Identification of GRFs in Rosaceae
The GRF Hidden Markov Model (HMM) configuration file (PF08879 and PF08880) was used to identify GRF members. Furthermore, the SMART tool was used to verify the existence of two characteristic conserved domains (QLQ and WRC) in the candidate genes. After manually removing the incomplete sequences from peach (Prunus persica), European pear (Pyrus communis), strawberry (Fragaria vesca), and apple (Malus domestica), respectively, 46 GRF genes were identified and named according to their position on the chromosome. The GRFs were unevenly distributed across the chromosomes of the four species. The 46 predicted GRF proteins ranged from 187 to 843 amino acid residues in length, and their relative molecular mass varied from 20.70 to 94.27 kDa (Table 1).
Phylogenetic and structural analyses of GRF genes
To get a better understanding of evolutionary relationships, a neighbor-joining phylogenetic tree was constructed among GRFs from Arabidopsis (9), soybean (22), rice (12) poplar (19), and four Rosaceae species (46). As shown in Fig. 1, 108 GRFs from the different species were divided into six subgroups ( I-VI) . The majority [~ 23% (25/108)] of GRF proteins belonged to clade I, 21 to clade II (without AtGRF), 24 to clade III, 15 to clade IV (without OsGRF), 13 to Clade V (without OsGRF), and 10 to clade VI (without GmGRF) (Fig. 1). No species-specific clades were discovered, and every subgroup contained Rosaceae GRFs, which shows that GRF proteins are evolutionarily conserved .
To better explore the evolutionary relationships and predict the function of GRF proteins, we investigated the exon-intron patterns and motif characteristics of GRFs. We found some structural features among clades or subclades according to the alignment and motif results. As shown in Fig. 2, a total of 10 conserved motifs of Rosaceae GRFs were found using the MEME online software. Each subgroup has three to nine conserved motifs, and the motif composition is similar within subgroups [36–40]. All of the GRF family members contained motif 1 or motif 2. Based on such a gene structural analysis, we determined that motif 1 and motif 2 correspond to the WRC and QLQ domains, respectively. Some motifs occur only in certain specific subgroups, which possibly contribute to functional diversity. For instance, motif 6 is unique to subgroup I, and motif 5 is unique to subgroup II, while motifs 3, 4, 7, 9, and 10 are concentrated in group III. Motifs 3, 4, and 9 are specific to group IV. Motifs 3 and 4 are specific to group VI. We also investigated the structure of GRF gene to further describe their evolutionary trajectory. All GRFs contained conserved QLQ and WRC domains in their N-terminal regions. However, group V has two WRC domains. These results indicate that they contain different exons numbers, varying from 3 to 12. The gene structures were similar or the same in each subgroup. Overall, phylogenetic relationships are strongly supported by gene structure and motif characteristics.
Chromosomal localization and collinearity analyses of GRF genes
It’s random for the distribution of the GRF genes in the four Rosaceae genomes. Unlike in previous studies, in the four Rosaceae, some chromosomes do not have a GRF gene. In M. domestica and P. communis, the GRF genes are mainly found on chromosomes 2 and 15. For P. persica and F. vesca, the GRF genes were principally distributed on chromosomes 2 and 7, respectively. Additionally, two GRF genes (MdGRF15 and MdGRF16) could not be mapped to any chromosome in the M. domestica genomes (Fig. 3).
Subsequently, in order to further and deeper infer the phylogenetic relationship between peach and other Rosaceae plants, the collinear relationships between three Rosaceae plants and peach (Fig. 3) was analyzed. We found that nine FvGRF genes were collinear with PpGRF genes, followed by MdGRFs (14) and PcGRFs (10). Moreover, the collinear relationship in the peach genome was also examined to elucidate the evolution and origin dynamics of PpGRF genes. Three pairs of PpGRF genes (PpGRF01/08, PpGRF03/07, PpbZIP05/10) have a collinear relationship and were generated by whole-genome duplication (WGD) or segmental duplication (Fig. 3 and Additional File 1: Supplementary Table S1). Furthermore, to determine the selection constraints on the duplicated PpGRF genes, we calculated the non-synonymous/synonymous substitution ratio (Ka/Ks) for each pair of duplicated genes. The Ka/Ks ratios of most PpGRF pairs were less than 1, the evidence shows that these PpGRFs had undergone purifying selection processes (Additional File 1: Supplementary Table S1).
UVB effect on peach leaf growth
During the whole groeing period, the peach trees were treated with 1.44 Kj m− 2 d− 1 UVB radiation . It can be seen from Fig. 4 that, during the growth period of the new shoot, UVB supplementary light treatment can inhibit the elongation of the new shoot. After harvesting for 120 days (September 7), the new shoot of the UVB group stopped lengthening, while the CK group continued to increase slowly. The average new shoot length (114.08 cm) was about 1.5 times that of the UVB group (69.45 cm). The thickness of the new shoot in the CK group increased by 0.3 cm, while the thickness of the new shoot in the treatment group increased by 0.29 cm; the difference was not statistically significant. Based on this finding, this dose of UVB treatment had little effect on the thickness of the new shoot.
