Genome-wide identification, and characterization of the CDPK gene family reveal their involvement in abiotic stress response in Fragaria x ananassa

Calcium-dependent protein kinases (CDPKs) are encoded by a large gene family and play important roles against biotic and abiotic stresses and in plant growth and development. To date, little is known about the CDPK genes in strawberry (Fragaria x ananassa). In this study, analysis of Fragaria x ananassa CDPK gene family was performed, including gene structures, phylogeny, interactome and expression profiles. Nine new CDPK genes in Fragaria x ananassa were identified based on RNA-seq data. These identified strawberry FaCDPK genes were classified into four main groups, based on the phylogenetic analysis and structural features. FaCDPK genes were differentially expressed during fruit development and ripening, as well as in response to abiotic stress (salt and drought), and hormone (abscisic acid) treatment. In addition, the interaction network analysis pointed out proteins involved in the ABA-dependent response to plant stress via Ca2+ signaling, especially RBOHs. To our knowledge, this is the first report on CDPK families in Fragaria x ananassa, and it will provide valuable information for development of biofortified fruits and stress tolerant plants.


Results
Identification of CDPK genes in strawberry. Bioinformatics methods were used to identify CDPK family genes in the strawberry (F. ananassa) genome. Genomic analysis was performed using the RNA-seq data from the Sequence Read Archive (SRA) (accession code SRP148865). The presence of typical CDPK domains, including the Ca 2+ binding, kinase, N-terminal, and autoinhibitory domains, were verified using the PROSITE tools to verify the reliability of candidate sequences. Using this procedure, nine possible new CDPK sequences in F. ananassa were identified and designated as FaCDPK3 to FaCDPK11 according to the proposed CDPK gene nomenclature (FaCDPK1 and FaCDPK2 sequences were previously described) 19 . All identified FaCDPK genes encoded proteins with amino acid numbers and molecular weights from 255 to 600 and from 28.7 to 67.1 kDa, respectively. The isoelectric point varied between 4.51 (FaCDPK9) and 9.09 (FaCDPK2). All identified sequences corresponded to the complete protein sequence and featured at least one kinase domain and one EF-hand domain as shown in Table 1. Five sequences included predicted myristoylation motifs at their N-terminus sites, and six contained at least one palmitoylation site, an indication of possible protein-membrane interaction. Bioinformatics prediction of subcellular protein localization also suggested that different members can localize into different subcellular compartments such as the chloroplast, cell membrane, nucleus, and cytoplasm (Table S1).
Gene structure and phylogenetic analysis. Structural divergence among the FaCDPKs, such as the presence/absence and positions of domains, may provide information regarding the evolutionary history of this gene family 27 . Therefore, eleven identified FaCDPK genes were analyzed to confirm the presence of typical CDPK domains. Four conserved motifs were identified (Fig. 1A). Six sequences included one motif (Motif 4) in two different regions, however with different block heights. The height of a block indicates the significance of the match (taller blocks are more statistically significant). In addition, one sequence included only three of the four conserved motifs, suggesting the occurrence of duplication and deletion events from the ancestral sequence that characterize modifications including domain deletion and insertion. www.nature.com/scientificreports/ The organization of the exons/introns and the number of introns present may also indicate the evolutionary history within a gene family 28 . An analysis of the structures of the eleven FaCDPKs revealed that four FaCDPKs (FaCDPK10, 5, 6, and 8) include three introns, two FaCDPKs (FaCDPK1 and 7) include six introns, FaCDPK4 and FaCDPK11 include seven introns. FaCDPK9, FaCDPK3 and FaCDPK2 included one, two, and three introns, respectively (Fig. 1B).
The evolutionary relationships of CDPK genes were also observed through phylogenetic trees. The phylogenetic trees were constructed using the Neighbor Joining grouping method with different substitutions models. The five trees obtained showed similar topology. The tree obtained with the substitution model p-distance is presented in Fig. 2, and the others as Figure S1. The eleven FaCDPK swere assigned to four main subfamilies according to tree topology (Groups A, B, C and D), which matched the AtCDPKs classification (Fig. 2). Eleven AtCDPKs and four FaCDPKs (FaCDPK7, 4, 11 and 10) belong to group A; twelve AtCDPKs and two FaCDPKs (FaCDPK9 and 5) to group B; eight AtCDPKs and two FaCDPKs (FaCDPK6 and 1) to group C; and three AtCD-PKs and three FaCDPKs (FaCDPK8, 3 and 2) to group D. The results of the phylogenetic analysis of the predicted FaCDPK protein sequences revealed that there is no egalitarian representation of F. ananassa and A. thaliana proteins in the four subgroups, since the ancestral CDPK gene was subjected to multiple gene duplication events.
