Explants were acquired from in vitro germination of gabiroba seeds inoculated in standard culture medium consisting of Woody Plant Medium salts (WPM) (Lloyd & McCown, 1980) with the addition of 30 g L^-1 of sucrose and 7 g L^-1 of agar, at pH 5.8. Seedlings were kept in a growth room at a temperature of 25 ± 2°C with a 16 hours photoperiod and photon irradiance of 43 µmol m^-2.-1, and were sub cultured every 40 days.

Seedlings at 40-days post germination were used for callus induction. The explants used were nodal segments of approximately 1 cm in length, which were inoculated in an induction medium consisting of WPM salts with the addition of 30 g L^-1 of sucrose, 7 g L^-1 of agar, and 0.9 mM of polyvinylpyrrolidone (PVP). The pH was adjusted to 5.8 prior to autoclaving at 120ºC for 20 minutes. Different concentrations of 2,4-dichlorophenoxyacetic acid (2,4-D: 4.52, 9.04 and 18.08 µM), indole-3-acetic acid (IAA: 5.7, 11.4 and 22.8 µM), indole-3-butyric acid (IBA: 4.92, 9.84 and 19.68 µM), naphthaleneacetic acid (NAA: 5.37, 10.74 and 21.48 µM), or 4-amino-3,5,6-trichloropicolinic acid (picloram: 4.14, 8.28 and 16.56 µM) were added to the culture medium, while a control was established without the addition of growth regulators. The treatments were maintained in the absence of light.

Evaluations were made every 30 days until reaching 90 days of inoculation (subcultures were not performed), at which time the cell mass percentage was evaluated according to the following scores: 1 = absence of callus, 2 = 25% of explant covered by callus, 3 = 50% of explant covered by callus, 4 = 75% of explant covered by callus, and 5 = 100% of explant covered by callus.

The results were submitted to analysis of variance by the Kruskal-Wallis non-parametric test, which was also used to compare the means at 5% probability (Kruskal & Wallis, 1952).

A growth curve was established based on the treatment with the highest average callus percentage in the shortest amount of time. Explants of plants sub cultured every 40 days under the same initial conditions were used by the fourth subculture for callus induction in the growth curve. Calli were obtained through treatment with induction medium supplemented with 4.14 µM of picloram. After inoculation, the explants were kept in a growth room at a temperature of 25 ± 2°C and without light.

Twenty-five replications were used to determine the fresh weight of calli, which were weighed every seven days until stabilization of the growth curve. Fresh matter gain as a function of each phase was defined by the ratio between the maximum weight reached by the callus and the weight of the callus in each of the phases.

The results were submitted to ANOVA, using the F test, and to regression analysis, also using the F test, both at 5% probability.

Cytochemical analysis and SEM were performed with calli that were induced parallel to those used for determining the growth curve so as not to alter the number of explants used for determining weight. The calli were collected at 14-day intervals until stabilization of the curve.

The samples for cytochemical analysis were fixed according to Pádua, Paiva, Labory, Alves and Stein (2013), sectioned at a thickness of 5 µm using a rotary microtome, and stained with 0.05% toluidine blue and 2.5% Lugol’s solution. Slides were viewed using a Zeiss photonic microscope connected to a digital camera. For SEM, samples were fixed according to Pádua et al. (2013), adapted the Bozzola and Russel (1998) protocol, and analyzed with a Zeiss EM 109 transmission electron microscope.

The presence of calli was observed in all treatments containing 2,4-D, IAA, IBA, NAA and picloram (Figure 1). The use of 4.14 µM picloran showed the best results for callus induction from nodal segments since it was possible to obtain calluses on 100% of the explant at 30 days. Doses of 8.28 and 16.56 μM picloram had similar results, but were only able to obtain 100% callus formation after 60 days, at the point oxidation began to be observed.

Figure 1
Evaluation of callus induction on nodal segments of gabiroba in WPM culture medium with different concentrations of 2,4-D, IAA, IBA, NAA, and Picloram at 30, 60, and 90 days. The percentage scores attributed to the calli were compared by the Kruskal-Wallis non-parametric test.

