Transcriptomic and Biochemical Analysis Reveal Integrative Pathways Between Carbon and Nitrogen Metabolism in Guzmania monostachia (Bromeliaceae) Under Drought

Most epiphytes are found in low-nutrient environments with an intermittent water supply. To deal with water limitation, many bromeliads perform crassulacean acid metabolism (CAM), such as Guzmania monostachia, which shifts from C3 to CAM and can recycle CO2 from the respiration while stomata remain closed during daytime and nighttime (CAM-idling mode). Since the absorbing leaf trichomes can be in contact with organic (urea) and inorganic nutrients (NO3−, NH4+) and the urea hydrolysis releases NH4+ and CO2, we hypothesized that urea can integrate the N and C metabolism during periods of severe drought. Under this condition, NH4+ can be assimilated into amino acids through glutamine synthetase (GS), while the CO2 can be pre-fixated by phosphoenolpyruvate carboxylase (PEPC). In this context, we evaluated the foliar transcriptome of G. monostachia to compare the relative gene expression of some genes involved with CAM and the N metabolism when bromeliads were submitted to 7days of drought. We also conducted a controlled experiment with an extended water deficit period (21days) in which bromeliads were cultivated in different N sources (urea, NH4+, and NO3−). Our transcriptome results demonstrated an increment in the expression of genes related to CAM, particularly those involved in the carboxylation metabolism (PEPC1, PPCK, and NAD-MDH), the movement of malate through vacuolar membrane (ALMT9), and the decarboxylation process (PEPCK). Urea stimulated the expression of PEPC1 and ALMT9, while Urease transcripts increased under water deficit. Under this same condition, GS1 gene expression increased, indicating that the NH4+ from urea hydrolysis can be assimilated in the cytosol. We suggest that the link between C and N metabolism occurred through the supply of carbon skeleton (2-oxoglutarate, 2-OG) by the cytosolic isocitrate dehydrogenase since the number of NADP-ICDH transcripts was also higher under drought conditions. These findings indicate that while urea hydrolysis provides NH4+ that can be consumed by glutamine synthetase-cytosolic/glutamate synthase (GS1/GOGAT) cycle, the CO2 can be used by CAM, maintaining photosynthetic efficiency even when most stomata remain closed 24h (CAM-idling) as in the case of a severe water deficit condition. Thus, we suggest that urea could be used by G. monostachia as a strategy to increase its survival under drought, integrating N and C metabolism.


Transcriptomic and Biochemical Analysis Reveal Integrative Pathways Between Carbon and Nitrogen Metabolism in Guzmania monostachia (Bromeliaceae) Under Drought
Ana Zangirolame Gonçalves * and Helenice Mercier Departamento de Botânica, Instituto de Biociências, Universidade de São Paulo (USP), São Paulo, Brazil Most epiphytes are found in low-nutrient environments with an intermittent water supply. To deal with water limitation, many bromeliads perform crassulacean acid metabolism (CAM), such as Guzmania monostachia, which shifts from C 3 to CAM and can recycle CO 2 from the respiration while stomata remain closed during daytime and nighttime (CAM-idling mode). Since the absorbing leaf trichomes can be in contact with organic (urea) and inorganic nutrients (NO 3 − , NH 4 + ) and the urea hydrolysis releases NH 4 + and CO 2 , we hypothesized that urea can integrate the N and C metabolism during periods of severe drought. Under this condition, NH 4 + can be assimilated into amino acids through glutamine synthetase (GS), while the CO 2 can be pre-fixated by phosphoenolpyruvate carboxylase (PEPC). In this context, we evaluated the foliar transcriptome of G. monostachia to compare the relative gene expression of some genes involved with CAM and the N metabolism when bromeliads were submitted to 7 days of drought. We also conducted a controlled experiment with an extended water deficit period (21 days) in which bromeliads were cultivated in different N sources (urea, NH 4 + , and NO 3 − ). Our transcriptome results demonstrated an increment in the expression of genes related to CAM, particularly those involved in the carboxylation metabolism (PEPC1, PPCK, and NAD-MDH), the movement of malate through vacuolar membrane (ALMT9), and the decarboxylation process (PEPCK). Urea stimulated the expression of PEPC1 and ALMT9, while Urease transcripts increased under water deficit. Under this same condition, GS1 gene expression increased, indicating that the NH 4 + from urea hydrolysis can be assimilated in the cytosol. We suggest that the link between C and N metabolism occurred through the supply of carbon skeleton (2-oxoglutarate, 2-OG) by the cytosolic isocitrate dehydrogenase since the number of NADP-ICDH transcripts was also higher under drought conditions. These findings indicate that while urea hydrolysis provides NH 4 + that can be consumed by glutamine synthetase-cytosolic/ glutamate synthase (GS1/GOGAT) cycle, the CO 2 can be used by CAM, maintaining

