Quantification and quality of RNA

In general, the average amount of RNA in mature leaves with PBZ (2 500 mg L^-1, 1X at 30 dpp) was 44.7 ng µL^-1 reaching a 260/280 ratio of 1.96 on average. The average amount of cDNA was 2103.6 ng µL^-1 and a 260/280 ratio of 1.87 (Table 1). In leaves with P-Ca (500 mg L^-1, 3X at 30 + 45 + 60 ddp) the amount of RNA was 42.3 ng µ L^-1, with a ratio of 1.95 (260/280), while that of cDNA was 2025 ng µL^-1 and a ratio of1.87. WithAG3, the amount of RNA was 36.1 ng µ L^-1 and of 2006.7 the cDNA content with a 260/280 ratio of 1.92 and 1.85, respectively. In leaves without regulator, we found an amount of 33.6 ng µ L^-1 RNA and 2006.3 ng µL^-1 cDNA, the 260/280 ratio was 1.89 and 1.85, respectively.

Irrespective of the experiment, the extracted amount of RNA in ‘Ataulfo’ mango leaves was sufficient to achieve an outstanding amount of cDNA once converted by retrotranscriptase, an enzyme that exists in nature as part of the virus replication mechanism, and uses a tRNA as oligonucleotide or primer (^{Falcon and Valera, 2007}).

The extracted RNA was of excellent quality, according to the ratio 260/280 obtained in the spectrophotometer. The ratio between the readings obtained at 260 nm and 280 nm (OD260/ OD280) provides an estimate of the purity of the nucleic acids. Nucleic acids have a maximum absorption at a wavelength of 260 nm. At this length therefore, the absorption is indicative of the concentration, whereas in the 280 nm length, the absorption shows proteins and phenol compounds. Nucleic acid extractions of good quality and purity should have a 260/280 ratio with values between 1.8-2, lesser amounts mean that there are presence of proteins, phenols or other contaminants that could interfere with the results of the tests to be performed (^{Wilfinger et al., 1997}; ^{Anonimo, 2007}). Likewise the amount of cDNA obtained was of good quality without contamination of compounds as mentioned above. This allowed the qPCR to be efficiently performed to quantify the expression of the MiFT gene in both experiments.

Temperature and precipitation

In 2013-2014, the minimum temperature in October and November remained between 23 and 24 °C, while the maximum temperature was between 32 and 35 °C, in December the minimum temperature dropped to 17 °C and in January and February remained between 15 and 18 °C. Rainfall was observed during October and November (122 and 161 mm, respectively), in december rainfall was 105 mm and there was little rainfall in January and February (Figure 1).

Figure 1 Maximum temperature, minimum and precipitation presented during the period of induction, initiation and floral sprouting in ‘Ataulfo’mango. 

Experiment 1. Expression of MiFT by period effect and orientation in the tree.The analysis of variance for the QR variable, relative amount of expression of the gene of interest MiFT, showed high significance for the intercept and months (Table 2). While the multiple comparison with Kruskal-Wallis of the MiFT gene expression according to the orientation, showed a relation for the west and east side (Table 3).

In Figure 2, the results obtained in the expression of the gene of interest MiFT in relation to the endogenous MiEF gene are shown. Regarding the time of sampling, the relative expression of MiFT was observed in all months sampled. The expression varied from 0.30 to 3.4 with respect to MiEF. The month of greatest expression was december followed by january and february. Regarding orientation, the relative expression of MiFT was on average 0.81, 1.9 and 0.54 on the north, east and west sides, respectively (Figure 2).

Figure 2 Relative expression of MiFT relative to MiEF in ‘Ataulfo’ mango leaves treated with PBZ. N, O and P, correspond to north, east and west of the tree. 

Considering the interaction it was found that in September the expression was MiFT was 0.25 and 0.53 in north and east side, respectively, in November the expression was 1.51, 0.8 and 0.4 in north, east and west side, in December MiFT was expressed only on the east side with 3.4 and west with 0.75, while in january the expression was 2.5 and 0.15 north and west sides and 3.1 on east side. In February, the expression was on the north, east and west sides of 0.25, 1.8 and 0.5, respectively.

