Isolation and sequences analysis of MiTFL1 genes
Four TFL1 homologous genes were identified from our previous transcriptome data. We further verified the sequences by RT-PCR and showed that these sequences were consistent with those obtained from the transcriptome data. The four genes were named MiTFL1-1, MiTFL1-2, MiTFL1-3 and MiTFL1-4, and their DNA sequence lengths were 1175 bp, 1054 bp, 962 bp and 1299 bp, respectively. All MiTFL1 genes contained four exons and three introns (Fig. 1A). The full coding sequences of the four MiTFL1s were 516 bp, 525 bp, 519 bp and 510 bp and encoded 172 aa, 175 aa, 173 aa and 170 aa, respectively. Nucleotide and amino acid sequence alignment analyses showed a higher similarity between MiTFL1-3 and MiTFL1-4 (90%) than between MiTFL1-1 and MiTFL1-2 (70%). The amino acid sequences of the MiTFL1-1, MiTFL1-2, MiTFL1-3, and MiTFL1-4 proteins exhibited 68.9%, 69.1%, 62.9% and 60.1% similarity with AtTFL1 (NP_196004.1) of Arabidopsis, respectively. In addition, all MiTFL1 proteins were identified as TFL1 proteins containing the crucial conserved amino acid residues of TFL1-like proteins (Fig. 1B), namely, histidine at position 85 (H85) and aspartic acid at position 140 (D140).
The PEBP gene family is divided into the TFL1, FT and MFT subfamilies. According to the phylogenetic tree analysis (Fig. 2), the MiTFL1-1, MiTFL1-2, MiTFL1-3 and MiTFL1-4 proteins were clustered with the TFL1 proteins of other species. Among the investigated proteins, the MiTFL1-1 protein was found to be closely related to the TFL1 proteins of apple, pear, apricot, plum, walnut, jujube and other fruit trees of Rutaceae and rose plants. The MiTFL1-2 protein was closely related to the TFL1 proteins of longan and grape and the CEN proteins of apple and cocoa. The MiTFL1-3 and MiTFL1-4 proteins were clustered together and were found to be closely related to the PvCEN protein of pistachio.
Expression analysis of MiTFL1s
The expression pattern of MiTFL1s in different tissues of mango, including juvenile leaves, mature leaves, mature stems and flowers, was determined by qRT-PCR (Fig. 3A). The results showed that MiTFL1-1, MiTFL1-2, MiTFL1-3 and MiTFL1-4 were more highly expressed in mature stems than in juvenile leaves and showed lower expression in mature leaves and flowers. MiTFL1-2 and MiTFL1-3 exhibited higher expression levels than MiTFL1-1 and MiTFL1-4 in all tested tissues.
To explore the expression patterns of the MiTFL1 genes at different flowering stages of mango, mature leaves of M. indica L. cv. ‘SiJiMi’ were collected from the vegetative growth period to the flowering period (November 2016-March 2017), and the results from their analysis are shown in Fig. 3B. The expression patterns of the four MiTFL1 genes in mature leaves differed among the different flowering stages. MiTFL1-1 and MiTFL1-2 were most highly expressed at the flowering induction period (early stage), followed by the floral differentiation period (late stage), whereas the expression level of MiTFL1-2 was significantly higher than that of MiTFL1-1. MiTFL1-1 and MiTFL1-2 were expressed at fairly low levels at other flowering development stages. The expression levels of the MiTFL1-3 and MiTFL1-4 genes were highest at the floral differentiation period (late stage), followed by the flowering induction period (early stage), and were lower at other stages. Among the tested genes, MiTFL1-2 presented the highest expression level at the flowering induction period, and MiTFL1-3 exhibited the highest expression level at the floral differentiation, inflorescence elongation and flowering periods.
Subcellular localization of MiTFL1s
To examine the subcellular localization of MiTFL1s, 35S::GFP-MiTFL1-1, 35S:: GFP-MiTFL1-2, 35S::GFP-MiTFL1-3, 35S::GFP-MiTFL1-4 and 35S::GFP-P1300 were separately transformed into onion epidermal cells, and the results are shown in Fig. 4. The fluorescent signal of the empty vector 35S::GFP-P1300 was observed in all the cells. The 35S::GFP-MiTFL1-1, 35S::GFP-MiTFL1-2, 35S::GFP-MiTFL1-3, 35S::GFP-MiTFL1-4 fusion proteins were visible only in the nucleus, which was stained with DAPI.
Phenotypic analysis of MiTFL1 overexpression in Arabidopsis thaliana
MiTFL1s delayed flowering in Arabidopsis thaliana
To explore the function of the MiTFL1-1, MiTFL1-2, MiTFL1-3 and MiTFL1-4 genes in the flowering process of mango, individual overexpression vectors of pBI121-MiTFL1s were constructed and transferred separately into WT Arabidopsis thaliana. Phenotypic observations of T3 generation homozygous plants were conducted, and WT and pBI121 empty vector-expressing Arabidopsis served as controls.
