Impact of Drought and Flooding on Alkaloid Production in Annona crassiﬂora Mart

: The Brazilian Cerrado is the second largest Brazilian biome. In recent decades, a reduction in rainfall has indicated an extension of the dry season. Among the many native species of the Cerrado of the Annonaceae family and used in folk medicine, Annona crassiﬂora Mart. has fruits of high nutritional value and its by-products are sources of bioactive compounds, such as alkaloids. The aim of the study was to investigate how water stress impacts the production of alkaloids. The study was carried out in a nursery, and the knowledge was ﬂood, ﬁeld capacity and drought. Gas exchange, chlorophyll a ﬂuorescence, antioxidant enzymes, total soluble sugars, starch, reducing sugars, sucrose, total alkaloids and liriodenine were analyzed. We observed that plants subjected to drought had an increase in the production of total alkaloids and liriodenine, without a reduction in photosynthetic metabolism. Plants kept under drought and ﬂood conditions dissipated higher peroxidase activity, while catalase was higher in ﬂooded plants. Starch showed the highest concentration in ﬂooding plants without differing from drought plants; the lowest trehalose concentrations were found in both drought and ﬂooding plants. The drought stimulated the synthesis of total alkaloids and liriodenine without reducing the primary metabolism, which suggests adaptation to Cerrado conditions. C.S.F.B.: writing—original supervision. G.F.: Conceptualization, curation, investigation, resources, writing—original writing—review supervision.


Introduction
The Brazilian Cerrado is the second largest Brazilian biome, considered one of the 25 global biodiversity hotspots, present in more than twelve states and occupying approximately 25% of the national territory, with native flora characterized by small and twisted trees [1].
This biome covers important aquifers and rivers [2] and is located in the central area of the country, being the origin of large hydrographic regions in Brazil and in the South American continent [3,4]. However, Lee et al. [5], Debortoli et al. [6] and Penereiro et al. [7] reported reduction of approximately 70 mm in annual precipitation in the Cerrado region between 1979 and 2006, indicating an extension of the dry season. Furthermore, although with great biological diversity, the biome is under continuous threat of extinction due to the expansion of agriculture and pastures, as occurs with other biomes [2,8].

Plant Material
Annona crassiflora Mart. seedlings were obtained in the municipality of Paraguaçu-Minas Gerais (April 2019) and transported to the nursery of the Department of Forest Science of the Faculty of Agronomic Sciences, Unesp-Botucatu (coordinates 22 • 51 latitude S and 48 • 26 longitude W), where they were submitted to a 6-month acclimation period. Transplantation to 5-L polyethylene pots was carried out in November 2019. To fill the pots, medium-texture Dystrophic Red Latosol was used [35,36], collected from the surface layer (0-20 cm in depth).
During the period in which the experiment was conducted, plants were submitted to humidity and temperature conditions shown in Figure 1.
From the water retention curve, the percentage of water needed for the soil to reach −0.01 MPa (Field capacity) (Control) and −1.5 MPa (Drought) was calculated, which corresponded to 16% and 8% of water, respectively. After placing pots under the established conditions, they were weighed daily and the evapotranspiration difference was replaced to maintain previous conditions (−0.01 MPa and −1.5 MPa). To maintain plants under flooding, pots were kept in flooded trays throughout the experiment.
The moisture values corresponding to the water retention tension obtained through the soil water retention curve are shown below (Table 1).

Gas Exchanges
Gas exchanges were monitored weekly in all treatments from 9:00 a.m. to 11:00 a.m., with the aid of a CO2 gas and infrared gas analyzer ("InfraRed Gas Analyzer-IRGA", model GSF 3000 Fl WALZ, Germany) with saturating light of 450 m −2 s −1 determined by means of a light curve. For monitoring, six replicates (1 plant per replicate) of each treatment were evaluated, taking measurements on the 2nd and 3rd fully expanded leaves.

Experimental Design
The experimental design was completely randomized, with three water stress levels (Flooding (>0.01 MPa); Field Capacity (−0.01 MPa) (Control) and Drought (−1.5 MPa)), with six replicates of two plants per plot. Plants remained in treatment (13-31 May 2020) until the permanent wilting point was reached. Two days after this stress condition was reached, the entire experiment was collected.
From the water retention curve, the percentage of water needed for the soil to reach −0.01 MPa (Field capacity) (Control) and −1.5 MPa (Drought) was calculated, which corresponded to 16% and 8% of water, respectively. After placing pots under the established conditions, they were weighed daily and the evapotranspiration difference was replaced to maintain previous conditions (−0.01 MPa and −1.5 MPa). To maintain plants under flooding, pots were kept in flooded trays throughout the experiment.
The moisture values corresponding to the water retention tension obtained through the soil water retention curve are shown below (Table 1).

