SOIL FERTILITY AFFECTS ELEMENTAL DISTRIBUTION IN NEEDLES OF THE CONIFER

: Araucaria angustifolia is a conifer species found in South American subtropical forests that comprises less than 3% of the native vegetation and little is known concerning the accumulation of nutritional elements in its needles. In this study, scanning electron microscopy (SEM) coupled with energy dispersive spectroscopy (EDS) was used to assess the elemental distribution in needles. Needles were selected from 28 month-old plants grown in a fertilization experiment supplied with: 1) N, P, and K; 2) N and P; and 3) N and K. In microanalysis, four types of specialized needle tissues (adaxial epidermis, palisade mesophyll, spongy mesophyll and abaxial epidermis) were evaluated for elemental composition (C, O, P, K, Ca, S and Al). When crystals were detected, the concentrations of 12 elements were determined (C, O, P, K, Ca, S, Al, Fe, Mg, Na, Si, and Cl). Under low soil P and K, these elements were found in low concentrations in the epidermis, mesophyll, and crystals. Under low soil P, Ca and K accumulated in the spongy mesophyll, while under low soil K only Ca accumulated in this tissue. In addition, low soil P or K availability


INTRODUCTION
Araucaria angustifolia (Bert.)O. Ktze, popularly known as araucaria, Brazilian pine, or Paraná pine, is one of the most important native species in southern Brazil.Originally it covered large continuous areas in southern Brazil (Paraná, Santa Catarina and Rio Grande do Sul states), extending north to São Paulo, Minas Gerais and Rio de Janeiro where they are found in small isolated patches, especially in colder and higher regions of these states.Some isolated individuals are also found in northern Argentina and eastern Paraguay.In the Atlantic Forest area, it is a component of the Mixed Ombrophilous Forest (also called the Araucaria Forest).However, today A. angustifolia is on the endangered species list (REIS et al., 2014).
The seeds (pinhão) of A. angustifolia have economic and social importance, especially in states where edible seeds obtained from natural stands provide hundreds of families with a main source of annual income.In addition, A. angustifolia also has ecological importance since its seeds are the main food for many mammals and birds in the winter and its presence creates a microclimate that favors the development of many species (REIS et al., 2014;PERALTA et al., 2016).Factors such as historical timber cutting and seed removal have hindered natural forest reestablishment.
Generally, A. angustifolia is nutrient demanding, especially for the nutrients K and P, which are among the most demanded by this plant (SIMÕES; COUTO, 1973).Therefore, under cultivation application of these two nutrients represents a considerable cost.In studies with A. angustifolia, elemental composition analysis has been used to assess nutritional status and plant/water relationships (GARBIN;DILLENBURG 2008;CASSANA;DILLENBURG, 2013).However, use of total needle elemental concentrations can have limited use in the study of plant metabolism because tissue types within the needles may have variable composition (REGVAR et al., 2013;CARVALHO et al., 2016).
Advanced microscopy and spectroscopy techniques have allowed for a more comprehensive assessment of the elemental composition of plant tissues (DINH et al., 2015;KRIEGER et al., 2017).In the evaluation of elemental accumulation, studies have concentrated on the elemental distribution in different tissues and plant cells (CONN; GILLIHAM, 2010) including crystals (HE et al., 2014).In such cases, SEM combined with EDS has been used ef ciently in the microanalysis of essential and potentially toxic elements in plant tissues (SILVA et al., 2014;HE et al., 2015;KRIEGER et al., 2017).However, there is a paucity of information on elemental accumulation in A. angustifolia needles.Thus, our study objective was to assess the elemental accumulation pro le of A. angustifolia needles grown in soil that varied in P and K nutrient availability.

