Genome-Wide Characterization of CDPK Gene Family in Apple (Malus Domestica) and Its Transcriptional Expression During Apple Fruit Development

Background: Calcium-dependent protein kinases (CDPKs) play important roles both in developments and response to stresses, via mediating Ca 2+ signal transduction in plants. To characterize the CDPKs in apple (Malus domestica), the apple CDPK gene family, together with those from pear (Pyrus bretschneideri), peach (Prunus persica), strawberry (Fragaria vesca), and and Arabidopsis thaliana, were analyzed at the genome-wide level in the present study. Results: A total of 116 CDPKs, consisting of 24 MdCDPKs, 28 PbCDPKs, 16 PpCDPKs, 14 FvCDPKs, and 34 AtCDPKs, was identied from apple, pear, peach, strawberry, and Arabidopsis, respectively. An integrated analysis of these CDPKs was performed on their chromosomal distribution, phylogenetic and collinearity relationships, characteristics of gene structures and conserved motifs. As a result, the CDPK gene family members were showed to be highly conserved both at their kinase and EF-hand domains. Among 209 gene-pairs with interspecies collinearity, there existed 22, 36, 21, and 25 ones between MdCDPKs and other CDPKs in Arabidopsis, pear, strawberry, and peach, respectively. And the evaluated Ka/Ks ratios were less than 1 between the CDPK gene pairs with collinearity relationships. Transcriptomic analysis demonstrated that among 24 members of the apple CDPK gene family, two up-regulatory ones (HF05266 and HF09216) and two down-regulatory ones (HF05471 and HF15429), were differentially expressed with signicance between the apple fruit developmental stage S4 (mature) and other stages (early growing-S1, mid growing-S2, and late growing-S3), respectively. Conclusions: whole genome important

activities, independence of the activation process mediated by calmodulins. In the case of lacking Ca 2+ signature, CDPK kinase activities are auto-inhibited by a special sequence, namely junction domain, between the CDPK kinase and EF-hand domains [6-8]. The highly variable N-terminus of some CDPK, containing speci c sites for myristoylation or palmitoylation, have been reported as membrane anchors [7,9]. Ca 2+ signals play versatile roles in regulating a various growth and developmental processes [10,11]. It is therefore likely that CDPKs, as a class of Ca 2+ sensors, have functions involved in the speci c processes of plant development [8]. Some such cases associated with CDPKs, have been reported on embryogenesis, seed development and germination [12,13], early stages of potato tuberization [14], pollen tube growth [15], and shoot growth [16].
Furthermore, CDPKs have been demonstrated to mediate adaptive regulations in response to a variety of abiotic and biotic stresses, such as cold, high salinity, drought, wounding, and pathogen infection [8,17]. Transcriptional upregulation of CDPKs has been identi ed in a variety of species encountered by abiotic stresses [7,[17][18][19]. Arabidopsis CDPK10, and tobacco NtCDPK2 are involved in modulating osmotic potential [20]. OsCDPK7, a rice CDPK, has an important role in the tolerance to both cold and salt stress. The transgenic plants with overexpressed levels of OsCDPK7, showed an enhanced tolerance to cold, drought and salt stresses [21]. A set of the cotton CDPKs (GhCPK8, GhCPK38, GhCPK54, and GhCPK55) could participate in the early signaling events in cotton responses to salt stress [19].
With the completion of genome sequencing, characterization of CDPK gene family at the genome-wide level, has been carried out in many a plants species, such as Arabidopsis [7], maize [22], barley [23], and upland cotton [19]. The identi ed CDPK families in these species are composed of varied numbers of members, inferring their diverse functions during evolution.
