The Genome-Wide Analysis of RALF-Like Genes in Strawberry (Wild and Cultivated) and Five Other Plant Species (Rosaceae)

The rapid alkalinization factor (RALF) gene family is essential for the plant growth and development. However, there is little known about these genes among Rosaceae species. Here, we identify 124 RALF-like genes from seven Rosaceae species, and 39 genes from Arabidopsis, totally 163 genes, divided into four clades according to the phylogenetic analysis, which includes 45 mature RALF genes from Rosaceae species. The YISY motif and RRXL cleavage site are typical features of true RALF genes, but some variants were detected in our study, such as YISP, YIST, NISY, YINY, YIGY, YVGY, FIGY, YIAY, and RRVM. Motif1 is widely distributed among all the clades. According to screening of cis-regulatory elements, GO annotation, expression sequence tags (EST), RNA-seq, and RT-qPCR, we reported that 24 RALF genes coding mature proteins related to tissue development, fungal infection, and hormone response. Purifying selection may play an important role in the evolutionary process of RALF-like genes among Rosaceae species according to the result from ka/ks. The tandem duplication event just occurs in four gene pairs (Fv-RALF9 and Fv-RALF10, Md-RALF7 and Md-RALF8, Pm-RALF2 and Pm-RALF8, and Pp-RALF11 and Pp-RALF14) from four Rosaceae species. Our research provides a wide overview of RALF-like genes in seven Rosaceae species involved in identification, classification, structure, expression, and evolution analysis.


Introduction
The rapid alkalinization factors (RALFs) belong to the small and cysteine-rich secreted peptides, which are involved in various processes of growth and development and play an important role in communication between cells in plants [1]. Generally, the molecular weight of RALF peptides ranges from 80 kDa to 90 kDa, except for a few special peptides which are larger [1]. Pearce et al. [2] showed that mature peptides were almost 50 aa and contain four highly conserved cysteines in the C-terminus, which are essential for activation and are likely involved in disulfide bridges. The mature peptides also contain other conserved regions, such as the YISY motif near the N-terminus, a GASYY motif between the first and second conserved cysteines, and a PYXRGCS motif that contain the third conserved cysteine residue [3]. The YIXY motif is the key regulatory functional element required for targeting RALF to its putative receptor. Meanwhile, when isoleucine (I) is replaced by alanine (A), this can cause a reduction in the alkalinization of suspension cells [2]. Campbell and Turner [4] identified 765 RALF proteins from 51 plant species and divided them into four clades. There are typical features of RALFs in clades I, II, and III, including the RRXL cleavage site and the YISY motif. Nevertheless, clade IV lacks the typical RALF trait, whereas the members of clade IV exhibit diverse expression profiles and

The Expression Profiles of RALF-Like Genes in Strawberry and Apple from RNA-Sequencing
To analyze the wide expression profiles, RNA-seq data were used, and the FASTQ files were downloaded from European nucleotide archive (https://www.ebi.ac.uk/ena). Here, Fragaria vesca, Fragaria × ananassa, and Malus × domestica (only these three species in the present study can be found the available RNA-seq data to analyze the expression profiles about diverse tissues and stresses) were considered when investigating the RALF-like gene expression of diverse tissues, organs, developmental stages, and stresses. The study accessions from NCBI were SRP096282, SRP098567, ERP004230, SRP035308, SRP018410, ERP013896, SRP111905, SRP125281, SRP091754, and SRP139480 [54][55][56][57][58][59][60][61], which were involved in the development of flower tissues, fruit tissues, seed tissues, and fungal stresses. Ultimately, the FPKM (fragments per kilobase per million) [62] of RALF-like genes was extracted from the GTF files and submitted to TBtools [52] to draw the heatmaps with the log 2 translation of FPKM.

