Comparison of phenolic prof iles and antioxidant activities in skins and pulps of eleven grape cultivars (Vitis vinifera L.)

LI Fu-xiang, LI Fu-hua, , YANG Ya-xuan , YIN Ran, MING Jian ,

1 College of Food Science, Southwest University, Chongqing 400715, P.R.China

2 Research Center of Food Storage & Logistics, Southwest University, Chongqing 400715, P.R.China

3 Ernest Mario School of Pharmacy, Rutgers University, NJ 008854, USA

Abstract Eleven grape cultivars were analysed to explore the variety differences of fresh grape phenolic prof iles. The results showed that free phenolics were predominant in grape skins and pulps, and showed the higher antioxidant activities than bound. In 11 cultivars, Muscat Kyoho extracts had the highest total phenolic content in skins (10.525 mg GAE g-1 FW) and pulps (1.134 mg GAE g-1 FW), and exhibited the highest DPPH radical scavening capacity (EC50=11.7 μg mL-1) and oxygen radical absorbance capacity (ORAC) value (190.57 μmol TE g-1 FW) of free phenolic in skin. In addition, the most abundant phenolics in grape skins were found to be f lavonoids such as kaempferol in Kyoho skin (541.2 μg g-1 FW), rutin, catechin and epicatechin in Muscat Kyoho skin (262.3, 86.3 and 70.0 μg g-1 FW, respectively). Furthermore, the principal component analysis showed a strong difference of phenolic prof iles with the cultivars, existing forms and distributions. Pearson correlation coeff ciient analysis showed a signif icant linear correlation between total phenolic content and antioxidant activity (P＜0.05). Therefore, both skins and pulps were rich sources of bioactive phenolic compounds, and Muscat Kyoho was the ideal source among all samples.

Keywords: grape phenolics, varietal diversity, antioxidant activity, principal component analysis

1. lntroduction

Grape (Vitis vinifera L.) ha s a long history of cultivation and consumption. Fresh grape is rich in nutrients (dietary f iber, minerals, etc.) and bioactive compounds (phenolic acids, f lavonoids, isof lavonoids, thiols, carotenoids, ascorbic acid, tocopherols, etc.) (Yang and Xiao 2013; Fabani et al. 2017). Phenolic compounds as one kind of phytochemicals have various nutritional and healthy properties, such as antibacterial, anti-inf lammatory, hypoglycemic, and hypolipidemic activities (Kemperman et al. 2013; Jara-Palacios et al. 2014; Shahidi and Ambigaipalan 2015). Extensive literatures investigated the phenolic content and antioxidant capacity of grape products, especially wine and raisin (Jacob et al. 2008; Fang et al. 2010; Williamson and Carughi 2010; Meng et al. 2011; De Castilhos et al. 2017). However, information on the phenolic prof ile of fresh grape is scarce.

The phenolic prof iles of grapes depend on various factors such as variety, maturity (Fanzone et al. 2011), genetic diversity (Bustamante et al. 2017), viticulture practices (De Pascali et al. 2014), soil characteristics (Cheng et al. 2015), environmental st r ess and vine health status (Rusjan et al. 2012). The phenolic composition of grapes strongly depends on grape varieties (Yang and Xiao 2013; Aubert and Chalot 2018). The (+)-catechin content of cv. Muscat de Hambourg (19.3 mg kg-1fresh weight (FW) of samples) was higher than that of cv. Italia (3.1 mg kg-1FW), the caftaric acid content of cv. Alphonse Lavallée (93.8 mg kg-1FW) was higher than cv. Centennial Seedless (25.6 mg kg-1FW) (Aubert and Chalot 2018). The grape quality and sensory properties are signif icantly affected by grape varieties (Aubert and Chalot 2018). Therefore, it is essential to study the inf luence of grape varieties on the phenolic prof iles.

Distribution of phenolic compounds in grape is uneven. About 64% of total free phenolic compounds are in the seeds, 30% are in the skin, and 6% are in the pulp. Phenolic compounds in the seeds, skin and pulp are represented by f lavan-3-ols, f lavono ls and hydroxycinnamic acids, respectively (Teixeira et al. 2013; Yilmaz et al. 2014).