INTRODUCTION
Most epiphytic bromeliads inhabit oligotrophic environments with intermittent water and nutrient availability (Benzing, 2008). Periods of severe drought are frequent in the canopy, and it is important to understand how each species copes with harsh environments. One of the principal morphologic/anatomic features of bromeliads to face water and nitrogen (N) scarcity is the formation of a tank by the arrangement of leaves, such as a rosette; this structure is capable of accumulating water and dissolved nutrients, including organic N source, such as urea provided by the excrement of animals (Benzing and Bennett, 2000;Cambuí and Mercier, 2006). Highly efficient absorbing trichomes are distributed along the leaf blade following a gradient in which the basal part has the largest density and also the highest degree of cellular structure specialization (Kleingesinds et al., 2018). Alternatively, the apex has fewer trichomes and a greater number of stomata, thus playing an important role to take up CO 2 from the atmosphere. A relevant adaptative physiological characteristic of many epiphytic tank bromeliads to face water stress is the crassulacean acid metabolism (CAM) photosynthesis because this process confers a higher carboxylation efficiency (Luttge, 2004) and a greater Rubisco (Ribulose-1,5-Biphosphate Carboxylase-Oxygenase) efficiency (Matiz et al., 2013;Gonçalves et al., 2020a).
To deal with nutrient scarcity of the canopy, the leaves of epiphytic tank bromeliads acquired the function of the roots of terrestrial plants. The leaf became the principal vegetative organ responsible for plant survival (Zotz, 2013), showing a spatial and functional division due to its ability to uptake nutrients, mainly in the basal region that is in contact with the tank solution, and to perform photosynthesis, mainly in the apical portion (Freschi et al., 2010;Takahashi and Mercier, 2011). Highly efficient uptake and great storage capacity of nutrients is another important physiological feature of the epiphytic tank bromeliads (Zotz and Asshoff, 2010). Effective leaf N uptake has been demonstrated for NH 4 + and urea, which are the preferable N sources for Vriesea gigantea and Guzmania monostachia (Inselsbacher et al., 2007;Gonçalves et al., 2020b). Urea, after its intracellular hydrolysis by urease enzyme, provides NH 4 + and CO 2 to the leaf tissues (Miflin and Lea, 1980). Low and high affinity N transporters have been detected along the leaf blade of G. monostachia. At the base, for example, a significant increase of gene expression of NH 4 + (AMT1.2) and urea (DUR3 and PIP1.2) transporters were measured (Gonçalves et al., 2020b). Considering N assimilation, the highest activity of glutamine synthetase (GS) was found at the apex, while urease can be found at the base, indicating the importance of the apical foliar portion for amino acid production (Gonçalves et al., 2020b). Transcriptional analysis of G. monostachia leaf showed that the apex and middle portion had differentially expressed genes related to amino acid metabolism and photosynthesis, while the base showed responses to nutrients, particularly N .
In its natural environment, G. monostachia can survive several months with the empty tank (Maxwell et al., 1995). To deal with water limitation, many bromeliad tanks exhibit CAM. This metabolism is a photosynthetic pathway that can be divided into four phases, in which CO 2 is obtained overnight when stomata open (phase I) and it may be assimilated by the enzyme phosphoenolpyruvate carboxylase (PEPC), providing oxaloacetic acid (OAA) and subsequently forming malate by the cytosolic malate dehydrogenase (NAD-MDH; Luttge, 2004). Afterwards, malate is transported into the vacuoles through the aluminum-activated malate transporter (ALMT) family (Pereira et al., 2018) where it can be accumulated. At dawn (phase II), both PEPC and Rubisco are able to fix carbon while the stomata are still open (Matiz et al., 2013). In the subsequent light period (phase III), malic acid is remobilized to the cytosol and decarboxylated through NAD-MDH and phosphoenolpyruvate carboxykinase (PEPCK), producing CO 2 that is consumed by Rubisco while stomata remain closed (Luttge, 2004). At dusk (phase IV), PEPC activity increases and Rubisco decreases while stomata reopen enabling CO 2 assimilation (Luttge, 2004). In G. monostachia, for example, the isoform ALMT9 is involved in the transport of both malate and fumarate to the vacuole (Pereira et al., 2018). During the subsequent light period, stomata remain closed while malic acid is remobilized from the vacuoles to the cytosol where it can be decarboxylated to phosphoenolpyruvate (PEP) and CO 2 which is used for refixation in the C 3 cycle (Luttge, 2004). Guzmania monostachia, as a facultative CAM species, can shift from C 3 to CAM under drought and absorbs CO 2 in the dark period (Medina et al., 1977). There is a variant of normal full CAM called CAM-idling in which stomata remain closed during night and day when water availability is unfavorable, like in a dry season. In this mode of CAM, respiratory CO 2 is refixed in the dark period and recycled to carbohydrate during the light period (Lüttge, 2006). Guzmania monostachia showed a capacity to express a reversible CAM-idling photosynthesis under drought (Freschi et al., 2010), and it was estimated that the accumulation of nocturnal acid malic can contribute to approximately 50% of the total carbon assimilated under water deficit (Pikart et al., 2018). During drought, atmospheric CO 2 absorption by the leaves is restricted due to stomata closure, causing a significant decrease in CO 2 availability. photosynthetic efficiency even when most stomata remain closed 24 h (CAM-idling) as in the case of a severe water deficit condition. Thus, we suggest that urea could be used by G. monostachia as a strategy to increase its survival under drought, integrating N and C metabolism. Interestingly, the delivery of 13 C-labeled urea to the leaves under water deficit enriched the 13 C-malate pool of the apical portion of V. gigantea leaves (Matiz et al., 2017), suggesting that this organic N source could be an important source of CO 2 for bromeliads under water stress (Matiz et al., 2017). In this context, our goal was to investigate the connection between the N and C metabolism through urea nutrition during drought periods considering the G. monostachia bromeliad as a plant-model. Since urea hydrolysis releases CO 2 and NH 4 + , we hypothesized that the interaction between drought and urea nutrition can activate the CAM pathway through the carboxylation metabolism (PEPC1, PPCK, and NAD-MDH), the movement of malate through vacuolar membrane (ALMT9), and the decarboxylation process (PEPCK), while NH 4 + can be assimilated into glutamine through the activation of the GS/GOGAT cycle. Ultimately, the cytosolic isocitrate desidrogenase (NADP-ICDH) can participate in this process, providing 2-oxoglutarate (2-OG) from isocitrate. This 2-OG is then imported into the chloroplasts, where it can be used as the carbon skeleton required for amino acid synthesis.

