Effects of Harvest Time on the Yield, Quality and Active Substance of Torreya grandis Nut and Its Oil.

Torreya grandis is an important economic tree species in China. It provides nutritional value and is important to the health care industry. There are ongoing issues with product quality which are primarily related to improper management and early harvest. This study was carried out during the fruit ripening processes to evaluate the influence of harvesting date on T. grandis quality, and to determine the optimal harvest period. The effects of harvest time on the variation of quality and nutritional parameters of T. grandis nuts and its oil were evaluated, and the optimal harvest period was determined. The results showed that harvest timing had a strong effect on both oil yield and quality. Prolonged ripening could induce higher levels of kernel rate, fruit inclusions, oil and nutritional quality. When the sample harvested in the mid-September, the kernel rate and oil content were increased by 1.88±0.31% and 6.65±0.47%, respectively, compared to samples harvested in the beginning of late-August. Similarly, the mid-September harvest resulted in total unsaturated fatty acids content of the oil being increased by 5.3±0.34%, the FFA and peroxide value being decreased by 40.7±0.15% and 76±0.08%, respectively, and total tocopherols and free amino acids were increased 7.5±0.24% and 47.3±0.15%, respectively, compared to the samples harvested on Aug. 25. The results indicated that the optimal harvest time of T. grandis fruits was mid-September as it was beneficial for improving the quality of T. grandis nut and its oil. It was suggested that T. grandis fruit should be harvested later.

producers prefer to pick the fruits all at once to reduce the harvest costs. In these cases, harvest time is suitable for some varieties but too early for other varieties.
It is known that early harvest induces significant quality and weight losses 16 , harvesting too late can make collecting seeds difficult, and mildew creates conditions whereby the nuts easily crack and fall to the ground which threatens quality and safety 17 . It is vital to properly time the fruit harvest to obtain high nut quality and yield 18 . Compared with camellia Camellia oleifera Abel. , olive Olea europaea L. , walnut Juglans regia L. , and other woody oil species, the research on the harvesting and maturation of T. grandis is minimal. There are no reports about the fruit traits, accumulation of oil content and nutrients at the different harvest periods.
Determination of the optimal fruit-ripening stage for the production of T. grandis nut represents a critical choice based on the best combination of quantity and quality. Usually, the optimum harvest time of nuts varies by years, depending on ecological conditions. Accordingly, determination of harvest time should be made by phenological observation. The quite flushing of husks, efficient rotation of fruit in the husk, taking its color of hard shell, and the dropping of nuts about 3/4 rate when shaken indicate the optimum harvest time for nuts 18,19 . As T. grandis ripening proceeds, nut characteristics such as weight, kernel rate, oil content, profiles of various phytochemicals, including fatty acids, protein, starch, soluble sugar, tocopherol and total reducing capacity, are changing. These changes influence the chemical composition and qualities of T. grandis nut and its oil 20 . This study evaluates the influence of harvest timing on T. grandis nutrient quality and identifies the optimal time for harvest, thereby providing a reference and a guide for actual production.

Sampling site
The study site was located in Fuyang 30 06 N, 119 95 E , Zhejiang Province Southeast China . The region is characterized by mountains and hills 105 m in altitude . It is a mild rainy region with a sub-tropical humid monsoon climate. Yearly precipitation is 1477.9 mm average of 30 years . Monthly mean temperature is 4.3 for January and 28.8 for July Fig. S1 . The agro-climatic data were supplied by the local weather bureau in Fuyang district of Hangzhou.

Sampling and processing
The T. grandis trees were managed and had no nutrient deficiencies or pest damage. 8-year-old trees were selected and labeled in groves. To determine the best harvest time in terms of oil yield and other fruit quality parameters, T. grandis fruit samples were harvested by hand five times in 2016: Aug. 25 mature green stage , Aug. 31, Sept. 6, Sept. 12 and Sept. 18 ripe fruits with epicarp longitudinal cracks , when the fruit was 30 mature until fully maturity. Only healthy fruits without any infection or physical damage were used. All the fruits were collected from the same 10 trees with a representative subsample about 100 g collected from each tree. Immediately after harvesting, fruits were sealed in plastic bags and transported to the laboratory where they were separated from the hulls, then cleaned nuts-in-shell and dried to 7 kernel moisture content 21 in fan-forced oven at 50 , and stored at 18 before analysis.

