Metabolic profile of Brazilian pine embryos and megagametophyte of stored seeds

Changes in availability of metabolites during seed deterioration might damage processes of synthesis and energy release for embryonic growth. This research aimed to determine which alterations occur in metabolic profile over storage of Araucaria angustifolia seeds and how these alterations are related to viability loss. Metabolic profile of samples stored at 60, 120, and 180 days, at ambient temperature, refrigerator (5°C), and freezer (-18°C), was analyzed by Fourier transform mid infrared (FTIR) spectroscopy. Additionally, soluble proteins of embryos and soluble carbohydrates and starch of megagametophytes were also quantified and related to dry matter and seed germination. Presence of primary and secondary metabolism compounds was identified in embryos (starch, proteins, lipids and phenolic). Both embryos and megagametophytes are composed principally of carbohydrate and starch. At 180 days of storage, only freezer-stored samples maintained a metabolic profile similar to freshly harvested samples, but seed viability was dramatically reduced. Storage in refrigerator can be an alternative to control the catabolism of reserve compounds in A. angustifolia seeds, retaining about 70% germination.


INTRODUCTION
Species Araucaria angustifolia (Bert.)O. Kuntze (Brazilian pine) is a native gymnosperm from Brazilian Atlantic Rain Forest, and has an emergent status in upper forest canopy (Korndörfer et al., 2008;Elbl et al., 2014).Because of high economic value of its wood, natural populations of A. angustifolia have been a target of progressive exploitation (Eira et al., 1994), and species has been classified as critically endangered by International Union for Conservation of Nature and Genetic Resources (IUCN, 2013).
A. angustifolia seed or pinion contains an embryo with two cotyledons, and reserve tissue called megagametophyte (Ferreira, 1981), which quickly lose physiological potential after harvest because they are recalcitrant (Espindola et al., 1994;Panza et al., 2002;Balbuena et al., 2011).Therefore, its use in seedlings production for areas recovery and commercial plantations, as well as human consumption, is difficult, being restricted to a few months per year.Although there are several means of clonal propagation of conifers, such as somatic embryogenesis, rooting of cuttings, and organogenesis, there are risks due to lack of genetic diversity in clonal populations (Bonga, 2015).Storage conditions must be adequate to minimize seminiferous metabolic activity and extend seed longevity, but even under optimal storage conditions, conifer seeds will deteriorate with time, especially for particular species which appear to be peculiarly susceptible to deterioration (Terskikh et al., 2008).
Considering importance of seed quality evaluation for a successful production (Oliveira et al., 2014), some studies have investigated a reduction of physiological quality of A. angustifolia seeds during storage (Piriz Carrillo et al., 2003;Caçola et al., 2006;Amarante et al., 2007;Garcia et al., 2014), and has been reported cell death and loss of viability caused by freezing of recalcitrant seeds.
Seed deterioration is accompanied by biochemical changes which affect the supply of energy required for germination.In fact, some authors studied chemical composition of A. angustifolia seeds, which may have from 2 to 7% lipids, 4 to 9% protein, 68 to 88% starch (Ramos and Souza, 1991;Piriz Carrillo et al., 2003;Astarita et al., 2003;Leite et al., 2008).However, researches on chemical composition of A. angustifolia seeds, in general, aim to assess its potential use as food, and there are few studies concerning metabolic profile changes during storage, including damage from freezing to metabolism of A. angustifolia seeds, but it is known that seeds starch content can be reduced from 73 to 22% after 18 months of storage in polyethylene bags, at 4°C (Piriz Carrillo et al., 2003).
Understanding variations in metabolic profile in function of viability loss may indicate strategies for maintaining quality of A. angustifolia seeds for long storage periods.Based on three popular forms of seed storage among regional producers (use for seedlings production) and consumers (use for food) of A. angustifolia seeds, the aim of this study was to determine which alterations occur in metabolic profile over storage time and how these alterations are related to viability loss.

