Insights into the Genes Involved in ABA Biosynthesis and Perception during Development and Ripening of the Chilean Strawberry Fruit

Hormones act as master ripening regulators. In non-climacteric fruit, ABA plays a key role in ripening. Recently, we confirmed in Fragaria chiloensis fruit that in response to ABA treatment the fruit induces ripening-associated changes such as softening and color development. In consequence of these phenotypic changes, transcriptional variations associated with cell wall disassembly and anthocyanins biosynthesis were reported. As ABA stimulates the ripening of F. chiloensis fruit, the molecular network involved in ABA metabolism was analyzed. Therefore, the expression level of genes involved in ABA biosynthesis and ABA perception was quantified during the development of the fruit. Four NCED/CCDs and six PYR/PYLs family members were identified in F. chiloensis. Bioinformatics analyses confirmed the existence of key domains related to functional properties. Through RT-qPCR analyses, the level of transcripts was quantified. FcNCED1 codifies a protein that displays crucial functional domains, and the level of transcripts increases as the fruit develops and ripens, in parallel with the increment in ABA. In addition, FcPYL4 codifies for a functional ABA receptor, and its expression follows an incremental pattern during ripening. The study concludes that FcNCED1 is involved in ABA biosynthesis; meanwhile, FcPYL4 participates in ABA perception during the ripening of F. chiloensis fruit.


Introduction
Chilean strawberry (Fragaria chiloensis (L.) Mill. subsp. chiloensis Staudt) is a native wild species from Chile that produces a fruit which is highly appreciated for its good organoleptic qualities such as aroma and taste, its exotic white color and remarkable biotic stress tolerance [1][2][3]. Albeit those extraordinary properties, the fruit softens very fast and intensively during ripening which reduces its post-harvest shelf life.
The Chilean strawberry fruit ripens with a non-climacteric pattern. Ripening is a complex process, during which a fruit develops into a fleshy, colorful and tasty fruit. For these biochemical and physiological changes, a series of metabolic pathways are switched on, such as the degradation of chlorophyll and starch, biosynthesis of pigments and volatile compounds and accumulation of sugars and organic acids [4]. It is important to highlight that the perfect coordination of those metabolic pathways is needed to develop a fruit of good quality [5]. In climacteric fruits, most of those metabolic pathways are regulated and coordinated by the hormone ethylene; however, in non-climacteric fruit ethylene does not have the same effect [5]. Several hormones have been analyzed as master ripening regulators in non-climacteric fruit, including abscisic acid (ABA) and auxins.
Hormones play essential roles in the control and coordination of several physiological events, including fruit ripening. In strawberry, hormonal changes include the rise of ABA levels along the ripening development, and this increment coincides with the drop of of ABA, induces the expression of genes related to fruit softening (FvePL) [27]. FvePYL2 transcripts reached a relatively high expression level at the red fruit stage, whereas other FvePYLs have diversified expression patterns [14].
In F. chiloensis, several pieces of evidence confirm that ABA is involved in its ripening as it induces the transcriptional changes required for fruit softening and color development [5,10,11,[28][29][30]. As ABA seems to have key participation in the ripening stimulation of F. chiloensis fruit, the molecular network involved in ABA metabolism in ripening fruit was analyzed. Currently, there is little evidence of the participation of NCEDs on F. chiloensis fruit [11]; however, there is none for PYR/PYLs. Additionally, these genes belong to a gene family, and several members could or could not be involved in the process. Therefore, the clarification of which gene family member takes part during fruit ripening will precisely determine the critical players for softening and color development. Hence, the expression level of genes involved in ABA biosynthesis and ABA perception was quantified in F. chiloensis fruit as a way to provide further molecular evidence of the participation of ABA on the ripening development of this fruit. As several NCEDs and PYR/PYLs family members have been identified in the fruit species yet only a few of them seem to participate during fruit ripening, this work aims to identify those members involved in the ripening of this non-climacteric fruit.