Gene Expression Related to CAM and N Metabolism
This experiment was conducted with adult G. monostachia (about 2.5 years) obtained previously by in vitro propagation as described in Rodrigues et al. (2016). Until bromeliads reached this maturevegetative age, they were kept in a greenhouse of the Department of Botany at the University of São Paulo (São Paulo, Brazil), where they received tap water daily and nutritive solution modified with half of the original concentration of macronutrients (Knudson, 1946) and with micronutrients (Murashige and Skoog, 1962) each, for 15 days. These plants were kept in pots (14 cm in diameter, 11 cm high) containing a mixture of Pinus sp. bark and commercial organic substrate (Tropstrato ® ). These maturevegetative bromeliads had around 20 leaves each and the oldest leaves measuring around 20 cm length.
Previously to the experiment, bromeliads had their tanks emptied and were transferred to the growth chamber where they were kept for 1 month under the following controlled conditions: 12-h photoperiod, photosynthetic photon flux of 250 μmol m −2 s −1 , air temperature of 27/22°C day/night, and relative humidity of 60/70% day/night. During this period, all plants received distilled water daily and were not fertilized. After 30 days, bromeliads were divided into two treatments (n = 3 replicates): (1) bromeliads were subjected to 7 days of water deficit (WD), and (2) bromeliads received distilled water for 7 days (W). While the water treatment received distilled water for 7 days, bromeliads of the water deficit treatment remained with the tank emptied throughout this period. Subsequently, apical and basal leaf portions of the 8th to 12th innermost nodes were harvested at 7 am and 7 pm, immediately frozen with liquid nitrogen, and stored at −80°C for RNA extraction.

CAM Modulation Under Different N Sources
In order to evaluate the interaction between CAM and nitrogen metabolism, the experiment was carried out with adult bromeliads in the same controlled conditions of temperature, light, relative humidity, and photoperiod as described above. After 30 days, bromeliads received nutrient solutions with a total of 5 mM of nitrogen, according to the following treatments (n = 8 replicates per treatment): (1) without nitrogen source (control), (2) Ca(NO 3 ) 2 (nitrate), (3) (NH 4 ) 2 SO 4 (ammonium), and (4) urea (urea). The nutrient solution was prepared with half of the original concentration of macronutrients (Knudson, 1946; modified without nitrogen sources) and with micronutrients (Murashige and Skoog, 1962). Each experimental bromeliad received 30 ml of the nutrient solution inside its tank once every 7 days for 30 days (totaling four applications).
Twenty-four hours after the last nutrient-solution application, bromeliads had their tanks emptied and were separated into two new treatments for CAM induction (n = 4 replicates per treatment): (1) receiving distilled water to keep the tank full of water (W) and (2) water deficit condition, in which bromeliads did not receive distilled water (WD). Bromeliads remained in these conditions for 21 days as this environmental condition was shown to be the best to induce the transient shift of the photosynthesis from C 3 to CAM-idling in the apical part of the leaves of G. monostachia . At the end of the experiment, the apical leaf portion of the 8th to 12th innermost nodes were harvested at 7 am in order to analyze the relative water content (RWC) in the leaves (RWC), the activity of the enzyme PEPC, the expression of the CAM-specific phosphoenolpyruvate carboxylase gene (PEPC1), and the tonoplast aluminum-activated malate transporter gene (ALMT9). For the quantification of nocturnal acid accumulation (ΔH + ), bromeliad leaves were collected at 7 am and 7 pm.

Relative Water Content
The RWC of the plants was determined by collecting 10 leaf discs of a known area, from which the values of fresh mass (MF), and, subsequently, the discs remained in distilled water for 24 h to obtain the turgid mass (MT). After this period, the discs were dried at 60°C for 48 h to determine the dry mass (DM). These results were used in the following equation: (Weatherley, 1950).

Phosphoenolpyruvate Carboxylase Enzyme Activity
The determination of PEPC activity was performed by the in vitro method as described by Nievola et al. (2001) and adapted for G. monostachia (Freschi et al., 2010;Pereira et al., 2013). In this determination, samples of the apical portion of leaves (1 g) were collected, macerated in liquid nitrogen and immersed in 5 ml of the extraction buffer (pH 8), composed of 200 mM Tris-HCl, 10 mM MgCl 2 , 5 mM DTT, 1 mM EDTA, 0.5% (w/v) BSA, and 10% (v/v) glycerol. The samples were centrifuged for 2 min at 15,000 g. The supernatant was collected and passed through a SEPHADEX PD10/G-25 column. PEPC activity was analyzed in 2 ml of reaction buffer containing 50 mM Tris-HCl (pH 8.0), 1 mM DTT, 10 mM MgCl 2 , 100 mM NaHCO 3 , 20 mM NADH, and 3 mM phosphoenolpyruvate (PEP). The reaction was started by adding 200 μl of the extract. The consumption of NADH was quantified in a spectrophotometer (340 nm) at the initial moment of the reaction and after 4 min. The results were expressed in micromoles of NADH consumed per minute per gram of dry mass (μmol min −1 g −1 ).