In leaves treated with PBZ, the expression of MiFT (gene of interest) was observed in all months of study, but only in leaves of sunny buds during the morning was expressed in greater quantity with respect to leaves of sunny shoots in the afternoon and completely shaded leaves. In leaves of shaded shoots (north side), the expression greater expression was in the months of November and January; but it was minimal compared to the east side. This indicates that the MiFT gene is probably regulated by light when expressed in greater quantity on the soil side and a minimum expression on the north side where the trees are always shaded by the union of tree tops (hedge-like formation).

It can be observed that in the month of September (before the application of PBZ), the expression was null in leaves with orientation west, but there was a minimal expression of the gene in sunny leaves of the east side; in the following months MiFT presented a minimal expression in leaves shaded probably by effect of the regulator PBZ, which resulted in that a small percentage of shaded shoots flourished.

In ‘Fuji’ apple tree (36° north latitude, 139° west longitude and 747 meter above sea level) found no difference in flowering locus T (MdFT1) gene expression (MiFT homolog) in apical buds of shaded and sunny buds but the expression of MiTFL1 (floral repressor) was higher in shaded shoots (^{Kittikorn et al., 2011}), results that do not coincide with our study, despite results in apple, were obtained in different species and tissue and shading was only 50% and in the present study conducted under field conditions very different from those of apple tree (21°north latitude 105° west longitude, 150 m), at least on the north side, the shade was about 90%.

The highest expression was found in December and January on a sunny side, which coincided with the presence of colder temperatures and probably with the time when induction and floral initiation occurred in mango, as reported in mandarin ‘Satsuma’ by Nakagawa et al. (2007) where the level of the orthologous gene flowering locus T (CiFT) increased with cold temperatures during floral induction time and in sweet orange tree LFY and AP1 increased during and after floral induction (^{Pillitteri et al., 2004}). On the other hand, ^{Nakagawa et al. (2012)} identified a gene as the flowering locus T (MiFT) on leaves of ‘Irwin’ mango trees that developed under inductive temperatures at flowering (15°C).

In our study we observed a greater expression of the gene mentioned in leaves of ‘Ataulfo’ in the months of December and January, months in which the temperatures were lower than in September, October or November (Figure 1).

In sunny leaves and with application of PBZ, it was found that MiFT was expressed in the majority of months evaluated; although the expression was highly increased in the month of December, January and February. The above indicates that the regulator favors gene expression probably through its effect on the inhibition of gibberellin synthesis; in addition to this, the months of december to february were the freshest compared to September, October and November. Similar results are reported in fresh orange leaf ‘Salustiana’ where PBZ stimulated the flowering of homologous genes to flowering locus T, CiFT (^{Muñoz et al., 2012}).

Type of growth emerged during the flowering season. During the flowering season, on the east side (sunny shoots), 66% of shoots had flowering, 12% had only vegetative shoots and 25% were inactive. On the west side (shaded in the morning and sunny in the afternoon), 40% of the shoots had flowering, 50% only vegetative growth and 12% were inactive. On the north side (shaded side) most shoots were inactive (76%), although 12% presented flowering (Figure 3).

Figure 3 Percentage and type of growth emerged during the flowering period in ‘Ataulfo’ mango shoots treated with PBZ and in different orientation in the tree. BF= floral bud; BV= vegetative shoot; BI= inactive bud. East (O), west (P) and north (N) of the tree, 2014. The bars at each point represent the average of 160 outbreaks per treatment ± standard error. 

In the flowering period, the percentage of flower buds was higher on the sunny side when PBZ was applied, which coincides with the greater expression of MiFT. In addition, it is likely that PBZ by inhibiting the synthesis of gibberellins favored the expression of the gene responsible for floral induction as reported by ^{Nakagawa et al. (2012)}. On the west side (sunny in the afternoon) and north (shade of approximately 90%) the percentage of buds with flowers was minimal, but flowers were present on both sides when PBZ was applied. The above indicates that the regulator induced, although in smaller quantity, the expression of the MiFT gene provoking the induction of flowering in this type of shoots and therefore the presence of flowers; however, the light is strongly regulating its expression.