Gas Exchanges
Gas exchanges were monitored weekly in all treatments from 9:00 a.m. to 11:00 a.m., with the aid of a CO 2 gas and infrared gas analyzer ("InfraRed Gas Analyzer-IRGA", model GSF 3000 Fl WALZ, Germany) with saturating light of 450 m −2 s −1 determined by means of a light curve. For monitoring, six replicates (1 plant per replicate) of each treatment were evaluated, taking measurements on the 2nd and 3rd fully expanded leaves. CO 2 assimilation rate (Anet, µmol CO 2 m −2 s −1 ), transpiration rate (E, mmol water vapor m −2 s −1 ) and stomatal conductance (gs, mol m −2 s −1 ) were determined. Water use efficiency (WUE, µmol CO 2 (mmol H 2 O −1 )) was calculated using the relationship between assimilated CO 2 and the transpiration rate (Anet/E). The apparent carboxylation efficiency was calculated according to the relationship between CO 2 assimilation rate and leaf intercellular CO 2 concentration (Anet/Ci, mol m −2 s −1 Pa −1 ).

Chlorophyll a Fluorescence
Chlorophyll a fluorescence was performed from 9:00 a.m. to 11:00 a.m. using a fluorometer (LED-Array/PAM-Module3055-FL) on 18 plants (six replicates of 1 plant each) and leaves were acclimated to a period of 30 min in the dark by covering them with aluminum foil; then, an actinic light pulse of 4500 µmol m −2 s −1 was applied to obtain Fm (maximum dark-adapted fluorescence) and Fm (maximum light-adapted fluorescence). In addition to the maximum leaf light-adapted and dark-adapted fluorescence, Fo (minimum dark-adapted fluorescence) and Fo' values (minimum light-adapted fluorescence) were also obtained.

Carbohydrate Concentration
Total soluble sugars were extracted from the leaf material obtained from a pool of samples of two plants per replicate (six replicates per treatment), according to Garcia et al. [42], with minor modifications, and starch was extracted according to Clegg [43]. The procedure to determine the concentration of total soluble sugars was performed according to Morris [44]; for starch, it was described by Yemm and Folkes [45]; for reducing sugars, it was determined by Miller [46]; and for sucrose, it was established by Passos [47], with minor modifications.

Activity of Antioxidant Enzymes, Hydrogen Peroxide and Lipid Peroxidation
The extraction of antioxidant enzymes was performed as described by Kar and Mishra [48] from leaf material obtained from a pool of samples of two plants per replicate (six replicates per treatment). The activities of superoxide dismutase, EC 1.15.1.1 and catalase EC 1.11.1.6 enzymes were determined by the method of Peixoto et al. [49]; the activity of the peroxidase EC 1.11.1.7 enzyme was established according to Teisseire and Guy [50]; and soluble proteins were quantified as described by Bradford [51].
The hydrogen peroxide content was determined by the method of Alexieva et al. [52] and lipid peroxidation was determined according to methodology proposed by Heauth and Packer [53], and both analyses were obtained using a pool of leaf material from two plants per replicate (six replicates per treatment).

Extraction of Total Alkaloids
Total alkaloids were extracted from the root material of 18 A. crassiflora plants (six replicates of 2 plants each); the material was stored in a greenhouse with forced air circulation at 30 • C for ten days, and subsequently ground to obtain 1 g of dry mass for each replicate. Alkaloids were extracted from roots previously dried using the acid-base method. After thorough grinding, the plant material was moistened with a saturated sodium carbonate (Na 2 CO 3 ) solution and left to dry for 48 h at room temperature. Alkaloids were extracted with chloroform (CHCl 3 ) by constant stirring for 1 h and then filtered and washed with distilled water. The CHCl 3 phases were extracted into a 1 M hydrochlo-ric acid (HCl) solution before being alkalinized to pH 9.5 with a saturated solution of Na 2 CO 3 . The alkaline solution was then re-extracted with CHCl 3 , dried with anhydrous sodium sulfate (Na 2 SO 4 ), filtered and evaporated at approximately 25 • C to obtain total alkaloids [25].