Apple (Malusdomestica Borkh.) is one of the major fruit crops produced in the world. Its genome has been sequenced in 2017 [24]. Thereafter, a new version data of apple genome (HANFU) was released in 2019 (https://github.com/moold/Genome-data-of-Hanfu-apple). These data provide a better likelihood to excavate the CDPK gene family and should facilitate the elucidation of CDPK properties and functions in apples. In the present research, a comprehensive analysis of evolution and function of apple CDPKs was carried out at the whole-genome level. In addition to apple CDPKs, other CDPKs in three species of the Rosaceae family, including pear (Pyrus bretschneideri), strawberry (Fragaria vesca), and peach (Prunus persica), together with the model plant Arabidopsis (A. thaliana), were retrieved from the individual plant species with available genome data. The phylogenetic, gene structures and protein motifs of the identi ed CDPK family members, accompanied by the collinearity analysis on these genes, provided some clues to the evolutionary relationships among these CDPKs. Furthermore, to examine the apple CDPKs' involvement in the development of apple fruits, the transcriptomic data from RNA-seq analysis on two apple strains at the different stages of fruit development [25], were re-quanti ed to address this question. The results may lead to a primary understanding of the apple CDPKs with both redundant and distinct functions, and further investigating the function of calcium signaling mediated by the speci c CDPK in regulating apple fruit development.

Phylogenetic tree construction
For multiple sequence alignment based on amino acid sequences, all of the identi ed CDPK proteins with the full-length sequences or only the conserved domains (i.e. Pkinase and EF-hand_7), were aligned via the online program Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/). The alignment result from the CDPK full-length sequences was further applied to the construction of phylogenetic tree with the Maximum Likelihood method in MEGA X.
Analysis of CDPK molecular characteristics, gene structures, conserved protein motifs The primary molecular characteristics, such as amino acid length, isoelectric point (IP) and molecular weight, were analyzed by the online program ExPASy (http://www.expasy.ch/tools). Chromosomal localization of CDPK genes were visualized on the program package MapInspect (http://www.plantbreeding.wur.nl/). Analysis of CDPK gene structures was performed using the GSDS server (http://gsds.cbi.pku.edu.cn/). Analysis of the conserved motifs among the CDPKs was carried out using the program MEME (http://meme-suite.org/tools/meme) with default parameters.

Collinearity analysis of CDPK genes
For collinearity analysis, the local databases of proteins were built and blasted using the program package MCScanX [34]. The criterion for collinearity relationships was referred to the previous method [35]. For micro-synteny analysis between the CDPK gene pairs, anking positions (i.e. a frame with 0.1-million base pairs of genomic sequences) located at both up-and downstream of the individual target genes, were used in a Blast comparison. Ka (nonsynonymous nucleotide substitutions) to Ks (synonymous nucleotide substitutions) ratios were analyzed via the local package DNASP5.
Expression analysis of MdCDPKs based on the transcriptomic data of apple fruits The RNA-Seq data of both yellow apple 'Blondee' (BLO) and red apple 'Kidd's D-8' (KID) fruits, were downloaded from NCBI with the accession number SRP062637 (http://www.ncbi.nlm.nih.gov/sra). This data set consists of 24 FASTQ les, sequenced for two apple strains (BLO and KID) at the four stages of fruit developments (i.e. early growing-S1, mid growing-S2, late growing-S3, and mature-S4). And each sequencing samples were assigned with three replicates [25]. The analysis of RNA-seq data was carried out following the previous publication [25], except that the HANFU apple genome released in 2019 (https://github.com/moold/Genome-data-of-Hanfu-apple), instead of the apple genome in 2017 [24], was used as the reference genome for reads alignment. Differentially expressed genes (DEGs) were ltered by the log 2 FoldChange value more or less than 1 with an adjusted p-value ≤ 0.05.

Phylogenetic and gene structural analysis of the CDPKs
To investigate the phylogenetic relationships and molecular evolutionary history of the sequences in the examined species, following the alignment of 116 CDPK proteins, a phylogenetic analysis was conducted and a phylogenetic tree was generated using the Maximum Likelihood (ML) method. The phylogenetic tree showed that the 116 CDPKs were clustered into ve main subgroups, among which the highest numbers of members were 33 in subgroups I and IV, followed by 31 and 16 in subgroups III and II, respectively (Fig. 2).