Gene Validation Experiments by Quantitative Real-Time PCR (RT-qPCR)
F. × ananassa cv. Benihoppe plants cultured in MS (Murashige and Skoog) solid medium for four weeks to perform, and 0.2 mM NAA (naphthylacetic acid), 0.2 mM ABA (abscisic acid), 0.2 mM GA 4 (gibberellin 4), 0.2 mM MeJA (methyl jasmonate), and 1.0 mM SA (salicylic acid) were sprayed on the tissue of cultured seedlings. NAA, ABA, and GA 4 will influence the growth of plants, and MeJA and SA are related to the stresses of plants. Three plants were used for one treatment, and the leaves of strawberry seedlings were immediately flash frozen using liquid nitrogen and stored in a −80 • C freezer after treating for 12, 24, and 48 h.
Total RNA was isolated according to the instructions of the RNA extraction kit (Tiangen, Beijing, China), and RNA was reverse-transcribed into complementary DNA (cDNA) using the PrimeScript RT reagent kit (TaKaRa, Dalian, China) for quantitative real-time PCR (RT-qPCR). RT-qPCR was performed on an ABI 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The gene-specific primers were designed by the Beacon Designer 7 program (Premier Biosoft, Palo Alto CA, USA). EF-1α (XM_004307362) was used as an internal control to normalize the expression level of target genes. All primers are listed in Table 1. The cycling conditions were maintained as follows; 95 • C for 4 min, 40 cycles at 95 • C for 25 s, 60 • C for 20 s, and 73 • C for 43 s to create a melting curve. The reaction was performed in a 20 µL reaction mixture including a diluted cDNA sample as a template, a SYBR premix ExTaq (2×) (TaKaRa, Kyoto, Japan), and primers. The relative expression level of genes was calculated by the 2 −∆∆CT method [63]. Table 1. Primers used to perform quantitative real-time PCR (RT-qPCR) for RALF-like genes.