Plant growth conditions
The study was conducted in a nursery (open sky) in the city of Curitiba,Paraná,, altitude 911 m) located at Federal University of Paraná (Sector of Agricultural Sciences).The soil used for seedling growth was a Cambisol collected from an area where A. angustifolia naturally occurs in Pinhais, Paraná.The soil, at eld moisture, was sieved (4 mm), homogenized, and sampled for analysis (Table 1).
Based on soil chemical analysis, dolomitic lime was applied (3.6 g per kg of soil with a relative total neutralization power of 75.2%) to increase base saturation to 70% (using the method of base saturation).The soil was then homogenized and incubated for two months (time when the reaction of limestone in the soil tends to stabilize).After this incubation period, the soil was again homogenized and 16 dm 3 were placed in 20 dm 3 containers (height = 30 cm; bottom diameter = 25 cm; top diameter = 35 cm).Three treatments with four replications based on CQFS/RS-SC (2004) were evaluated: 1) control, soil with high P and K availability (N-P-K = 22.2, 18.3, and 12.5 g per container); 2) -K, soil with low K availability (N-P-K = 22.2, 18.3, and 0 g per container); and 3) -P, soil with low P availability (N-P-K = 22.2, 0, and 12.5 g per container).The complete amounts of P and K were mixed into the soil.Half of the total N was initially mixed into the soil; the other half was top-dressed in a split manner at 12 (50 %), 14 (30 %) and 16 (20 %) months after seedlings were transplanted.The sources used to supply N-P-K were urea (45 % N), super triple phosphate (41 % P 2 O 5 ), and KCl (60 % K 2 O), respectively.
In December 2013, a seven month old seedling (~20 cm tall) was transplanted into each container.Containers were randomly distributed on a plastic surface to avoid contact and interference from surrounding soil.Weed control was done manually and irrigation (2 L .plant -1 at 9:00 AM) occurred during periods without adequate rainfall.
Twenty-one months after seedlings were transplanted (September 2015), trees had an average height of 88, 75, and 60 cm for the control, -K, and -P treatments, respectively.At this time, Mehlich-1 extractions showed that the respective soil K and P levels were 0.33 cmol c . dm -3 (or 128 mg .dm -3 K) and 50 mg .dm -3 for the control treatment, 0.06 cmol c . dm -3 (or 23 mg .dm -3 K) and 35 mg .dm -3 for the -K treatment and, 0.40 cmol c . dm -3 (or 156 mg .dm -3 K) and 1.4 mg .dm -3 for the -P treatment.In addition, needle elemental composition con rms the low soil K and P availability (Table 2).