And subgroup V has the least members (3), all of which are from AtCDPKs. As shown in Fig. 2, the CDPKs from individual species were grouped into different clades rather than a single one. Additionally, their numbers varied within different subgroups. Out of 24 MdCDPKs, nine were located in subgroup I, seven in subgroups IV, and four in subgroups II and III, respectively (Fig. 2). The other Rosaceae species also exhibited similar patterns in their CDPK distributions, among which 9, 4, and 4 members from pear, strawberry, and peach, were included into subgroup I, respectively. Accordingly, 5, 2, and 2 in subgroup II; 5, 3, and 6 in subgroup III; 9, 5, and 4 in subgroup (Fig. 2). 34 AtCDPKs were dispersed across subgroups I~V, with the most (13) in subgroup III and the least (3) in subgroups II and V (Fig. 2). Additionally, it appears that AtCDPKs were clustered with each other in the ve subgroups, compared to the CDPK clustering across Rosaceae species. Moreover, three AtCDPKs were grouped into distinct the subgroup V separated from CDPK homologs from the other species examined (Fig. 2).
Based on sequence alignment, it was found out that all of characteristic domains of CDPK family (i.e. a domain of protein kinase for CDPK activities and four EF-hands for calcium-binding) were presented among 113 out of 116 CDPKs (Fig. 2). Among the remaining 3 CDPKs, AT2G35890 and HF28950 have only two EFhands (i.e. the 1st and the 2nd ones), whereas Pb001308, without the 1st one, has three EF-hands at the Cterminus, respectively ( Fig. 2 and Fig. 3). And there showed a high conservation among the EF-hand domains of the CDPKs identi ed (Fig. 3).
To characterize their gene structural diversity, the exon-intron organizations of the CDPKs were analyzed (Fig. 4A). The number of exons was diverse, with a minimum of one (i.e. HF00526, HF28950, Pbr027545, Pbr033411, Pbr033416, and Prupe.3G035400) and a maximum of 12 (AT2G17890, AT4G04710, AT4G36070, and AT5G66210). Generally, the CDPK gene structures within each subgroup of the phylogenetic tree, showed a similar pattern, supporting their phylogenetic relationships (Fig. 4A). An exception is that the six CDPKs with a single exon are clustered into a distinct clade within subgroup II, which consists of other CDPKs with seven exons (8 members), six exons (1member), or two exons (1 member). However, CDPKs from a speci c clade within a subgroup, apparently have the same numbers of exons (Fig. 4A). In addition, motif analysis by MEME demonstrated that most representatives of the motifs in CDPKs from the same subgroup, showed a conservation in both motif distribution and composition, coordinating with their distribution across various subgroups in the phylogenetic tree (Fig. 4B).

Collinearity analysis of CDPKs
To investigate the gene duplication that promotes the evolution of CDPK gene family among the species examined, multiple-round analysis of collinearity relationship was carried out between each pair of species. A total of 245 CDPK gene-pairs with collinearity relationships were identi ed, consisting of 36 intraspeciespairs and 209 interspecies-pairs across each pair of species (Table 2, Fig. 5, Additional le 1: Table S1 and Additional le 2: Figure S1). Among 36 gene-pairs with intraspecies collinearity, 10, 13, 1, and 12 ones were blasted out from apple, pear, peach, and Arabidopsis, respectively (Fig. 5A, 5B and 5C, Table 2 and Additional le 1: Table S1). And no intraspecies-pair of CDPKs was found in strawberry (Fig. 5D). Among 209 gene-pairs with interspecies collinearity, there existed 22, 36, 21, and 25 ones between apple and four other species (i.e. Arabidopsis, pear, strawberry, and peach), respectively (Fig. 5, Table 2 and Additional le 1: Table S1). Both pictorial micro-synteny of 10 MdCDPK gene-pairs and 22 CDPK gene-pairs between apple and Arabidopsis (Fig. 6), were demonstrated to support their synteny relationships.