Oil extraction and quality analysis
T. grandis nuts were pressed by a vertical hydraulic oil press 6YY-190, Luoyang, China . The following quality evaluation parameters were tested: the kernel percent, kernel moisture content, kernel oil yield, fatty acid profile, free fatty acids FFA , peroxide value, reducing capacity, tocopherols, soluble sugar content, starch content and the kinds and contents of free amino acids.
Kernel percent was calculated from the total nut and kernel weights after drying. Kernel moisture was determined by differencing weight before and after drying. Kernel oil content was determined using the Soxhlet method whereby seed samples were finely ground in household food grinder and extracted with petroleum ether boiling range 30 60 in a B811-Soxhlet apparatus Buchi, Switzerland for 8 h at a constant temperature of 70 . Protein, starch and total soluble sugar content were determined according to the procedure described by Wang 22 . FFA content and peroxide value were carried out following the analytical methods described in ISO International Organization for Standardization 660 and 3960, respectively.

Fatty acid profile
The fatty acid profile was analyzed by gas chromatography. Methyl esters were prepared in compliance with the standard method to determine the fatty acid profile, and oil was analyzed by GC-2010 plus gas chromatography Shimadzu, Japan equipped with a flame-ionization detector FID detector. The gas-chromatographic condition used were based on the previous study 23 . The gas chromatograph is equipped with an autosampler, a split/splitless injector, and a FID. A fused silica capillary column FAMEWAX 50 m length 0.25 mm i.d. and 0.25 µm film thickness, Restek, America was used. The carrier gas was nitrogen with 2 mL/min flow rate, and the temperature of injection block and detector was 220 . The GC oven heating was initiated at 150 . That temperature was maintained for 1 min and then increased to 190 at a rate of 5 /min and those conditions were maintained for 20 min. The injection volume was 1 µL, the split ratio was 1:10. Commercial mixtures of fatty acid methyl esters Sigma Aldrich, Milan, Italy were used as reference data for the relative retention times. The results were expressed as relative area percent of the total.

Determination of tocopherols
Oil samples weighing 1 g to an accuracy of 0.01 g were dissolved with n-hexane in a 10-mL brown volumetric flask and filtrated through a 0.22-µm PTFE filter for HPLC analysis. Tocopherols were measured using an Agilent 1290 high-performance liquid chromatography Agilent Technologies, Santa Clara, CA, USA equipped with a silica column 250 4.6 mm i.d., 5 µm, Waters Co., Milford, MA, USA and a fluorescence detector Agilent Technologies Inc., Santa Clara, CA, USA . Mobile phase was n-hexane and isopropanol with the volume ratio of 98.5:1.5. The flow rate was 1.2 mL/min, and the sample size was 10 µL. The column temperature was 30 . The fluorescence detection was operated with the excitation wavelength at 295 nm and the emission wavelength at 333 nm. The α-, β-, γ-, and δ-tocopherols were identified and quantified by comparing with the standards MilliporeSigma, Billerica, MA . The results were expressed as milligrams of tocopherol per 100 gram of oil.

Reducing capacity assay
Folin-Ciocalteu reagent has a meaningful disadvantage since the reacting not only with the phenolic compounds but also with vitamin C, some amino acids, peptides, reduction sugars, organic acids, Maillard reaction products and some more compounds, therefore several researchers advise to interpret obtained results as reducing capacity instead of total phenolics content 24,25 . The Folin-Ciocalteu method was used to measure sample reducing capacity in the research as reported by Mahmood Ghasemnezhad et al. 26 with some variations. The extracting solution was prepared by solid-liquid extraction using ground kernels 5 g and methanol-water 30:70, v/v, 50 mL at 70 for three times. 1 mL of extract was added to 5 mL of Folin-Ciocalteu s phenol reagent 5 w/v and to 4 mL of sodium carbonate 7.5 w/v . The mixture was kept in the dark for 30 min after 1 min of vortex oscillation. Finally, the absorptions of the solution at 765 nm were measured. Reducing capacity were expressed as mg gallic acid/g of kernel.

Determination of free amino acids
The composition and contents of free amino acid in T. grandis nut were determined using L-8800 Amino Acid Analyzer Hitachi, Japan as reported by Mo R.H. et al. 27 .
All analyses were performed at least in triplicate except where there was insufficient equipment capacity, in which case appropriate repeat analyses were performed.

Statistical analysis
All data were processed and analyzed by Microsoft Office Excel 2010 and SPSS 19.0 statistical software. Results were tested for statistical significance by one-way ANOVA. Differences were considered statistically significant at the p 0.05 level.