Synteny analysis was carried out to investigate the overall structural variation within the genome, such as fission and fusions of chromosomes. FaCDPKs were found to maintain the synteny with CDPKs from the four Arabidopsis chromosomes (Fig. 3B). Two genes linked to each other by one line in Fig. 3B are defined as syntenic genes. Individual chromosomes from the two genomes were mostly connected by lines of the same color, indicating an evolutionary similarity between these genomes. Most of the CDPK genes were preserved during polyploidization, indicating their evolutionary importance. CDPK genes were also evenly-distributed within www.nature.com/scientificreports/ these genomes, and synteny analysis further indicated that the syntenic CDPK gene pairs were widely-distributed within the genomes. Evolutionary pattern prediction based on the calculation of synonymous (Ks) and non-synonymous (Ka) substitution rates provides information regarding the type of gene pair selection during the divergence process, such as purifying, positive, and neutral selection 29 . Ka/Ks ratios below, equal to, and above 1.00 indicate purifying, neutral, and positive selection types 30 . Here, the Ka, Ks, and Ka/Ks of the orthologous gene pairs were calculated based on the phylogenetic tree analysis results (Table S2). Time between 22 orthologous pairs of FaCDPK genes was also determined. Interestingly, 12 out of 22 orthologous FaCDPK gene pairs yielded a Ka/Ks ratio < 1.00, indicating purifying selection. However, the remaining 10 FaCDPK genes exhibited a Ka/Ks ratio > 1.00, which suggests a positive selection of the orthologous pairs. Furthermore, the divergence time and duplication time ranged from 0.07 to 8.4 (Ks values) and 2.4 to 280 million of years ago (MYA), respectively.
Expression of FaCDPK genes during the development of the F. ananassa. In order to elucidate the regulatory roles of FaCDPKs during the development and ripening of strawberry fruits, RT-qPCR expression analysis was performed. Highest accumulation of FaCDPK3, FaCDPK4, and FaCDPK10 transcripts was found in the final maturation stage (FR), whereas for the other genes the accumulation occurred in the initial stages (SG or BG) (Fig. 4). OnlyFaCDPK3 showed increased expression during fruit maturation, and no FaCDPK1expression was detected in young fruits.
Expression of FaCDPK genes under abiotic stress conditions. CDPKs have been reported as essential factors in regulating plant tolerance to biotic and abiotic stresses. The gene expression profiles of FaCDPKs under salt and drought stress conditions were evaluated (Fig. 5). The expression of nine FaCDPKs was altered in response to stress treatments, and the expression levels of six these genes (FaCDPK1, FaCDPK3, FaCDPK5, FaCDPK6, FaCDPK8, and FaCDPK9) increased upon salt treatment, whereas three genes (FaCDPK4, FaCDPK10 and FaCDPK11) were overexpressed under drought conditions. The highest differences were observed for FaCDPK1 and FaCDPK3, which showed 15-and 30-fold higher expression levels in salt and drought stress than control fruit, respectively. In contrast, lower expressions of FaCDPK5, FaCDPK6, FaCDPK7 and FaCDPK8 were detected under drought stress. This suggests that different genes are involved in responses to salt and drought stresses.
The regulation of tolerance to biotic and abiotic stresses is mediated by CDPKs. In some cases, this is accompanied by the modulation of ABA signaling. In order to better understand this relationship, the expression of FaCDPK transcripts was evaluated in strawberry fruits treated with ABA and an ABA synthesis inhibitor (NDGA). The expression of FaCDPK4 and FaCDPK11 genes increased upon ABA treatment (Fig. 6). The www.nature.com/scientificreports/ www.nature.com/scientificreports/ expression of the FaCDPK1 and FaCDPK7 genes was also improved by ABA and NDGA treatment, however the highest accumulation of transcripts of these genes was observed in fruits treated with NDGA (FaCDPK1 showed 15-fold higher expression than that of control fruits).