Identification of FcNCED/CCDs Family Genes and Analysis of Deduced Amino Acid Sequences
Through the annotation process by different databases, it was possible to identify two NCEDs and two CCDs sequences (FcNCED1, FcNCED3, FcCCD1 and FcCCD4) into the F. chiloensis transcriptome (Table 1). These sequences were named according to their closest relative sequence within the F. vesca genome. The analysis of their coding sequences recognized the characteristic domain of NCED/CCD proteins 'Retinal pigment epithelial membrane protein' (RPE65) in all sequences, the domain responsible for the binding of Fe 2+ . An evolutionary relationship study between different rosaceas species (F. x ananassa, F. vesca, Malus domestica, Rosa chinensis, Rosa rugosa and Prunus avium), Arabidopsis thaliana and F. chiloensis provided a phylogenetic tree that clearly separates NCEDs from CCDs into different subfamilies, and CCD comprises two clades (CCD1 and CCD4) ( Figure 1). The translated amino acid sequences of F. chiloensis share a high sequence identity with F. vesca and F. x ananassa, confirming the parental relationship with this last species.
A comparative alignment was performed between the deduced protein sequences of FcNCEDs and FcCCDs and that of the crystal ZmVP14 (maize viviparous14, a key enzyme in the biosynthesis of ABA) (PDB code: 3NPE) ( Figure 2) [31]. The alignment showed that the sequences share cleavage domain 2 (CV2) that interacts with 9 cis-violaxanthin and partly CV1 (Figure 2; green boxes). In addition, 2 of the 6 residues described to bind the 9-cis bond of violaxanthin and the near methylcyclohexane are highly conserved among the sequences. This is the case for the 2 conserved Phe (F52 and F96) (Figure 2; red arrows). Meanwhile, the other 4 residues vary in amino acid type; however, their hydrophobic character is maintained: Met (M357) is changed to Ile or Leu, Val (V403) for Ale or Phe, Trp (W426) for Met or Ile and finally Pro (P427) for Ala ( Figure 2; red arrows). On the other hand, all the residues that coordinate Fe 2+ (H223, H272, H337 and H515) ( Figure 2; light blue arrows) and the amino acids stabilizing the His (E455) (Figure 2; blue arrows) are highly conserved between the sequences. The residues that bind the isoprene chain are also highly conserved between the sequences, except for FcCCD4. Only Glu (E189) is conserved throughout all sequences; meanwhile, replacements of Ala (A139) for Val, Ile (I140) for Phe, Leu (L143) for Phe, Asp (D190) for Ser, Met (M270) for Leu and Phe (F289) for Met are shown in FcCCD4 (Figure 2; green arrows). A comparative alignment was performed between the deduced protein sequences of FcNCEDs and FcCCDs and that of the crystal ZmVP14 (maize viviparous14, a key enzyme in the biosynthesis of ABA) (PDB code: 3NPE) ( Figure 2) [31]. The alignment showed that the sequences share cleavage domain 2 (CV2) that interacts with 9 cis-violaxanthin and partly CV1 (Figure 2; green boxes). In addition, 2 of the 6 residues described to bind the 9cis bond of violaxanthin and the near methylcyclohexane are highly conserved among the sequences. This is the case for the 2 conserved Phe (F52 and F96) (Figure 2; red arrows). Meanwhile, the other 4 residues vary in amino acid type; however, their hydrophobic character is maintained: Met (M357) is changed to Ile or Leu, Val (V403) for Ale or Phe, Trp (W426) for Met or Ile and finally Pro (P427) for Ala ( Figure 2; red arrows). On the other hand, all the residues that coordinate Fe 2+ (H223, H272, H337 and H515) (Figure 2; light blue arrows) and the amino acids stabilizing the His (E455) (Figure 2; blue arrows) are highly conserved between the sequences. The residues that bind the isoprene chain are also highly conserved between the sequences, except for FcCCD4. Only Glu (E189) is conserved throughout all sequences; meanwhile, replacements of Ala (A139) for Val, Ile (I140) for Phe, Leu (L143) for Phe, Asp (D190) for Ser, Met (M270) for Leu and Phe (F289) for Met are shown in FcCCD4 (Figure 2; green arrows).  Table S1.
There are 5 residues which have been described to bind the methylcyclohexane group, and only one of them is conserved among the FcNCED/CCD sequences: Met (M295). For the remaining four residues, high variability is detected; however, they maintain a high degree of hydrophobicity character ( Figure 2; orange arrows). The residues that bind the carotenoid section between C9 and C15 are also highly conserved between the sequences, being Phe in all sequences (F97, F336 and F514) except for FcCCD4 which in replacement of F336 there is a Met (Figure 2; purple arrow).
Importantly, two amino acid residues have been described as key to differentiate between NCED and CCD (obtained from [31]). Those residues are Leu (L95 in 3NPE crystal) and Val/Ala or Ile (V403 in 3NPE crystal) which are replaced by a Trp (W95) and a Phe (F403) in CCDs (Figure 2; asterisks). The sequences of FcNCED3 and FcCDD1 agree with bibliography; however, FcNCED1 and FcCCD4 present different expected residues in location 95. FcNCED1 presents a Phe (F95) instead of the expected Leu for NCED activity, and FcCCD4 displays a Leu (L95) instead of the expected Trp for CCD. Red arrows indicate the binding site for 9cis-violaxanthin at 9-cis bond and the near methylcyclohexane; light-blue arrows indicate binding site for Fe 2+ ; blue arrows indicate amino acids holding the histidines which binds Fe; green arrows indicate binding site for 9-cis-violaxanthin at isoprene chain; orange arrows indicate binding site for 9-cis-violaxanthin at methylcyclohexane; purple arrow indicates binding site for 9-cis-violaxanthin at carotenoid section between C9-C15; black arrows indicate amino acids which interact with cell membrane; and asterisks indicate amino acids that differentiate NCED and CCD (information obtained from [31]). Sequences were aligned using MAFFT version 7. Red arrows indicate the binding site for 9-cis-violaxanthin at 9-cis bond and the near methylcyclohexane; light-blue arrows indicate binding site for Fe 2+ ; blue arrows indicate amino acids holding the histidines which binds Fe; green arrows indicate binding site for 9-cis-violaxanthin at isoprene chain; orange arrows indicate binding site for 9-cis-violaxanthin at methylcyclohexane; purple arrow indicates binding site for 9-cis-violaxanthin at carotenoid section between C9-C15; black arrows indicate amino acids which interact with cell membrane; and asterisks indicate amino acids that differentiate NCED and CCD (information obtained from [31]). Sequences were aligned using MAFFT version 7.