Identification, Coding Capacity, Classification, and Structural Features of RALF-Like Genes in Rosaceae Species
The genome-wide search for and verification of the seven Rosaceae species reference genomes indicated that there are 124 sequences with the RALF domain, with 41 from cultivated strawberry, 13 from wild strawberry, 20 from apple, nine from Japanese apricot, 17 from sweet cherry, 14 from peach, and 10 from European pear, in addition to 39 Arabidopsis RALF-like protein sequences that were retrieved from Phytozome [23] and previous research [4], among these 163 sequences, there are five common protein pairs (Pp-RALF1 and Pp-RALF2, Fv-RALF5 and Fa-RALF10, Fa-RALF7 and Fv-RALF6, Fa-RALF12 and Fv-RALF10, and At-RALF7 and At-RALF37).
The RALF-like proteins' primary structure information is listed in Table S1. The molecular weight ranges from 7 to 121 kDa, the number of amino acids ranges from 64 to 747, and the minimum theoretical isoelectric point (pI) is 4.42 and the maximum is 10.31. In the seven Rosaceae species, there were 81% (101/124) RALF-like proteins with a pI higher than 7, and others lower than 7. Therefore, most are basic amino acids. There are~86% (107/124) of these proteins which have a signal peptide, which is essential for secreted proteins. The majority of RALF-like proteins located extracellularly are also involved in other positions, such as cell membrane, cytoplasm, endoplasmic reticulum, mitochondrion, and the nucleus, which implies these proteins mainly function outside the cell.
The results of CPC [45] showed that Fa-RALF22 and Fa-RALF23 have no coding capacity, and 32  have a weak noncoding capacity, 58 have the weak coding ability, and 32 have a clear coding capacity  (Table S1).
According to the phylogenetic analysis (Figure 1), these RALF-like proteins were divided into four clades (I-IV), and clade I was divided into nine subclades (A-G, S1, and S represents the single sequence) and clade III was divided into two subclades (A and B). The majority (~57% (71/124)) of RALF-like proteins in Rosaceae species belong to clade I, 24 RALF-like proteins to clade II, 26 to clade III, and three to clade IV. Clade I-D, clade I-F, clade II, and clade III-B contain all the Rosaceae species, clade I-A contains five Rosaceae species, and clade I-C and clade I-E contain six Rosaceae species. Clade I-G and clade III-A contain four Rosaceae species. Others contained one Rosaceae species, which implied that most RALF-like proteins identified in the present study have conserved functional structures.
Genes 2020, 11, x FOR PEER REVIEW 6 of 20 species, there were 81% (101/124) RALF-like proteins with a pI higher than 7, and others lower than 7. Therefore, most are basic amino acids. There are ~86% (107/124) of these proteins which have a signal peptide, which is essential for secreted proteins. The majority of RALF-like proteins located extracellularly are also involved in other positions, such as cell membrane, cytoplasm, endoplasmic reticulum, mitochondrion, and the nucleus, which implies these proteins mainly function outside the cell.
The results of CPC [45] showed that Fa-RALF22 and Fa-RALF23 have no coding capacity, and 32 have a weak noncoding capacity, 58 have the weak coding ability, and 32 have a clear coding capacity (Table S1).
According to the phylogenetic analysis (Figure 1), these RALF-like proteins were divided into four clades (I-IV), and clade I was divided into nine subclades (A-G, S1, and S represents the single sequence) and clade III was divided into two subclades (A and B). The majority (~57% (71/124)) of RALF-like proteins in Rosaceae species belong to clade I, 24 RALF-like proteins to clade II, 26 to clade III, and three to clade IV. Clade I-D, clade I-F, clade II, and clade III-B contain all the Rosaceae species, clade I-A contains five Rosaceae species, and clade I-C and clade I-E contain six Rosaceae species. Clade I-G and clade III-A contain four Rosaceae species. Others contained one Rosaceae species, which implied that most RALF-like proteins identified in the present study have conserved functional structures.  Combining the alignment from Jalview Version 2 (muscle with defaults) [43] ( Figure S1) and the analysis of motifs retrieved from MEME [44] (Figure S2), we found some structural features among clades or subclades according to the alignment and motif results.
A majority of the RALF-like proteins have four conserved cysteines, except 22 sequences from the Rosaceae species (sequences marked with green background in Table S1). There are 51 sequences which have the YISY motif, the RRXL cleavage site, and four conserved cysteine residues the typical features of mature RALF genes and 45 from the Rosaceae species. All the sequences in clade I-F and most sequences of clade I-D, clade I-E, and clade II (seven from Rosaceae species in Clade I-D, eight from Rosaceae species in Clade I-E, 18 from Rosaceae species in Clade II) have typical features. These mature RALF-like genes may be more meaningful for future research, whereas other sequences have none or one of these features. Searching these sequences, some YISY motifs may have some mutations, such as YISP, YIST, NISY, YINY, YIGY, YVGY, FIGY, YIAY, and RRXL, which may mutate to RRVM.
These protein sequences were searched for the presence of six motifs (Figure 2), and motif1 was found to exist in 141 sequences and 110 sequences from Rosaceae species. The second and third most common motifs are motif3 (84 sequences with 71 from Rosaceae species) and motif2 (75 sequences with 63 from Rosaceae species). Motif1 distributed widely among all the clades. Combining the alignment from Jalview Version 2 (muscle with defaults) [43] ( Figure S1) and the analysis of motifs retrieved from MEME [44] (Figure S2), we found some structural features among clades or subclades according to the alignment and motif results.
A majority of the RALF-like proteins have four conserved cysteines, except 22 sequences from the Rosaceae species (sequences marked with green background in Table S1). There are 51 sequences which have the YISY motif, the RRXL cleavage site, and four conserved cysteine residues the typical features of mature RALF genes and 45 from the Rosaceae species. All the sequences in clade I-F and most sequences of clade I-D, clade I-E, and clade II (seven from Rosaceae species in Clade I-D, eight from Rosaceae species in Clade I-E, 18 from Rosaceae species in Clade II) have typical features. These mature RALF-like genes may be more meaningful for future research, whereas other sequences have none or one of these features. Searching these sequences, some YISY motifs may have some mutations, such as YISP, YIST, NISY, YINY, YIGY, YVGY, FIGY, YIAY, and RRXL, which may mutate to RRVM.
These protein sequences were searched for the presence of six motifs (Figure 2), and motif1 was found to exist in 141 sequences and 110 sequences from Rosaceae species. The second and third most common motifs are motif3 (84 sequences with 71 from Rosaceae species) and motif2 (75 sequences with 63 from Rosaceae species). Motif1 distributed widely among all the clades.

