSORT (https://wolfpsort.hgc.jp/) and TargetP-2.0 (http://www.cbs.dtu.dk/services/TargetP/). Prediction of N-terminal chloroplast targeting peptides was performed using ChloroP 1.1 software (http://www.cbs.dtu.dk/services/ChloroP/).

Total RNA isolation was carried out according to Hernández et al. (2005) using 1.5 g of frozen olive tissue. RNA quality verification, removal of contaminating DNA, and cDNA synthesis were performed as described by Hernández et al. (2009).

Gene expression levels of the olive HPT and MPBQ MT were analyzed by quantitative real-time PCR (qRT-PCR) using a CFX Connect real-time PCR System and iTaq Universal SYBR Green Supermix (BioRad, California, USA) as described by Hernández et al. (2019). Primers for gene-specific amplification of OeHPT and OeMPBQ MT were designed using the Primer3 program (http://bioinfo.ut.ee/primer3/) and the Gene Runner program ( Supplementary Table 1 ). The housekeeping olive ubiquitin2 gene (OeUBQ2, AF429430) was used as an endogenous reference to normalize (Hernández et al., 2009). The qRT-PCR data were calibrated relative to the corresponding gene expression level at 12 WAF ‘Picual’ mesocarp for developmental studies, at 13 WAF ‘Arbequina’ mesocarp of FI treatment for water deficit study, and zero time for each stress treatment and cultivar, respectively, as calibrator. The relative expression level of each gene was calculated following the 2^-ΔΔCt method for relative quantification (Livak and Schmittgen, 2001). The data are presented as means ± SD of three biological replicates, each having two technical replicates per 96 well plate.

Lyophilized mesocarp tissue (500 mg DW) was processed in a glass tube with 3 mL of hexane and the mixture was shaken for 15 min including 1 min vortexing. After centrifugation at 2000 rpm for 10 min at RT, the hexane phase was recovered, and the tissue was extracted twice. The hexane phases were mixed, and hexane was added to complete an exact volume of 10 mL in a graduated flask. An aliquot of the extracts was filtered and injected into an HPLC. The tocopherols were determined according to Abdallah et al. (2015), using a Supersphere Si60 Lichrocart 250-4 HPLC cartridge (Merck), and eluted with hexane:2-propanol (99:1) at a flow rate of 1 mL/min. A fluorescence detector (Shimadzu RF535) was used, setting the excitation and emission wavelengths at 290 and 330 nm, respectively. The injection volume of the sample was 20 µL. Peaks of α-, β-, γ- and δ-tocopherol were identified using a mix standard solution of all the four compounds. The quantitative evaluation was performed by external standardization using an α-tocopherol calibration curve. The data are presented as means ± SD of three biological replicates, each having three technical replicates.

Two sequences were identified from the olive transcriptome (Muñoz-Mérida et al., 2013) and the olive (var. sylvestris) genome (Unver et al., 2017), which displayed an elevated degree of similarity to the Arabidopsis HPT and MPBQ MT genes (Collakova and DellaPenna, 2001; Cheng et al., 2003). Based on these sequences, specific primer pairs were designed and utilized for PCR amplification, along with an aliquot of an olive fruit (13 WAF) cDNA library (cv. Picual). Two full-length cDNA clones were isolated and named OepHPT and OepMPBQ MT, with sizes of 1296 and 1264 bp, respectively. They exhibited ORFs encoding predicted proteins of 408 and 346 amino acid residues, which correspond to calculated molecular masses of 45.7 and 38.8 kDa, respectively, and pI values of 9.6 for OepHPT and 9.1 for OepMPBQ MT. Alignment of the olive HPT deduced amino acid sequences with Arabidopsis HPT indicated that they displayed 60% identity ( Figure 1A ), while olive MPBQ MT shared 75% identity with Arabidopsis MPBQ MT ( Figure 1B ).