Seed sampling, storage and periodicity of assessments
A. angustifolia mature cones (megastrobili) containing seeds were collected from a natural population located in the region of Painel (27° 55' of latitude south, 50° 06' of longitude west and an average altitude of 1144 m), southern Brazil.Cones were collected from 25 mother trees randomly selected with approximately 50 m away from each other, with a total of 80 cones.Seed sample was homogenized and divided into four replicates, from which a fraction was removed for each storage condition: ambient temperature (laboratory), refrigerator (5 ± 1°C and 45 ± 5% of humidity relative), and freezer (-18 ± 1°C and 90 ± 5% of relative humidity), in sealed transparent plastic bags with porosity of 0.015 μm.As a reference, for storage period of samples at ambient (laboratory) conditions, variations in temperature and relative humidity to city of study ranged from -3°C and 33%, until 30°C and 99%, respectively, with average temperature and relative humidity of 15°C and 80% between months of July and December (Epagri/Ciram, 2013).Analyses were performed in freshly harvested seeds and every 60 days during storage over 180 days.As a control, viability and dry matter were conducted using whole seeds; mid-infrared vibrational spectroscopy (FTIR) and soluble proteins were conducted with embryos tissue, since embryo reserves are the first hydrolyzed during deterioration process; soluble carbohydrates and starch, major compounds of A. angustifolia seeds were extracted from megagamethophytes in order to compare them with FTIR results.

Dry matter analysis
Dry matter content was determined with four replicates of three seeds for each storage time and storage condition, where were transversely cut, then weighed (wet weight), dried in an oven at 105°C ± 3°C, for 24 h and reweighed to determine dry matter.

Germination test
Four replicates of 25 seeds were surface-decontaminated with sodium hypochlorite solution (2%, v/v) for three minutes and subsequently sown in trays containing moist sand up to 60% of field capacity, deposited in germination chamber at 25°C, with constant luminosity.Germination evaluation was performed after 65 days, considering germinated seeds with root protrusion (at least 10 mm of primary root).

Fourier transform infrared mid spectroscopy (FTIR)
A pool of 10 embryos/replicate (dry weight) were macerated into liquid nitrogen in a mortar with pestle, and analyzed by an infrared spectrometer, equipped with a single reflection ATR (Golden Gate) system in potassium bromide matrix.128 scans/sample were collected in spectral window of 4000 to 500 cm -1 .Spectra were normalized, baseline corrected in the region of interest (3000 to 600 cm -1 ) and processed with the aid of OPUS software (Bruker Corporation).Analyses were performed in triplicate.
Soluble carbohydrates and starch were extracted through methodology described by McCready et al. (1950), using a pool containing megagametophytes from 10 seeds per replicate.1 g of dry biomass was macerated in mortar and pestle with liquid nitrogen, followed by triple extraction with ethyl alcohol 80%.Supernatants were collected for quantification of soluble carbohydrates and residue was used for starch extraction by adding perchloric acid (52%).After filtering aliquots of soluble carbohydrates and starch in fiberglass filter, quantification was proceeded by colorimetric analysis using a spectrophotometer in range of 490 nm by phenol-sulfuric method (Dubois et al., 1956), with D-glucose as standard (Glucose 0 to 100 µg ml -1 , R 2 = 0.9873, y = 0.0214 x + 0.0839), in triplicate.

Experimental design and statistical analysis
Experiment was conducted in a completely randomized design, in split plot, with three storage conditions (ambient, refrigerator and freezer) and four storage periods (0, 60, 120 and 180 days).Analysis of regression and variance was performed, and test of means by Tukey test at 5% probability using statistical program SAS 2009 (SAS Institute Inc.).Total set of data of processed spectra was subjected to multivariate statistical analysis, by applying methods of principal components (PCA), using statistical package The Unscrambler 9.1 (CAMO Software Inc.).

Dry matter content
Dry matter content of freshly harvested seeds was 548 mg g -1 , with a reduction for all storage conditions.More pronounced discrepancies among storage conditions were detected after 120 days of storage, with a greater reduction in dry matter in seeds stored at ambient temperature (Figure 1).

Seeds germination
Freshly harvested seeds showed 92% of root protrusion (Figure 2).Regression analysis showed a significant decreasing tendency in germination percentage, especially for samples stored in freezer, for which there was complete loss of viability at 60 days of storage.Samples maintained at ambient temperature showed complete loss of viability at 180 days of storage.Refrigerator condition was superior in maintaining seed viability, showing 70% germination at 180 days of storage.