Chromosome Location, Duplication Events, and Divergence Rates
As depicted ( Figure S3), there is an expanding distribution among Rosaceae species according to the chromosomal location results. In wild strawberry, there are six chromosomes, except Fvb7, in which 13 RALF-like genes were located. In cultivated strawberry, there are 21 chromosomes which have RALF-like gene sites. The following numbers of RALF-like genes were identified in specific chromosomal locations: 10 in apple, 8 in sweet cherry, 6 in peach, and 9 in European pear. In Japanese apricot, which has eight chromosomes, Pltd and Unplaced Scaffold regions, RALF-like gene sites were found on six chromosomes, whereas Pm-RALF9 is located in Un region. As shown, most of the RALF-like genes are distributed around the telomere, and others around the centromere.
The ka and ks were calculated by DnaSP [49] using all the RALF-like CDs sequences identified. As is calculated, the ka range is 0 to 4.8, ks range is 0 to 4.1, and the ka/ks is 0 to 22.1. The data distribution map was displayed through box plots which were constructed via Origin (OriginLab Corporation, Northampton, MA, USA) and the outlines were eliminated automatically ( Figure 4). The ka and ks values were stably distributed (Figure 4a,b), which implies there were stable

Chromosome Location, Duplication Events, and Divergence Rates
As depicted ( Figure S3), there is an expanding distribution among Rosaceae species according to the chromosomal location results. In wild strawberry, there are six chromosomes, except Fvb7, in which 13 RALF-like genes were located. In cultivated strawberry, there are 21 chromosomes which have RALF-like gene sites. The following numbers of RALF-like genes were identified in specific chromosomal locations: 10 in apple, 8 in sweet cherry, 6 in peach, and 9 in European pear. In Japanese apricot, which has eight chromosomes, Pltd and Unplaced Scaffold regions, RALF-like gene sites were found on six chromosomes, whereas Pm-RALF9 is located in Un region. As shown, most of the RALF-like genes are distributed around the telomere, and others around the centromere.
The ka and ks were calculated by DnaSP [49] using all the RALF-like CDs sequences identified. As is calculated, the ka range is 0 to 4.8, ks range is 0 to 4.1, and the ka/ks is 0 to 22.1. The data distribution map was displayed through box plots which were constructed via Origin (OriginLab Corporation, Northampton, MA, USA) and the outlines were eliminated automatically ( Figure 4). The ka and ks values were stably distributed (Figure 4a,b), which implies there were stable evolutionary counts and mutation frequencies between and among these species. The ka/ks orients the evolution direction. As depicted (Figure 4c), all the average mean values of genes pairs were lower than one, which implies that purifying the selection may promote evolution. The values of ka, ks are listed in Table S2.
Genes 2020, 11, x FOR PEER REVIEW 9 of 20 evolutionary counts and mutation frequencies between and among these species. The ka/ks orients the evolution direction. As depicted (Figure 4c), all the average mean values of genes pairs were lower than one, which implies that purifying the selection may promote evolution. The values of ka, ks are listed in Table S2.

Gene Ontology and Cis-Regulatory Elements Analysis
Among 124 sequences (RALF-like genes in Arabidopsis not included), 17 GO terms were found in 51 RALF-like genes. Approximately 85% (138/163) of RALF-like genes in the present study (conclude the sequences from Arabidopsis thaliana) were enriched in the GO category of cellular component (CC). The gene number of each GO term and functional class were counted and depicted in Figure 5. The GO categories of the biological process (BP) and molecular function (MF) accounted for 74% (120/163) and 33% (54/163), respectively. The calcium-mediated signaling (GO: 0019722), plasmodesma (GO: 0009506), protein phosphorylation (GO: 0006468), hormone activity (GO: 0005179), cell-cell signaling (GO: 0007267), and the integral component of the membrane (GO: 0016021) are major GO categories.