Identification of FcPYR/PYL Family Genes and Analysis of Amino Acid Sequences
The search for PYR/PYL genes within the F. chiloensis transcriptome allows the identification of one PYR and five PYLs sequences (FcPYR1, FcPYL2, FcPYL4, FcPYL8, FcPYL9 and FcPYL12) ( Table 2). The sequences were named in accordance with F. vesca sequences. Their coding sequences contain the conserved domain described as 'Polyketide cyclase/dehydrase and lipid transport', which is present in most of the sequences, except for FcPYL2. The phylogenetic analysis grouped them into three subfamilies (I-III). As the nomenclature used to name each subfamily is not consistent in the literature, we decided to name them as reported by [32], based in A. thaliana. FcPYR1 and FcPYL2 belong to subfamily III; FcPYL4 and FcPYL12 are members of subfamily II (separated into two clades); meanwhile, FcPYL8 and FcPYL9 belongs to subfamily I, also known as the AtPYL8-like subfamily ( Figure 3).  (5), Prunus persica (6) and F. vesca (6). The phylogenic tree was built using the neighbor-joining method with a bootstrap consensus tree inferred from 5000 replicates. Bootstrap values are indicated in the figure; values over 50 indicate clade robustness. The sequences were grouped into three subfamilies (I-III) named as reported by [32].The list of protein sequences employed in the analysis, including Genbank accession numbers, is shown in Table S2. Prunus persica (6) and F. vesca (6). The phylogenic tree was built using the neighbor-joining method with a bootstrap consensus tree inferred from 5000 replicates. Bootstrap values are indicated in the figure; values over 50 indicate clade robustness. The sequences were grouped into three subfamilies (I-III) named as reported by [32].The list of protein sequences employed in the analysis, including Genbank accession numbers, is shown in Table S2.
A comparative alignment was performed between PYR/PYLs amino acid sequences of F. chiloensis and that of AtPYL2, whose 3D structure has been analyzed (PDB code: 3KAZ) ( Figure 4) [33]. The alignment indicated the existence of important domains for these ABA receptors. The four connecting loops (CL1-CL4) which have been described as the interaction domains with ABA are highly conserved within the sequences, except for CL1 which is present in some of them (FcPYR1, FcPYL4 and FcPYL8) and CL2 which is absent in FcPYL2. Further, CL2 has been described as a gate and latch structure [34].  (6) were aligned with another 23 sequences from F. x ananassa (1), M. domestica (5), R. chinensis (5), Prunus persica (6) and F. vesca (6). The phylogenic tree was built using the neighbor-joining method with a bootstrap consensus tree inferred from 5000 replicates. Bootstrap values are indicated in the figure; values over 50 indicate clade robustness. The sequences were grouped into three subfamilies (I-III) named as reported by [32].The list of protein sequences employed in the analysis, including Genbank accession numbers, is shown in Table S2. Red arrows indicate the binding site for carboxyl group of ABA; blue arrows indicate hydrogen bonds with the carboxyl group of ABA; green arrows indicate binding site for methyl group of ABA; yellow arrows indicate binding site for dimethyl group of ABA; asterisks indicate amino acids present in the hydrophobic pocket; and the orange arrow indicates the binding site for PP2C protein (information captured from [33,34]). Sequences were aligned using MAFFT version 7.
Furthermore, the sequences contain the amino acid residues involved in ABA and PP2C binding. The protein sequences coded by FcPYR1, FcPYL4 and FcPYL8 present a highly conserved Lys residue (K52) that binds to the carboxyl group of ABA (Figure 4; red arrow). All sequences possess a highly conserved Glu residue (E135) that also binds to a carboxyl group of ABA, except for FcPYL12 which presents a Gln (Figure 4; red arrow). All sequences share 2 highly conserved residues, Glu (E86) and Asn (N161), which interact through hydrogen bonds with the carboxyl group of ABA (Figure 4; blue arrow).
An amount of 5 of the 6 residues that bind to the methyl group of ABA are highly conserved between the sequences: Phe (F54), Val (V75), Leu (L79), Pro (P80) and Phe (F153); however, the sixth residue, Val (V157), may vary for a Leu or Ile among the different FcPYR/PYL proteins (Figure 4; green arrow). Some of these residues are also involved in the hydrophobic pocket (F54, V75, L79 and F153), in addition to His (H107), Leu (L109) and Tyr (Y112) which hold a high degree of conservation; however, other residues are modified in this hydrophobic pocket, Ala (A81) for Gly, Val (V157) for Leu or Ile and Val (V158) for Ile (Figure 3; asterisk).
There are 3 residues that have been identified to bind the dimethyl group of ABA (Ala 81, Ser 84 and Val 102) (Figure 4; yellow arrow), which are mostly conserved among F. chiloensis sequences. Few changes exist as there is a Gly instead of Ala81 in FcPYL12, and Val102 may vary for Ile or Met. Finally, a highly conserved His among sequences (H107) allows the interaction of PYR/PYL with PP2C ( Figure 4; orange arrow).