FTIR spectroscopy analyses
FTIR spectroscopy allows detection of 19 to 22 relevant peaks in spectral window between 3000 to 600 cm -1 (Figure 3).Bands with higher intensities for all samples were detected in spectral window 1200 to 800 cm -1 , which is associated with occurrence of typical carbohydrates signals (Cerná et al., 2003;Kuhnen et al., 2010).In this spectral region, bands are primarily due to axial deformation of links C-O, C-C, and C-O-C and the presence of COH groups (Schulz and Baranska, 2007;Kuhnen et al., 2010).Particularly, peaks were detected in band 1022 cm , since region between 1650 and 1500 cm -1 resulted from axial deformations of C = O group and angular deformations of NH 2 group regarding the amines (Lambert et al., 2001;Schulz and Baranska, 2007;Kuhnen et al., 2010).Lipids constituent in A. angustifolia embryos were identified by presence of bands in 2926, 2856 and 1742 cm -1 . These metabolites are associated to axial deformation of functional group C = O typically found in fatty acids in 1740 and 1440 cm -1 (Silverstein, 1994;Schulz and Baranska, 2007), as well as signals between 3000 to 2800 cm -1 , related to axial and angular deformation of methyl group (-CH 3 ) and/or methylene (CH 2 ) (Lambert et al., 2001;Kuhnen et al., 2010).
Various signals in spectral range 900 to 690 cm -1 were observed in samples, suggesting the presence of compounds having aromatic rings in their structure, such as (poly)phenols, which may result from deformation of link = CH of aromatic compounds (Lambert et al., 2001).As for the bands observed in spectral window 1140 to 1150 cm -1 , they can represent the presence of terpenoids (Schulz and Baranska, 2007).
Principal components 1 (PC1) and 2 (PC2) revealed a clear discrimination of samples, being effective to explain 97% (96% explained by component PC1 and 1% by PC2) of variance in spectral data of FTIR (Figure 4).

Biochemical changes during storage
Protein content of embryos form freshly harvested seeds  ) and refrigerator (4.8 mg g -1 ) (Figure 5).However, storage in freezer was efficient in maintaining soluble protein levels, which remain similar (P ≤ 0.05) to freshly harvested samples until the end of experiment (180 days).Content of soluble carbohydrates in megagametophyte of freshly harvested seeds was 58 mg g -1 , and this amount remained stable (P ≤ 0.05) up to 180 days of storage only for freezer condition (Figure 6A).At 180 days, soluble ) was similar (P ≤ 0.05) to that observed for refrigerator (40 mg g -1 ). Freshly harvested seed showed 509 mg g -1 of starch (Figure 6B).At 180 days, only samples stored in refrigerator showed similar starch content (P ≤ 0.05) to that observed in freshly harvested samples and there was a significant reduction of 28% in starch content of samples stored in freezer for 180 days.For ambient condition, the megagametophyte tissue was fully consumed at 120 days, precluding analyses of soluble carbohydrates and starch; therefore, it is inferred that soluble carbohydrates and starch content of samples was 0 mg g -1 at 120 days and 180 days of storage at ambient condition.