The Expression Information from Expression Sequence Tag Data
As listed in Table S3, according to the results from ESTs, we reported that twelve plant tissues (flower, bud, fruit, fruit mesocarp, fruit core, fruit cortex, phloem, xylem, leaf, shoot internodes, young root, and young shoot), six abiotic stresses (cold, water, drought, heat, and salt), and two biotic stresses (Venturia inaequalis and Choristoneura rosaceana) are involved ( Table 2).

The Expression Information from Expression Sequence Tag Data
As listed in Table S3, according to the results from ESTs, we reported that twelve plant tissues (flower, bud, fruit, fruit mesocarp, fruit core, fruit cortex, phloem, xylem, leaf, shoot internodes, young root, and young shoot), six abiotic stresses (cold, water, drought, heat, and salt), and two biotic stresses (Venturia inaequalis and Choristoneura rosaceana) are involved ( Table 2). Table 2. The expression profile of RALF-like genes from seven Rosaceae species from EST data.

The RNA-Sequencing Expression of RALF-Like Genes in Strawberry and Apple
The downloaded RNA-seq data were used to analyze the expression profiles of RALF-like genes and the basic information of these datas listed in Table S4. In wild strawberry, Fv-RALF1, Fv-RALF2, Fv-RALF3, Fv-RALF5, and Fv-RALF12 were more highly expressed than other genes ( Figure S5a). Fv-RALF1 shows a lower expression level in the ghost development stages and wall 4, wall 5 stages, which means it may affect the development of ghost and wall. Fv-RALF2 and Fv-RALF3 all have different expression profiles in the cortex compared with other tissues, which implies the potential functions on cortex development. Fv-RALF5 and Fv-RALF12 both have lower expression in embryo than other tissues and may influence the development of the embryo. Fv-RALF4, Fv-RALF6, Fv-RALF7, Fv-RALF8, Fv-RALF9, Fv-RALF10, and Fv-RALF11 all have higher expression levels in specific tissues ( Figure S5d). Fv-RALF4, Fv-RALF9, and Fv-RALF10 are more highly expressed pollen-specific genes. These genes may play an important role in the development of pollen. The receptacle fruit at green stage (15 days post-anthesis) and white (turning) stage (22 days post-anthesis) with achenes removed were used to analyze the RALF-like gene expression profile ( Figure S5b). As depicted, Fv-RALF1, Fv-RALF5, Fv-RALF12, and Fv-RALF13 all show higher expression levels in some ways. To retrieve the expression profile of RALF-like genes in achene and receptacle fruit, the RNA-seq data of one red-fruited and two natural white-fruited strawberry varieties in two tissues and three ripening stages were used ( Figure S5e). Fv-RALF1, Fv-RALF2, Fv-RALF5, Fv-RALF10, and Fv-RALF12 were more highly expressed in some cultivars and tissues, and showed variable expression profiles, which means they may have an effect on the development of achene and receptacle fruit. The powdery mildew-infected leaf sample from Hawaii was used to analyze the expression level under fungal stress ( Figure S5c). Fv-RALF12 always showed a high expression levels, though this does not exclude it from having an essential function in resistance from infection by powdery mildew. However, Fv-RALF1 is more highly expressed one day after infection, which may represent a response to infection.
In cultivated strawberry, the RNA-seq data of the development achene and receptacle fruit were also analyzed ( Figure S6a). Fa-RALF1, Fa-RALF18, and Fa-RALF39 are more highly expressed at the white receptacle stage, whereas, Fa-RALF14, Fa-RALF24, and Fa-RALF33 exhibit a higher level of expression in leaf than other tissues, and Fa-RALF29 and Fa-RALF35 may be highly expressed root-specific genes. In diverse development stages of cultivated strawberry ( Figure S6b), we reported that Fa-RALF1, Fa-RALF10, and Fa-RALF33 were all more highly expressed at the turning stage compared to other stages. Fa-RALF6, Fa-RALF17, Fa-RALF26, and Fa-RALF32 were more highly expressed at the green fruit stage.
According to the RNA-seq data of apple, we reported that Md-RALF7 and Md-RALF8 may have an effect on the development of flower ( Figure S7a). Md-RALF3, Md-RALF4, Md-RALF13, and Md-RALF16 are more highly expressed in root tip ( Figure S7b). Md-RALF3, Md-RALF4, and Md-RALF13 were more highly expressed than other isotypes in during infection by Alternaria alternata (Figure S7c). Md-RALF3 responds to infection after 12 h. Md-RALF4 responds to the infection after 18 h. The expression level of Md-RALF13 remained high until 72 h, in which it began to decline after.