Expression Analysis of ABA Biosynthesis Related Genes in F. chiloensis
The relative expression level of ABA biosynthesis related genes, FcNCED and FcCCD, was quantified by RT-qPCR in different vegetative tissues. A high relative expression level of FcNCED1 transcripts was noticed in roots, and a middle expression level in runners and low level in the other vegetative tissues were analyzed ( Figure 5A). In the case of FcCCD1 transcripts, a high expression level was detected in flowers and a low expression level in the other vegetative tissues ( Figure 5B). On the other hand, transcripts levels for FcNCED3 and FcCCD4 were extremely low in the vegetative tissues under study. Each value corresponds to the mean ± SE of three independent RNA extractions and qPCR analysis using three technical replicates. Asterisks indicate significant differences compared to calibrated tissue (* p < 0.05, *** p < 0.001, or **** p < 0.0001; one-way ANOVA with Dunnet correction post hoc). Each value corresponds to the mean ± SE of three independent RNA extractions and qPCR analysis using three technical replicates. Asterisks indicate significant differences compared to calibrated tissue (* p < 0.05, *** p < 0.001, or **** p < 0.0001; one-way ANOVA with Dunnet correction post hoc).
In the development of fruit tissue, the relative accumulation of FcNCED1 transcripts displays an increment during development with a maximum value at the C3 stage ( Figure 6A). FcNCED3 shows an early increment at the C2 stage which decays over time ( Figure 6B). FcCCD1 displays the same expression profile as FcNCED3 ( Figure 6C). Finally, FcCCD4 is mainly expressed at C3 and C4 stages with no significant differences among those stages ( Figure 6D).

R PEER REVIEW
10 of 22 Heatmap analysis of non-calibrated qPCR values for ABA biosynthesis genes indicates that the expression level of FcNCED1 was higher than for the other genes in fruit samples (Figure 7). In comparison, the expression level in fruit tissues followed the order FcNCED1 Heatmap analysis of non-calibrated qPCR values for ABA biosynthesis genes indicates that the expression level of FcNCED1 was higher than for the other genes in fruit samples (Figure 7). In comparison, the expression level in fruit tissues followed the order

Expression Analysis of ABA Receptor Genes during Development of F. chiloensis Fruit and Vegetative Tissues
The expression of FcPYR/FcPYLs was quantified by qPCR analysis in different vegetative tissues in order to provide a molecular evidence of ABA perception (Figure 8). Among the six different genes analyzed, only five of them were transcribed in the different vegetative tissues under analysis; an extremely low expression level was detected for FcPYR1. A high relative expression level of FcPYL2 and FcPYL12 was detected in roots and flowers and a middle to low expression in runners, leaves and stems. A relatively high level of transcripts accumulation was also observed for FcPYL4 in roots. In the case of FcPYL8 and FcPYL9, there were no significant differences in transcripts accumulation among the vegetative tissues analyzed. Heatmap analysis was built with non-calibrated expression data and indicated that the most expressed gene among vegetative tissues is FcPYL8, followed by a middle-high expression level in the case of FcPYL4 and FcPYL12 and a middle level for FcPYL9 ( Figure  9).