Protein-protein interaction network analysis. The above results indicate a possible involvement of
FaCDPK4 and FaCDPK11 genes in the abiotic stress response through interactions between CDPK and other proteins. Network interaction analysis is a powerful method to study the gene function 31 . An interaction network for CDPK proteins was constructed based on the orthologous gene of Fragaria x vesca using STRING 10 database tools. Accordingly, CDPKs were found to be involved in signaling pathways related to plant growth and development and abiotic stress response. This analysis showed that FvCPK2 (homologous to FaCDPK4 from Fragaria x ananassa) is associated with several respiratory oxidase homolog proteins (RBOH), which promote ROS scavenging (Fig. 7). FvCPK2 also interacts with calcineurin B-like protein (CBL)-a positive regulator of salt and drought responses-as well as thiamine pyrophosphokinase (TPK) and guard cell S-type anion channel (SLAC), which are involved in energy production and stomatal responses. Similarly, FvCDPK29 (homologous toFaCDPK11) interacts with RBOH proteins. In addition, FvCDPK29 interacts with serine decarboxylase (SDC), an enzyme related to the formation of membrane phospholipids and probable transcription factor (POSF), a transcription factor gene. FvCDPK29also interacts with LRR receptor-like serine/threonine-protein kinase (GSO), a positive cell division regulator and tubulin-folding cofactor (TFCA), which play roles in modulating tubulin folding. Thus, network analysis results indicate a possible involvement of FvCDPK29 in the maintenance of cell integrity. Given the homology between these CDPKs to FaCDPK4 and FaCDPK11, it is possible that these FaCDPKs play similar roles.

Discussion
The first calcium-dependent protein kinase (CDPK) was identified over 30 years ago in pea 32 . CDPKs have since been identified in several plants and some protozoa 20 . In Arabidopsis thaliana and Oryza sativa (rice), 34 and 31 genes of this family have been identified, respectively 18,20 . However, only two CDPK-coding sequences were identified in strawberry (F. ananassa) thus far. Here, we analyzed genomic and transcriptomic data from F. ananassa, and revealed the presence of nine novel CDPK genes. Due to the scarcity of DNA and RNA sequence data for this species, further CPDK variants other than those described here are likely to be identified as more sequencing data are made available in sequence databases. CDPK structure contains at least one variable N-terminal domain, a protein kinase domain, a junction domain, a calmodulin-like Ca 2+ binding domain and a C-terminal domain. All CDPKs analyzed here included complete sequences, including the N variable domain, the C terminal domain, at least one kinase domain for target phosphorylation, and one EF-hand domain. However, seven of the nine novel sequences exhibited a pattern of four EF-hand motifs, which allow Ca 2+ binding as previously reported for Arabidopsis, rice and wheat CDPKs 4 . This configuration, featuring a tandem of EF-hand domains able to bind up to four calcium ions, possibly originated from a single motif via two cycles of gene duplication and fusion, and enabled cooperativity between the calcium binding sites. From an evolutionary point of view, this structural organization in the form off www.nature.com/scientificreports/ The expression level of PIRUV_DESCARB, DBP and HISTH4 was used as reference gene to normalize each reaction. The data was from four biological and three technical replicates. Data were subjected to one-way ANOVA test and when significant (p ≤ 0.05) were submitted to a comparison of means by Tukey test at 5% of probability. Different letters indicate significant differences between treatment.
Scientific RepoRtS | (2020) 10:11040 | https://doi.org/10.1038/s41598-020-67957-9 www.nature.com/scientificreports/  www.nature.com/scientificreports/ our repetitive structures may be interpreted as a way to obtain different domain movements, and thereby diversify molecular recognition 33 . The EF hand motifs mostly exist in pairs to increase structural stability. The presence of a single EF-hand motif, as observed in FaCDPK3 and FaCDPK8, was reported to drastically reduce the sensitivity of the kinase for calcium-binding, and the subsequent calcium-induced conformational change 6 . For example, a 60-fold reduction in the binding affinity for Ca 2+ ions was demonstrated for a CDPK with a single EF-hand motif in Plasmodium falciparum 34 . Taking into account the wide variety of CDPK genes in plants, investigation of the factors determining the isoforms responding to different signals, the effects on protein phosphorylation and physiological responses is important. The signaling specificity may be determined by the subcellular location of the CDPK proteins, which can be regulated at translational and post-translational levels. In silico predictions indicate that FaCDPKs are localized in several cell compartments, including the chloroplast, nucleus, peroxisome and cytoplasm. Most of FaCDPKs were also predicted to include N-myristoylation and/or palmitoylation motifs. CDPK genes which have a N-myristoylation motif tend to