SEM-EDS microanalysis
In September 2015, needles were collected (uninjured and free from symptoms of disease) from the central region of branches in the third whorl from the crown apex.The needles were washed in deionized water and ~2 mm wide cross-sections from the central region of the needles were removed with a stainless steel blade.Needle sections were xed in FAA medium (5% glacial acetic acid and 70% [v/v] ethanol) for 24 h in the dark at 4°C.After xation, sections were dehydrated in an ascending ethanol series of 70, 80, 90, 95, and 100% for 1 hour and dehydrated to critical point with liquid CO 2 .The sections were then mounted on aluminum stubs coated with double-sided copper tape.The microanalyses were performed using SEM (Vega3 LM, Tescan) with EDS (X-Max N , Oxford).The SEM instrument was a thermionic emission system with a tungsten heated lament as the electron source with a resolution of 3 nm, and microscope functions were controlled via VegaTC software.The EDS instrument contained an 80 mm 2 silicon drift detector, and data processing was conducted with AZTech (Advanced) software.
Four points of needle cross sections (adaxial epidermis, palisade mesophyll, spongy mesophyll and abaxial epidermis; Figure 1) were selected for determination of elemental composition.Subsequently, the elemental composition of crystals found in needle mesophyll was determined.The images were obtained with accelerating voltage of 15 kV in low vacuum.Microanalyses by EDS were conducted with accelerating voltage of 15 kV, 35º takeoff angle, and 120 s per time point.In general, a maximum of twelve elements were detected (C, O, Na, Mg, Al, Si, P, S, Cl, K, Ca and Fe); these elements were analyzed based on K-α lines and the use of internal standards (C.Vit, SiO 2 , NaAlSi 3 O 8 , MgO, Al 2 O 3 , SiO 2 , GaP, FeS 2 , NaCl, KBr, CaSiO 3 and Fe, respectively).4.20 1 Before analysis soil was re-sieved (2 mm) and dried at 40ºC; Sand, silt, and clay (hydrometer method); pH (CaCl 2 0.01 mol .L -1 ); Ca 2+ , Mg 2+ , Al 3+ (extracted with KCl 1 mol .L -1 ); H + Al 3+ (calcium acetate 0.5 mol .L -1 extraction); organic carbon (OC) (volumetric method by potassium dichromate); K + and P (Mehlich-1 extraction); base saturation (BS); cation exchange capacity (CEC) at pH 7; and Al 3+ saturation (m).All analyzes were based on Embrapa (2011). 1 Unit of elements is mg .kg -1 and the analytical procedure was: ~200 mg of needles (oven dried, ground and passed through a 1 mm sieve) was digested in 6 ml of HNO 3 + water mixture (ratio 2:1, v/v) and 2 ml H 2 O 2 (30 %) using a microwave heating system (MARS, CEM Microwave Technology Ltd); once digestion was complete, the tubes were cooled and brought to a volume of 25 mL by adding Milli-Q water; subsequent total elemental analysis was carried out by ame emission spectroscopy (DM-62, Digimed) for K, while the other elements (Ca, Mg, P, Fe, Mn, Zn, Cu, B, Si, Cr, Al, Ba, Ti) were determined using an optical emission spectrometer with inductively coupled plasma (ICP-OES) (720-ES, Varian).
To estimate the number of crystals found within A. angustifolia needles, the following procedure was performed: ve squares (0.1 x 0.1 mm) were arranged diagonally across each needle section; the number of crystals within each square was counted; the mean and standard deviation of crystal numbers were obtained considering ve squares in four replicates per treatment (totaling 20 squares per treatment).

Statistical analysis
Relative elemental concentration data from specialized tissues of A. angustifolia needles were analyzed in a completely randomized design in split plot with four replications.The three fertility treatments (control, -K and -P) were considered as main plots and the four types of specialized tissues (adaxial epidermis, palisade mesophyll, spongy mesophyll and abaxial epidermis) were considered sub-plots.Data were subjected to analysis of variance and means were compared by Tukey test at 5% probability.Statistical analyzes were performed with Sisvar software (FERREIRA, 2011).