Expression Analysis of ABA Receptor Genes during Development of F. chiloensis Fruit and Vegetative Tissues
The expression of FcPYR/FcPYLs was quantified by qPCR analysis in different vegetative tissues in order to provide a molecular evidence of ABA perception (Figure 8). Among the six different genes analyzed, only five of them were transcribed in the different vegetative tissues under analysis; an extremely low expression level was detected for FcPYR1. A high relative expression level of FcPYL2 and FcPYL12 was detected in roots and flowers and a middle to low expression in runners, leaves and stems. A relatively high level of transcripts accumulation was also observed for FcPYL4 in roots. In the case of FcPYL8 and FcPYL9, there were no significant differences in transcripts accumulation among the vegetative tissues analyzed. Heatmap analysis was built with non-calibrated expression data and indicated that the most expressed gene among vegetative tissues is FcPYL8, followed by a middle-high expression level in the case of FcPYL4 and FcPYL12 and a middle level for FcPYL9 (Figure 9).
During the development and ripening of F. chiloensis fruit there are significant changes in the expression level of all FcPYL/PYR genes under study ( Figure 10). In FcPYL genes, there is a clear increment in the level of transcripts along fruit development and ripening, although different profiles can be observed. In the case of FcPYL2, there is a clear increment in transcripts from stage 3, with maximum values at stage 4. In the case of FcPYL4, there is also a sudden rise at stage 3; however, it decreases after that. In the case of FcPYL9, the rise in transcripts is observed from stage 2, reaching the maximum value at stage 3 and then decreasing at stage 4. In the case of FcPYL12, a rise was observed at stage 2, and the level was maintained until the end of ripening. In the case of FcPYR1, a completely different expression profile was observed, with a rise in expression between stages 1 and 2, followed by a reduction at stages 3 and 4. Heatmap analysis indicated that FcPYL8 is the most expressed gene in ripe fruit tissues, followed by FcPYL2, FcPYL12, FcPYL4 and FcPYL9 ( Figure 10). The expression level of FcPYR1 is very low compared to the other PYL genes.      in transcripts is observed from stage 2, reaching the maximum value at stage 3 and then decreasing at stage 4. In the case of FcPYL12, a rise was observed at stage 2, and the level was maintained until the end of ripening. In the case of FcPYR1, a completely different expression profile was observed, with a rise in expression between stages 1 and 2, followed by a reduction at stages 3 and 4. Heatmap analysis indicated that FcPYL8 is the most expressed gene in ripe fruit tissues, followed by FcPYL2, FcPYL12, FcPYL4 and FcPYL9 (Figure 10). The expression level of FcPYR1 is very low compared to the other PYL genes. Each value corresponds to the mean ± SE of three independent RNA extractions and three technical replicates. Asterisks indicate significant differences compared to C1 stage (* p < 0.05, ** p < 0.01, *** p < 0.001, or **** p < 0.0001; one-way ANOVA with Dunnet correction post hoc).

During Fruit Development ABA Levels Correlate with ABA Biosynthesis Genes and ABA Receptor Genes
ABA levels during the development and ripening of F. chiloensis fruit have been reported earlier [11], showing an increment as fruit ripens, with higher levels at stages three and four. Pearson correlation values indicate a direct correlation between ABA levels during fruit development and the expression level of ABA biosynthesis genes such as FcNCED1 (r= 0.86) and FcCCD4 (r = 0.99) (Figure 11). This indicates that FcNCED1 and FcCCD4 may be involved in the biosynthesis of ABA during the ripening of F. chiloensis fruit. In addition, the increment in ABA is coincident with the increased expression of FcPYL2 (r = 0.70) and FcPYL4 (r = 0.64), suggesting their participation in ABA sensing.
Positive correlations were found between the expression levels of the different genes:  Each value corresponds to the mean ± SE of three independent RNA extractions and three technical replicates. Asterisks indicate significant differences compared to C1 stage (* p < 0.05, ** p < 0.01, *** p < 0.001, or **** p < 0.0001; one-way ANOVA with Dunnet correction post hoc).

During Fruit Development ABA Levels Correlate with ABA Biosynthesis Genes and ABA Receptor Genes
ABA levels during the development and ripening of F. chiloensis fruit have been reported earlier [11], showing an increment as fruit ripens, with higher levels at stages three and four. Pearson correlation values indicate a direct correlation between ABA levels during fruit development and the expression level of ABA biosynthesis genes such as FcNCED1 (r= 0.86) and FcCCD4 (r = 0.99) (Figure 11). This indicates that FcNCED1 and FcCCD4 may be involved in the biosynthesis of ABA during the ripening of F. chiloensis fruit. In addition, the increment in ABA is coincident with the increased expression of FcPYL2 (r = 0.70) and FcPYL4 (r = 0.64), suggesting their participation in ABA sensing.