localize in the plasma membranes 35,36 . This myristoylation may function as part of a primary signaling process that directs these proteins to a cell membrane binding site. In addition, palmitoylation, another type of lipid modification, was also reported to be necessary for stability of membrane association 35 . N-myristoylation and palmitoylation motifs were observed in five and six of the FaCDPKs evaluated in our study, respectively. However, the association of CDPKs with the membrane is complex, and may be affected by the presence of other motifs. For example, in wheat (Triticum aestivum L), TaCPK3 and TaCPK15 without myristoylation regions were associated with membranes 37 , whereas in maize (Zea mays) ZmCPK1 with an N-myristoylation pattern, was found in the cytoplasm and nucleus 38 . Other studies have shown that the subcellular locations of CDPKs can be restricted to a single compartment or widely-distributed throughout the cell. CDPKs have already been observed in the plasma membrane, cytoplasm, nucleus, endoplasmic reticulum, mitochondria, chloroplasts, oily bodies, peroxisomes and the Golgi complex 39 . Therefore, additional in vitro studies are necessary to determine the subcellular localization of FaCDPKs. Genes originating from a common ancestor share similar protein function. Thus, a syntenic analysis was performed, and a phylogenetic tree was constructed in order to understand the evolutionary relations between the CDPKs present in F. ananassa and A. thaliana. All identified FaCDPK homologs in Arabidopsis showed high sequence similarity, indicating a high level of conservation and a common ancestor for each potential orthologous pair. Four groups (A, B, C, and D), each with at least one counterpart from Arabidopsis, were identified in the phylogenetic tree. This finding implies a functional conservation of CDPKs between the two plants. Similar evolutionary classification was also found in other species, such as grape, pineapple, cassava, and rice [40][41][42][43] . Here, FaCDPK4, FaCDPK7, and FaCDPK11 were present with AtCDPK23 in subgroup A. AtCDPK23 plays important roles in drought and saline stress responses 44 , suggesting possible involvement of FaCDPK4 and FaCDPK11 in response to these stress conditions in strawberry. GrCDPK16 has been reported to be in the same phylogenetic group as AtCDPK23 and associated with plant tolerance to drought stress in cotton 45 . FaCDPK6 shares 92% sequence identity with AtCDPK30. Hence, FaCDPK6 may be involved in hormonal signaling pathways in strawberry. CDPK genes are also involved in resistance to different pathogens. In potato, StCDPK7 was reported to promote resistance against Phytophthora infestans infection 46 . StCDPK7 was also found in the same phylogenetic www.nature.com/scientificreports/ group as AtCDPK1. Moreover, heterologous overexpression of AtCPK1 in Arabidopsis thaliana was reported to upregulate disease resistance genes responsible for broad-spectrum protection against pathogen infection 47 . Here, the finding of FaCDPK5 in the same phylogenetic group as AtCDPK1 suggests that FaCDPK5 may be involved in response to pathogen infection in strawberry. Further investigation is necessary to determine whether homologs possess similar functions. Structural divergence, i.e. the presence and position of domains, as well as the organization of exons/introns can indicate evolutionary history within a gene family, and are also closely related to protein function 27,28 . As in barley and poplar 48,49 , the number of introns varied between one and seven in strawberry as well, indicating similarities in CPDK gene structure between different species.. Genes with similar intron phases share a common ancestor. A change in the intron phase reflects a divergence of homology between genes during evolution. The insertion or deletion of a small DNA fragment can alter the transcript, and thus lead to a change in gene function 50 . In Arabidopsis, the diversity of RNA structures was shown to arise due to a high number of deletions and additions of introns from the ancestor 28 . According to Mitall and co-workers 51 , variation in intron frequency in maize, rice, and sorghum may be one of the reasons for the functional diversity of CDPKs, as this variation increases the possibility of alternative splicing (AS) and exon shuffling. However, literature information on AS or RNA editing of the CDPK transcripts is very limited. Nishiyama and co-workers 52 reported an AS of a CDPK gene, where EF-hand motif sequences were modified. Almadanin and co-workers 53 showed that OsCPK13, OsCPK17, OsCPK18, and OsCPK19 produced AS transcripts in rice, which encode truncated proteins lacking whole or partial domains with different biological functions. Finally, Ding and co-workers 54 detected two transcripts of CsCDPK1 from C. sinensis using RT-qPCR. The identification of AS variants in CDPK transcripts is particularly important for the development of genetically modified crops with improved biotic and abiotic stress tolerance. Therefore, future studies to identify splicing variants from CDPK genes, and the impact on their functional diversity are of relevance in this regard.