The phenolic compounds in fruits are in both free and bound forms. Bound phenolics, are mainly associated with the cell wall material, and cannot be digested through the stomach and small intestine but reach the colon completely, where they are released to exhibit bioactivity with health benef its. Most of studies investigated the free phenolics of grape (Sun et al. 2002; Teixeira et al. 2013; Yilmaz et al. 2014; Belviso et al. 2017; Aubert and Chalot 2018). The bound phenolic prof iles of grapes still remain unclear.

The objective of this study is to: (1) investigate the composition and distribution of free and bound phenolic compounds, as well as their antioxidant activities in skins and pulps of 11 grape cultivars; (2) evaluate the correlation between the phenolic compounds and antioxidant activities by the principal component analysis (PCA) and the Pearson correlation coeff icient method.

2. Materials and methods

2.1. Materials

1,1-Diphenyl-2-picrylhydrazyl (DPPH), 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), as well as phenolic standards were purchased from Sigma-Aldrich (St. Louis, MO, USA), and the purity of each compound was ＞98%. 2,2′-Azobis (2-methylprop ionamidine) dihydrochloride (ABAP) was purchased from Tokyo Chemical Industry (Tokyo, Japan). Acetonitrile (HPLC grade) and formic acid (HPLC grade) were obtained from Tianjin Shield Fine Chemicals Company (Tianjin, China). Other chemicals were of analytical grade.

2.2. Grape samples

G rapes with a characteristic scent were harvested in August 2017 from the Experimental Farm of Southwest University, Chongqing, China. The pictures, f lavor characteristics and other information of the collected grapes were summarized in Table 1. Fresh grapes were picked randomly distributed throughout different vines at the appropriate ripeness, and then packed in black polyethylene bags. F or each cultivar, skins (30 g) and pulps (75 g) were separately packed, and then stored in the dark at -40°C until analysis within three weeks.

2.3. Determination of the soluble solid content (SSC), titratable acidity (TA) and SSC/TA ratio

The SSC, TA and SSC/TA ratio were measured using the method described by Ali et al. (2016) with a slight modif ication. The SSC was measured with PAL-2 digital refrectometer (Atago Co., Japan) and expressed as the percent. TA was determined by the titration of grape juice against 0.01 mol L-1NaOH, and given as the percent of tartaric acid. The SSC/TA ratio was calculated by dividing SSC with corresponding TA value of the grape sample.

2.4. Extraction of free and bound phenolics

Grape skins and pulps were prepared according to the method described by Sun et al. (2002) and Wolfe et al. (2003) with a slight modif ication. Brief ly, the fresh pulp s (25 g) or skins (10 g) were extracted with 50 mL chilled acetone solution (80%, v/v), respectively. The mixture was homogenized for 10 min at 10 000 r min-1by a high speed dispersator (XHF-D; Ningbo Scientz Instrument Co., China). The supernatant was collected. The residue was re-extracted twice with the same procedures. All f iltrate was collected and evaporated at 45°C until approximately 90% of the f iltrate was evaporated by a rotovapor (RE-52AA; Shanghai Yarong Instrument Co., China). The free phenolic extracts were recovered with water to a f inal volume of 25 mL. The soluble free phenolic extracts were stored at -40°C until analysis within three weeks. The residues from the above soluble free phenolic extraction were digested with 20 mL (2 mol L-1) NaOH for 90 min with shaking in darkness. The mixture was acidif ied to p H=2 with concentrated HCl, and extracted for f ive times with 100 mL ethyl acetate. The ethyl acetate fraction was evaporated at 45°C to dryness. The bound phenolic extracts were reconstituted in 10 mL of water and stored at -40°C until analysis within three weeks.

Table 1 The grape varieties investigated

2.5. Determination of the phenolic content

The contents of free and bound phenolics in grapes were measured by the Folin-Ciocalteu colorimetric method as reported previously (Singleton et al. 1999) with a slight modif ication. Brief ly, 0.2 mL of the appropriate dilutions of extracts and 0.8 mL water were mixed. Then, 0.2 mL Folin-Ciocalteu reagent was added. After 6 min, the reaction was neutralized with sodium carbonate (2 mL, 7%, w/v), and 1.6 mL water was then added. The samples were given the resting time for 90 min in darkness. The absorbances of samples were measured at 760 nm by a UV-vis spectrophotometer (UV-722; Shanghai Jinghua Instrument Co., China). The gallic acid was the standard. The phenolic content was expressed as milligram gallic acid equivalents per gram of grape in fresh weight (mg GAE g-1FW).