RT-qPCR Analysis of Nineteen RALF-Like Genes in Cultivated Strawberry Related to Hormones
Nineteen RALF-like genes of cultivated strawberry under different hormone stresses were analyzed by RT-qPCR ( Figure 6). Herein, we reported that

Expression Profile of RALF-Like Genes in Seven Rosaceae Species
As reported, RALF genes are not only involved in several biotic and abiotic stresses but also influence the growth and development of plants. Here, we found that the expression profiles are mostly similar with the researchers reported, such as the drought response [13], root growth [64], hormone response [18], fungi response [18][19][20], cell wall development [15,17], and the flower development (especially the development of pollen tube) [14,16]. In the present study, we combined the results of RNA-seq, EST data, cis-regulatory elements analysis, and RT-qPCR to identify several genes which have identical expression profiles in two or more analyses, which may give some instructions of the research of RALF-like genes among Rosaceae species.
Fv-RALF10 exhibits drought responses according to the analysis of EST data and cis-regulatory elements. RNA-seq consistent with cis-regulatory element analysis involved in Fv-RALF1, Fv-RALF2, Fv-RALF12, Fv-RALF15 (seed, embyro, and endosperm), which have no research reported, may be the special findings in our study among Rosaceae species. The RALF genes most reported were the role they play in root growth and pollen tube development [14,16]. Here, the comparison between RNA-seq and EST data reveal that Md-RALF3 may have influence on the growth of young root, and Md-RALF7 and Md-RALF8 may affect the development of flowers. As reported that MeJA may influence the expression of RALF genes [18]; here, we tried more hormones, which may influence the tissue development and stresses response, the results from RT-qPCR and cis-regulatory elements showed that Fa-RALF1, Fa-RALF2, Fa-RALF12, Fa-RALF13, Fa-RALF14, Fa-RALF18, Fa-RALF19, Fa-RALF24, Fa-RALF25, Fa-RALF29, Fa-RALF31, Fa-RALF33, and Fa-RALF40 have identical expression profiles, in other words, they may have influence on the growth, development and stresses response of Rosaceae species.