The growth curve of calli on nodal segments of gabiroba exhibited a sigmoid pattern, with a gain of fresh matter during the period of cultivation. Five phases of growth were observed during the analyzed period (56 days from explant inoculation) due to the high callus growth rate (Figure 2).

The lag phase occurred up the 7^th day of culture, resulting in a weight gain of 9% of the total. The exponential growth phase occurred between the 7^th and 28th day of cultivation, and resulted in a weight gain of 66%. The linear growth phase, between the 28^th and 35^th day, resulted in a 17% gain of fresh mass. The deceleration period began on the 35^th day and lasted until the 42^nd day, resulting in 8% growth. Callus growth reached the stationary phase on the 42^nd day post-inoculation, after which a reduction of 5% of the fresh weight was observed, indicating the beginning of the decrease phase.

Figure 2.
Growth curve of calli on nodal segments of gabiroba inoculated in WPM culture medium supplemented with 4.14 µM of picloram. I – lag phase (0-7 days); II – exponential phase (7-28 days); III – linear growth phase (28-35 days); IV – deceleration phase (35-42 days); V – stationary phase (after 42 days).

Cytochemical and SEM analyses revealed that calli induced with 4.14 μM picloram in nodal segments of gabirobeira possessed isodiametric cells at 14 and 28 days after inoculation (Figure 3A and B;Figure 4A and C), which corresponds to the exponential growth phase according to the growth curve.

Figure 3.
Scanning Electromicrograph of calli on nodal segments of gabiroba inoculated in WPM culture medium supplemented with 4.14 µM of picloram. Calli (A) at 14 days showing isodiametric cells (arrow); (B) at 28 days with isodiametric (arrow) and elongated (*) cells; (C) at 42 days with isodiametric (arrow) and elongated (*) cells; and (D) at 56 days with elongated cells (*).

At 14 days, during the exponential growth phase, the ratio between the number of elongated cells and the number of isodiametric cells was low. However, at 28 days, still in the exponential stage, and at 42 and 56 days, in the deceleration and decrease phases, respectively, the ratio between the number of elongated and isodiametric cells increased, thus characterizing a negative relationship between increasing numbers of elongated cells and decreasing numbers of isodiametric cells over time (Figure 3B, C and D).

Figure 4.
Cytochemical analysis of calli in WPM culture medium supplemented with 4.14 µM of picloram on nodal segments of gabiroba – A, C, E, G; calli stained with toluidine blue - B, D, F, H; calli stained with Lugol’s solution – A and B. Calli at 14 days showing: A) cells with prominent nuclei (N) and intense cell division, and B) presence of starch grains (arrow). C and D) Calli at 28 days showing: C) cells in division and the formation of pro-embryogenic mass with cells with prominent nuclei (N), and D) presence of starch grains (arrow). E and F) Calli at 42 days showing: E) elongated cells and F) cells without the presence of starch grains (arrow). G and H) Calli at 56 days showing: G) elongated and dispersed cells and H) cells without the presence of starch grains. Bar: 100 µm.

Results of the cytochemical analysis of calli were consistent with those of the SEM analysis. At 14^th and 28^th days, isodiametric cells could be seen undergoing division and positioned juxtaposed, while on the 28^th day cellular agglomerates were observed forming pro-embryogenic masses (PEM) that were intensely stained with toluidine blue. However, at 42^th and 56^th days after inoculation, elongated and dispersed cells were observed (Figure 4E and G).

Staining with Lugol’s solution revealed the presence of starch grains in the callus cells at 14^th and 28^th days (Figure 4B and D), but none at 42^th and 56^th days (Figure 4F and H), indicating greater embryogenic potential at beginning of cultivation (at 14^th and 28^th days).

Tissue culture techniques, including callus induction, are important viable alternatives for the vegetative propagation of native Cerrado species (Pinhal et al., 2011). In the present study, nodal segments of gabiroba were found to be efficient sources of explants for the induction of calli with the presence of pro-embryogenic masses.