Nocturnal Acid Accumulation
The apical leaf portion was analyzed for the degree of CAM expression according to the nocturnal acid accumulation (ΔH + ) using the titratable acidity methodology (Matiz et al., 2017). For this, the leaves of the bromeliads were cut into small fragments of 1 cm 2 , and 0.1 g was macerated in liquid nitrogen, homogenized in 500 μl of MCW solution, and incubated at 60°C for 30 min. The MCW solution was prepared in a proportion of 12:5:1 of methanol:chloroform:water, respectively. Subsequently, the tubes received 500 μl of distilled water and were centrifuged for 10 min at 16,000 g. After centrifugation, 1.5 ml of the supernatant was collected and completely dried at 60°C. After this procedure, the supernatant was collected and stored at −20°C until the titratable acidity procedure.
For the titratable acidity procedure, a solution of phenolphthalein (10 mg in 10 ml of ethanol) was diluted in 3 ml of boiled distilled water to be pipetted into an Elisa ® plate. In the microplate, the following volumes per sample were pipetted in this order: 100 μl of boiled distilled water, 100 μl of phenolphthalein, 5 μl of the respective sample, and 100 μl of the curve solution. All samples received all points on the curve in order to calculate their H + concentrations.
Curve points were obtained as follows:

Gene Expression
The genes encoding the expression of the CAM-specific PEPC enzyme (PEPC1) and the tonoplast aluminum-activated malate transporter (ALMT9) which presented the most discrepant gene expression in the transcriptome of G. monostachia  were selected for the primer design using the PerlPrimer software 3 (Supplementary Table S2). Among all PEPC isoforms found in the transcriptome, PEPC1 was selected because its genomic sequence was more similar to the genomic sequences of the CAM-specific PEPC enzyme comparing to other CAM species (Beltran and Smith, unpublished data). In addition, ALMT9 isoform was selected because it showed to be involved in malate transport to the vacuole in G. monostachia (Pereira et al., 2018). The reference genes (beta-amylase, BAM; betahydroxysteriod-dehydrogenase decarboxylase, HSDD; Supplementary Table S2) were selected from the G. monostachia transcriptome as described by Gonçalves et al. (2020b).
Total RNA of bromeliad leaves was extracted from 300 mg Mix (Applied Biosystems ® ). The analyses of gene expressions were performed using BestKeeper (Pfaffl et al., 2004).

Statistical Analyses
The response variables of the RWC, the activity of the PEPC enzyme, the nocturnal acid accumulation (ΔH + ), and the expression of the genes PEPC1 and ALMT9 were compared using ANOVA, and Tukey HSD post hoc tests were used for pair-wise comparisons. All statistical analyses were performed with the statistical platform R (R Core Team, 2020).

Gene Expression Related to CAM
Some genes related to CAM showed markedly greater expression in the leaf apex of plants subjected to WD, especially PPCK1 and ALMT9 (Figure 1) when compared to the base. PEPC1 showed a high gene expression throughout the entire leaf; however, its highest expression was observed at 7am under WD (Figure 1). NAD-MDH showed greater number of transcripts at 7am independent of the leaf portion, especially when the plant was subjected to WD (Figure 1). The highest expression of PEPCK was observed at the leaf apex at 7am under WD (Figure 1).

Gene Expression Related to N Metabolism
Both NR and NiR had the highest relative gene expression at 7pm in the apex of the bromeliad leaves subjected to WD (Figure 1). Similarly, the quantity of urease transcripts in the apical leaf portion were the highest at 7pm under drought (Figure 1). GS2 showed greater gene expression at the leaf apex compared to the base, independent of the treatments and time of harvest; however, for the base the GS2 expression was stronger at 7pm, independent of the water availability (Figure 1). Similar to GS2, GS1 showed higher gene expression at the leaf apex, especially at 7am, when plants were subjected to WD (Figure 1). The gene expression of Fd-GOGAT was greater at the leaf apex compared with the base, while at the basal part its expression at 7pm under WD was more intense compared with the other treatments (Figure 1). The NADH-GOGAT showed higher gene expression at 7pm with the apex showing a greater quantity of transcripts than the basal part (Figure 1). The transcripts of NADP-ICDH were more abundant in the apical leaf portion compared to the base, mainly under water deficiency.

CAM Modulation Under Different N Sources
Bromeliads that were subjected to WD and received urea, ammonium or did not receive N showed the lowest RWC in the leaves, while the plants that were fertilized with nitrate had the highest water content (Supplementary Table S1; Figure 2). Hydrated plants that were fertilized with nitrate showed the highest RWC and, after 21 days of WD, those bromeliads kept 50% of their water content (Supplementary Table S1; Figure 2). Consistent with these results, we observed that the PEPC enzyme activity and the nocturnal acid accumulation were higher in bromeliads that acquired urea and were submitted to WD, while bromeliads that took up nitrate with WD showed lower PEPC activity and nocturnal acid accumulation (Supplementary Table S1; Figures 3, 4). In addition, bromeliads that received ammonium or did not receive N at all and were subjected to WD presented intermediate PEPC activity and nocturnal acid accumulation (Supplementary Table S1; Figures 3, 4). On the other hand, bromeliads showed the lowest PEPC activities and nocturnal acid accumulation when they received water regardless of nutritional treatment (Supplementary Table S1; Figures 3, 4).
Concerning the gene expression of PEPC1, a significant increase in its expression was observed in bromeliads that acquired urea under WD, followed by plants that were fertilized with ammonium under WD (Supplementary Table S1; Figure 5A). On the other hand, plants that received nitrate or did not receive N and were subjected to WD showed lower PEPC1 gene expressions comparing to bromeliads with other nutritional treatments ( Figure 5A). Additionally, hydrated plants presented the lowest PEPC1 gene expressions regardless the nitrogen they received (Supplementary Table S1; Figure 5A). Consistently, ALMT9 showed its highest expression in bromeliads that acquired urea and were submitted to WD, and the lowest expression was observed in plants that were nourished with nitrate followed by WD (Supplementary Table S1; Figure 5B). Bromeliads that received ammonium or did not have contact with any N sources and were submitted to WD showed intermediate ALMT9 gene expressions (Supplementary Table S1; Figure 5B). Likewise, the gene expressions of ALMT9 were lower in hydrated plants comparing to WD plants (Supplementary Table S1; Figure 5B).