2.6. Phenolic prof iles analysis by HPLC-DAD

Phenolic compounds were determined by a Shimadzu HPLC system that performed as described previously (Iii et al. 2010; Sandhu and Gu 2010). The HPLC system was consist of Shimadzu LC-20AD pump, Shimadzu SIL-20A autosampler, and a Shimadzu SPD-M20A diode array detector (DAD). The grape phenolic extracts were analysed on a Hypersil BDS C18column (5 μm, 250 mm×4.6 mm i.d.; Thermo Fisher Scientif ic, New York, USA) with the column temperature 40°C and the f low rate of 0.7 mL min-1. The gradient system included: A, water including 0.1% formic acid; B, acetonitrile. The elution program was the following: 10% B at 0-5 min; 10-40% B at 5-50 min; 40-90% B at 50-55 min; 90% B at 55-62 min; 90-10% B at 62-65 min; 10% B at 65-75 min. The injection volume was 20 μL. The wavelength range of scanning was 190-800 nm, and wavelength at 280 nm was analyzed. Phenolic compounds were identif ied by comparing the retention time in specif ic wavelength spectra with those of authentic standards.

2.7. Antioxidant activity analyses

Free radical scavenger assayThe free radical scavenging capacity of grape extracts were measured based on the method (Brand-Williams et al. 1995; Li et al. 2013) with a slight modif ication. Brief ly, 1 mL extract was mixed with the stock solution of DPPH (5 mL, 0.1 mmol L-1) freshly prepared in ethanol. After keeping in the dark at 20°C for 30 min, the absorbance was then measured at 517 nm by a UV-vis spectrophotometer (UV-722; Shanghai Jinghua Instrument Co., China). The absorbance of freshly prepared DPPH solution was measured prior to analysis. The distilled water and ascorbic acid were used as the blank and standard, respectively. DPPH antioxidant capacity of the phenolic extracts was expressed as the percent of DPPH-free radical scavening activity using the eq. (1):

The extract concentration providing 50% of DPPH free radicals scavenging activity (EC50) was calculated by plotting the scavenging activity percentage against the extract concentration. Where, Asample, Ablankand ADPPH, are the absorbance of sample, distilled water and DPPH solution, respectively.

Oxygen radical absorbance capacity assayThe oxygen radical absorbance capacity (ORAC) of extracts were measured by the method reported by Wolfe et al. (2008) with a slight modif ication. In brief, grape extracts were diluted by 75 mmol L-1phosphate buffer (p H 7.4) with the appropriate concentration (20 μL/well) to a 96-well black microplate in triplicate. The microplate was then incubated at 37°C for 10 min, and mixed with 0.96 mmol L-1f luorescein (200 μL/well). Reactions were incubated at 37°C for 20 min, then adding 119.4 mmol L-1ABAP (20 μL/well). Fluorescence intensity was recorded immediately at excitation wavelength of 485 nm and emission of 520 nm for 35 cycles every 4.5 min by Multi-Mode Microplate Reader (Synergy H1 MG; Biotek, Vermont, USA). Phosphate buffer and trolox were used as the blank control and standard, respectively. The ORAC value was expressed as μmol Trolox equivalents of one gram grape in fresh weight (μmol TE g-1FW) and calculted by the eqs. (2) and (3):

Where, AUC is the area under the f luorescence vs. time curve, ?1is the f irst f luorescence recording value, ?nis the n th f luorescence recording value, CT is the time of interval measure.

2.8. Statistical analyses

The statistical signif icance evaluation of measured differences were analyzed using one-way analysis of variance (ANOVA), the differences between means were performed by Tukey's multiple comparison test, and signif icance was def ined as P＜0.05 by the SPSS software 20.0 (SPSS Inc., Chicago, IL, USA). All data w ere reported as mean±SD for three replications, but not f ield replicates. The charts were created using the Sigmaplot software (Graph Pad Software, San Diego, CA, USA).

Correlation coeffcient analysisThe Pearson's coeff icient of correlation analysis was determined between the contents of individual phenolic compounds, free and bound phenolics determined. Correlation signif icance was def ined at the 0.01 or 0.05 level by SPSS Software 20.0 (SPSS Inc., Chicago, IL, USA).