According to Titon, Xavier, Otoni, and Motoike (2007), picloram is used for the induction of calli or the maintenance of cultures in suspensions, as well as for inducing the formation of embryogenic calli. Most studies have indicated that calli are achieved primarily through the use of 2,4-D (Bajpai, Kalim, Chandra, & Kamle, 2016; Fehér, 2015; Rocha et al., 2016; Winkelmann et al., 2015), including in the induction of calogenesis in Campomanesia rufa leaf explants (Sant’Ana, Paiva, Reis, Silva, & Silva, 2018), which was not found in the present study. Satisfactory results have been achieved using picloram for embryogenic callus induction in Eucalyptus grandis x E. urophylla (Moura et al., 2017) and Acca sellowiana (Cangahuala-Inocente, Caprestano, & Pierre, 2007), both from the same family, Myrtaceae, as gabiroba.

It is essential to know the behavior and development of the cells that are in intense multiplication after the induction of callus formation in order to maintain embryogenic characteristics. The callus growth curve of the present study presented five developmental stages of cell growth: lag, exponential growth, linear, deceleration and stationary (Santos, Ferreira, & Sarubo, 2010). The analysis of the growth curve of the present study revealed rapid growth and development of calli in comparison with that of other species (Nogueira et al., 2008; Santos, Paiva, Paiva, & Paiva, 2008; Santos et al., 2010; Santos & Souza, 2016; Stein et al., 2010; Vasconcelos et al., 2012), demonstrating that growth pattern varies among explants of different species. In the present case, the rapid growth and development of calli exhibiting embryogenic characteristics is extremely advantageous for producing plants on a large scale in a short period of time.

The exponential growth phase, characterized by maximum cell division (Stein et al., 2010), and the linear growth phase, with a decrease in cell division and an increase in cell size (Santos et al., 2008), exhibit varying temporal patterns among species. In leaf explants of Coffea canephora var. Conilon, growth occurred between the 16^th and the 34^th day after inoculation (Santos et al., 2010). For calli obtained from leaf explant of inga (Inga vera Willd. Subsp. Affinis (DC.) T.D. Penn), this phase started on the 40th day and continued until the 50^th day (Stein et al., 2010). The linear growth phase of calli of leaf explants of murici-pequeno (Byrsonima intermedia A. Juss.) occurred between the 40^th and 60^th days of culture (Nogueira et al., 2008). In calli obtained from both leaf and nodal segments of Coffea canephora L. cv. Apoatã, linear growth was obtained from the 63^rd to 70^th day (Santos et al., 2008).

In the present study, the deceleration period was observed at 35 days, and was characterized mainly by a reduction in nutrient content in the culture medium, drying of the agar, and accumulation of toxic substances in the growth medium (Smith, 2012). Thus, it is appropriate for calli to be transferred from the culture medium prior to the 28^th day in order to maintain exponential growth.

According to the results presented in the growth curve (Figure 2), the cytochemical (Figure 4) and SEM (Figure 3) analyses identified isodiametric cells at 14 and 28 days after inoculation, which corresponded to the exponential growth phase, as determined by the growth curve. Isodiametric cell shape is characteristic of meristematic cells, calli composed of cells of its size, with small dimensions and a dense cytoplasm with high embryogenic potential (Peña-Ramírez et al., 2011; Vejsadová, Matiska, Obert, Ürgeová, & PreŤová, 2016). The presence of elongated cells, seen on the 28^th day, indicates vacuolization, which, along with the rupture of cell membranes, is one of the first signs of cell death (Hatsugai, Yamada, Goto-Yamada, Hara-Nishimura, & Beers, 2015).

Another indication of embryogenic competence is the presence of starch grains produced prior to embryo formation, which serves to initiate and sustain embryo development (Hazubska-Przybył, Kalemba, Ratajczak, & Bojarczuk, 2016). Analysis of starch content during development of somatic embryos of goiabeira-serrana (Acca sellowiana (O. Berg.) Burret) revealed an initial content six times greater than that at the end of embryogenic development (Cangahuala-Inocente, Steiner, Maldonado, & Guerra, 2009). In the present study, starch grains could be observed until the 28^th day of culture, after which cell masses lose their embryogenic characteristics (Figure 4F and H).

Embryogenic structures do not continue their development during the stationary phase because cells go into decline and lose their intracellular organization due to the process of senescence (Santos et al., 2013). Therefore, transfer to a new culture medium can maintain the continuity of development, resulting in a greater proliferation of cells with embryogenic potential.