DISCUSSION
Our results indicate that drought had a positive effect on the C 3 -CAM shift at the apical leaf part, increasing the expression of genes related to carboxylation metabolism (PEPC1, PPCK, and NAD-MDH), movement of malate through vacuolar membrane (ALMT9) and decarboxylation process (PEPCK; Figure 6). Together with water deficiency, urea hydrolysis FIGURE 2 | Relative water content in the apical leaf portion of G. monostachia submitted to the following treatments for 1 month: nutrient solution without nitrogen source (control), Ca(NO 3 ) 2 (nitrate), (NH 4 ) 2 SO 4 (ammonium), and urea (urea). Twenty-four hours after the last application of the nutrient solutions, bromeliads were separated into two new treatments for 21 days: (1) receiving distilled water (water) and (2) water deficit condition in which the bromeliads did not receive distilled water (water deficit). Bars indicate standard error. Letters indicate statistical differences between nutritional treatments, and asterisk ( * ) indicates differences between water vs. water deficit (ANOVA/Tukey HSD post hoc test, α = 0.05). Letters are showing the statistical differences among nutritional treatments also under water treatments. Since the nutritional vs. water treatment was statistically significant (Supplementary Table S1), we represented each nutritional treatment with a single letter. seemed to provide CO 2 for malate production and accumulation into vacuoles since we observed that urea stimulated the expression of PEPC1 and ALMT9 genes (Figure 6). At the same time, urea hydrolysis provides NH 4 + (Miflin and Lea, 1980) that can be assimilated into Gln since our results showed increased GS1 gene expression under water deficiency (Figure 6). Ultimately, NADP-ICDH can supply the carbon skeleton to Fd-GOGAT functioning (Figure 6). Because G. monostachia is an epiphytic bromeliad subjected to the intermittence of water and nutrients in its natural environment, we reinforce that urea nutrition obtained in nature by the symbiotic interaction of bromeliads with anurans may bring advantages since the CO 2 can maintain the photosynthetic efficiency even when most stomata remain 24 h closed, while NH 4 + can maintain amino acid and protein production.