On both sunny and shaded sides a large number of shoots remained inactive, which could be due to the fact that the environmental conditions during November, when induction and floral initiation in mango usually occur (^{Pérez et al., 2009}), were not adequate for flowering since there were temperatures above 20 °C and presence of rain; however, with the application of PBZ, an adequate percentage of floral shoots were obtained on sunny side, which coincided also with the greater expression of MiFT in December, when thentemperature was lower to 20 °C.

Experiment 2. Expression of MiFT by effect of growth regulators.The relative expression in leaves with P-Ca was 0.02% in September, 1.1% in November and increased in December to 4.7% and decreased in January with 1.1%, in October and February the expression was 0 (Figure 4). In leaves applied with PBZ the relative expression of MiFT in relation to MiEF was 0.7% in September, in November the expression was 2.8% increasing in december to 30.3 and 22.2% in January, in February it was 6.5%. In control leaves (without regulator) the expression was 0.06% in September, zero in October, November and December, 0.87% in January and zero in February; while in leaves with AG3, the expression was 0 in all months evaluated except for february which had 0.02% expression.

Figure 4 Relative expression of MiFT over MiEF in ‘Ataulfo’ mango leaves with and without growth regulator. The bars at each point represent the average of three replicates per treatment ± standard error. 

Results similar to those obtained in PBZ were presented in leaves treated with P-Ca, another inhibitor of gibberellin synthesis, although the expression was lower compared to PBZ. In leaves applied with AG3, the expression was practically null, indicating that this regulator inhibited the expression of MiFT. The results agree with those obtained in sweet orange ‘Salustiana’ (^{Muñoz et al., 2012}) and ‘Irwin’ mango (^{Nakagawa et al., 2012}), where it is reported that gibberellins repress the expression of genes related to flowering as CiFT and MiFT, respectively.

Type of emerged growth. Regarding the type of growth emerged during the flowering period, a percentage of flower buds of 36, 51, 0 and 12% were found as a result of P-Ca, PBZ, AG3 and without control (control), respectively. Trees treated with AG3 had the highest percentage of vegetative shoots. The remainder of the outbreaks of each treatment

remained inactive with 57, 39, 74 and 73% in P-Ca, PBA, AG3 and control, respectively (Figure 5).

Figure 5 Percentage and type of growth emerged during the flowering period in bud of ‘Ataulfo’ mango by regulators effect. 2014. The bars at each point represent the average of 160 outbreaks per treatment ± standard error. 

During the flowering period, the percentage of flower buds was higher when PBZ was applied, followed by those treated with P-Ca, although the percentage was lower compared to PBZ. Which coincides with the expression of MiFT in the two regulators. Probably by inhibiting the synthesis of gibberellins, PBZ and P-Ca favored the expression of the gene responsible for floral induction as reported by ^{Nakagawa et al. (2012)}, these authors also demonstrated that MiFT is regulated by gibberellins by finding that the content of active gibberellins such as AG1 and AG4 was increased by downstream regulation of the MiGA_3-ox and MiGA _20-ox, genes, inhibiting the expression of MiFT in leaves and therefore flowering on ‘Irwin’ mango trees with high yield.

On the other hand, in all treatments a large number of outbreaks remained inactive, although the percentage was 23 and 47% higher in outbreaks treated with AG3 and control with regard to P-Ca and PBZ. ^{Bangerth (2009)} mentions that in fruit trees the perenniality is given by the meristemos that remain vegetative, contrary to the annual plants where all meristems flourish at the same time with the inevitable death of the plants; however, in this study the majority remained inactive, which could be due to the fact that the environmental conditions during November, when induction and floral initiation in mango (^{Pérez et al., 2009}), were normally not adequate for flowering (temperatures above 20 °C and presence of rain).

However, despite these environmental conditions, with the application of inhibitors of gibberellins, but especially with PBZ, an adequate percentage of floral shoots was achieved. In outbreaks with AG 3, there was no flowering, which coincided with the lack of expression of MiFT, which corroborates the antagonistic effect of this regulator on the flowering of mango and other fruit trees as mentioned by ^{Nakagawa et al. (2012)} in mango ‘Irwin’ and ^{Muñoz et al. (2012)} in sweet orange ‘Salustiana’.