Results and Discussion
3.1 The kernel moisture, kernel yield and oil content under different harvesting times The quality and economic indicators of T. grandis fruits at different harvest timepoints were determined Table 1 . The results showed that the kernel moisture content, kernel yield and oil content of T. grandis were significantly different p 0.01 , which indicated that the fruit maturity had a substantial impact on the yield of T. grandis nut and its oil. The kernel moisture content decreased from 41.98 0.07 on Aug. 25 to 35.73 0.09 on Sept. 12, and the moisture content further decreased to 30.41 0.04 on Sept. 18 which was more than 5 lower than the moisture value determined six days prior. The kernel yield increased by 1.71 from Aug. 25 to Sept. 6, after which time kernel yield did not significantly change p 0.05 . The non-significant differences among harvest dates suggested that the kernels had attained their maximum size before the earliest harvest date Aug. 25, 2019 .
On the contrary, kernel oil content had a substantial increase in the early stages of maturity before Sept. 6 followed by a significant increase in the latter part of maturation from 42.32 on Sept. 6 to 48.4 on Sept. 18 . These data inconsistent were inconsistent with the results obtained by other researchers about grape berry development 28,29 . Oil content is one of the important indexes to evaluate oil plants quality, and the amount of oil content directly affects economic value of the oil plants. Usually high oil content is associated with high edible quality and good storage stability. Several factors such as cultivar, species, growing conditions, maturity, and extraction method affect plant oil yield 28 . Previous studies have shown that the crispy taste of T. grandis nuts is positively correlated with its oil content 3,30 . Oil accumulation in fruits increased consistently with postponement of harvest time. This increase may be attributed to continued activity of the triglyceride-forming biosynthesis pathway until the fruit reaches full maturation.

Fatty acid pro le of nut oil on different harvest date
The fatty acid profile of oil is its most useful chemical property. The primary fatty acid compositions and contents of oil from the T. grandis nut at the different harvest times is shown in Table 2. The results indicate that T. grandis nut oil was rich in unsaturated fatty acid UFA nearly 90 of the total fatty acids , especially linoleic acid and oleic acid. High oleic-linoleic oils are of interest because of their superior stability and nutritional importance. Linoleic acid is recognized as one of the most significant polyunsaturated fatty acids PUFA in human food because it has preventive effects on heart diseases 31,32 . T. grandis nut oil also contained higher proportion of cis-5,11,14-eicosatrienoic acid above 8 . In the predominating fatty acid, the oleic acid content increased p 0.05 with a delay of harvest time from 30.76 0.41 on Aug. 25 to 34.85 0.45 on Sept. 18. Both linoleic acid and cis-5,11,14-eicosatrienoic acid contents decreased between Aug. 25 and Sept. 18, 2019, but there was no significant difference p 0.05 , and there was a negative correlation between the content of oleic acid and linoleic acid. Research has shown that oleic and linoleic acid content are related to maturity or the degree of filling 33 , and the content of linoleic acid decreases and the oleic acid increases during seed development 34 . Between Aug. 25 and Sept. 18, 2019, palmitic acid and cis-11,14eicosadienoic acid decreased slightly, while stearic acid increased slightly, but these changes were not significant p 0.05 . The amounts of fatty acids in the T. grandis nut oil were in the order of PUFA monounsaturated fatty acid MUFA saturated fatty acid SFA for all five harvesting times, this difference was a result of higher levels of linoleic acid and the lower amount of oleic acid in the T. grandis nut oil sample. Total UFAs increased with postponement of harvest time, MUFA were not significant at the beginning of the ripening process of the T. grandis fruit Aug. 25th Aug. 31st and significant increase at the end last three points of harvesting, Sept. 6th to Sept. 18th . This is consistent with the results of the author s research group on camellia oil 35 .
The PUFA constituted over 50 of the oil, while the MUFA content was relatively lower, ranging from 31.34 0.41 to 35.44 0.04 . The SFA content was low about 11 and did not show an apparent change p 0.05 during ripening. The dietary fats abundant in PUFA can prevent chronic health conditions such as atherosclerosis, coronary heart disease, and high blood pressure. Furthermore, cis-5,11,14-eicosatrienoic acid is a special fatty acid also known as sciadonic acid, which often is found in the leaves and seeds of Gymnosperms, its name comes for its rich in a plant of Sciadopitys verticillata golden pine in Japan 36 . Previous research showed that sciadonic acid presents a strong role in lipid regulation in rats. Moreover, sciadonic acid could inhibit the activity of fatty acid synthase and 6-phosphate glucose dehydrogenase in liver and plasma of rats 37,38 . Thus, based on the contents of PUFA and sciadonic acid, the oil from T. grandis nuts is a higher nutritional and functional edible oil, and has a great value for exploitation and utilization. The fatty acid composition of seed oils varies widely among different plant species, and often the occurrence of unusual fatty acids can be used for the differentiation of particular plant families 39 . Therefore, sciadonic acid can also be used as an essential fatty acid to distinguish T. grandis oil from other plant oils.