Quantification of Total Alkaloids and Liriodenine
To determine the total alkaloid content, the 18 samples were stored at room temperature, re-solubilized with CHCl 3 and transferred to quartz cuvettes. The absorbance of each solution was obtained by spectrophotometer at 254 nm wavelength using liriodenine as the standard for the elaboration of the standard curve (y = 0.0881x − 0.0112, R 2 = 0.9949).
After obtaining the extract, liriodenine was quantified using ultra-high-performance liquid chromatograph (UHPLC-Thermo Fisher-Scientific ® , Waltham, MA, USA) with a gradient pump and UV-Vis detector using C 18 reverse phase column (150 × 4, 6 mm and 5 µm in particle diameter). The mobile phase was 30:70 water (pH 3.5 with acetic acid) and cratic isomethanol, with a flow rate of 1 mL/min, keeping the column temperature at 30 • C. Detection was carried out in UV at 254 nm. For liriodenine quantification, calibration curves were performed by analyzing the stock solution series (y = 0.3595x − 0.0011; R 2 = 0.9989 for samples with up to 10 µg of liriodenine in the extract and y = 0.3658x + 1.142; R 2 = 0.9992 for samples with more than 10 µg [25].

Statistical Analysis
Data were submitted to analysis of variance (ANOVA) using the SigmaPlot software Version 12 and means were compared by the Tukey test at 5% (p < 0.05) [54]. To present the biochemical variables, a radar chart was used. Input variables were initially standardized as a result of the different units, using the scale command from the basic package of the R computing environment, which centers the mean at zero and changes the scale to standard deviation [55].

Results
Annona crassiflora plants showed, in general, that the ability to adapt to water restriction conditions (drought: −1.5 MPa) reflected in the increase in specialized metabolism, unlike what occurred under flooding conditions (Figures 2 and 3, Table 2). In this experiment, an increase in the production of total alkaloids without the occurrence of reductions in the photosynthetic metabolism of plants (gas exchange and chlorophyll a fluorescence) was observed when A. crassiflora plants were kept under drought stress conditions (Figures 4 and 5).   Table 2.    In this context, plants kept under water restriction showed greater carboxylation efficiency of the Rubisco enzyme (Anet/Ci) compared to plants kept under flooding ( Figure  4). However, in both conditions, the carbon assimilation rate (Anet) was lower in relation to plants without water restriction (Field Capacity) and without differences in relation to Ci (data not shown). The other gas exchange variables did not show significant differences (stomatic conductance (gs), transpiration (E), vapor pressure deficit (VPD) and water use efficiency (WUE). The chlorophyll a fluorescence was also impacted by treatments, and plants kept under drought conditions had higher maximum quantum yield (Fv/Fm), effective quantum yield (фPSII), potential quantum efficiency values (Fv′/Fm′) and lower fraction of energy dissipated in the form of heat (D) compared to plants kept under flooding conditions (Figure 5). Photosystem II energy that cannot be dissipated and used in the photochemical phase (Ex), photochemical quenching (qP), non-photochemical quenching (NPQ), electron transport rate (ETR), quantum yield of unregulated non-photochemical energy loss in photosystem II (фNO) and quantum yield of regulated non-photochemical energy loss in photosystem II (фNPQ) did not show significant differences. At the same time that plants showed adaptation to water restriction conditions characterized by responses observed in primary metabolism, the production of total alkaloids and liriodenine was also increased. Plants kept under drought conditions produced higher concentrations of total alkaloids in relation to those kept in soil with maximum water availability (field capacity), while saturated soil did not cause significant variations in total alkaloids but reduced the liriodenine concentration in relation to drought soils.
In this context, plants kept under water restriction showed greater carboxylation efficiency of the Rubisco enzyme (A net /Ci) compared to plants kept under flooding ( Figure 4). However, in both conditions, the carbon assimilation rate (A net ) was lower in relation to plants without water restriction (Field Capacity) and without differences in relation to Ci (data not shown). The other gas exchange variables did not show significant differences (stomatic conductance (gs), transpiration (E), vapor pressure deficit (VPD) and water use efficiency (WUE).
The chlorophyll a fluorescence was also impacted by treatments, and plants kept under drought conditions had higher maximum quantum yield (Fv/Fm), effective quantum yield (фPSII), potential quantum efficiency values (Fv /Fm ) and lower fraction of energy dissipated in the form of heat (D) compared to plants kept under flooding conditions ( Figure 5). Photosystem II energy that cannot be dissipated and used in the photochemical phase (Ex), photochemical quenching (qP), non-photochemical quenching (NPQ), electron transport rate (ETR), quantum yield of unregulated non-photochemical energy loss in photosystem II (фNO) and quantum yield of regulated non-photochemical energy loss in photosystem II (фNPQ) did not show significant differences. Starch and trehalose were affected depending on the conditions in which plants were kept, while total sugars, reducing sugars and sucrose did not show significant differences. Starch was found in higher concentrations in plants kept under flooding but without differing from plants kept under drought, and the lowest trehalose concentrations were found both in leaves of plants kept under both drought and flooding, indicating that this sugar may have been translocated to roots and used in order to neutralize the damage caused by stress ( Figure 2, Table 2).
In general, the enzymatic system acted satisfactorily, preventing membrane damage, since there was no difference in lipid peroxidation and hydrogen peroxide among treatments, possibly indicating that the antioxidant enzymes inhibited the activity of reactive oxygen species. In plants kept under drought soil and flooding, higher peroxidase activity (POD) was observed, while catalase activity (CAT) was higher only in plants kept under flooding. Thus, flooded plants required greater enzymatic activity ( Figure 2, Table 2).