Apart from apple and Arabidopsis, CDPKs collinearity were also identi ed between the other species (Additional le 1: Table S1 and Additional le 2: Figure S1). Noticeably, the individual CDPK gene-pairs with collinearity relationships were not only distributed within the same phylogenetic subgroups, but also identical in their exon-intron patterns ( Fig. 2 and Fig. 4), such as HF05471-Pbr001322, HF20170-Pbr010307, and AT3G10660-AT5G04870 in subgroup I; Pbr033416-Pbr033411, AT1G35670-AT4G09570, and HF39191-Pbr040137 in subgroup II; Pbr024654-Prupe.5G110500, HF17744-Pbr031892, and HF04323-Pbr039714 in subgroup III; HF15429-Pbr021635, HF01706-Pbr036114, and gene25220-Prupe.7G064300 in subgroup IV; AT2G17890-AT4G36070, AT2G17890-AT5G66210, and AT4G36070-AT5G66210 in subgroup V. This result supported the evolutionary relationships between the identi ed CDPKs. To assess the evolutionary rates among these CDPK gene-pairs, the Ka (nonsynonymous nucleotide substitutions) to Ks (synonymous nucleotide substitutions) ratios were calculated (Table 2 and Additional le 1: Table S1). The Ka/Ks values ranged from 0.022 to 0.751 for the gene-pairs between apple and other species, while from 0.085 to 0.399 for those within apple. It is noticeable that there is no gene-pair with Ka/Ks values >= 1, inferring that the duplicated CDPKs within the species examined have been undergone purifying selection.

Quantitative analysis of MdCDPKs expression during apple fruit development
To examined the expression patterns of apple MdCDPKs, the expression data set, based on the transcriptomic analysis of two apple strains at their four stages of fruit development [25], was applied to addressing this question. After the quantitative analysis of transcriptomic data, the whole expression levels of 24 MdCDPKs were visualized via heatmap plotting (Fig. 7). As showed in Fig. 7, ve out of 24 MdCDPKs (i.e. HF03960, HF05458, HF13700, HF20185, and HF29516) presented no any expression amounts at the transcriptional level. In contrast, the remaining 19 MdCDPKs were constitutively expressed with various patterns in related to both the apple strains (BLO-yellow fruit skin vs. KID-red fruit skin) and the different stages of fruit development (S1~S4, Fig. 7). The expansion patterns of these MdCDPKs were in the trends with three types: (I) higher expression levels, (II) lower expression levels throughout four stages of fruit development, and (III) apparent difference in expression levels at the different stages of fruit development. The expression of six MdCDPKs (HF00526, HF05266, HF10624, HF17744, HF36202, and HF39191) were characterized with higher expression levels (i.e. pattern I), while three MdCDPKs (HF04060, HF20170, and HF28950) were expressed at lower levels (pattern II). Ten other MdCDPKs (HF01706, HF04323, HF05471, HF06540, HF09216, HF10630, HF13801, HF14253, HF15429, and HF39491) showed the difference in expression levels at four stages of fruit development (i.e. pattern III), among which the expression levels of HF01706, HF04323, HF09216, and HF39491, were in a trend of gradual elevation from the stage S1 to S4, with expression peaks at the stage S4 for both of the apple strains (i.e. BLO and KID). On the contrary, the transcriptional expression in a reverse pattern were found out from another group of MdCDPKs (HF05471, HF06540, HF10630, HF13801, HF14253, and HF15429), with higher amounts at the S1, compared to those at other stages (Fig. 7).