Elemental distribution in epidemis and mesophyll
The P, K, and Ca relative concentrations in the four specialized tissues are shown in Figure 2. The -P microanalysis values are related to the low total concentration of that nutrient in needles (Table 2).Unlike the present study where there was no variation among the specialized tissues, Conn and Gilliham (2010) reported that monocots accumulate P in the mesophyll, while dicots accumulate P in the epidermis.In addition, -P was the only treatment in which this element was not detected in the spongy mesophyll and abaxial epidermis (Figure 2a).This suggests that under de ciency conditions, P was either discharged directly from the xylem to the palisade mesophyll and distributed to the adaxial epidermis (KARLEY et al., 2003), or the opposite occurred without preferential distribution to other tissues.Clearly, further studies with SEM-EDS are required to further understand P dynamics in A. angustifolia needles.
In A. angustifolia grown under -P, the K relative concentration in the spongy mesophyll of needles was higher than other specialized tissues and when compared to the control and -K treatments (Figure 2b).The -P treatment also increased K concentration in the needles (Table 2).Thus, it is possible that low P favored the xylem K of oad to the mesophyll through the vascular sheath (KARLEY et al., 2003).Findings may be related to the low plant growth in soil with low P availability.However, it is important to note that the K relative concentrations determined by EDS were low (being near that of P); this was in opposition to total concentrations (Table 2) where K was ~13 times higher than P concentration.This result may indicate that part of the K was lost during tissue xation since this element is mainly in free forms in plant cells (TAIZ et al., 2015).
Considering the variation among treatments, Ca results from EDS microanalysis (Figure 2c) followed the same trend as total Ca concentrations in needles (Table 2).In other words high concentrations were noted in the -K treatment.The competition between Ca and K for absorption through the root system (where the lack of one favors absorption of the other) is well known in the literature and has been reported by Simões and Couto (1973) for A. angustifolia.Regarding Ca accumulation in spongy mesophyll, Rios et al. (2012) found that Ca accumulated in the palisade mesophyll of Brassica rapa leaves when the nutrient solution availability was high, whereas under low Ca availability the distribution in leaves was homogeneous.However, Ca accumulation sites can vary among species: mesophyll for dicots (CONN; GILLIHAM, 2010); and epidermis and mesophyll for the FIGURE 2 Relative concentration (%) of P (a), K (b) and Ca (c) in the adaxial epidermis (AdE), palisade mesophyll (PM), spongy mesophyll (SM) and abaxial epidermis (AbE) in needles of A. angustifolia grown under low P (-P), low K (-K) and control conditions with high soil P and K availability (Con).Uppercase letters (nutritional status effect) followed by the same letters and lowercase letters (specialized tissue effect) followed by the same letters do not differ by Tukey test (p <0.05).
Zn/Cd hyperaccumulator species Noccaea caerulescens (DINH et al., 2015).The C and O relative concentrations in specialized A. angustifolia needle tissues followed this descending order: adaxial epidermis> abaxial epidermis> mesophyll palisade> spongy mesophyll and, spongy mesophyll> abaxial epidermis>mesophyll palisade>adaxial epidermis, respectively (Figure 3a, b).The differentiation in C and O concentrations between the epidermis and mesophyll is mainly due to the biochemical composition of each specialized tissue type.Regarding the composition, the epidermis consists of structural biomolecules (i.e., lignin, cellulose and pectin) which are rich in C and higher than that found in the mesophyll (REGVAR et al., 2013).Biologically, it can be inferred that these results are congruent with the functionality of each cell type.Accordingly, higher C concentrations of biomolecules confer resistance to the outer layers of needle tissue (epidermis, especially the adaxial epidermis) thereby protecting the inner layers (mesophyll) which are responsible for photosynthesis (TAIZ et al., 2015).
The S relative concentrations in the specialized tissues of A. angustifolia needles ranged from 0.05% to 0.42%, being higher in the palisade mesophyll compared to the other specialized tissues (Figure 3c).These results indicate that in palisade mesophyll there is a greater accumulation of organic molecules containing S. Likewise, Mastroberti and Mariath (2003) reported an accumulation of phenolic compounds and a large number of chloroplasts in A. angustifolia; both generally contain S in their composition.Conn and Gilliham (2010) also reported higher S concentrations in the mesophyll compared to the epidermis of monocot and dicot leaves.
In general, the Al relative concentrations ranging from 0.12% to 0.30% were not signi cantly affected by specialized tissue type or nutritional status (Figure 3d).However, further studies are needed to evaluate the distribution of Al in needles of this species under other conditions (e.g., a greater diversity of soils).Considering the total concentration of Al in needles, observed values (Table 1; on average 94 mg .kg -1 ) were similar to those reported by Reissmann et al. (1976) (on average 136 mg .kg -1 ), which are relatively low concentrations.In plants that tolerate high Al concentrations in leaves (such as tea Camellia sinensis (L.) O. Kuntze a mean Al concentration: 6,866 mg .kg -1 ), it is often accumulated in the epidermis (CARR et al., 2003).For a wide range of species that accumulate metals in leaves, Leitenmaier and Küpper (2013) observed that the principal accumulation site is the epidermis.Under the most common condition (accumulation in the epidermis), it is likely that metals cannot be discharged from the rib xylem to the mesophyll symplastically, thus, they are carried by rib extensions to the epidermis apoplastically (KARLEY et al., 2003).