Discussion
The increase in alkaloid production in A. crassiflora plants kept under drought stress conditions seems to be related to their ability to adapt to the Cerrado conditions, which has well-defined drought periods [56], since the photosynthetic process was preserved, ensuring both primary and specialized metabolism. This ability to adapt to the Cerrado conditions seems to be specific, since under flooding, reductions in primary metabolism were evident, affecting the specialized metabolism, especially the synthesis of alkaloid liriodenine.
The fact that drought stress causes increases in specialized metabolism substances such as alkaloids has been shown by several authors, such as Ghorbanpour and Hatami [57] in work with Hyoscyamus niger; Kleinwächter et al. [58] with thyme (Thymus vulgaris); Kleinwächter and Selmar [59] with spices and medicinal plants; and Liu et al. [29] with Catharanthus roseus. Specifically with the genus Annona, Castro-Moreno et al. [60] found Starch and trehalose were affected depending on the conditions in which plants were kept, while total sugars, reducing sugars and sucrose did not show significant differences. Starch was found in higher concentrations in plants kept under flooding but without differing from plants kept under drought, and the lowest trehalose concentrations were found both in leaves of plants kept under both drought and flooding, indicating that this sugar may have been translocated to roots and used in order to neutralize the damage caused by stress ( Figure 2, Table 2).
In general, the enzymatic system acted satisfactorily, preventing membrane damage, since there was no difference in lipid peroxidation and hydrogen peroxide among treatments, possibly indicating that the antioxidant enzymes inhibited the activity of reactive oxygen species. In plants kept under drought soil and flooding, higher peroxidase activity (POD) was observed, while catalase activity (CAT) was higher only in plants kept under flooding. Thus, flooded plants required greater enzymatic activity ( Figure 2, Table 2).