FFA and peroxide value on different harvest time
FFA and peroxide value are important indexes to evaluate the quality of oil plants. The amounts of FFA and peroxide value in T. grandis nut oil at different fruit maturity times were analyzed Fig. 1 . The effect of harvesting time on the FFA and peroxidation value were significant p 0.05 , with both factors declining with delayed harvesting time. This result was consistent with previous research on camellia seeds 17 . The FFA and peroxide value of T. grandis sharply decreased from 0.39 0.06 mg/g and 2.92 0.30 g/100 g at first harvest to 0.23 0.02 mg/g and 0.70 0.41 g/100 g at the last harvest, respectively. Combined with the fact that riper T. grandis fruits have higher oil content, we could draw the conclusion that more FFA in cell tissues are used to synthesize triglyceride as T. grandis fruit matures, which leads to the decrease of FFA This result is consistent with prior study 40 . Also, ripening fruits contained more antioxidants that entered the oil, thereby reducing the oxidation of the oil to some extent 41 .

Protein, soluble sugar and starch contents on different harvest times
Protein is one of the main nutritional components in T. grandis nuts, and starch and soluble sugar are the standard indexes to evaluate the quality of oil seed. With the T. grandis fruit maturing, the protein content in its nut increased p 0.05 from 23.3 0.08 on Aug. 25 to 25.3 0.12 on Sept. 18. As shown in Fig. 2, there were no significant differences in terms of starch and soluble sugar amounts for different harvest times p 0.05 . The soluble sugar content was 4.24 0.03 on Aug. 25, and then it increased to 6.06 0.32 on Sept. 18 with the delay of harvest time. On the contrary, the starch content dropped from 5.95 0.21 on Aug. 25 to 5.50 0.07 on Sept. 18. It was interesting that the increased in soluble sugar content coincided with a decline in starch content during ripening; the two factors were negatively correlated r 0.8773, 95 CI 0.718 1.000, p 0.05 , which suggesting that the increased soluble sugar mainly came from the hydrolysis of starch.

The composition and content of tocopherol on different harvest times
Tocopherol is an important natural antioxidant, which can effectively block free radical chain reactions, prevent and reduce lipid peroxidation damage of cell membranes to protect cell membranes, remove free radicals and prevent cancer in vivo 42 . It is present in plants in four forms which possess different molecular structures and functions: α-, β-, γand δ-, of which α-tocopherol has the strongest physiological activity.
The tocopherol composition and contents of T. grandis nut collected at the different maturation dates are presented in Table 3. α-tocopherol and β-tocopherol were identified from T. grandis nut, the content of β-tocopherol more than 100 mg/100 g was about four times higher than  α-tocopherol when concentrations are above 23 mg/100 g , but no γ-tocopherol and δ-tocopherol had been detected, which is consistent with previous study 4 . The total tocopherol contents gradually increased from 127.5 0.04 mg/100 g on the first date to 137.1 0.02 mg/100 g on the last date. The levels of α-tocopherol and β-tocopherol increased with ripening, and reached the highest values at Sept. 18, which increased 11.51 and 7.28 from the first fruit picking respectively. Besides the influence of the harvest time, also other factors should be taken into account. For instance, in acorns, genotype and the year have a significant impact on the concentration of tocopherols 43 . Commonly, plant seeds are rich in homologues αand γbut little or no βand δ-44 , but T. grandis nut oil is rich in β-tocopherol. There was rarely reported about the predominance of β-tocopherol on other plants, only in last decade some meaning reports in acorns 43 , coffee beans 44 , wheat germ 45 , apple seeds 46 , European cranberrybush seeds 47 , kerkir seeds 48 had been published. Therefore, in T. grandis β-tocopherol is a main tocopherol homologue which can serve as a characteristic indicator of T. grandis nut oil. It can be used as an unconventional natural source of the rare tocopherol homologue.

