A. brachiata cutting technique
For the cutting of rejuvenated mother plants, herbaceous branches of A. brachiata were collected from a clonal garden in the field with specimens approximately two years old, originating from seedlings produced by stem cuttings, frequently pruned. And for the cuttings from adult mother plants, we collected semi-woody branches of 20 adult mother plants of natural occurrence in the understory of a Mixed Ombrophylous Forest. For both experiments, plants were in an area belonging to Embrapa Florestas in Colombo, Paraná, Brazil, and the collection of plant material took place in the four seasons of 2019.
Stem cuttings measuring 8 ± 1 cm in length were made from the plant material collected and disinfected with 0.5% sodium hypochlorite. After disinfestation, the bases of the cuttings were treated with different concentrations of indole butyric acid (IBA) in a hydroalcoholic solution (50% v/v) for 10 seconds: 0, 2000, 4000, 6000, and 8000 mg L− 1. Planting was carried out in polypropylene tubes (53 cm³) filled with fine-grained vermiculite, and the cuttings were kept in a climate-controlled greenhouse with intermittent nebulization.
The experiment was established in a completely randomized design, in factorial scheme 5x 4, with five concentrations of IBA and four seasons of the year, with four replications of 20 cuttings per experimental unit. After 45 days of experiment installation, we evaluated rooting percentage, the number of roots per cuttings, length of the three largest roots per rooted cutting (cm), callogenesis percentage, leaf retention, shoots and mortality (cuttings with necrotic tissues).
Anatomical analyses
Cuttings’ bases of rejuvenated and adult mother plants were collected at the installation and during the evaluation of the cutting experiment in the summer. After fixation in FAA 70 (formaldehyde, acetic acid, and 70% ethanol) for 24 hours (Johansen 1940), the samples were stored in 70% ethanol.
The material was sectioned into 0.5 cm pieces and embedded in historesin (Leica®). The blocks were fixed in wooden tutors, and we obtained transverse cuts in a rotating microtome (Olympus® CUT 4055) with a thickness of 8 µm (Kraus and Arduin 1997). The sections were adhered to slides and stained with toluidine blue dye. Then, we analyzed and documented the slides using a Zeiss photomicroscope with a Sony® Cyber-shot P72 digital camera attached.
Biochemical analyses
Vegetative material and aqueous extract preparation
At the time of installation and evaluation (at 45 days) of each experiment in the four seasons of the year, 20 to 30 cuttings of rejuvenated and adult mother plants were randomly separated (composite sample), dried in microwaves for 5 minutes, crushed in an electric grinder and stored in a freezer (-20°C) until analyses.
To prepare the aqueous extracts, 20 mg of plant material was weighed in microtubes, and 2 mL of ethanol:water (1:1) was added. For extraction, the material was vortexed for 30 seconds, sustained in the Thermomixer® equipment for 1 hour at 60°C and 450 rpm, and next filtered. The extracts were prepared in triplicate for the determination of total phenolic compounds. Samples moisture was determined in triplicate by gravimetric method, drying in an oven at 105°C until constant mass. The results are expressed on a dry basis.
Total phenolic compounds and proteins
The content of total phenolic compounds was determined by Folin-Ciocalteu spectrophotometric method, as described by Singleton and Rossi (1965), with minor modifications. An aliquot of 0.15 mL of aqueous extract was added in a volumetric flask, followed by 0.5 mL of Folin-Ciocalteu reagent and 5 mL of distilled water. The solution was vortexed for 1 minute, and then 2 mL of 15% Na2CO3 aqueous solution was added. The solution was stirred for 30 seconds and adjusted to 10 mL with type 2 water. The solution was kept in dark for 2 hours, and the absorbance was determined in a spectrophotometer at 760 nm. The standard curve was obtained with gallic acid (3,4,5-trihydroxy benzoic acid) at concentrations from 0.25 to 10 mg L-1, and the results expressed in mg equivalent to gallic acid (mg EAG g-1 dry mass).