Discussion
The increase in alkaloid production in A. crassiflora plants kept under drought stress conditions seems to be related to their ability to adapt to the Cerrado conditions, which has well-defined drought periods [56], since the photosynthetic process was preserved, ensuring both primary and specialized metabolism. This ability to adapt to the Cerrado conditions seems to be specific, since under flooding, reductions in primary metabolism were evident, affecting the specialized metabolism, especially the synthesis of alkaloid liriodenine.
The fact that drought stress causes increases in specialized metabolism substances such as alkaloids has been shown by several authors, such as Ghorbanpour and Hatami [57] in work with Hyoscyamus niger; Kleinwächter et al. [58] with thyme (Thymus vulgaris); Kleinwächter and Selmar [59] with spices and medicinal plants; and Liu et al. [29] with Catharanthus roseus. Specifically with the genus Annona, Castro-Moreno et al. [60] found the highest liriodenine concentration in Annona lutescens roots at the end of the dry season (about 377 µmol/g), which is the first report of the presence of liriodenine in Annona crassiflora roots under water stress (about 80.9 µg/g of total alkaloids and 13.47 µg/g of liriodenine). Some periodic collections of Annona species tissues in an annual cycle allow us to point out that although the biosynthesis of alkaloids is distributed throughout the plant, the roots generally accumulate the greatest number of alkaloids and produce a higher yield regardless of the phenological stage of plants [61].
In general, water restriction conditions lead to stomatal closure and reduced CO 2 absorption and, as a consequence, there is a considerable decrease in the consumption of NADPH + H + for CO 2 fixation via the Calvin cycle, which generates an excess supply and accumulation of this equivalent reducer. Thus, metabolic processes are directed towards the synthesis of highly reduced compounds such as isoprenoids, phenols and alkaloids with the use of these accumulated reducing agents [62]. In this context, although there were no significant differences in stomatal conductance in A. crassiflora, reductions in CO 2 assimilation rates (A net ) were observed due to water restriction (drought) and flooding, which corroborates the results of Simonneau et al. [63] and Oliveira and Gualtieri [64], respectively, and may have led to lower CO 2 fixation in the Calvin cycle [65]. However, only A. crassiflora plants kept under drought conditions showed greater synthesis of total alkaloids in relation to those kept under field capacity, in addition to having higher carboxylation efficiency (A net /Ci) and higher liriodenine concentration in relation to flooded plants, which indicates that these responses were evident when plants were under water restriction.
The high carboxylation efficiency in A. crassiflora indicates adaptation for survival in environments with periods of low water availability, which is justified by the fact that the species is native to the Cerrado, unlike the results obtained by Mantoan et al. [66] with A. emarginata, showing a decrease in the carboxylation efficiency under irrigation suspension conditions, which may be related to the fact that A. emarginata is a species present in the Atlantic Forest, an environment with greater water availability. In addition, in plants submitted to water stress events, increased damage to the photosynthetic apparatus is observed, causing changes in chlorophyll a fluorescence patterns, which changes the light energy dissipation pathways and increases plant stress [67]. However, in this experiment, plants kept under drought conditions did not show reductions in the chlorophyll a fluorescence pattern, indicating that there was no significant damage to the photosynthetic apparatus. Under flooding, damage is evidenced by low chlorophyll a fluorescence values (фPSII, Fv/Fm, Fv /Fm ) and high energy dissipation in the form of heat (D) (Figure 5), which resulted in lower carboxylation efficiency of the enzyme ribulose 1,5-bisphosphate carboxylase (rubisco) (Figure 4). Thus, flooding directly affected the photosynthetic apparatus in A. crassiflora, reducing its efficiency regardless of stomatal conductance, as proposed by Parolin and Wittmann [68] and Oliveira and Gualtieri [64]. In studies with Annona glabra, this reduction in photosynthetic efficiency did not occur, which reinforces its characteristics of adaptation to restinga, a highly flooded environment [69].
The highest effective quantum yield (фPSII), maximum quantum yield in the dark (Fv/Fm) and potential quantum efficiency (Fv /Fm ) values observed in A. crassiflora plants kept under drought conditions (similar to plants kept under field capacity) indicate that the energy generated may have been destined both for the production of carbon skeletons used in primary metabolism and for specialized metabolism [58,70]. On the other hand, in plants submitted to flooding conditions, the low quantum yield in the dark (Fv/Fm), low potential quantum efficiency (Fv /Fm ) and greater heat dissipation of the antenna (D) indicate a photoprotection mechanism to minimize damage to the photosystem [65].
Thus, in addition to the increase in NADPH + H that can be used for the synthesis of specialized metabolites such as alkaloids, an increase in the production of free radicals from the energy generated in the system is observed [58], and therefore, in stress situations, increases in the production of hydroxyl radicals (OH), superoxides (O 2 − ), hydrogen peroxide molecules (H 2 O 2 ) and singlet oxygen ( 1 O 2 ) are observed, originating from redox reactions that can be in the form of free radicals or in the molecular form of a non-radical [71,72]. Lipoperoxides are the result of the interaction between free radicals and fatty acids in cell membranes, and when this process occurs, the cell membrane integrity is compromised, resulting in the production of carboxylic compounds such as monoaldehyde [73]. To protect itself and try redox homeostasis, the plant has antioxidant enzymes, such as superoxide dismutase (SOD), catalase (CAT) and peroxidases (POD) [74].