The specificity and gene expression patterns are clearly altered during plant development stages 55 . Here, FaCDPK1 was not detected in young fruits. This result is in agreement with those reported by Llop-Tous 24 , who observed that FaCDPK1 was only expressed at the final stage of fruit development (28 DAA). However, transcript levels increased as fruits whitened and during the maturation stage. Here, such increased expression during fruit ripening was observed only for FaCDPK3. Li L. et al. 45 also observed similar results in cotton (Gossypium raimondii), in which only GrCDPK11 expression increased during fruit development. Expression levels of the other GrCDPKs remained unchanged or decreased. Strawberry ripening process involves a series of molecular and physiological events leading to dramatic modifications in fruit size, color, texture, flavor, and aroma. Therefore, understanding the mechanisms underlying this process is essential to achieve optimal fruit quality as well as long shelf life. Transcriptomic, metabolomic, and gene silencing studies have shown that this process in non-climacteric fruits such as strawberry is ABA-dependent 56 , and triggered by a signaling cascade involving calcium (Ca 2+ ) and calcium-dependent protein kinases (CDPKs) 15 .Increased expression of FaCDPK3 here indicates a possible role during strawberry maturation. However, further studies are necessary to address ABA-dependence of this process.
In addition to the stage of development, different stimuli also determine which isoforms are activated or inactivated and their function. For example, OsCDPK4 confers salt and drought tolerance to rice 11 , whereas ZmCDPK1 plays a vital regulatory role in abiotic stress response of maize 38 . Drought and soil salinity negatively impact plant growth and crop production. Therefore, a better understanding of molecular response against these stresses is required to develop more resistant cultivars. Strawberry is generally considered moderatelysensitive to high salt levels. However, the cv. Camarosa used in this study is considered tolerant to drought and salt stress compared to other cultivars 57,58 . For example, the stress levels used in this study did not result in necrosis, extensive dehydration or death of plants ( Figure S2). On the other hand, plants showed response to the inflicted stress as evidenced by RT-qPCR analysis results showing increases in expression levels of FaCDPK1, FaCDPK3, FaCDPK4, FaCDPK5, FaCDPK6, FaCDPK8, and FaCDPK9 upon salt treatment, whereas those of FaCDPK4, FaCDPK10 and FaCDPK11 increased expression levels under drought conditions. This result suggests that FaCDPK proteins play a role in the tolerance of this cultivar to salinity and drought stress, and corroborates the relationships between sequence, structure and protein function. FaCDPK4, FaCDPK10 and FaCDPK11 were grouped in the same phylogenetic subgroup (A) as AtCDPK23, which was reported to be involved in drought tolerance of Arabidopsis 44 . Although osmotic stress is caused by both salt and drought stresses, excessive accumulation of Na + and Cl − ions may also occur under high salinity due to reduced water availability in the soil. Na + and Cl − accumulation is also detrimental to biochemical processes, and leads to chlorosis and necrosis. FaCDPK4 was the only CDPK gene overexpressed under both experimental conditions. FaCDPK4 may thus be associated with physiological responses triggered by both stresses, such as regulation of stomatal movement and repression of cell growth and photosynthesis. As an example, AtCPK1 mediates salt and drought stress through decreased production of H 2 O 2 and malondialdehyde, and increased accumulation of proline 59 . Since most of the CDPK genes were differentially-expressed between these stress conditions, and due to the varying substrate specificities of CDPK isoforms 46 , this result suggests that strawberry plants are able to respond differentially to these stresses, and the kinase cascade formed by CDPK proteins may have a role in this differential response.
Although plants use closely-related response mechanisms against salinity and drought, the effects these stresses on plants are not identical. Changes in the osmotic potential observed in both stresses seems to trigger the induction of different CDPKs. For example, tolerance to drought and salt stresses in Morus atropurpurea were reported to be related to the interaction of MaCDPK1 with ABA response elements (MBS elements, ABRE and GARE-motif) 60 . However, the CDPK copies and the induction of phosphorylation cascade may differ between these two stresses. In Poncirus trifoliata, overexpression of PtrCDPK10 resulted in enhanced drought stress tolerance compared to the wild type due to the interaction of PtrCDPK10 with ascorbate peroxidase (PtrAPX), which led to a reduction in ROS accumulation 61  www.nature.com/scientificreports/ adaptation to salt stress in rice through phosphorylation of OsGF14. In grapevine, VaCPK21 may act as a positive regulator involved in the response to salt stress 62 . MaCDPK3 63 and AtCDPK27 perform similar functions in banana and Arabidopsis, respectively 64 . Plant responses to stress conditions are known to occur through various signal transduction networks, and abscisic acid (ABA) plays a crucial role in these responses. ABA mainly functions as a regulator of plant water balance and osmotic stress tolerance 65 . CDPK isoforms were also shown to differ in their Ca 2+ affinity, and calcium influx generated by ABA is one of the determinants of the induction of CDPKs 15 . As an example, Zhang H. et al. 22 performed an analysis of the transcripts of 21 BnaCDPK genes in canola seedlings following different treatments under abiotic stress conditions, and found that different CDPK genes may participate in the signaling processes of a single stress factor, and a single CDPK gene likely plays a role in multiple stress responses. In rice, OsCDPK4 was reported to play a positive role in tolerance to saline and water stresses through membrane protection against lipid peroxidation 11 ; where as OsCPK12 was found to increase salt tolerance by reducing ROS levels, suggesting that different copies work on multiple signaling pathways to positively regulate salt tolerance 66 .