C3/CAM-Idling Comparison Through RNA-Seq
Our bromeliads were submitted to drought, and this environmental condition was shown to be the best to induce the transient shift of the photosynthesis from C 3 to CAM-idling in the apical part of the leaves of G. monostachia (Freschi et al., 2010;Mercier et al., 2019). However, the molecular mechanisms underlying the stress induction of the transition from C 3 to CAM remain poorly understood. Recently, this was observed for the facultative C 3 -CAM plant Mesembryanthemum crystallinum (Aizoaceae), whose guard cells themselves can shift from C 3 to CAM, increasing expression of PEPC1 and PPCK1, which encode key enzymes of CAM photosynthesis (Kong et al., 2020).
Comparison of the differential patterns of mRNA abundance in the apical leaf portion between C 3 (W treatment) and CAM-idling (WD treatment) of G. monostachia tissues suggests that the C 3 -CAM switch may be determined by the greater expressions of some genes involved in the carboxylation metabolism (PEPC1, PPCK, and NAD-MDH), the movement of malate through vacuolar membrane (ALMT9) and the decarboxylation process (PEPCK), independent of harvest time (7am or 7pm). It is interesting to observe that in the basal portion, although there was an increase in the number of PEPC1 transcripts in drought condition, there was practically no expression of PPCK1. This result indicates that the enzyme PEPC1 could not be intensely activated through the phosphorylation process, and, consequently, the CO 2 fixation could not happen through the CAM pathway in this part of the leaf. Besides that, the basal part of the leaves has fewer stomata in relation to the middle and apex portions and this morphology may cause some degree of restriction of CO 2 influx FIGURE 3 | Activity of the phosphoenolpyruvate carboxylase (PEPC) enzyme in the apical leaf portion of G. monostachia submitted to the following treatments for 1 month: nutrient solution without nitrogen source (control), Ca(NO 3 ) 2 (nitrate), (NH 4 ) 2 SO 4 (ammonium), and urea (urea). Twenty-four hours after the last application of the nutrient solutions, bromeliads were separated into two new treatments for 21 days: (1) receiving distilled water (water) and (2) water deficit condition in which the bromeliads did not receive distilled water (water deficit). Bars indicate standard error. Letters indicate statistical differences between nutritional treatments, and asterisk ( * ) indicates differences between water vs. water deficit (ANOVA/Tukey HSD post hoc test, α = 0.05). Letters are showing the statistical differences among nutritional treatments also under water treatments. Since the nutritional vs. water treatment was statistically significant (Supplementary Table S1), we represented each nutritional treatment with a single letter. (Freschi et al., 2010). Previous studies have demonstrated that there is an upregulation of PEPC1 expression (7am, 12pm, and 5 pm) in the leaf apex of G. monostachia submitted to 20d of water deficit. However, it was interesting to note that there is also a diurnal modulation of decarboxylation activity of PEPCK with higher gene expression at 7am and an increase in the enzyme activity at noon, avoiding a futile cycling of organic acid during CAM .
Considering that the plants were under nutritional restrictions and the base part was not in contact with N, the RNA-seq analysis showed a higher NR and NiR transcript accumulation following water deficit at 7pm in the foliar apex. In addition to the primary role in N uptake and assimilation, the enzyme nitrate reductase has also been described to possess a nitrite-NO reductase activity that uses nitrite to produce nitric oxide (NO; Astier et al., 2018). This signaling molecule could be involved with stomatal closure in response to plant acclimation to drought (Mioto et al., 2014). Detached leaves of G. monostachia exposed for 7 days to polyethylene glycol showed that NO increased exclusively in the apical part of the leaves and the stomata remained closed even in the night period (CAM-idling mode; Mioto and Mercier, 2013). In Sedum album, another facultative C 3 -CAM plant (Crassulaceae), an increase in NR activity enzyme and a corresponding increase in endogenous NO levels were observed under drought treatment. This rise in NO level occurred during C 3 -CAM transition, correlating well with CAM expression (Habibi, 2020).
Photosynthesis and N metabolism are closely interconnected with the latter being a sink for ATP, reduced power and carbon skeletons produced during photosynthesis (Lea and Ireland, 1999). NH 4 + , whether converted from nitrate or generated by urea hydrolyzation, is assimilated via the GS/glutamine-2-oxoglutarate aminotransferase (GOGAT) cycle, which can be considered very important to maintain NH 4 + at low concentrations in the leaves, helping to counteract ROS formation due to the consumption of reductants, such as Fdred and NADH (Lea and Miflin, 2003). In the context of nutritional deficiency, as the transcriptome of G. monostachia was performed, the base of the leaves presented markedly lower transcript levels of GS2, GS1, Fd or NADH-GOGAT, NADP-ICDH and urease genes. GS1 transcripts increased in the apical part in response to water stress, suggesting some importance in the G. monostachia CAM-idling shift. McNally et al. (1983), studying some C 4 -type and CAM plants, observed that the foliar GS1 enzyme can contribute 35-80% of total GS-activity. Previous work with epiphytic tank bromeliads, such as V. gigantea and G. monostachia, demonstrated that assimilation of NH 4 + took place mainly in the apical leaf part FIGURE 4 | Nocturnal acid accumulation (ΔH + ) in the apical leaf portion of G. monostachia submitted to the following treatments for 1 month: nutrient solution without nitrogen source (control), Ca(NO 3 ) 2 (nitrate), (NH 4 ) 2 SO 4 (ammonium), and urea (urea). Twenty-four hours after the last application of the nutrient solutions, bromeliads were separated into two new treatments for 21 days: (1) receiving distilled water (water) and (2) water deficit condition in which the bromeliads did not receive distilled water (water deficit). Bars indicate standard error. Letters indicate statistical differences between nutritional treatments, and asterisk ( * ) indicates differences between water vs. water deficit (ANOVA/Tukey HSD post hoc test, α = 0.05). Letters indicate statistical differences between nutritional treatments, and asterisk ( * ) indicates differences between water vs. water deficit (ANOVA/Tukey HSD post hoc test, α = 0.05). Letters are showing the statistical differences among nutritional treatments also under water treatments. Since the nutritional vs. water treatment was statistically significant (Supplementary Table S1), we represented each nutritional treatment with a single letter.
Frontiers in Plant Science | www.frontiersin.org 9 October 2021 | Volume 12 | Article 715289 A B FIGURE 5 | Gene expression (fold change) of the (A) phosphoenolpyruvate carboxylase gene (PEPC1) and (B) the tonoplast aluminum-activated malate transporter gene (ALMT9) in the apical leaf portion of G. monostachia submitted to the following treatments for 1 month: nutrient solution without nitrogen source (control), Ca(NO 3 ) 2 (nitrate), (NH 4 ) 2 SO 4 (ammonium), and urea (urea). Twenty-four hours after the last application of the nutrient solutions, bromeliads were separated into two new treatments for 21 days: (1) receiving distilled water (water) and (2) water deficit condition in which the bromeliads did not receive distilled water (water deficit). (Takahashi and Mercier, 2011;Gonçalves et al., 2020b). The enzyme GS1 is located in the cytosol and plays an important role in the primary NH 4 + assimilation from the environment (Sakakibara et al., 1996). In pine and other conifers NH 4 + , which is the predominant source of N absorbed for these trees, is assimilated in the cytosol through GS1, and glutamine (Gln) and glutamate (Glu) biosynthesis occur in the chloroplast (Suárez et al., 2002). Tomato GS1/GOGAT cycle was related to FIGURE 5 | Bars indicate standard error. Letters indicate statistical differences between nutritional treatments, and asterisk ( * ) indicates differences between water vs. water deficit (ANOVA/Tukey HSD post hoc test, α = 0.05). Letters indicate statistical differences between nutritional treatments, and asterisk ( * ) indicates differences between water vs. water deficit (ANOVA/Tukey HSD post hoc test, α = 0.05). Letters are showing the statistical differences among nutritional treatments also under water treatments. Since the nutritional vs. water treatment was statistically significant (Supplementary Table S1), we represented each nutritional treatment with a single letter.