The enzyme that acts first is SOD, catalyzing the dismutation of two O 2 •− radicals, generating H 2 O 2 and O 2 . Then, CAT, which is one of the main enzymes acting in the elimination of H 2 O 2 generated during photorespiration and β-oxidation of fatty acids, converts two H 2 O 2 molecules into water and molecular oxygen. Subsequently, POD, located in the cytosol and vacuoles, catalyzes reactions that use H 2 O 2 as oxidants, so this reactive oxygen species (ROS) is also eliminated, even when SOD activity is low [74,75].
Within this complex enzyme system, no significant superoxide dismutase (SOD) activity was observed in A. crassiflora; however, CAT activity was higher in flooded plants, which seems to be a specific characteristic of saturated soils [76] and indicates an attempt to reduce H 2 O 2 accumulated by stress in order to avoid damage to lipids, proteins and nucleic acids and ensure flooding tolerance [77]. POD activity was also observed, both in plants kept in flooding and in drought, which indicates the continued elimination of reactive oxygen species (ROS) from the system to avoid damage to cells. In this context, POD seems to have been efficient, since no differences were observed in the hydrogen peroxide or lipoperoxide concentration between treatments and control ( Figure 2 heat map), which could mean higher malondialdehyde concentrations, Which, in turn, would represent damage to cell membranes [74].
It is noteworthy that to avoid the deleterious effect caused by reactive oxygen species (ROS) in plant tissues, especially in root regions, plants can show greater activity of antioxidant enzymes, especially POD, and accumulate amino acids in roots. Thus, with the greater allocation of amino acids to roots, plants can increase their nitrogen reserve by synthesizing alkaloids [57], which would explain their higher production in A. crassiflora roots, especially when submitted to drought stress.
Roots are the main alkaloid production organ in plants of the genus Annona [25] and alkaloid accumulation in certain situations, as observed in A. crassiflora roots, may indicate osmotic adjustment due to the accumulation of precursor osmolytes, such as amino acids, carbohydrates and sugars such as starch and trehalose [78]. The accumulation of osmolytes can alter the water potential and favor water absorption even in soils with water restriction and generate greater alkaloid [79], which may have occurred with A. crassiflora. Thus, even under reduced soil water conditions, gas exchange and fluorescence were not negatively affected, and alkaloid production still occurred.
Regarding the presence of osmolytes, starch interconversion into other sugars that act as osmoprotectors can be considered, with the ability to influence the carbon allocation for the entire plant, mitigating the negative effect of stress caused by water restriction. Furthermore, in photosynthetic cells, starch can be synthesized and temporarily stored in chloroplasts, and this "transient" starch is synthesized and degraded within a 24-hour period [80][81][82]. Flooding caused an increase in starch concentration in A. crassiflora leaves, as reported in wheat (Triticum spp.) [83,84], which may be a result of the rapid inhibition of plant growth at the beginning of flooding, leading to lower consumption of sugars produced by photoassimilation, in addition to reduction in photosynthesis [85][86][87].
Another aspect to be observed regarding the capacity of A. crassiflora plants to tolerate abiotic stress was the presence of trehalose both in flooded plants and in those under drought stress. As this non-transport-reducing disaccharide acts as a biostimulant in stress tolerance, its lower concentrations in leaves may indicate its translocation to roots and use in order to neutralize the damage that could be caused by stress, since two glucose molecules are generated from the hydrolysis of trehalose [88]. In addition, trehalose synthesis is induced from some stress condition in order to protect enzymes, proteins and lipid membranes against denaturation under stress situations, also playing an osmoprotective role [78,89].
In summary, when A. crassiflora plants are under stress situations, e.g., due to lack of water, CO 2 assimilation tends to decrease and as a consequence, a smaller amount of NADPH 2 is consumed within the Calvin cycle. Thus, much of the energy produced should be dissipated, and, despite the action of non-photochemical mechanisms, such as photorespiration and the xanthophyll cycle, these are potentiated in this situation, and numerous electrons are still transferred to molecular oxygen, generating ROS. Under this situation, plants activate their antioxidant system (SOD, POD and CAT), thus blocking the harmful effect of ROS, leading to a strong increase in the reduction potential of the reducing equivalent (NADPH 2 ), which can be directed to the synthesis of specialized metabolites [62], which seems to have occurred with A. crassiflora plants. Furthermore, starch and trehalose play a role in mitigating the effects of reduced soil water availability to ensure root water absorption (osmoregulation). Thus, A. crassiflora plants under drought stress showed increases in the content of total alkaloids, specifically liriodenine.

Conclusions
A. crassiflora plants are affected by flooding and drought conditions. Drought generates a stimulus signal for the synthesis of total alkaloids and liriodenine without reducing primary productivity, which denotes rusticity and adaptation of the species to the Cerrado conditions. On the other hand, flooding stress is harmful to the photosynthetic apparatus, which does not result in increased alkaloid production and reduces liriodenine production.