Four independent lines of MiTFL1-1-overexpressing (OE-1#13, OE-1#22, OE-1#25 and OE-1#29) and three independent lines of MiTFL1-2-overexpressing (OE-2#24, OE-2#45 and OE-2#55) were selected for functional analysis. Semiquantitative RT-PCR analysis showed that MiTFL1-1 and MiTFL1-2 can be expressed normally in the MiTFL1-overexpressing transgenic plants but not in the empty vector-expressing transgenic or WT plants (Figs. 5A1 and 5B1). All independent lines of MiTFL1-1 and MiTFL1-2 showed delayed bolting and flowering: these processes occurred at 28.7–32.5 and 33.3–42.2 days, respectively, in these lines and at 24.9–25.3 and 28.5–28.8 days, respectively, in the control plants (Figs. 5A and 5B, Table 1). All transformant lines with MiTFL1-1 and MiTFL1-2 showed normal bolting similar to that observed in the WT plants. Additionally, compared the WT plants, the heights of the MiTFL1-1 and MiTFL1-2 plants were significantly increased, but the rosette leaves were not significantly affected (Table 1).
Table 1
Flowering phenotype analysis of WT, pBI121, MiTFL1-1-overexpressing (OE-1) and MiTFL1-2-overexpressing (OE-2) transgenic plants
ID | Days to bolting (d) | Days to flowering (d) | No. of rosette leaves | Plant height (cm) |
WT | 25.3 ± 0.1 | 28.8 ± 0.2 | 8.1 ± 0.2 | 24.1 ± 0.8 |
pBI121 | 24.9 ± 0.2 | 28.5 ± 0.3 | 7.9 ± 0.2 | 24.7 ± 0.4 |
OE-1#13 | 32.3 ± 0.5* | 42.2 ± 1.1* | 8.7 ± 0.2 | 35.7 ± 1.2* |
OE-1#22 | 28.8 ± 0.4* | 33.3 ± 0.4* | 8.1 ± 0.3 | 31.9 ± 1* |
OE-1#25 | 29.1 ± 0.6* | 33.9 ± 0.5* | 8.4 ± 0.3 | 30.1 ± 0.5* |
OE-1#29 | 28.7 ± 0.4* | 33.4 ± 0.6* | 8.9 ± 0.3 | 29.6 ± 0.6* |
OE-2#24 | 31.6 ± 0.5* | 37.2 ± 0.7* | 8.6 ± 0.3 | 36.1 ± 1.2* |
OE-2#45 | 32.5 ± 0.3* | 38.8 ± 1.2* | 8.8 ± 0.3 | 38.2 ± 1.5* |
OE-2#55 | 29.7 ± 0.3* | 34.6 ± 0.5* | 8.5 ± 0.3 | 31.8 ± 1.0* |
NOTE: The analysis was performed using four MiTFL1-1-overexpressing and three MiTFL1-2-overexpressing independent transgenic lines. The bolting time and rosette leaves were measured when the bolting height was 0.5-1 cm. The flowering time was considered the time when the first flowers opened. The plant height was measured 15 days after flowering. The error bars represent the ± SDs. The asterisks indicate significant differences (Student’s t-test: *P < 0.05). |
Three independent lines of MiTFL1-3-overexpressing (OE-3#19, OE-3#23 and OE-3#42) and MiTFL1-4-overexpressing (OE-4#24, OE-4#45 and OE-4#55) were selected for functional analysis. A semiquantitative RT-PCR analysis demonstrated that MiTFL1-3 and MiTFL1-4 were abundantly expressed in the transgenic lines but absent in WT and pBI121 transgenic Arabidopsis plants (Figs. 6A1 and 6B1). The MiTFL1-3-overexpressing and MiTFL1-4-overexpressing transgenic plants showed normal bolting, but their bolting time was significantly delayed compared with those of the WT and pBI121 transgenic lines under long-day (LD) conditions (Figs. 6A and 6B, Table 2). The inhibitory effect of MiTFL1-4 on flowering was lower than those of the other three MiTFL1 genes. The plant heights of some MiTFL1-3 and MiTFL1-4 transgenic lines showed significant differences, but the heights of some of the plants did not significantly differ from those of the control lines. The rosette leaves were not significantly affected in any of the plants (Table 2).