Continued
FIGURE 6 | A simplified schematic showing our hypothesis about nitrogen metabolism and its interaction with carbon metabolism in a water stress-induced CAMidling bromeliad, G. monostachia, submitted to urea nutrition. Urea is taken up at the leaf basal part and after its hydrolysis generates NH 4 + and CO 2 . At the green leaf part and mainly in the apex portion during the night, the CO 2 may be assimilated through the crassulacean acid metabolism (CAM) pathway. Urea nutrition associated with water stress positively stimulated PEPC1 and ALMT9 gene expressions and PEPC1 activity, resulting in a higher nocturnal vacuolar acidity. NH 4 + , in turn, mainly after entering a cell of the apical leaf region may be assimilated during the light-dark period in the cytosol mediated by the enzyme GS1 forming glutamine (Gln). Therefore, glutamine and glutamate (Glu) biosynthesis may occur in separate compartments. The mitochondria may supply the carbon skeleton for Fd-GOGAT functioning. The key C-compound, 2-oxoglutarate (2-OG), may derive from sugar respiration reactions. In this pathway, citrate is used to generate the isocitrate required for 2-OG synthesis, allowing the net Glu production in the chloroplast and then other amino acid formation. Abbreviations for enzymes: PEPC1, phosphoenolpyruvate carboxylase 1; PPCK, phosphoenolpyruvate carboxylase kinase; NAD-MDH, NAD-malate dehydrogenase; PEPCK, phosphoenolpyruvate carboxykinase; Rubisco, ribulose-1,5-bisphosphate carboxylase/oxygenase; GS1, glutamine synthetase 1; Fd-GOGAT, glutamine-2-oxoglutarate aminotransferase; and NADP-ICDH, NADP-isocitrate dehydrogenase. Abbreviations for transporters: ALMT9, tonoplast aluminum-activated malate transporter 9; DUR3, urea transporter; AMT3.1, ammonium transporter; TIP2.1, tonoplast intrinsic protein; PIP1.2, plasma membrane intrinsic protein; AA, amino acid. stress responsiveness since their expression levels were modified under drought, cold or heat stress treatments (Liu et al., 2016). On the other hand, GS2 and Fd-or NADH-GOGAT transcripts of G. monostachia did not show any increase under water deficiency, and GS2/Fd-GOGAT genes were highly expressed in the apical leaf part. It is well known that the chloroplastidic GS2/Fd-GOGAT cycle plays important roles in the assimilation of N from NO 3 − in leaf and in plant survival under photorespiratory conditions which lead to the generation of excess NH 4 + accumulation (Krapp, 2015), while NADH-GOGAT is crucial for NH 4 + assimilation in heterotrophic tissues (Selinski and Scheibe, 2019). For Mesembryanthemum crystalinum, after the transition from C 3 to CAM the plants still maintain the capacity to photorespire (Whitehouse et al., 1991). In a previous study using CAM-induced leaves of G. monostachia, it was observed that glycolate oxidase 1 (GLO1) mRNA abundance, which is a gene considered a photorespiration indicator, presented higher expression at noon in CAM-induced plant compared to the dawn and dusk times . The shift from C 3 to CAM-idling in G. monostachia leaf apex could be related to the increase in the NADP-ICDH gene expression, which encodes the cytosolic enzyme isocitrate dehydrogenase. This enzyme allows a net synthesis of Glu, providing the key organic acid 2-oxoglutarate (2-OG), which is then transported into plastids allowing GOGAT functioning (Hodges, 2002). Since 2-OG could play a significant role modulating the flow of C to N metabolism depending upon environmental conditions (Foyer et al., 2011), it is possible that changes in its synthesis could reflect the decrease of internal CO 2 levels caused by day/night stomatal closure, which is typical of the CAM-idling photosynthesis.