The Conserved Domain search revealed that the olive HPT protein belongs to the PLNO2878 superfamily, with homogentisate phytyltransferase activity, whereas olive MPBQ MT belongs to the PLNO2490 superfamily, with MPBQ methyltransferase activity. In addition, Pfam analysis indicated the presence of an UbiA prenyltransferase domain between amino acids 137-396 for OepHPT, and a methyltransferase and an UbiE methyltransferase domain between amino acids 123-215 and 232-262, respectively, in the case of OepMPBQ MT.

Various distinctive conserved motifs were identified in the alignment of the HPT deduced amino acid sequences ( Figure 1A ). Among them, OepHPT contains prenyl-DP and divalent cation binding motifs typical for polyprenyltransferases (Lopez et al., 1996; Collakova and DellaPenna, 2001). In addition, an Asp-rich motif is also present in the olive HPT sequence, which is involved in substrate binding (Huang et al., 2014). Regarding olive MPBQ MT ( Figure 1B ), three conserved domains (SAM I, II, and III) were detected in the alignment, which are characteristic of S-adenosylmethionine-dependent methyltransferases (Kagan and Clarke, 1994; Joshi and Chiang, 1998). SAM I and II correspond to the binding sites for substrate S-adenosylmethionine, while SAM III corresponds to the binding site of catalytic products.

Regarding the subcellular localization, analysis of the deduced olive HPT and MPBQ MT protein sequences with subcellular localization prediction software such as ProtComp, WoLF PSORT, or TargetP suggests that both proteins could be located in the chloroplast. In addition, an N-terminal transit peptide with the characteristic features of chloroplast targeting peptides was detected by ChloroP software in both sequences ( Figure 1 ), with a predicted cleavage site after Cys at residue 39 and Ser at residue 65 for OepHPT and OepMPBQ MT, respectively. In line with these observations, experimental evidence of chloroplast localization for HPT and MPBQ MT from sweet potato in tobacco leaves has been reported (Ji et al., 2016).

As shown in Supplementary Figure 1 , both olive HPT and MPBQ MT proteins are highly hydrophobic based on hydropathy plotting (Kyte and Doolittle, 1982), indicating that they are membrane proteins. Concerning the membrane topology, transmembrane predictions based on a hidden Markov model (TMHMM) analysis of the olive HPT and MPBQ MT proteins were generated ( Supplementary Figure 2 ). OepHPT showed nine putative transmembrane domains as reported for other plant HPT proteins (Wunnakup et al., 2018), whereas, in the OepMPBQ MT sequence, only one transmembrane domain was found in the C-terminal region at positions 316-335. This transmembrane domain was previously identified in the Arabidopsis MPBQ MT and it has been suggested that the protein is anchored to the inner chloroplast envelope membrane at this position (Motohashi et al., 2003).

Two unrooted phylogenetic trees based on deduced amino acid sequences of known and characterized plant HPT ( Figure 2A ) and MPBQ MT ( Figure 2B ) were generated to investigate the phylogenetic relationship of olive HPT and MPBQ MT, respectively. In agreement with previous findings (Chaudhary and Khurana, 2009; Ji et al., 2016), plant HPT and MPBQ MT could be classified into two separate monocot and dicot-specific clades, with OepHPT and OepMPBQ MT being positioned, respectively, in the branch accompanied by other HPT and MPBQ MT from dicots plants. It seems that HPT and MPBQ MT evolved differently between dicots and monocots species.

Altogether, sequence analysis of the olive HPT and MPBQ MT displayed the distinguishing characteristics typical of the HPT and MPBQ MT families and along with phylogenetic analysis indicated that they code for homogentisate phytyltransferase and methylphytylbenzoquinol methyltransferase enzymes, respectively.