The reducing capacity on different harvest times
The antioxidant activity of a substance is directly related to its reducing capacity. The stronger the reducing capacity is, the stronger the antioxidant activity. Therefore, the antioxidant activity of a substance can be illustrated by measuring its reducing capacity. Figure 3 lists the reducing capacity in the T. grandis nuts harvested at the five harvest dates. The results showed that the reducing capacity initially increased reaching a peak value of 17.87 0.03 mg/g on Sept. 6, then declined gradually during the later ripening stages, though values were still higher than the initial value. With the harvest time delay, the color of T. grandis aril changed from light red to crimson which might be related to oxidation of phenols; phenols were oxidized to quinones which could be polymerized to produce colored substances that caused tissue browning 49 . The degree of browning is usually positively correlated with the content of phenols. This falling of reducing capacity at the late harvest date was mainly attributed to the decline in astringent substance, for example, phenolic substances occurring as part of the physiological maturation of the fruits 18 . In addition, the content of phenolic substances in fruits is closely related to the edible qualities of fruits such as texture, fragrance and astringent and bitter taste 50 . The higher level of phenolics could bring strong astringency, which affected the taste and flavor of the T. grandis nuts. But at the same time, phenolic compounds can raise the stability, sensory and health properties of oil, the amount of which was an important factor affecting the quality of oil 20 . It is known that the content of antioxidant compounds in plant materials such as phenolics varies greatly with, for example, processing, genotype, harvest time, growing environment and conditions.

Effects of harvest time on free amino acids
Amino acids are the primary composition unit of protein and the decomposition product of protein, which are also one of the components of fruit quality and participate in the synthesis of other quality characteristic components and flavor substances 51 . Since the content of free amino acids is closely related to food processing and food flavor, we studied the composition and content of free amino acid in the T. grandis nut. Table 4 shows that there were 17 free amino acids in the T. grandis nut, the total content of  these free amino acids is more than 12 , among which the contents of tyrosine, phenylalanine, glutamate and valine were higher than 1 . Humans require seven types of amino acids that we cannot naturally synthesis: threonine, valine, methionine, leucine, isoleucine, phenylalanine and lysine. In our analysis, tyrosine content was highest in the T. grandis nut, while the content of histidine was the lowest. With the delay in harvesting time, the content of all essential amino acids except isoleucine and leucine showed an upward trend. Amino acid compositions are nutritionally important and are influence the taste of plant fruits 52 . Amino acids which have flavor properties can be divided into sweet e.g. glycine, alanine, serine, threonine, proline, histidine and glutamine , bitter e.g. valine, leucine, isoleucine, methionine, tryptophan and arginine and delicious e.g. lysine, glutamine, aspartic acid and asparagine and aromatic amino acids e.g. phenylalanine, tyrosine and cysteine 53 . As shown in Fig. 4, among the flavor amino acids in T. grandis nuts, the content of sweet amino acids and aromatic amino acids were relatively high more than 30 and 20 of the total free amino acids, respectively . Both the sweet and aromatic amino acids showed rapid increases during the early stage from 4.42 and 2.75 on Aug. 25 to 6.79 and 5.68 on Sept. 18, respectively. By contrast, the content of bitter amino acids decreased with the delay of picking time and drop more slowly early from Aug. 25 to Sept. 6 , since then it has fallen rapidly to 1.56 . These results indicate that a late harvest times is beneficial to the flavor of T. grandis nuts.
These results were in agreement with previous research that there was an improvement in flavor quality of pecans after they reached full size, but that the flavor was subject  to impairment, mostly through the onset on oxidation, staleness and rancidity 54 .

Conclusion
The study showed that T. grandis nuts were a rich source of tocopherols, mainly homologue β-tocopherol. which was a unique finding, due to previous studies often reporting low abundance of this homologue in other plant materials. Therefore, T. grandis nut can be used as an unconventional natural resource of the rare tocopherol homologue.
Based on the results of different harvest time in this study, we concluded that the kernel yield of T. grandis had no clear change in later harvesting. With the exception of moisture, protein, and starch contents, the content levels of most analytes increased with prolonged ripening time, reaching their highest values on Sept. 18. The acid values and peroxide values declined with prolonged ripening time, and reached their lowest values at Sept. 18. This showed that T. grandis kernels were no longer gaining weight, but their nutrients were still accumulating, and at that point, their nutrition and quality indicators were high. Therefore, prolonged ripening is helpful to the accumulation of dry matter and nutrients in T. grandis nuts. This results are helpful for the choosing suitable harvest time for T. grandis fruit to maximize their beneficial effects for human health.

Conflicts of Interest
The authors have no conflicts of interest to declare.