Principal component analysisPrincipal component analysis (PCA) was applied to check for similar characteristics of samples, as well as to associate response variables by SPSS software 20.0 (SPSS Inc., Chicago, IL, USA). Data that contained 11 objects×12 or 16 variables were processed using the covariance matrix with autoscaling. Where the 11 objects were the number of samples, the variables were the phenolic compounds and total phenolic contents determined by HPLC-DAD, and antioxidant activity (DPPH and ORAC). The initial matrices were 11×16, 11×16, 11×16, and 11×12 for free and bound phenolics of skin and pulp samples, respectively.

3. Results

3.1. Physicochemical properties of the 11 grape varieties

Muscat Kyoho showed the highest soluble solid content (20.18%), Shine Muscat showed the lowest SSC (12.83%), and the SSC of other grape varities ranged from 14.35 to 18.93% (Appendix A). The titratable acidity (TA) of all the grape varities ranged from 0.301% (Red Bharati) to 0.553% (Moldova). Red Bharati exhibited the highest ratio of SSC to TA (58), and the ratio of the rest grape varities ranged from 28 (Moldova and White Olympia) to 54 (Muscat Kyoho).

3.2. Total phenolic content (TPC) of the grape skins and pulps

As shown in Table 2, most of the phenolic compounds were detected in the skins, whereas very low concentrations in the pulps. The TPC of grape skins (1.296-10.525 mg GAE g-1FW) were higher than those of pulps (0.189-1.134 mg GAE g-1FW). Muscat Kyoho had the highest TPC in its pulp and skin, while Shine Muscat had the lowest TPC value. The TPC of skins were 6.5 to 17.8 times higher than those of pulps. Free phenolic compounds were the dominant both in pulps (about 85-98% of TPC) and skins (about 78-97% of TPC).

Table 2 Phenolic contents of different parts of grape varieties (mg GAE g-1 FW)

3.3. ldentif ication and quantitation of phenolic compounds by HPLC

As shown in Table 3 (skins) and Table 4 (pulps), a total of 13 phenolic compounds were detected, including seven phenolic acids, f ive f lavonoids, and resveratrol.

Phenolic compounds in grape skinsAs shown in Table 3, the most abundant phenolics in grape skins were f lavonoids (accounts for 79% of free phenolics and 56% of bound phenolics). The f lavonoids of grapes were strongly different in grape varieties. This result was in agreement with that in Di Lecce et al. (2014). For free f lavonoids, kaempferol was the most abundant f lavonoids in Kyoho (541.2 μg g-1FW), while rutin (262.3 μg g-1FW), catechin (86.3 μg g-1FW) and epicatechin (70.0 μg g-1FW) were the most abundant in Muscat Kyoho. Compared to free f lavonoids, the concentrations of bound f lavonoids were lower. For free phenolic acids, the highest concentration of gallic acid was in White Olympia (81.8 μg g-1FW). For bound phenolic acids, gallic acid and caftaric acid were in all varieties except Beni Fuji and Hutai-8, while caffeic acid and p-coumaric acid were detected in all varieties. The most abundant bound resveratrol was in Summer Black (172.6 μg g-1FW). The contents of resveratrol were signif icantly different (P＜0.05), which was in agreement with the result reported by Okuda and Yokotsuka (1996).

Phenolic compounds in grape pulpsThe composition of phenolic compounds in grape pulps was similar to that of skins. However, the content of phenolic compounds in grape pulps was lower than that of skins. Phenolic acids were the most abundant phenolics in grape pulps (accounts for 53% of free phenolics and 54% of bound phenolics), which was in accordance with the report (Panteli? et al. 2016). For free phenolic acids, gallic acid was the most abundant in Gold Finger (44.9 μg g-1FW), and caftaric acid was the most abundant in Hutai-8 (192.3 μg g-1FW). For free f lavonoids, catechin was the most abundant in Red Bharati (134.5 μg g-1FW), and epicatechin was the most abundant in Muscat Kyoho (45.5 μg g-1FW). Resveratrol was only determined in Summer Black and Moldova in minor quantities (1.9 μg g-1FW). For bound f lavonoids, catechin was the most abundant in Red Bharati (9.8 μg g-1FW), rutin and isoquercitrin were not determined in the grape pulps. It suggested that the phenolic compounds of grape skins and pulps were different between free and bound forms.