Olive HPT and MPBQ MT transcript levels were determined in distinct olive organs and tissues from ‘Picual’ and ‘Arbequina’, the two main cultivars for oil production, using qRT-PCR ( Figure 3 ) to investigate their physiological function. Both genes exhibit higher expression levels in green tissues from both cultivars such as young drupes, green mesocarp, and leaves compared to mature mesocarp or young seeds. Analogous observations have been reported in the case of oil palm HPT (Ling et al., 2016) and tomato MPBQ MT (Quadrana et al., 2014). The high transcript levels detected for both genes in green tissues are consistent with the elevated tocopherol content characteristic of photosynthetic organs, given the protective role of tocopherols against photooxidation and photoinactivation (Havaux et al., 2005). In addition, all these data indicate a spatial regulation of HPT and MPBQ MT genes in olive considering that they were differentially expressed in all organs and tissues studied.

Afterward, the expression levels of olive HPT and MPBQ MT genes in the olive fruit were studied in more detail. Specifically, seeds and mesocarp tissue during the development and ripening of olive fruit from the cultivars Picual and Arbequina were analyzed.

Concerning developing seeds ( Supplementary Figure 3 ), gene expression analysis revealed that HPT and MPBQ MT genes exhibited an increase in their transcript levels during the whole period of olive fruit development and ripening in both cultivars, showing maximum values for both genes at 35 and 31 WAF for ‘Picual’ and ‘Arbequina’, respectively. The detected increase of HPT and MPBQ MT transcripts during seed development has been previously observed in other plants, as in the case of the rice and blackberry HPT (Wang et al., 2013; Yang et al., 2020), and the HPT and MPBQ MT from oat (Gutierrez-Gonzalez and Garvin, 2016). However, in the case of olive, this increase in HPT and MPBQ MT expression levels detected in the seed is not in accordance with the slight continuous decrease of tocopherol content observed in oils from Picual and Arbequina cultivars obtained at different stages of development and ripening (Pérez et al., 2019), likely because the contribution of the seed to the final composition of the VOO is very minor (Hernández et al., 2016).

A similar investigation was performed in the olive mesocarp of Picual and Arbequina cultivars ( Figure 4 ). The observed pattern in HPT and MPBQ MT expression levels during olive mesocarp development and ripening ( Figure 4B ) is in line with that of the total tocopherol content and the (α+γ)-tocopherol/(β+δ)-tocopherol ratio ( Figure 4A ), respectively, with a peak at 28 WAF in the case of ‘Picual’, coinciding with the beginning of the ripening period. In the case of the cultivar Koroneiki, a decrease in their expression levels has been described for both genes (Georgiadou et al., 2015). All these data point out that in olive fruit the expression of HPT and MPBQ MT genes seems to be temporally regulated.

Concerning other fruits, a reduction of HPT expression levels has been reported during fruit ripening in apple and the pulp of Citrus species (Seo et al., 2011; Rey et al., 2021b), whereas a significant up-regulation was observed for oil palm HPT at late stages of mesocarp ripening (Ling et al., 2016). On the contrary, MPBQ MT from tomato fruit and the pulp of Citrus species exhibited comparable transcript levels along development and ripening (Quadrana et al., 2014; Rey et al., 2021b).

Our study was expanded to other olive cultivars characterized by a low (190 and 285 mg/kg oil for ‘Klon-14’ and ‘Abou Kanani’, respectively) or high (820 and 1610 mg/kg oil for ‘Piñonera’ and ‘Dokkar’, respectively) tocopherol content in their oils (Pérez et al., 2019), using mesocarp tissue corresponding to four different representative stages of fruit development and ripening ( Figure 5 ). Unlike OeMPBQ MT which shows similar expression levels for each stage of these four cultivars ( Figure 5B ), the transcript levels of OeHPT were higher in those cultivars with a higher tocopherol content such as ‘Piñonera’ and ‘Dokkar’ ( Figure 5A ). In contrast, in ‘Klon-14’ and ‘Abou Kanani’, characterized by a low content of tocopherols, their expression levels were lower. These data suggest that HPT, but not MPBQ MT, could be involved in the regulation of the tocopherol biosynthetic pathway in olive mesocarp at the transcriptional level, which is consistent with its role in catalyzing the first committed step of this route. However, the variation in the OeHPT and OeMPBQ MT transcript levels during olive mesocarp development and ripening ( Figure 5B ) was not parallel to that of the total tocopherol content and the (α+γ)-tocopherol/(β+δ)-tocopherol ratio ( Figure 5A ), respectively. This fact indicates that not only HPT but also other factors such as the availability of HGA and PDP as precursors regulate the flux of the tocopherol biosynthetic pathway (Pellaud and Mène-Saffrané, 2017).