Discussion
Carotenoid cleavage dioxygenases (CCDs) are a family of enzymes that catalyzes the cleavage of carotenoids into smaller apocarotenoids, molecules with relevant properties including pigments, flavor and aroma compounds and plant growth regulators such as vitamin A, ABA and strigolactone [35,36]. The CCD gene family comprises CCD and NCED sub- Figure 11. Pearson correlation analysis between ABA levels during fruit development, ABA biosynthesis genes and ABA receptor genes. Raw expression values of each gene and ABA level data from the four development and ripening stages of F. chiloensis fruit were analyzed in order to find correlations.

Discussion
Carotenoid cleavage dioxygenases (CCDs) are a family of enzymes that catalyzes the cleavage of carotenoids into smaller apocarotenoids, molecules with relevant properties including pigments, flavor and aroma compounds and plant growth regulators such as vitamin A, ABA and strigolactone [35,36]. The CCD gene family comprises CCD and NCED subfamilies. Importantly, the NCED enzyme can catalyze the cleavage of the 11, 12 double bond of violaxanthin (C40) or neoxanthin (C40) to form xanthoxin (C15), the precursor of ABA. The reaction carried out by NCED is considered the rate-limiting step in ABA biosynthesis.
In Fragaria chiloensis, two NCEDs and two CCDs were identified within the transcriptome prepared from fruit samples [37]. In general, the number of CCD genes identified in the genome of different fruit species is reduced compared to the number described in non-fruit species (30 in oilseed rape; 19 in tobacco) [35]. In the genome of watermelon (Cucumis lanatus), melon (Cucumis melo) and cucumber (Cucumis sativus) a recent report indicated the existence of 10, 9 and 9 NCED/CCD protein sequences, respectively [35]. In all these cucurbitaceae species there are four NCED sequences and five or six CCDs. In litchi (Litchi chinensis), 15 LcCCO genes were identified in its genome, 3 NCEDs and 12 CCDs [36]. In tomato, seven CCD sequences have been described within its genome [38]. In comparison, in Arabidopsis a total of nine members of the CCD family have been reported, divided into four CCDs and five NCEDs [39]. Therefore, the number of CCD members in F. chiloensis is reduced compared to other fruit species; however, as the search has been restricted to a fruit transcriptome there may be other sequences waiting to be discovered.
Several studies suggested that different CCD subfamilies exhibit different roles within the plant kingdom. For example, CCD1 plays important roles in the aroma and flavor of horticultural products as it catalyzes the formation of α-ionone, β-ionone and geranylacetone [40,41]. CCD4 contributes to color formation in flower petals and fruit peel and additionally aroma production [42][43][44][45][46][47][48]. CCD7 and CCD8 seem to participate in the biosynthesis of the hormone strigolactone, which could control shoot branching, reproductive development and plant responses to drought and salt stress [49][50][51][52]. Finally, the NCED subfamily is involved in ABA biosynthesis and closely involved in fruit development and ripening. In the commercial strawberry, FaNCED1 has been demonstrated to be crucial for ABA biosynthesis, as RNAi constructs promoted a dramatic reduction in ABA content that reduced the development of a red color [12], and the phenotype was reverted by the application of exogenous ABA. Similarly in grape [53], sweet cherry [54] and litchi fruit [55], ABA accelerated the accumulation of anthocyanins by increasing the expression of NCED.
In litchi LcCCD4s, LcCCD1, LcNCED1 and LcNCED2 might participate in postharvest storage of the fruit and peel coloration [36]. The expression of LcNCED1 in fruits was consistent with the accumulation of ABA during the ripening of litchi [56]. In peach and grape fruits, PpNCED1 and VvNCED1 transcripts increase at the early stages of ripening, which initiates ABA biosynthesis and ABA accumulation [57]. In Citrus clementina, the expression of CcNCED5 increases at color break and remains high at the ripe stage in parallel to ABA levels, suggesting a role in the ripening of mandarin fruit [58].
In melon, the expression of CmCCD1 was upregulated during the development of the fruit and seems to participate in aroma formation [59]. In addition, CmCCD1 was upregulated by ABA and other stress conditions such as drought and salt [35]. The expression of CmNCED3 was upregulated in melon leaves under a series of abiotic stressors (salt, cold and drought), indicating that it plays an important role in stress [35]. Finally, CmNCED5s are highly expressed in flowers and play crucial roles in flower growth and development [35].
From the FcNCED/CCD sequences identified in F. chiloensis, all of them were detected in fruit tissues, albeit with different profiles. This is not surprising, as the sequences were obtained from a fruit transcriptome. Only FcNCED1 and FcCCD1 were detected in vegetative tissues: FcNCED1 in roots and runners and FcCCD1 mostly in flowers.
The phylogenetic analysis grouped FcNCED1 in the same branch with FaNCED1 and FvNCED1. Our data indicate that the expression level of FcNCED1 increases as the ripening of the fruit is taking place. This also correlates with the increment in ABA observed in the fruit during development. Interestingly, this gene is mainly expressed in fruit tissues, with a low expression level in roots. All this evidence indicates that FcNCED1 is the orthologous of FaNCED1, a gene which has been demonstrated to be involved in ABA biosynthesis during the ripening of F. x ananassa fruit whose repression by RNAi avoided the biosynthesis of ABA [12]. Importantly, the expression of FcNCED1 has been reported to be induced in fruit by ABA treatment [11].