Crystal formation and precipitation of elements
During microanalysis of A. angustifolia needles with contrasting nutritional status, various crystals were seen in photomicrographs (Figure 4).It was noted that needles from the -K treatment had a higher density of crystals (3,400±900 crystals mm 2 ) compared to other treatments.For the -P treatment (1,800±800 crystals mm 2 ), needles were found to have more crystals than the control treatment (900±300 crystals mm 2 ).
There is evidence in our study for the regulation of Ca availability in needles since the high density of crystals in needles -K treatment (Figure 4b) was due to a higher total Ca concentration (24,624 mg .kg -1 ) in needles (Table 2).In addition, the higher crystal density in the -P treatment (vs.control; Figure 4a) can be explained by variation in the total Ca concentration (-P = 9,022 mg .kg -1 ; control = 7,660 mg .kg -1 ) in the needles (Table 2).In needles of a conifer species Picea abies (L.) Karst., Fink (1991) found that the water soluble Ca concentration did not vary with element availability despite a wide variation in Ca-oxalate concentration (acid soil: 1,269 mg .kg -1 Ca; calcareous soil: 8,712 mg .kg -1 Ca) after extraction with HCl.In the same study, the authors found Ca crystals occurring in the extracellular portion of the intercellular spaces along the cell wall.Pritchard et al. (2000) found and Al (d) in the adaxial epidermis (AdE), palisade mesophyll (PM), spongy mesophyll (SM) and abaxial epidermis (AbE) in needles of A. angustifolia grown under low P (-P), low K (-K) and control conditions with high soil P and K availability (Con).Uppercase letters (nutritional status effect) followed by the same letters and lowercase letters (specialized tissue effect) followed by the same letters do not differ by  2011), as in this study.
In general, the main elements contained in the crystals were C, O, and Ca.These three elements comprised 98% of the crystals on average (Figure 5; Table 3).With respect to C and Ca, it was possible to verify crystal differences according to the nutritional status of A. angustifolia.In the -K treatment, the relative concentrations of C and Ca of two crystals (numbers 2 and 3) were lower and higher, respectively, compared to the crystals (1, 2, and 3) of the other treatments.Therefore, needles with higher Ca concentration (Table 2) and crystal density (Figure 4) had crystals richer in Ca.These results support the importance of Ca crystal formation in regulating the availability of this element et al., 2014).This indicates that A. angustifolia is probably an oxalic plant regarding Ca physiological nutrition (WHITE, 2005).
Regarding other elements detected in crystals, only P, K, and Mg (Table 3) exhibited an in uence of nutritional status.In the -P treatment, none of the crystals showed P in their constitution.In the -K treatment, only one crystal was found to contain K and was the only treatment where Mg was detected in the needle crystals.Thus, due to low soil P and K there was no signi cant incorporation of these elements in crystal structures which corresponds with the low total concentrations in A. angustifolia needles (Table 2).On the other hand, plants grown under -K (Table 3) incorporated Mg and Si into the crystal structure due to an increase in the total Mg (and secondarily Si) concentration in the needles.Although the concentrations of P, Mg, and Si in crystals are low and similar to those seen in cellular structures, it is likely that the crystal formation process can in uence the balance of metal availability at the cellular level.This may also be true for S since this nutrient was found in all crystals (Table 3).Although other studies veri ed the presence of P, K, Mg, and S in crystals, some plant  species (i.e., Piper arboreum var.arboreum, P. callosum, P. tuberculatum, Acacia robeorum, Acacia stipuligera, Acacia stellaticeps, and Fagus sylvatica) have high concentrations of these elements (SILVA et al., 2014, HE et al., 2014, KRIEGER et al., 2017).
Although Al was detected in all analyzed crystals (Table 3), one of the needle crystals from the -K treatment was comprised of 2.79% Al.In addition, Fe (3.82%) and Cl (0.67%) were detected only in the crystal with a high Al content.This suggests that some crystals are multielement compounds.In the case of Al, the sequestration of this potentially toxic element in the crystal structure can be bene cial to the plant, considering that it would be one means of inactivating Al and decreasing (at least partially) the harmful effects to leaf physiology.Thus, Al sequestration in crystals can be a means for native acid soil species (e.g., A. angustifolia) to tolerate increased soil Al availability.Furthermore, the accumulation of several other elements (Mn, Cu, V, Ti, Ba, Sr, Cd, Pb, Fe, Zn) in crystals has been reported (HE et al., 2014;HE et al., 2015).
Altering photomicrographs (40% brightness decrease and 80% contrast increase) obtained by SEM in low vacuum allowed for better visualization of crystals in A. angustifolia needles (Figure 4d,f).This was only possible because these photomicrographs were obtained under low vacuum without metallization.Image collection in low vacuum causes a chromatic contrast between materials of different composition (e.g., Ca oxalate crystals and plant tissues).When the sample is metallized this chromatic contrast is drastically reduced (Figure 1) due to the high atomic number of the metal used in metallization (Au).Thus, the change in the chromatic properties of the photomicrographs only increases the natural contrast between crystals and needle tissues; this contrast is somewhat similar to highlighting crystals by bleaching plant tissue and visualizing with polarized light microscopy (LERSTEN;HORNER, 2011;SILVA et al., 2014).