Expression Profile of RALF-Like Genes in Seven Rosaceae Species
As reported, RALF genes are not only involved in several biotic and abiotic stresses but also influence the growth and development of plants. Here, we found that the expression profiles are mostly similar with the researchers reported, such as the drought response [13], root growth [64], hormone response [18], fungi response [18][19][20], cell wall development [15,17], and the flower development (especially the development of pollen tube) [14,16]. In the present study, we combined the results of RNA-seq, EST data, cis-regulatory elements analysis, and RT-qPCR to identify several genes which have identical expression profiles in two or more analyses, which may give some instructions of the research of RALF-like genes among Rosaceae species.
Fv-RALF10 exhibits drought responses according to the analysis of EST data and cis-regulatory elements. RNA-seq consistent with cis-regulatory element analysis involved in Fv-RALF1, Fv-RALF2, Fv-RALF12, Fv-RALF15 (seed, embyro, and endosperm), which have no research reported, may be the special findings in our study among Rosaceae species. The RALF genes most reported were the role they play in root growth and pollen tube development [14,16]. Here, the comparison between RNA-seq and EST data reveal that Md-RALF3 may have influence on the growth of young root, and Md-RALF7 and Md-RALF8 may affect the development of flowers. As reported that MeJA may influence the expression of RALF genes [18]; here, we tried more hormones, which may influence the tissue development and stresses response, the results from RT-qPCR and cis-regulatory elements showed that Fa-RALF1, Fa-RALF2, Fa-RALF12, Fa-RALF13, Fa-RALF14, Fa-RALF18, Fa-RALF19, Fa-RALF24, Fa-RALF25, Fa-RALF29, Fa-RALF31, Fa-RALF33, and Fa-RALF40 have identical expression profiles, in other words, they may have influence on the growth, development and stresses response of Rosaceae species.
In strawberry, the FaRALF-33-like gene, which is homologous to FvRALF-33-like (GenBank accession No.: XM_011460413.1) and AtRALF33 (named At-RALF1 here), responded to infection by three different fungi [20]. Local BLASTP [46] was used to search the similar genes among our 163 sequences, using FvRALF-33-like and AtRALF33 as the query sequences. We revealed that Fa-RALF1 (GenBank accession No.: XP_011458715) and Fv-RALF5 are the best-match sequences. As shown (Figure 7), except for Fa-RALF1, other sequences all possess the typical features and the integral YISY, YYNC, RCR motifs. Fv-RALF5 belongs to clade II, and among the clade, there are seventeen other mature RALF genes (Pc-RAL2, Md-RALF4, Pc-RALF6, Md-RALF11, Pa-RALF10, Pm-RALF5, Pp-RALF5, Fa-RALF35, Fa-RALF29, Fa-RALF5, Pp-RALF6, Pm-RALF7, Pa-RALF1, Pc-RALF7, Md-RALF7, Md-RALF12, and Fa-RALF10) from Rosaceae species, and no the RCR motif was found in Pc-RALF6. Therefore, the 16 RALF-like genes, except for Pc-RALF6, may have similar functions as Fv-RALF5, At-RALF1, and FaRALF-33-like genes. As depicted in the phylogenetic tree, we reported that mature Fv-RALF5 belongs to clade II, and among the clade, there are seventeen other mature RALF genes (Pc-RAL2, Md-RALF4, Pc-RALF6, Md-RALF11, Pa-RALF10, Pm-RALF5, Pp-RALF5, Fa-RALF35, Fa-RALF29, Fa-RALF5, Pp-RALF6, Pm-RALF7, Pa-RALF1, Pc-RALF7, Md-RALF7, Md-RALF12, and Fa-RALF10) from Rosaceae species, and no the RCR motif was found in Pc-RALF6. Therefore, the 16 RALF-like genes, except for Pc-RALF6, may have similar functions as Fv-RALF5, At-RALF1, and FaRALF-33-like genes. As depicted in the phylogenetic tree, we reported that mature RALF genes in clade II, except Pc-RALF6, At-RALF4, and At-RALF1, were similarly divided within the same clade, which is consistent with previous reports. Stegmann et al. [65] already investigated that RALF33 (AT4G15800, named At-RALF1 here), RALF23 (AT3G16570, named At-RALF4 here), and RALF34 (AT5G67070, named At-RALF14 here) negatively regulate immunity, whereas RALF32 (AT4G14010, named At-RALF5 here) did not. Fv-RALF5 clustered with Fa-RALF10, which implies that they have a similar function. They may also respond to the fungal infection. Furthermore, according to the RNA-seq results and cis-regulatory element analysis, we reveal that they may have influence on the development of receptacle and be involved in some hormone stresses (ABA, MeJA, etc.) and low temperature stress. Compared with the expression profiles of Arabidopsis in Campbell and Turner [4], we found that the clade IV (clade I-A and clade III-A concluded here) are not consistent with our research mostly. The expression profiles are involved in hormone, fruit development, flower development, leaf growth, salt stress, powdery mildew response, and achene development.
The genes mentioned above are essential to the growth, development, and stress response. In particular, the mature RALF genes were already found to exhibit obvious expression differences. The GO terms give a clear function cluster of RALF-like genes, which involved in the pollen tube development, root development, negative regulation of growth, cell wall growth, hormone activaty, brassinosteroid-mediated signaling pathway, and so on. According to the classification, we found they are very consistent with previous researches about the functions of RALF genes [1, [14][15][16][17][18]64]. Fungal infection response and pollen tube development are important for Rosaceae species, and the RALF-like gene expression in seed and fruit among Rosaceae species specially.