The alignment of deduced amino acid sequences indicated that FcNCEDs and FcCCDs share important domains related to activity. The amino acid residues identified as a part of cleavage domains are mostly conserved. Furthermore, 2 amino acids seem to discriminate between NCED and CCD proteins: L95 and V403 from ZmVP14. The corresponding residues in FcNCED3 and FcCCD1 are in agreement with a hypothetic differentiation for activity; however, in the case of FcNCED1 and FcCCD4 the residue at location 95 is not the expected. FcNCED1 contains a Phe instead of the expected Leu, and FcCCD4 contains a Leu instead of Trp. Nevertheless, the changes in amino acids do not interfere with their character, as Phe, Leu and Trp are nonpolar hydrophobic residues. Interestingly, the same Phe identified in FcNCED1 exists in FvNCED1 and AtNCED3, and in the case of FcCCD4, the Leu in position 95 also exists in FvCCD4, RrCCD4, PaCCD4 and MdCCD4. Therefore, the discrimination between these two activities requires further studies.
The search for ABA receptor sequences within the F. chiloensis fruit transcriptome provided the identification of one PYR and five PYLs sequences. The translated amino acid sequences displayed the connecting loops required for the interaction with ABA and most of the residues involved in ABA and PP2C binding. The phylogenetic analysis grouped them into three subfamilies that correlates well with the number of subfamilies reported in many species. As in Arabidopsis, subfamily II includes FcPYL4 in one clade and FcPYL12 in another. Interestingly, this second clade is missing in monocot species and additionally in some dicots such as tomato and citrus [32]; however, it is present in F. chiloensis, a Rosaceae species. The three PYR/PYL families have arisen during evolution; the latest to emerge is subfamily III, which is present only in angiosperms [40]. Subfamilies I and II are composed of monomeric receptors with high basal activity and require only a low level of ABA to induce PP2C inhibition, whereas subfamily III receptors are dimeric in solution and have low basal activity [60].
All FcPYR/PYLs were expressed in F. chiloensis fruit tissues, and all of them except for FcPYR1 were expressed in vegetative tissues. FcPYL8 was the most expressed gene in vegetative tissues. Different expression profiles were obtained for FcPYR/PYLs in ripening fruit. Of interest is the expression of FcPYL2 and FcPYL4 along the ripening progress, as their expression followed the increment in ABA levels reported in the fruit, suggesting their participation in ABA perception. Nevertheless, FcPYL2 does not contain CL2 which is crucial for ABA interaction.
The expression of the PYR/PYL gene family has been analyzed during ripening in different fruit species. In cucumber, CsPYL2 was expressed at a high level during fruit ripening, with a peak of expression between turning and ripe stages, coincident with the highest level of ABA content in the fruit, suggesting its participation in ABA perception during fruit ripening [61]. In Chinese white pear, from the 11 PbrPYL genes identified in its genome, 7 of them, which were distributed within the 3 subfamilies, were expressed in ripe fruit (PbrPYL1/4/5/6/7/8/9) [61]. In grapes (Vitis vinifera), from the 8 PYL genes identified in the genome, 2 of them have a particular expression profile of interest for fruit ripening, as VvPYL1 and VvPYL8 significantly increased from fruit set until the ripening stage [62]. Those two genes can be classified in subfamily I. In a citrus genome, 11 PYL sequences have been identified [32]. Expression analysis in fruit performed in only six CsPYLs has shown that CsPYL9 belonging to subfamily I was the most expressed during sweet orange development and ripening [63]. In tomato, a climacteric fruit, two of the fourteen SlPYL genes identified displayed an expression peak at breaker stages when ripening is starting and were reported as candidates to regulate fruit ripening [64]. These two genes belong to subfamilies I and II, according to our nomenclature. A recent study confirmed that among the 14 SlPYLs, the expression level of SlPYL1 was the highest, and the expression patterns of SlPYL1, SlPYL4 and SlPYL9 agreed with the ABA accumulation in fruit during development [65]. Interestingly, SlPYL9 which is closest to monomeric AtPYL4 and therefore belongs to subfamily II, seems to have a specific role in fruit development and ripening. The overexpression of SlPYL9 accelerated the ripening of the fruit; meanwhile, RNAi lines showed a delay in ripening [65].
It has been proposed that PYR/PYL genes grouped into the same subfamily may perform similar functions; however, as described, members of the three subfamilies have been identified with a role during ripening in several fruit species (F. chiloensis, tomato, Chinese pear, grapes and citrus), and therefore, this statement is not fully true in the case of fruit ripening.
In conclusion, 2 out of the 10 genes analyzed were finally selected for being involved in ABA biosynthesis and ABA perception in F. chiloensis fruit: FcNCED1 and FcPYL4. FcNCED1 is expressed with an increasing pattern during its ripening, which correlates with the ABA levels reported for the fruit. In addition, the gene also increments its expression in response to ABA. On the other hand, FcPYL4 from subfamily II displays an expression profile which may explain ABA perception. According to the findings, these two genes, FcNCED1 and FcPYL4, might participate during the softening and color development of the F. chiloensis fruit.