CONCLUSION
Under low soil P and K, these elements were found in low concentrations in the epidermis, mesophyll, and crystals in needles of South American conifer Araucaria angustifolia.Under low soil K conditions, a higher Ca concentration increase occurs in the spongy mesophyll and Therefore, the availability of P and K in soil affected the accumulation of elements in needles and the type of tissue and formation of crystals were key to the nutrient dynamics in needles of this South American conifer.

FIGURE 1
FIGURE 1 Photomicrograph of a needle cross section of Araucaria angustifolia indicating the microanalysis points.The image was obtained via scanning electron microscope in high vacuum and after Au plating.The elemental microanalyses were obtained with energy dispersive spectroscopy.AdE = adaxial epidermis; PM = palisade mesophyll; SM = spongy mesophyll; and AbE = abaxial epidermis.

FIGURE 3
FIGURE 3 Relative concentration (%) of C (a), O (b), S (c)and Al (d) in the adaxial epidermis (AdE), palisade mesophyll (PM), spongy mesophyll (SM) and abaxial epidermis (AbE) in needles of A. angustifolia grown under low P (-P), low K (-K) and control conditions with high soil P and K availability (Con).Uppercase letters (nutritional status effect) followed by the same letters and lowercase letters (specialized tissue effect) followed by the same letters do not differ by Tukey test (p <0.05).

FIGURE 4
FIGURE 4 Photomicrographs of needles (in cross section) of A. angustifolia.Plants were grown under low P (a; d), low K (b; e) and control conditions (with P and K) (c; f).a-c: photomicrographs taken with a scanning electron microscope at low vacuum.d-f: are the same photomicrographs a-c with brightness reduction (-40%) and increased contrast (80%).

FIGURE 5
FIGURE 5 Photomicrographs of A. angustifolia needle mesophyll.Plants were grown under low P (a; d), low K (b, e) and control conditions (with P and K) (c; f).a-c: photomicrographs taken with a scanning electron microscope at low vacuum.d-f: are the same photomicrographs a-c with brightness reduction (-40%) and increased contrast (80%).The spectrum indications (1, 2 and 3 [crystals] and 4 [starch granules]) are the points selected to determine elemental composition.

TABLE 1
Granulometric and chemical properties of the soil.

TABLE 2
Mean and standard deviation of the concentration of elements in needles of Araucaria angustifolia low P (-P), low K (-K), and control (high soil P and K) treatments.1 Tukey test (p <0.05).Pinus palustris Mill.) which were not impacted by resource availability (e.g., nitrogen, water, or atmospheric CO 2 level).Note that the crystals found were probably Ca-oxalate since Ca-carbonate crystals usually do not resist the sample xation process when the solution contains acetic acid (LERSTEN; HORNER, CERNE | v. 23 n. 2 | p. 257-266 | 2017 Ca and Mg are formed.Under low P conditions, K and Ca accumulated in the spongy mesophyll with a small increase in crystal formation.