Classification, Duplication Events and Evolution of RALF-Like Proteins in Seven Rosaceae Species
In several researches, the classification and evolution of RALF-like proteins were analyzed. Cao et al. [9] characterized RALF proteins in Arabidopsis, rice, poplar, and maize, which were divided into ten groups, and elucidated that tandem duplication played a dominant role in the expansion of RALF gene family and RALF-like proteins mainly through purifying selection, which is consistent with the conclusions of our research. However, we found that the dominant duplication event was WGD or segmental, not tandem. Sharma et al. [11] divided RALF-like proteins of Arabidopsis, rice, maize, and soybean into seven groups, and the RALF-like genes in 51 plant species also diverged into four major clades [4]. In our work, four clades were also divided among RALF-like genes between Arabidopsis and seven Rosaceae species.
The number of RALF-like genes was not consistent with the previous. Cao et al. [9] identified 33 RALF-like genes in Arabidopsis; Sharma et al. [11] identified 39; Campbell and Turner [4] identified 37; and, in our research, we retrieved 39 from analysis. Due to the reference genomes we selected not being identical with Campbell and Turner [4], we identified 13, and not 9, RALF-like genes in Fragaria vesca, 20, and not 33, in Malus × domestica, and 14, and not 13, in Prunus persica. Meanwhile, Fragaria × ananassa and not Arabidopsis had the most RALF-like genes (41). The classification in our study among RALF-like genes in Arabidopsis largely confirmed previous reports.
The alignment of RALF-like genes was implemented using Jalview Version 2 (muscle with defaults) [43], and the residue conserved within the mature peptide region of the four major clades was demonstrated using WebLogo 3 [66]. As depicted in Figure S8 compared with the analysis of Campbell and Turner [4], clade I-C, clade I-G, clade II, and clade IV were similar to clade I and II (A), clade I-D, clade I-E, clade I-F to clade III, clade I-A to clade IV, clade III-B to clade II (B), and clade III-A to clade IV, whereas clade I-B did not share any similarity to other clades and has none of the features of RALF genes. Campbell and Turner [4] reported that the genes in clade IV were not mature RALF genes. They also reported the RALF-like genes in clade I-A and clade III-A to be RNA-like genes, which is consistent with our research.

Conclusions
This study provides extensive information on the RALF-like genes of seven Rosaceae species, and involved, including their identification, an analysis of their coding capacity, evolutionary relationships, duplication events, motifs, domains, GO annotation, cis-regulatory elements, and EST data, as well as use of RNA-seq and RT-qPCR to characterize expression profiles. These RALF-like genes are divided into four clades, and purifying selection plays a main role in the evolutionary process. In particular, the 24 RALF genes coding for mature proteins among the Rosaceae species deserve to be further researched in regard to hormone response, fungal infection, growth, and development.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4425/11/2/174/s1, Table S1: The basic information of RALF-like genes among Arabidopsis thaliana and seven Rosaceae species ; Table S2: The ka, ks and ka/ks values among Arabidopsis and seven Rosaceae species; Table S3: The EST data analysis results;  Table S4: The RNA-seq datasets' information; Figure S1: The conserved protein structures analyzed in the research. The conserved alignment results from Jalview. The blue dotted line marks the YISY motif, the single arrow down represents the cysteine, the green line marks the RRXL cleavage site, and the shear shows the cutting site; Figure  S2. The motif location and arrange from MEME; Figure S3