Urea Can Positively Modulate C3-CAM Shift Under Water Deficit
Guzmania monostachia, as an example of epiphytic tank bromeliad, is a good model to study the effects of organic nitrogen nutrition on bromeliad development and, particularly, on photosynthetic plasticity when exposed to water deficit (Rodrigues et al., 2014(Rodrigues et al., , 2016. In the low-nutrient epiphytic environment, which frequently presents nitrogen and water scarcity, organic compounds such as urea could represent an important nutritional resource that provides both nitrogen and carbon (Matiz et al., 2017). Our investigation reveals for the first time that urea can positively modulate the C 3 -CAM shift when the leaves were under drought stress. Leaves from urea treatment showed the highest increases in the expressions of PEPC1 and ALMT9 genes comparing with inorganic N sources (NO 3 − or NH 4 + ) and, consequently, the greatest levels of PEPC activity and nocturnal acid accumulation. Therefore, it seems that the change from C 3 to CAM mode was stimulated by this organic N source. On the other hand, plants supplied with NO 3 − were less vulnerable to water loss under drought conditions while showed lower expressions of PEPC1 and ALMT9 genes. Nitrate is known to increase the endogenous cytokinin level, which may regulate negatively the PEPC activity and decrease CAM in G. monostachia leaves (Pereira et al., 2013(Pereira et al., , 2018. Furthermore, NO 3 − uptake could be accompanied by K + absorption in Arabidopsis thaliana as a charge balance, which can enable osmotic adjustments and may keep the RWC in this plant (Wang et al., 2012;Rubio et al., 2014). Therefore, the maintenance of water content may delay or down-regulate the CAM.
It is well known that abiotic stresses, such as drought, induce stomata closure and, consequently, limit uptake of atmospheric CO 2 into the leaf tissue (Chen et al., 2018;Gobara et al., 2020). In this context, urea hydrolysis through foliar urease activity may partially compensate for the CO 2 reduction by providing it, in addition to NH 4 + (Matiz et al., 2017. This mechanism may represent an important physiological adaptation to obtain C and N at the same time. For G. monostachia, urea and NH 4 + are the preferred N forms of uptake (Gonçalves et al., 2020b). The absorption of urea as a whole molecule took place mainly at the basal region of the leaf through the transporters of low (TIP2.1) and high (DUR 3) affinity and may be stored in the vacuole before the water deficit imposition Gonçalves et al., 2020b). If we consider the CAM-idling photosynthetic mode, whose gas exchange with the atmosphere largely ceases and respiration provides the CO 2 (Winter, 2019), this strategy of using urea as a preferential source of N becomes even more relevant. Upper leaf parts of the C 3 tank epiphytic bromeliad V. gigantea which close their stomata during the day under drought had a greater increase in urease activity after being supplied with urea at nighttime when compared to daytime. At the same time, a higher PEPC activity was found, indicating that at least part of the CO 2 produced by urea hydrolyses was fixed during the night (Matiz et al., 2017). Furthermore, a cytochemical detection of CO 2 was performed in the apex leaf portion after urea application during the night, and CO 2 deposition was detected in a greater quantity in the cytoplasm near the chloroplasts than in well-watered leaves (Matiz et al., 2017).
It seems that urease activity occurs during both the day and night in G. monostachia leaves since we observed an increase in the expression of urease and GS1 genes under water scarcity. Thus, we consider that the first step of NH 4 + assimilation may take place in the cytosol of this bromeliad through GS1 enzyme, giving raise to glutamine (Gln) concentration that could enter the chloroplast. Subsequently, the enzyme GOGAT transfers the amide group of Gln to 2-OG, generating two molecules of glutamate (Glu), one of which may be used to produce other amino acids through the action of transaminases (Forde and Lea, 2007). Recently, in A. thaliana, Gln biosynthesis via GS1 was found to be related to increased tolerance to abiotic stress (Ji et al., 2019). Net NH 4 + assimilation requires a source of 2-OG, which is generated through partial respiration of sugars. The provision of carbon skeletons for GOGAT functioning could be supplied when the citrate is exported from the mitochondria and transformed to isocitrate and then into 2-OG by the cytosolic NADP-ICDH enzyme. A number of observations corroborate the idea that cytosolic NADP-ICDH is the major isoform in green leaves and plays an important role in N-assimilation (Foyer et al., 2011). Levels of 2-OG can reflect C/N status and may play a signaling role in the co-ordination of carbon and nitrogen metabolism (Hodges, 2002).
Previous work using G. monostachia detached leaves showed that NH 4 + was more effective than nitrate to positively modulate the increase in CAM expression under water stress. The expressions of ALMT and PEPC genes were intensified, and a higher nocturnal malic acid accumulation was observed. On the other hand, NO 3 − supplied at various concentrations had little effect on the transcript number of ALMT and PEPC genes (Pereira et al., 2018). In the present investigation, we observed that when urea was supplied as the sole N source to the tank of G. monostachia plants, the nocturnal acid malic concentration was around five times greater than that in the presence of NH 4 + . After urea hydrolysis in the cytoplasm, the availability in CO 2 may increase its incorporation by PEPC1, while more oxalacetate acid (OAA) and malate (through the action of NAD-MDH) can be produced, increasing acidity at night. Alternatively, during the day the CO 2 generated by urea hydrolysis could be assimilated by ribulose-1,5-biphosphate carboxylase/oxygenase (Rubisco) in chloroplasts. In contrast, NO 3 − treatment had no effect on C 3 -CAM shift, curiously showing the higher RWC after 21 days of water deficiency. Therefore, urea may have a positive effect on both C and N metabolism of G. monostachia, mainly if we consider a CAM-idling plant with almost no access to CO 2 from the atmosphere. For some tank bromeliads, urea can contribute 30% of the total plant N (Romero et al., 2010;Gonçalves et al., 2016). In nature, this organic N source could be excreted by various animals (for instance, anurans) that visit the bromeliad's tank searching for shelter or a place to breed, and the leaf basal part assumed the role of roots in N absorption and urea hydrolysis (Takahashi and Mercier, 2011).

CONCLUSION
Considering our biochemical and molecular findings (Figure 6), we conclude that drought had a positive effect on the C 3 -CAM shift in the apical leaf part, increasing the expression of genes related to carboxylation metabolism (PEPC1, PPCK, and NAD-MDH), movement of malate through vacuolar membrane (ALMT9) and decarboxylation process (PEPCK). Furthermore, our investigation revealed that urea can positively modulate the CAM-idling photosynthesis, which is characterized by the closure of stomata during the day and night. Our results reinforced that, when G. monostachia stay in long periods of drought, urea stimulates the expression of PEPC1 and ALMT9, while Urease transcripts increase under the same conditions. At the same time, urea hydrolysis provides NH 4 + , which can be assimilated into Gln since our results showed increased GS1 gene expression under water deficiency. We suggest that the link between C and N metabolism in G. monostachia may occur through the supplying of 2-OG by isocitrate dehydrogenase since the number of NADP-ICDH transcripts was also higher under drought.
Thus, we suggest that urea nutrition provided in nature most by the symbiotic interaction of bromeliads with anurans may bring advantages to bromeliad survival under severe water stress, providing NH 4 + , which ultimately can produce amino acids and proteins, while CO 2 can maintain photosynthetic efficiency, even when most stomata remain closed for 24 h.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: https://www.ncbi.nlm. nih.gov/, Bioproject ID PRJNA532595.

AUTHOR CONTRIBUTIONS
AG conceived, designed, and performed the experiments, analyzed the data, and drafted the manuscript. HM conceived and designed the experiments, analyzed the data, and drafted the manuscript. All authors contributed to the article and approved the submitted version.