Here, FaCDPK4 and FaCDPK11presented the highest levels of expression against water stress, and these genes were also the copies with best response to ABA treatment. This suggests that these genes play important roles in water stress response, and the response mechanism is ABA-dependent. This relationship was also observed in Arabidopsis, in which mutants lacking AtCDPK10 expression were impaired in their ability to inhibit ABAinduced stomatal opening. In a study by Zhu et al. 67 , AtCPK4 and AtCPK11 were shown to positively regulate ABA signaling through the phosphorylation of stress-responsive transcription factors ABF1 and ABF4. ABA was also shown to activate calcium-dependent protein kinase (ACPK1) in the grape mesocarp by means of a complex mechanism involving the influx of calcium to the cytosol 68 . Therefore, the variations in ABA-induced Ca 2+ signal discharge (concentration, size, temporal distribution of spikes or waves) may result in specific signal transduction to induce the expression of specific CDPK genes. Consequently, the induced CDPK protein would play specific roles depending on the inflicted stress type through phosphorylation of specific substrates 46 .
We previously observed that salt and drought stress increase the content of phenolic and anthocyanin compounds as well as the expression of genes in related metabolic pathways 25,69 . ABA content was also observed to increase in strawberry fruits under saline and drought stresses; however, ABA biosynthesis related genes (FaNCEDs, FaCYP707As, FaGTs, and FaBGs) were mostly upregulated only in drought stressed fruits 25 . Similarly, only the drought stress increased the concentrations of ascorbic acid, sugars, and methylsyringin, whereas salt stress induced the synthesis of amino acids in the fruits 26,70 . A mechanism of salt and drought response in strawberry fruits based on these previous results and the gene expression data from this study is shown in Fig. 8. Saline stress response was found to be mediated by several FaCDPKs, in particular FaCDPK1 and FaCDPK3 through an ABA-independent mechanism. In contrast, FaCDPK4 and FaCDPK11 were found to be involved in drought stress tolerance, and utilize ABA as the signaling molecule to increase the content of phenolics, anthocyanins, ascorbic acid, and sugar compounds. The crosstalk between CDPKs, ABA, phenylpropanoids and sugar was recently reviewed by Vighi et al. 15 , and the influence of ABA in the ascorbic acid metabolism of strawberry fruits was discussed by Li et al. 71 .
The results above suggest an involvement of FaCDPKs in the response to salt and drought stresses. Several other studies have reported that CDPKs play important roles in tolerance to abiotic stress 72,73 . However, the underlying biochemical mechanisms are still unknown. Determining protein-protein interactions is an essential step towards understanding the mechanisms involved in the development of tolerance to abiotic stresses. Stress response involves the production of reactive oxygen species (ROS), and may induce the flow of calcium into the cell. Both signaling molecules induce downstream signaling events that result in the activation of Ca 2+ sensor proteins including CDPK as well as the ROS-producing enzyme andr espiratory oxidase homolog proteins (RBOH) 15 .
FaCDPK4 and FaCDPK11 genes were likely involved in response to abiotic stress, especially drought, and the interaction network based on the co-expression analysis revealed that FaCDPK4 and FaCDPK11were possibly associated with RBOHs. RBOHs are integral membrane proteins that generate superoxide anions, which are then converted to H 2 O 2. Moreover, RBOH proteins contain two calcium-binding EF hand motifs, and multiple phosphorylation sites at their N-termini that participate in regulation of enzyme activity. In addition to their roles in responses to abiotic stresses, RBOH are play fundamental roles in several metabolic processes such as cellular growth and hypersensitivity responses 74 . RBOHs are differentially expressed according to the tissue and developmental stage 75 . Moreover, different stimuli determine which isoforms will be activated/inactivated 7 . For example, in Arabidopsis, AtRBOHD was found to be responsible for the generation of ROS in response to pathogen infection 77 . In strawberry, FvRBOHA and FvRBOHD are required for accumulation of ROS during plant defense response against cold stress 78 . Similarly, the expression levels of VvRBOHA, VvRBOHB, and VvR-BOHC1 were significantly increased in grapes upon salt and drought stress treatments. In addition, VvRBOHB was strongly upregulated by exogenous ABA treatments 79 . Here, FaCDPK4 and FaCDPK11were found to be responsive to drought and ABA treatments as well, as shown in Fig. 8.This suggests that stress response involves the cellular uptake/synthesis of ABA and Ca 2+ , and both of these signaling molecules may induce downstream signaling events that result in increased production of H 2 O 2 by RBOH, and the consequent induction of ascorbic acid to mitigate the production of ROS 80 .