Hormone profiles after UVB exposure in peach
Based on the Previous studies, there have various statistics demonstrated that GRFs play important roles in regulating leaf growth [8, 13–15, 41]. The shoot tip is the growth point of the new shoot, and its hormone content determines the growth of the new shoot. At this facility, the new shoots of peach grow vigorously from June to September, so in this experiment, the shoot tip tissue was taken on August 7 and frozen for storage for hormone content analyses. The results of the hormone content analyses are shown in Fig. 5: Under UVB treatment, the content of IAA varies greatly, and the IAA content of the shoot tip tissue of CK group (1.978917 pmol/g) is 4.9 times that of UVB treatment group (0.40339 pmol/g). At the same time, the GA3 content was increased in the UVB treatment group (0.59663 pmol/g) relative to the CK group (0.22767 pmol/g). In our hormone levels analysis, IAA content showed a significant difference (P < 0.01). The GA3 content also showed a significant difference (P < 0.05). The content of abscisic acid (ABA), jasmonic acid (JA) and trans-Zeatin-riboside (TZR) were not significantly different when comparing CK group. Therefore, we speculate that the UVB-mediated growth inhibition of new shoot is closely related to the IAA and GA3 contents in the shoot tip.
Prediction of the miR396 target site and how its expression changes following UVB exposure
Differently to Arabidopsis, all GRFs in peach have a sequence that is partially complementary to miR396, which is located at the WRC domain. To delve deeper into the evolutionary relationship among GRFs, we analyzed the WRC structures of all GRFs in the four Rosaceae. According to Fig. 6, we found that all GRFs of the four Rosaceae contain this part, which is complementary to the miR396 share a bulge at position 7 (counting from the 5 end of the miRNA). From this point of view, the evolution of GRFs in Rosaceae is highly conservative. We have identified a series of differentially expressed miRNAs that have been predicted to be responsive to low-dose UVB. To identify the expression profiles of miRNA396, we utilized transcriptome data of Illumina RNA-Seq reads that were generated and analyzed by Li et al . We found that, in peach, miR396 was down-regulated after UVB irradiation, but this did not reach statistical significance. Also, we isolated another mirR319, which regulates TCP4, induces miR396, and represses GRF activity, which was also non-significantly downregulated following UVB irradiation  (Additional File 2: Supplementary Table S2).
Analysis of cis-elements in the promoter sequences of PpGRF genes
To further explore the involvement of the PpGRF genes in light responsiveness, hormone signaling pathways, plant growth and development, their promoter sequences were analysed using PlantCARE software . All cis-elements in the promoter regions of the PpGRF gene family members are shown in Fig. 7. There are at least six commonly occurring light-responsive elements (LREs): AE-box, ATCT-motif, G-box, GT1-motif, GATA-motif and I-box, which have been demonstrated to be essential for the regulation of light mediated transcriptional activity [43–45]. The results indicated that all 10 PpGRF promoter regions contained two or more LREs among which G-box is the most abundant element (Additional File 3: Supplementary Table S3). In this area we qualificated many other important potential cis-elements, such as GARE motif and P-box (GA responsive element), a CAT-box (cis-acting regulatory element related to meristem expression), AuxRR-core and TGA-element (auxin-responsive element), an RY element (seed-specific regulation) and a TC-rich repeats (defense and stress responsiveness) [46–51]. GARE, P-box, CAT-box, AuxRR-core, TGA-element, RY element and TC-rich repeats were distributed within 3, 3, 7, 2, 3, 1 and 5 PpGRF promoter regions, respectively (Additional File 4: Supplementary Fig S1).
Expression pattern of GRF genes in peach
Although we identified GRF genes in the peach genome, the functions of these genes remain largely unknown. To sum up, GRFs were highly expressed in growing tissues . Next, we studied the patterns of expression about the PpGRFs in root, stem, leaves, the shoot tip and hypocotyl. Except for PpGRF8, these genes were differentially expressed in different tissues (data not shown). Most of these were up-regulated, especially in the shoot tip and root (Fig. 8). For example, the expression levels of PpGRF2, 3, 4, 5, 6, 7, 9, and 10 were the highest in the shoot tip. However, PpGRF1 showed relatively strong expression in the root. Of the analyzed genes, PpGRF3, PpGRF9, and PpGRF10 were more preferentially expressed in the shoot tip; more than 20 times greater than that in other tissues. To further investigate the potential functions of PpGRFs, we also evaluated the effects of GA3 and UVB on the expression of the PpGRF gene (Fig. 9). In abiotic stresses, over half of the total genes were up-regulated, with PpGRF1, 4, 5, 6, 7, and 10 being up-regulated after GA3 treatment; PpGRF2, 3, 4, 5, 6, and 7 were up-regulated after UVB treatment. On the other hand, several PpGRF genes (including PpGRF2, PpGRF6, and PpGRF7) were suddenly down-regulated at 9 h after GA3 treatment (Fig. 9). Also, PpGRF5, PpGRF9, and PpGRF10 were suddenly down-regulated at 6 h after UVB treatment. Based on the expression patterns of PpGRFs, we propose that they likely have multiple functions in regulating growth and responsiveness to abiotic stresses.