DISCUSSION
Germination results showed refrigerator storage favored retention of seeds viability for 180 days, while storage in freezer leads to a rapid loss of cell viability since freezing causes embryos death (Garcia et al., 2014), and ambient condition favored a continuous decrease in viability over time.Lower temperatures such as those in refrigerator (5°C) contributes to viability maintenance, as observed by other authors, with 61% germination in A. angustifollia seeds stored under 4°C (Fowler et al., 1998), with low respiratory rates when stored at temperatures ranging from 2 to 10°C (Amarante et al., 2007).
Starch and lipids are among compounds stored in greater quantities in embryos of mature seeds of A. angustifolia (Farias-Soares et al., 2013), but the FTIR spectroscopy results demonstrated that proteinaceous compounds are, quantitatively as important as lipids, and phenolic constituents are also present in large quantities.Regardless of condition, storage resulted in changes in metabolic profile compared to freshly harvested samples.FTIR and chemometric analysis showed these changes resulted from structural differences conferred, especially by presence of carbohydrates represented by starch (1021 cm -1 ), proteins (1649 cm -1 ), and lipids (2930 cm -1 ). Sample distribution along PC2 was responsible for explaining only a small part of sample variance, with a greater factor contribution of phenolic compounds of A. angustifolia embryos, represented by presence of signals at 971 cm -1 throughout experimental period.These and similar methods, including high-resolution nuclear magnetic resonance (NMR) spectroscopy provided a wealth of information on a variety of primary and secondary metabolites of conifer seeds (Terskikh et al., 2005).
Construction of a descriptive model based in principal components analyses allowed identification of profiles of greater metabolic similarity between storage periods and conditions.A standard for samples grouping in intermediate storage periods (60 and 120 days) was detected, indicating minor metabolic discrepancies.Such clustering suggests samples stored for 60 days at ambient temperature are similar, in metabolic composition level, to those samples stored in refrigerator for 120 days and in freezer for 180 days.In fact, further analysis demonstrated protein, carbohydrates and starch contents decreased, in first moment, for ambient samples; in a second moment, protein content decreased for refrigerator samples; and for freezer samples, there was no reduction of protein and carbohydrate contents during storage period, that is, these compounds were not deteriorated or neither hydrolyzed in a preparatory germination metabolism.Soluble protein is accumulated in A. angustifolia embryos, especially in final stages of embryogenesis in embryonic axis region (Astarita et al., 2003;Silveira et al., 2008).However, during seed deterioration after harvest protein degradation occurs (McDonald, 1999;Murthy et al., 2003), which may impair early development of seedlings and subsequent field establishment.Except by freezer samples, whose death was caused by freezing, soluble proteins availability during storage seems to play an important role for embryonic viability maintenance, since the reduction of proteins levels (Figure 5) followed same pattern as the reduction observed in seed germination (Figure 2).
A progressive decline in seed viability may also be caused by peroxidation of storage lipids due to the reduction of nutrient reserves, and also because of the generation of toxic products of peroxidation (Terskikh et al., 2008).Soluble carbohydrates are also essential to seed viability maintenance during storage, acting in soaking water mechanisms and in embryo protection against desiccation and pathogen attack (Barbedo and Marcos Filho, 1998).Although high, carbohydrates content in freshly harvested seeds was 32% and 20% lower than observed by Ramos and Souza (1991) and Piriz Carrillo et al. (2003), respectively, but chemical composition can vary as a function of factors such as genotype, maturity stage, climatic conditions of collection site, and nutrition of mother plant.
There was no degradation of starch in seeds stored in refrigerator (5°C), despite other authors have reported the reduction of 23 and 12% in starch content after storage of A. angustifolia seeds for a period of 180 days at temperature of 5 and 4°C, respectively (Ramos and Souza, 1991;Piriz Carrillo et al., 2003).But in seeds of forest species Caesalpinia echinata (brazilwood), both contents of soluble carbohydrates and starch decrease during storage (Hellmann et al., 2008).It has been reported that viability loss of A. angustifolia is accompanied by a decrease in starch content during artificial aging (Ramos and Carneiro, 1991) and natural aging in storage (Ramos and Souza, 1991).Starch is an important compound in both megagametophyte and embryo of A. angustifolia seeds, since this component is present in higher quantity (quantitative analysis) and because glycidyl components were most important for sample separation (chemometric analysis).
Changes observed in seeds reserve compounds directly reflected in dry matter content.In freshly harvested seeds, dry matter content was similar to that observed when seeds reach maturity (500 mg g -1 ) (Astarita et al., 2003).As mentioned, ambient condition reflected earlier the changes in protein and carbohydrates (reduction at 60 days), and starch contents (reduction at 120 days), which were also reflected in a lower dry matter content shown at 180 days.These changes probably are associated to early germination events, when they occur in the beginning of reserve digestion and assimilation as energy for posterior embryonic development during germination process.On the other hand, storage in freezer proved to be more adequate to maintain protein and carbohydrates content of A. angustifolia seeds.This fact may be important for feeding purposes.
Seeds maintained in refrigerator showed a reduction in both protein content, carbohydrate content and germination during experimental period, suggesting an important positive relation between reserve availability and seed viability.However, after storage for 180 days at 5°C, some authors found no relation between reduction of viability (from 99% to 18% approximately), and protein content of seeds, which remained unchanged during storage (Ramos and Souza, 1991).Ultimately, results presented here demonstrated metabolic changes begin later during storage at controlled conditions (refrigerator and freezer) indicating that the higher the average temperature during storage, the higher speed of biochemical changes in both embryos and megagametophytes tissues of A. angustifolia.
In summary, analyses performed were effective to characterize the presence of primary (starch, proteins and lipids) and secondary metabolism compounds (phenolic) in A. angustifolia embryos.Changes in metabolic profile were reflected in seed viability, varying according to storage condition.Seed storage at -18°C (freezer) delayed hydrolysis of reserve metabolites after seed harvest, and this preservation of some compounds is vital for human as these seeds are edible, but this storage condition did not retain seed viability.Therefore, storage at 5°C (refrigerator) favored prolongation of seeds biochemical quality, minimizing damage to viability and proving to be a viable alternative to storing A. angustifolia seeds for 180 days.These results may aid further research aiming at characterizing other metabolic events that occur during deterioration of A. angustifolia seeds, such as those involving ultrastructural changes and consequences of cellular oxidative stress.
. Protein constituents in A. angustifolia embryos were mainly observed by band in 1648 cm -1