Plant Material
White Chilean strawberry fruit (F. chiloensis (L.) Mill. subsp. chiloensis f. chiloensis Staudt) and vegetative tissues (leaves, flowers, runners, stems and roots) were collected from plants growing in a commercial field in Purén, The Araucanía Region, Chile. The plants have been propagated through stolons and correspond to the cultivated species from Purén [3]. Vegetative tissues such as leaves, runners, stems and roots were obtained from a set of five plants; meanwhile, fruits and flowers were collected from plants growing in the same field lot. The fruit was segregated into four different stages: C1, C2, C3 and C4, as recently reported [11]. An amount of 3 pools of fruit tissue were prepared, including 5 fruits from each stage. Fruit and vegetative tissue samples were frozen in liquid nitrogen, converted into a powder with the help of a mortar and pestle and then stored at −80 • C until use.

F. chiloensis Gene Sequences
Genes annotated as NCED/CCD and PYR/PYL were selected from the Genome Database of Rosaceae [66] using Malus domestica, Prunus persica, Rosa chinensis and F. x ananassa as reference species. Further, a comparative mapping of selected sequences and a F. chiloensis transcriptome was carried out as a way to identify their orthologs [37]. Sequences were aligned using MAFFT version 7 (https://mafft.cbrc.jp/alignment/server/, accessed on 7 October 2021). The search provides one ortholog sequence for PYR, five for PYL, two for NCED and two for CCD. The sequences were named based on the best alignment against the NR database.

Phylogenetic Analysis and Motif Analysis
Phylogenetic analyses were performed to analyze the evolutionary relationship of NCED-CCD and PYR-PYL sequences. The analyses were conducted by MEGA software v10.1. Sequence alignment was performed through the CLUSTAL W method and the phylogenetic tree by the neighbor-joining algorithm with 5000 bootstrap replicates.
The amino acid sequences deduced from F. chiloensis genes were analyzed using Pfam prediction, searching for conserved evolutionary domains related to ABA.

RNA Extraction and Expression Analysis by Real Time PCR (qPCR)
The procedure described in [11] was employed to extract RNA and to perform cDNA synthesis from fruit/tissue samples. Specific primers for qPCR analysis are listed in Table 3. Further, qPCR reactions were carried out in triplicates and employing three independent cDNA preparations from each biological sample. Relative expression levels correspond to the mean of three biological replicates ± SE, normalized against the expression level of FcDBP (DNA binding protein) (constitutive gene) [67]. The C1 fruit stage was employed as a calibrator for fruit samples, and the tissue with the lowest expression was employed as a calibrator for the analysis of vegetative tissues. Asterisks indicate significant differences among fruit stages or different tissues. Table 3. Nucleotide sequence of the primers used in the qPCR analyses (Tm of 60 • C).

Gene
Sequence (

Heatmap Analysis
Heatmap analysis was generated using the open-source software R-graphic interface R-Studio (1.3.1093). The diagrams and dendrograms were obtained with the packages "gplots" and "corrplot" which are included in the R library.