Collectively, these results point out that olive HPT and MPBQ MT participate in the biosynthesis of the tocopherols present in VOO and their gene expression levels are cultivar-dependent, as reported from five Greek cultivars (Georgiadou et al., 2019).

It is well described that drought stress increases tocopherol content in the leaves of several plant species, including olive juvenile trees (Baccari et al., 2020). The effect of three distinct RDI treatments on the oil accumulation, fatty acid profile, and fatty acid desaturase gene expression levels in ‘Arbequina’ fruit mesocarp has been previously studied by our group (Hernández et al., 2018). However, data related to the effect of water stress on the expression levels of olive genes belonging to the tocopherol biosynthetic pathway are lacking. In the present work, higher expression levels of HPT and MPBQ MT genes were found along all the fruit developmental and ripening processes in ‘Arbequina’ mesocarp from olives subjected to 30RDI and 60RDI treatments, which produced substantial levels of water stress, in comparison to FI treatment ( Figure 6 ). This result is in agreement with the increase of tocopherol content reported by García et al. (2017) in oils obtained from the same olive fruit samples when comparing FI and RDI treatments (255, 327, and 321 mg/kg oil; for FI, 60RDI and 30RDI treatments, respectively). Therefore, the tocopherol biosynthetic pathway appears to be transcriptionally up-regulated in water-stressed olive mesocarp, with OeHPT and OeMPBQ MT genes increasing their expression levels and causing an increment in the tocopherol content in ‘Arbequina’ mesocarp and, consequently, in the corresponding oils. In line with these data, the transcription of genes involved in tocopherol biosynthesis in olives from the Koroneiki cultivar was up-regulated in response to lower rainfall during consecutive seasons (Georgiadou et al., 2016).

An increase in the HPT expression levels has also been detected in leaves from other plants under drought stress, such as Cistus cetricus (Munné-Bosch et al., 2009), lettuce (Ren et al., 2011), sweet potato (Ji et al., 2016), or barley (Schuy et al., 2019). Furthermore, it has been reported that drought-induced expression of HPT from Solanum chilense in tobacco plants results in the accumulation of α-tocopherol and increases tolerance to drought stress by delaying foliar tissue damage (Espinoza et al., 2013), demonstrating the role of HPT in drought stress resistance. On the contrary, no significant effect on MPBQ MT transcript levels has been described in plants under drought stress to date.

To investigate the effect of several abiotic stresses on the expression levels of the olive HPT and MPBQ MT genes in mesocarp tissue from ‘Picual’ and ‘Arbequina’, olive tree branches holding olive fruit at the turning stage (28 WAF) were incubated for 24 h modifying the standard conditions (25°C with 12 h light/12 h dark cycle) dependent on the effect to be examined. No significant changes in the olive HPT and MPBQ MT transcript levels were observed in the mesocarp tissue when olive fruit were incubated under the above-mentioned standard conditions ( Supplementary Figure 4 ).