To further validate if these MdCDPKs were differentially expressed with signi cance among the four stages of fruit development, or the two apple strains, the transcriptional data were screened for the MdCDPKs by the criterion: log 2 FoldChange > 1 and p-value < 0.05 in the present study. As a result, the MdCDPKs, differentially expressed with signi cance, were presented under the twelve types of comparison (Fig. 8). For both the apple strains BLO (Fig. 8A, 8C and 8E) and KID (Fig. 8J, 8K and 8L), the up-regulatory genes (HF05266 and HF09216) and down-regulatory genes (HF05471 and HF15429) were differentially expressed with signi cance under the comparison S4 vs. S1, S4 vs. S2, and S4 vs. S3. Interestingly, irrespective of their difference in expression fold-changes, the signi cantly up-regulatory genes (HF05266 and HF09216) and down-regulatory genes (HF05471 and HF15429), were also found out by inter-strain comparison, including KID_S4 vs. BLO_S1, BLO_S2, and BLO_S3 (Fig. 8B, 8D and 8F), or BLO_S4 vs. KID_S1, KID_S2, and KID_S3 (Fig. 8G, 8H and 8I), respectively. In addition, the signi cantly down-regulatory gene HF20170 was presented under the comparison S4 vs. S2 for KID (Fig. 8K) or KID_S4 vs. BLO_S1 (Fig. 8B), KID_S4 vs. BLO_S2 (Fig.   8D), while the signi cantly up-regulatory HF00526, HF04323, and HF01706 were under the comparison S4 vs. S2 for BLO (Fig. 8C), S4 vs. S2 for KID (Fig. 8K), and S4 vs. S3 for KID (Fig. 8L), respectively. Therefore, among the MdCDPKs with the expression pattern III at transcriptional level (Fig. 7), ve members (HF01706, HF04323, HF05471, HF09216, HF15429) showed signi cantly differential expression between the speci c stages of fruit development. And the previous described MdCDPKs with the expression pattern I (HF00526, HF05266) or II (HF20170) should be sorted to the pattern III, though these CDPKs appeared the higher or lower expression amounts throughout the four stages of fruit development due to the limited resolution by the heatmap (Fig. 7).

Discussion
CDPKs play essential roles in modulating a variety of developmental processes, and abiotic stress responses, via mediating Ga 2+ signatures [6-8]. Although absent in in animals or yeast, CDPKs are widely presented in plants, green algae, and certain protozoa [5]. In the present study, a total of 116 candidate CDPKs were identi ed in the ve species examined, all of which have been validated with the presence of the conserved domains (i.e. kinase and EF-hands, Fig. 2 and Fig. 3). Among these identi ed CDPKs, the maximal number of members was presented in Arabidopsis (34), followed by pear (28), apple (24), and peach (16), while the minimal number in strawberry (14). And the number of CDPKs in Arabidopsis, is consistent with those in the previous reports [7]. One of Arabidopsis CDPK (AT2G35890) has been reported to have a truncated C terminus containing two EF-hands, instead of four EF-hands in other AtCDPKs [7]. Analogously, CDPKs with an incomplete of four EF-hands, were showed in one MdCDPK (HF28950 with two EF-hands) and PbCDPK (Pb001308 with three EF-hands), respectively ( Fig. 2 and Fig. 3).
Although with a common characteristic of uneven distribution on chromosomes, the CDPKs in four species of the Rosaceae family do not exhibit a cluster of CDPKs on the local region of a chromosome, which could be observed on the short arm of Arabidopsis chromosome no.4 (Fig. 1). CDPK families in various species consist of a large number of members. For instance, 34 genes encoding CDPKs have been revealed from Arabidopsis genome [7], and 40 ones in maize (Zea mays) [22], 30 in poplar (Populus trichocarpa) [26], 19 in cucumber (Cucumis sativus) [27], 27 in barley (Hordeum vulgare) [23], 18 in melon (Cucumis melo) [28], 98 in upland cotton (Gossypium hirsutum) [19]. The generation of CDPK multigene family were likely resulted from the whole genome duplication (WGD) [29]. Following the gene duplication, the subsequent evolution was associated with both redundant and distinct functions of CDPK members. According to molecular clock analysis, it was estimated that the diversi cation of CDPKs in land plants occurred between 268 and 340 MYA (million years ago) [29]. Due to the timing point of divergence between vascular and non-vascular plants (350 ~ 400 MYA), there presented a likelihood that CDPKs in land plants were involved in an adaptation to terrestrial environments [29,30].