Besides the RBOH proteins, a calcineurin B-like protein (CBL) was highlighted in the interaction network of FaCDPK4. CBL are calcium sensor proteins that integrate and interact specifically with a family of protein kinases (CIPKs) 81 . CBL-CIPKs are involved in immune and abiotic stress responses in addition to plant growth and development 33,82,83 . CBL-CIPKs mediated stress response was also shown to be ABA-dependent in several studies 84,85 .
FaCDPK4 also showed a similar expression pattern to that of thiamine pyrophosphokinase (TPK), a protein responsible for thiamine pyrophosphate activation, and an essential cofactor of numerous enzymes participating in the metabolism of carbohydrates and amino acids 86 89 . Another protein identified in the FaCDPK4 interaction network is the guard cell S-type anion channel (SLAC). SLAC plays important roles in stress and hormone signaling as well as plant growth and development 90 . SLCA1 is also an essential protein for stomatal aperture, and activated by ABA and cytosolic Ca 2+ . SLAC1 channels are activated by CDPKs as well 91 . As an example, A. thaliana guard cells with silencedCDPK3 and CDPK6genes showed reduced activation of the ion channels of Ca 2+ and ABA, and stomatal closing was partially impaired 92 .
Similar to that of FaCDPK4, the most prominent proteins in the protein interaction network of FaCDPK-11were RBOHs. However, several other proteins were also highlighted, such as the serine decarboxylase (SDC). SDC is responsible for conversion of serine to ethanolamine, a phospholipid precursor in plants 93 . Phospholipids are essential components of biological membranes and signal transduction cascades in plants 94 . Phospholipids also influence the sensitivity of CDPK to calcium. Interaction of kinase with phospholipids was shown to lower its requirement for calcium in maize 95 . Moreover, FaCDPK11interacted with the probable transcription factor (POSF)/basic leucine zipper (bZIP). bZIPs are involved in various cellular processes related to plant development, environmental signaling and stress response 96 . bZIP is part of a complex signaling network involving ABA and CDPK in abiotic stress response. In the presence of ABA, the PYR class of receptors sequesters PP2C, which in turn releases SnRK 2 . As a result, SnRK 2 activates CDPK, and results in a phosphorylation cascade involving bZIP 97 . In Arabidopsis, three CDPKs (CDPK4, CDPK6, and CDPK33) efficiently phosphorylate bZIP 98 .
FaCDPK11 may also be associated with the LRR receptor-like serine/threonine-protein kinase (GSO). Receptor-like kinases are transmembrane proteins involved in multiple physiological processes, including development, stress responses, hormone perception, and disease resistance 99 . Under abiotic stress conditions, GSO perceives the extracellular ligands, and activates downstream pathways via phosphorylation of intracellular serine/threonine kinase domains 100 . GSO also plays important roles in the ripening process together with ABA. FaRIPK1 silencing inhibited ripening and reduced the expression of several genes involved in softening, sugar production, pigmentation, and ABA biosynthesis and signaling in strawberry plants 101 .
In summary, all proteins highlighted in the FaCDPK4 and FaCDPK11 interaction networks have well-documented roles in responses of plants to stress via Ca 2+ signaling. These responses directly or indirectly involve the phytohormone ABA. The only protein in the FaCDPK11 interaction network with no evident crosstalk with ABA and stress response is the tubulin-folding cofactor (TFCA). TFCA is involved in microtubule biogenesis, and thereby directly influences cell growth 102 . In Arabidopsis, TFCA requirement was demonstrated for the formation of trichome cells 103 . However, additional in vivo studies are necessary to validate the existence of this interaction predicted by in silico interaction network analysis.