To analyze total proteins, 0.5 g of plant material was added into a digestion tube along with 0.5 g of catalyst (6.3 g of sodium selenite, 4.0 g of copper sulfate pentahydrate, and 48.5 g of anhydrous sodium sulfate). Then, 5.0 mL of concentrated sulfuric acid was added and kept in the digester block (Tecnal®) at an initial temperature of 100ºC, raising the temperature by 50ºC every 30 minutes until reaching 350ºC, where it remained for a minimum of 3 hours. After complete digestion and cooling of the samples, 15 mL of distilled water was added. Sample distillation was performed in a nitrogen distiller (Tecnal® TE036-1) and titration with hydrochloric acid (0.1M). Total protein concentration was determined according to the Kjeldahl method (AOAC, 1984) and expressed as mg g-1 fresh mass.
Metabolomic analysis
For the metabolomic analysis, we evaluated rejuvenated and adult cuttings at the experiment installation. The determination of polar metabolites was performed according to Fiehn et al. (2000), with minor modifications.
For extraction, we weighed 20 mg of sample ground in liquid nitrogen, added 300 µL of pre-cooled methanol at -20°C, homogenized it in Vortex®, and kept it in a 2mL Eppendorf® in the freezer. Then, 30 µL of nonadecadonic acid (2 mg mL− 1 stock in CHCl3 – used for qualitative internal control for the lipid phase) and 30 µL of 13C6-sorbitol (used for quantitative internal standardization for the polar phase) were added. The sample was stirred for 15 min at 70°C and then cooled to room temperature. Afterwards, 200 µL of CHCl3 was added, stirring for 5 min at 37°C. Next, 400 µL of water was added to the sample, homogenized in Vortex®, and centrifuged for 7 minutes at 12,000 rpm.
Two aliquots of the polar phase (top) were transferred to two 1.5 mL conical Eppendorff® tubes and dried in Speed Vac® for 16 hours without heating. Finally, a set of samples was derivatized in the sequence: methoxyamine hydrochloride in pyridine (50 µL, 20 mg mL− 1, 40° C, 90 min) and MSTFA (N-methyl-N-(trimethylsilyl)-trifluoroacetamide, 50 µL, 40°C for 90 min) and analyzed by gas chromatography coupled to mass spectrometry (GC-MS, Focus, PolarisQ, Thermo Fisher®). 1 µL solution was injected by 1:25 flow division, and compounds were separated using a DB-5ms chromatographic column (30 m x 0.25 mm x 0.25 µm, Agilent®). The analysis was performed with constant flow and vacuum compensation with injector temperature at 230°C, transfer line at 250°C, and helium carrier gas at 1.0 mL min− 1. The oven's internal temperature programming was 70°C with an isotherm of 1 min, heating at 320°C at a rate of 8°C min− 1, and a final isotherm of 18 min. The mass spectrometer operation was in positive mode with electron ionization at 70 V and ion source temperature at 200°C.
In the AMDIS software (Automated Mass Spectral Deconvolution, NIST), the compounds were identified using mass spectra, a comparison of the retention index (IR), according to the Golm Metabolome Database platform (Hummel et al. 2010). Metabolite intensities (i.e., peak heights) were normalized by sample weight and internal standard (13C6-Sorbitol), using fragments (m/z) characteristic of each compound.
Statistical analysis
For cutting technique, variances homogeneity were tested for by Bartlett's test (p < 0.05). Data were submitted to square root transformation for all variables that did not meet homogeneity assumptions, followed by an analysis of variance (p < 0.01 and p < 0.05); the variables that showed significant differences by the F test had their means compared by the Tukey test. We performed all analyses in the statistical software Assistat version 7.7.
For metabolites, data were analyzed using factor analysis, multivariate analysis of variance (MANOVA), and discriminant analysis (Marriott 1974; Kachigan 1991; Johnson and Wichern 1998; Anderson 2003; Hair Junior et al. 2005; Ferreira 2018). In the factor analysis, we considered that the factors should have an eigenvalue greater than 1 to be included. The extraction method used was principal components, with varimax rotation and variables standardization (correlation matrix).
MANOVA was carried out for the groups of variables obtained in the extracted factors, aiming to compare the propagules obtained from the different mother plants (rejuvenated and adult) using the average of the four seasons of the year. Pillai's Trace, Roy's Largest Root, Wilks' Lambda, and Hotteling's Trace tests were used to assess the significance of treatments. In case of significant differences between the types of material, the discriminant analysis was performed using the standardized variables and the mother plants of structures, canonical correlation, F test for each original variable, and the means of the treatments were evaluated.