The relationships of 116 CDPKs indicated by the phylogenetic tree were further supported by the similar gene structure and protein-motif patterns within each subgroup. It is noticeable that 34 AtCDPKs were clustered in pairs with each other in all ve subgroups of the constructed phylogenetic tree (Fig. 2). In contrast, the clustering CDPK pairs between two different lineages, such as Pbr010295-HF20185, gene05409-Prupe.3G035400, Pbr024654-Prupe.5G110500, were presented among the Rosaceae species. The result suggests that AtCDPKs are relatively less close to those in the Rosaceae species.
According to the constructed phylogenetic tree (Fig. 2), 24 MdCDPKs were dispersed into the subgroups I (nine MdCDPKs), II (four), III (four), and IV (seven), respectively. In many cases, the MdCDPK genes within the same subgroup, do not necessarily present the similar expression patterns at the transcriptional level ( Fig. 2, Fig. 7 and Fig. 8). One such case is the signi cantly up-regulatory MdCDPK gene (HF09216) and down-regulatory one (HF15429) during the four stages of apple fruit development, although both HF09216 and HF15429 are from the subgroup IV (Fig. 2, Fig. 7 and Fig. 8). Moreover, another MdCDPK gene (HF29516) within the subgroup IV, presented no transcriptional expression (Fig. 7). Likewise, within the subgroup III, HF17744 and HF04323 showed the expression patterns I and III, respectively, whereas HF03960 and HF13700 were in transcriptional silence (Fig. 7). The results indicate that these CDPKs have undergone genetic variant events since the evolution of the plant lineage, and potential functional diversi cation such that single paralogous gene may confer different speci cities.
Plants have substantially higher gene duplication rates compared with most other eukaryotes. These plant gene duplicates are mostly derived from tandem, segmental and whole genome duplications. However, the in uence of duplication mechanism on CDPK gene family in the examined species, has not been thoroughly investigated. To uncover the contribution of gene duplications to the evolution of the CDPKs, their collinearity relationships were assessed. Collinearity analysis showed that there presented 36 intraspeciespairs and 209 interspecies-pairs across each pair of species, including 10, 13, 1, and 12 of intraspecies collinearity within apple, pear, peach, and Arabidopsis, and 22, 36, 21, and 25 of interspecies collinearity between apple with Arabidopsis, pear, strawberry, and peach, respectively (Table 2 and Additional le 1: Table S1). The synteny blocks on the individual chromosome with CDPK duplication occurred ( Figure 5 and Additional le 1: Figure S1), were in accordance with the large-scale duplication events. Furthermore, according to the analysis on chromosomal locations, the majority of CDPKs were unevenly distributed across individual genomes (Fig. 1). Altogether, it is indicated that the WGD duplications might have played an important role in the CDPK gene expansion, leading to structural and functional novelty during evolution of the species lineages.
The Ka/Ks ratio is considered an indicator for determining the type of selection pressure [31]. In the present research, all of the evaluated Ka/Ks ratios between CDPK gene pairs with collinearity relationships were less than 1 (Table 2 and Additional le 1: Table S1), indicating these genes have undergone purifying selection to different extents since duplication.