It has been shown that tocopherols play a key role in the low-temperature adaptation of plants such as maize (Leipner et al., 1999) or Arabidopsis (Maeda et al., 2008). When olive fruit was incubated at a low temperature (15°C) a significant increment in the expression levels of HPT and MPBQ MT genes in the mesocarp from both cultivars was noticed ( Figure 7A ). In ‘Picual’, a strong transient increase was observed for the transcript levels of both genes with maximum values after 0.5-1 h and 3-6 h of treatment for HPT and MPBQ MT, respectively. In contrast, in ‘Arbequina’ mesocarp a continuous increase was detected along all the incubation time, being higher in the case of the OeMPBQ MT gene. A similar up-regulation of HPT and MPBQ MT transcript levels has been described in mandarin fruit and grapefruit during long-term cold storage (Rey et al., 2021a; Rey et al., 2021c). In the same way, when Chlamydomonas reinhardtii cells were incubated at low temperature, an increase in HPT expression levels was observed (Gálvez-Valdivieso et al., 2015). In addition, HPT has been demonstrated to be essential for cold tolerance and low-temperature adaptation, since Arabidopsis and rice vte2 mutants were hypersensitive to cold stress (Maeda et al., 2006; Wang et al., 2017; Zhang et al., 2018).

Regarding high temperature, when olive fruit from both cultivars were incubated at 35°C the transcript levels of HPT and MPBQ MT showed a severe reduction until undetectable values were reached ( Figure 7B ). Both genes started to decrease their transcript levels after 6 h of incubation in the case of ‘Picual’, while in ‘Arbequina’ the diminution began sharp and rapidly from the start of the treatment. Interestingly, it has been reported that total tocopherol concentrations in olive fruit at the end of the oil accumulation period were generally higher when the air temperature was increased by 4°C in young trees from Arbequina and Coratina cultivars (Hamze et al., 2022). However, these authors suggest that the observed increase in total tocopherols appeared to be related to a reduction in fruit oil concentration with heating. In line with our results, C. reinhardtii cells incubated at high temperature exhibited a slight decrease in HPT expression levels (Gálvez-Valdivieso et al., 2015), and leaves of Cinnamomum camphora under high temperature showed a down-regulation of the MPBQ MT transcript levels (Wang et al., 2023).

Previous studies in plant tissues demonstrated that the tocopherol content is light regulated (Lichtenthaler, 2007). To examine whether darkness affects HPT and MPBQ MT expression levels in olive mesocarp from Picual and Arbequina cultivars, olive branches were incubated at 25°C for 24 h in the darkness. Expression analysis showed a substantial decline of OeHPT and OeMPBQ MT transcript levels in both cultivars mostly during the first 3-6 h of treatment, remaining with small values the rest of the experiment ( Figure 8A ). In agreement with these results, a down-regulation of the HPT expression levels has also been observed in dark-grown tomato fruit and grapefruit (Gramegna et al., 2019; Rey et al., 2021c). In addition, Gálvez-Valdivieso et al. (2015) reported a fast and strong reduction of HPT transcripts in C. reinhardtii cells transferred to darkness. In the case of MPBQ MT, very low transcript levels were found in the leaves of Arabidopsis plants grown in the dark (Motohashi et al., 2003). These data indicate that HPT and MPBQ MT are regulated by light.

Finally, the impact of wounding on the HPT and MPBQ MT gene expression levels was investigated for the first time. In particular, this effect was studied in the mesocarp of olive fruit from Picual and Arbequina cultivars subjected to mechanical damage from branches incubated at standard conditions. In both cultivars, OeHPT and OeMPBQ MT transcript levels exhibited a slight transient induction, showing peak values after 1 h of treatment and followed by a decrease, especially marked in the case of MPBQ MT ( Figure 8B ). Notably, a role for tocopherol biosynthesis in Arabidopsis basal immunity to bacterial infection has been recently proposed (Stahl et al., 2019). Specifically, a substantial increase of HPT gene expression levels was detected in Arabidopsis leaf in response to inoculation with Pseudomonas syringae, and an enhanced susceptibility toward P. syringae was observed in Arabidopsis vte2 mutant plants. In contrast, no significant differences in HPT and MPBQ MT transcript levels were found in sweet potato leaves infected with the bacterial pathogen Pectobacterium chrysanthemi compared to the control (Ji et al., 2016).