Table 2
Flowering phenotype analysis of WT, pBI121, MiTFL1-3-overexpressing (OE-3) and MiTFL1-4-overexpressing (OE-4) plants
ID | Days to bolting (d) | Days to flowering (d) | No. of rosette leaves | Plant height (cm) |
WT | 24.8 ± 0.3 | 28.1 ± 0.3 | 8.1 ± 0.2 | 24 ± 2.9 |
pBI121 | 24.7 ± 0.3 | 27.4 ± 0.2 | 8.2 ± 0.4 | 25.3 ± 0.8 |
OE-3#19 | 37.0 ± 0.5* | 47.3 ± 0.6* | 8.5 ± 0.2 | 38.9 ± 1.1* |
OE-3#23 | 30.9 ± 0.6* | 38.9 ± 1.9* | 8.4 ± 0.3 | 36.1 ± 1.2* |
OE-3#42 | 30.2 ± 0.6* | 34.7 ± 0.7* | 8.3 ± 0.2 | 26.3 ± 2.0 |
OE-4#12 | 28.0 ± 0.7* | 31.8 ± 0.6* | 7.6 ± 0.3 | 26.9 ± 0.7 |
OE-4#16 | 26.8 ± 0.3* | 30.7 ± 0.3* | 8.2 ± 0.2 | 26.5 ± 0.3 |
OE-4#24 | 29.4 ± 0.8* | 32.9 ± 0.8* | 7.7 ± 0.3 | 31.2 ± 1.0* |
NOTE: The analysis was performed using three MiTFL1-3-overexpressing and three MiTFL1-4-overexpressing independent transgenic lines. The bolting time and rosette leaves were measured when the bolting height was 0.5-1cm. The flowering time was considered the time when the first flowers opened. The plant height was measured 15 days after flowering. The error bars represent the ± SDs. The asterisks indicate significant differences (Student’s t-test: *P < 0.05). |
MiTFL1-1 and MiTFL1-3 affect the phenotype of Arabidopsis
The MiTFL1-1-overexpressing (Fig. 7B) and MiTFL1-3-overexpressing transgenic lines (Fig. 7C) exhibited similar abnormal phenotypes compared to the WT lines (Fig. 7A). In the transgenic plants, some carpels developed into new inflorescences (Figs. 7B-a and 7C-a), and some flower structures lacked petals (Figs. 7B-b and 7C-b), which was different from the results obtained for the WT plants (Figs. 7A-a and 7A-b). Two types of silique variations were found in the transgenic plants compared with the WT plants (Figs. 7A-b, 7A-c): in some siliques, the fruit stalk continued to lengthen from the flower during formation (Figs. 7B-c and 7C-c), and in other siliques exhibited curved growth and shorter siliques (Figs. 7B-c and 7C-c) compared with those of the WT plants. In addition, the inflorescences of the transgenic plants were also significantly different from those of the WT plants due to the variations in flower morphology (Figs. 7A-d, 7B-e and 7C-e). The results also showed whorled leaves growing on the lateral branches of transgenic Arabidopsis thaliana but not in the control plants (Figs. 7A-e, 7B-f and 7C-f).
Expression patterns of endogenous genes in transgenic Arabidopsis expressing MiTFL1s
To determine whether MiTFL1 gene overexpression in transgenic Arabidopsis changed the expression of some flowering-related genes, such as AtFT, AtFD, and AtAP1 homologue genes in Arabidopsis, the aboveground portion of T3 generation homozygous transgenic Arabidopsis thaliana was collected 30 days after planting and subjected to qRT-PCR analysis (Fig. 8). AtACTIN2 was used as the internal reference gene. A similar expression pattern was found for the AtFT, AtFD, and AtAP1 transcripts in Arabidopsis after the overexpression of each of the four MiTFL1 genes (Figs. 8A-8D). The expression levels of the AtFT and AtAP1 genes were significantly lower in all MiTFL1-overexpressing transgenic lines than in the WT plants. However, the AtFD gene was significantly increased in many transgenic lines with the exception of MiTFL1-2-overexpressing line OE-2#45.
Proteins that interact with MiTFL1s
The yeast two-hybrid system was used to screen the proteins interacting with MiTFL1s and verify their interactions. The bait vector pGBKT7-MiTFL1 was constructed by double enzyme digestion, and no autoactivation or toxicity was detected (shown in Fig. S2). Yeast cells with bait plasmids were combined with the cDNA homogenization library of ‘SiJiMi’ to screen for the positive clones. Because the yeast two-hybrid system has a high false positive rate, three proteins selected from the library were selected for further point-to-point verification on DDO/X and QDO/X/A media. The three proteins were basic helix-loop-helix protein 13 (bHLH13), bHLH162 and 14-3-3D, as shown in Fig. 9. The cells with the candidate protein bHLH13 in the pGADT7 recombinant vector turned blue and exhibited normal growth on QDO/X/A solid medium, which indicated that the protein interacts with MiTFL1-2 and MiTFL1-3 proteins. bHLH162 can interact with MiTFL1-1, MiTFL1-2 and MiTFL1-4, whereas 14-3-3D only interacts with MiTFL1-1 and MiTFL1-2.