3.4. Antioxidant activities

The DPPH radical scavenging capacityThe results of DPPH radical scavenging capacity of grape extracts were shown in Fig. 1-A and B for pulps and skins, respectively.

For skins, the free phenolics of Muscat Kyoho presented the highest radical scavenging capacity (EC50=11.7 μg mL-1), while Shine Muscat showed the lowest antiradical capacity (19.2 μg mL-1). However, for the bound phenolics, Shine Muscat presented the highest radical scavenging capacity (19.3 μg mL-1), followed by Muscat Kyoho (23.3 μg mL-1). In general, the DPPH free radical scavenging ability of free phenolics was about 1.8 times stronger than bound phenolics in skins.

For pulps, the free phenolics of Beni Fuji presented the highest antiradical capacity (EC50=14.1 μg mL-1) and stronger than ascorbic acid (EC50=17.9 μg mL-1), followed by White Olympia (EC50=17.1 μg mL-1) and Red Bharati (EC50=17.5 μg mL-1), while Shine Muscat performed the lowest scavenging capacity (EC50=33.9 μg mL-1). For bound phenolics, Moldova presented the lowest EC50value of 23.2 μg mL-1. The DPPH free radical scavenging ability of free phenolics were about 1.6 times higher than those of bound phenolics. The free phenolics and bound phenolics in skins exhibited higher antiradical capacity than these in pulps. Generally, Muscat Kyoho, Beni Fuji and Red Bharati took advantages in scavenging DPPH free radical.

The oxygen radical absorbance capacity (ORAC)As shown in Fig. 2, for skins, the average of the ORAC values of free phenolics was 72.13 μmol TE g-1FW, and 10.44 μmol TE g-1FW for bound phenolics.

Table 3 Contents of phenolic acids, selected flavonoids, and resveratrol of grape skins (μg g-1 FW)1)

For the free phenolics, the variety of Muscat Kyoho had the highest ORAC value (190.57 μmol TE g-1FW); Kyoho, Moldova, Hutai-8 were the second (103.33 μmol TE g-1FW, 93.15 μmol TE g-1FW, 87.16 μmol TE g-1FW), and Shine Muscat presented the lowest ORAC value (9.62 μmol TE g-1FW). For the bound phenolics, Beni Fuji presented the highest ORAC value (16.14 μmol TE g-1FW); Gold Finger showed the lowest (4.19 μmol TE g-1FW).

For pulps, the average ORAC values of free phenolics was 10.18 μmol TE g-1FW, and 1.50 μmol TE g-1FW for bound phenolics. For the free phenolics, Muscat Kyoho (16.24 μmol TE g-1FW), Hakuho (12.51 μmol TE g-1FW) and Hutai-8 (12.47 μmol TE g-1FW) had the highest ORAC value. There were no signif icant differences in the ORAC values of the rest eight grape pulps (P＜0.05). For the bound phenolics, Kyoho had the highest ORAC value (4.73 μmol TE g-1FW), followed by Red Bharati, Muscat Kyoho and Summer Black, and others had no signif icant differences (P＜0.05).

In general, the ORAC of free phenolics were stronger than bound phenolics both in grape pulps and skins. The free and bound phenolics in skins showed higher ORAC values than these in pulps. Muscat Kyoho showed the highest ORAC, subsequently followed by Kyoho, Red Bharati, Hutai-8 and Moldova.

4. Discussion

Consump tion of fresh fruits and vegetables rich in phenolic compounds was benef icial for human health. Grapes were rich in phenolic acids and f lavonoids (Yilmaz et al. 2014). Fresh grapes are commonly consumed with pulp, but skins and seeds were seldom eaten. Our result showed that the most abundant phenolics in grape skins were f lavonoids, and in grape pulps were phenolic acids. Similar results were reported before (Di Lecce et al. 2014; Panteli? et al. 2016). The contents of free phenolics in grapes were higher than the bound phenolics, and the total phenolic contents in the skins were higher than in the pulps. These results were consistent with the report previously (Panteli? et al. 2016). The antioxidant activities assayed by DPPH and ORAC showed similar trend to the contents of phenolic compounds. This result was in accordance with the reports (Yilmaz et al. 2014; Cosmulescu et al. 2015; Panteli? et al. 2016). Muscat Kyoho had the highest content of total phenolic compounds, and the strongest antioxidant capacity. It means that Muscat Kyoho may have the greatest health benef its among the 11 grapes.