In this study, nine new copies of CDPKs were identified in F. ananassa genome via genome structure analysis and phylogenetic characterization. The expression profiles of the identified FaCDPKs during fruit development and ripening indicated that only FaCDPK3 expression increased during fruit ripening. FaCDPK genes were also found to be involved in the response to osmotic stresses: FaCDPK1 and FaCDPK3 in response against salt, and FaCDPK4 and FaCDPK11 against drought. The expression levels of the latter two genes were also significantly influenced by ABA treatment, suggesting the involvement of this phytohormone in the drought response mediated by these proteins. Moreover, the protein-protein interaction network analysis highlighted proteins involved in the ABA-dependent response to plant stress via Ca 2+ signaling (RBOHs in particular). Future functional  Synteny analysis and chromosome localization. FaCDPK genes were mapped to strawberry chromosomes based on information of Fragaria vesca available at the National Center for Biotechnology Information (NCBI), obtained by using BLASTP with the reference E < 1e -5 . The map was drafted using Mapchart software (https ://www.wagen ingen ur.nl/en.htm).The syntenic blocks were used for constructing a synteny analysis map within the strawberry (Fragaria vesca) genome and Arabidopsis. Diagrams were generated using the Circos program (version 0.69) (https ://circo s.ca/). Based on the phylogenetic tree analysis results, the synonymous (Ks) and non-synonymous (Ka) nucleotide substitutions between orthologous and paralogous gene pairs were calculated using ClustalW and PAL2NAL.
Plant materials and stress treatments. The study was conducted on a greenhouse, using the cultivar Camarosa, as it is extensive used in south Brazil, and because of its good agronomic performance. The seedlings were planted and grown in 9 L pots, containing soil (Ultisoil) and vermiculite (in a proportion of 3:1). The fertilization and irrigation were performed according to Galli, V. et al. 107 . The relative humidity in each pot was monitored using a hygrometer order to maintain between 16 and 19%, without water leaching.
The experiment was composed by five treatments with six replicates per treatment and ten plants per replicate. The treatments were as follows: control/normal irrigation (C), corresponding to 100% crop evapotranspiration (ETc); drought stress (DS), corresponding to 85% crop ETc; the salt stress (SS), with 80 mM NaCl; 200 µM of abscisic acid (ABA); and 50 µM of nordihydroguaiaretic acid (NDGA). ABA and NDGA treatments were imposed by spraying the foliage of seedlings. An approximate pool of ten ripe fruits (fully red, according to Jia, H. et al. 56 ) by replicate were harvested.
All samples were immediately frozen in liquid nitrogen and stored at -80 °C until further analysis.
Gene relative expression analysis. Total RNA isolation, cDNA synthesis and RT-qPCR amplification were performed as described by Galli, V. et al. 107 . The reference genes PIRUV_DESCARB (pyruvate decarboxylase), DBP (DNA binding protein), HISTH4 (histone H4) were used to normalize transcription levels, as proposed by the authors. The primers used are presented in Table 2. The relative expression data were calculated according to the 2 -ΔΔCq method. The analysis was performed for four biological replicates and three analytical replicates. The Heat map representation was used to estimate the fold-change of gene expression using the software Mev 4.8.1 (https ://sourc eforg e.net/proje cts/mev-tm4/files /mev-tm4/). The results were submitted to analysis of variance (ANOVA) and when significant (p ≤ 0.05) were submitted to a comparison of means by Tukey test at 5% of probability. Statistical analyzes were performed using SAS software.  FaCDPK1  GGT GAA ACA GTG AGA GCA AGGC  AAA TCC CTC TTG TAT GCC AGC   FaCDPK2  TGG ATT TCA CCG AGT TTG TTG  AGG CTA ATC TTC CCG TCT TTGT   FaCDPK3  CAC CAA ACC CAA ACT CAA CATT  TAT GGC GGT CAG AGT AGT AGG AGT   FaCDPK4  TCG GGT TGT CAG TGT TCA TCG  CTC AAA GTC AAT CTC TCC CTC CAA   FaCDPK5  GCA GGC AGC AGA TGT AGA TAA TAG C  AAC ATC CTC TAC ACC AAA CTC CTC A   FaCDPK6  TGA CAA GGA TGG AAG TGG CTAT  CGC CCA TCC TTG TCA GTG T   FaCDPK7  AAG CAA CTG AGA GCA ATG AAT AAG C TTG CCT GAC TTC CGT TTC TGAT   FaCDPK8  CTG CTA AGG ATG AGA GTT CGCC  ACA CTC CAT ACA TCA GCC TCCG   FaCDPK9  CGA CAA AGA TGG AAG CGG G  CCA TTT CCC TTC CTC ATC ATTG   FaCDPK10  TGA TGT GGA TGG AAA TGG AACC  TAA TCG TTT GTT CGT CAC CCAT   FaCDPK11  TCC AAC ACA ACA ACC TCA GAA AGT  CTT CCG AGT TCC TTG CCC A