Owing to the homologous members of CDPK families, it is in concert with no phenotypic effects when one gene in an organism is knocked out. One possible explanation is that the effect of knocking out a gene is compensated by its duplicate copy. With the availability of transcriptome (RNA-seq) as one of alternative methods, the analysis of multigene families, as such CDPK gene family, would be promoted in unravelling the functions of a particular CDPK. Based on the transcriptomic pro ling, Li et al. have reported that three CDPKs (Gb_11259, Gb_22778, Gb_26648) were differentially expressed with the signi cance between the two stages of ovule development in Ginkgo biloba [32]. In the present study, the transcriptomic data from RNA-seq analysis on two apple strains at the different stages of fruit development [25], were used for addressing the expression patterns and possible functions of 24 MdCDPKs. Apart from ve MdCDPKs with no expression amounts, 19 MdCDPKs, were characterized with three expression patterns according to their heatmap clustering: pattern I and pattern II, with relatively higher and lower expression levels throughout four stages of fruit development, respectively, and pattern III with apparently different expression levels at four stages of fruit development (Fig. 7). Furth signi cance analysis (log2FoldChange > 1 and p-value < 0.05) on the differential expression of MdCDPKs, showed that four MdCDPKs, including two up-regulatory expression genes (HF05266 and HF09216) and two down-regulatory genes (HF05471 and HF15429), were differentially expressed with signi cance under the comparison S4 vs. S1, S4 vs. S2, and S4 vs. S3, respectively (Fig. 8). It is inferred that HF05266 and HF09216, acting as the calcium sensors, may jointly regulate certain downstream targets at the developmental stage S4 in both apple strains (BLO-yellow fruit skin and KID-red fruit skin), whereas HF05471 and HF15429 could be functional at the stages S1, S2, and S3. And other MdCDPKs with signi cantly differential expression, such as the down-regulatory gene HF20170, the up-regulatory HF00526, HF04323, and HF01706, were presented only between some of the particular stage comparison (i.e. S4 vs. S2 for KID or KID_S4 vs. BLO_S1, KID_S4 vs. BLO_S2, S4 vs. S2 for BLO, and S4 vs. S3 for KID), indicating their speci c roles at the particular stage of apple fruit development.
It was reported previously that two CDPKs in alfalfa, MsCK1 and MsCK2, were differentially expressed under cold stress, with MsCK1 down-regulated and MsCK2 down-regulated, respectively [33], re ecting their speci c functions involved in stress responses. With respect to ve non-expressed MdCDPKs (i.e. HF03960, HF05458, HF13700, HF20185, and HF29516), there are two assumptions: (1) they are pseudogenes; (2) they are expressed only at speci c organs or tissues rather than fruits, or in response to speci c stimuli during developmental processes.

Conclusions
A total of 116 CDPKs in four Rosaceae species (i.e. apple, pear, strawberry, and peach) and Arabidopsis, was characterized at the genome-wide level, including chromosomal distribution, gene structures and conserved motifs, and phylogenetic and collinearity relationships. These CDPKs were showed to be highly conserved both at their kinase and EF-hand domains. Moreover, the WGD and subsequent purifying selection, might have played an important role in the CDPK gene expansion, leading to structural and functional novelty during evolution of the species lineages. Transcriptomic analysis provides an overview for expression patterns of 24 MdCDPKs at the four stages of apple fruit development. In many cases, the MdCDPK genes within a phylogenetic group could show the different expression patterns at the transcriptional level, suggesting that these MdCDPKs have undergone genetic variant events and potential functional diversi cation such that single paralogous gene may confer different speci cities. Furthermore, some of MdCDPKs with the fruit developmental stage-speci c alteration in expression levels, might be coordinated with their peculiar functions in each case.

Declarations Acknowledgements
We would like to thank Dr. Junshan Gao at Anhui Agricultural University for his assistance during the genomic data analysis on this research.
Authors' contributions DHL designed the research project; YD and XXZ, JT, and LFZ carried out the gene family analysis; DHL, YD and XXZ analyzed the transcriptomic data of apple fruit; DHL wrote the paper. All authors read and approved the nal manuscript.

Funding
The research, including the design of the study, data collection, analysis and interpretation, and writing manuscript, was supported by the grants from the Natural Science Foundation of Anhui Province, China (No. 1808085MC58). The funders had no other roles during the experiment and preparation of the manuscript.

Availability of data and materials
The gene sequence information mentioned in this study could be found in Table 1, included in the manuscript.
Transcriptome information used in this research can be downloaded from the NCBI Sequence Read Archive (http://www.ncbi.nlm.nih.gov/sra) with the accession number SRP062637.
Other data produced during this work are included in the manuscript and its supplementary les.      Heatmaps of expression patterns of 24 MdCDPKs for two apple strains (BLO and KID) at the four stages (S1-S4) of apple fruit development. LD1, LD2, and LD3 are triplicates of the RNA-seq samples.