Table 4 Contents of phenolic acids, selected flavonoids, and resveratrol of grape pulps (μg g-1 FW)1)

The PCA correlation plots were shown in Fig. 3, the main inf luential variables responsible for the clustering were identif ied using the loading plots (Fig. 4).

The phenolic compounds of grapes strongly depended on varieties, as showed in Figs. 3-A and 4-A. Compared to other varieties, Muscat Kyoho showed the highest contents of total free phenolics, free syringic acid and some free f lavon oi ds: (+)-catechin, (-)-epicatechin, rutin, isoquercitrin and kaempferol, and the strongest antioxidant activities.

The phenolic compounds of grapes were different in skins and pulps. Compared to other free phenolics in skins, free rutin, catechin, isoquercitrin, (-)-epicatechin, and kaempferol were the dominant (Figs. 3-A and 4-A). While for pulps, free phenolic acids such as p-coumaric acid and 3,4-dihydroxybenzoic acid were the dominant (Figs. 3-C and 4-C).

Fig. 1 The EC50 values of free and bound phenolics in skins (A) and pulps (B) of different grape varieties in the DPPH (1,1-diphenyl-2-picrylhydrazyl) free radical scavenging experiments (mean±SD, n=3). Values with different capital or lowercase letters within the same bar are signif icantly different at P＜0.05.

Fig. 2 The ORAC (oxygen radical absorbance capacity) values (μmol TE g-1 FW) of free and bound phenolics in grape skins and pulps (mean±SD, n=3). Values with different capital or lowercase letters within the same bar are signif icantly different at P＜0.05.

The phenolic compounds in skins and pulps were different between the free and bound forms. For free phenolics in skins, Kyoho skin was characterized with p-coumaric acid, vanillic acid and 3,4-dihydroxybenzoic acid (Figs. 3-A and 4-A). For bound phenolics in skins, Kyoho skin was characterized with phenolic acids, catechin, rutin, and resveratrol (Figs. 3-B and 4-B). For free phenolics in pulps, Kyoho pulp was characterized with gallic acid and rutin (Figs. 3-C and 4-C). For bound phenolics in pulps, Kyoho pulp showed high contents of catechin and (-)-epicatechin (Figs. 3-D and 4-D).

Fig. 3 Loading plots of free and bound phenolics in grape skins (A and B, respectively) and pulps (C and D, respectively). FPC, the total free phenolic contents determined by HPLC-DAD; BPC, the total bound phenolic contents determined by HPLC-DAD. GA, gallic acid; CAT, caftaric acid; DHB, 3,4-dihydroxybenzoic acid; VA, vanillic acid; CAF, caffeic acid; SYA, syringic acid; p-CMA, p-coumaric acid; C, (+)-catechin; EC, (-)-epicatechin; RT, rutin; IQE, isoquercitrin; KAE, kaempferol; RES, resveratrol; DPPH, the EC50 values of DPPH free radical scavening assays; ORAC, the oxygen radical absorbance capacity values.

Fig. 4 Principal component scores plots of free and bound phenolics in grape skins (A and B, respectively) and pulps (C and D, respectively).

Correlations between the total phenolic contents and antioxidant activities were analyzed, the results were shown in Appendix B. There were signif icant correlations between the contents of phenolics determined by HPLC-DAD and the ORAC values in skins (P＜0.05), the bound phenolic contents and ORAC values in pulps (P＜0.01). There was no signif icant linear correlation between DPPH and ORAC values except for the free phenolics in skins (P＞0.05).

In general, the phenolic compounds of grapes strongly depended on grape varieties and positions (skins and pulps), and free and bound forms. There were strongly linear correlation between the total phenolic contents and antioxidant activities.

5. Conclusion

Fresh grape is rich of bioactive phenolic compounds. The phenolic compounds of grapes strongly depend on grape varieties. Muscat Kyoho has the most abundant phenolics and the strongest antioxidant capacity. The phenolic compounds of grapes are mainly in skins. Flavonoids are the most abundant phenolics in skins, and phenolic acids are the most abundant in pulps.

Acknowledgements

This work was supported by the National Key R&D Program of China (2016YFD0400203) and the National Natural Science Foundation of China (31471576).

Appendicesassociated with this paper can be available on http://www.ChinaAgriSci.com/V2/En/appendix.htm