Chemical Fingerprint Analysis and Content Determination of Horned Gallnut and Bellied Gallnut in Galla Chinensis

To establish an objective and comprehensive methodology to analyse the connections and differences between horned gallnut (HG) and bellied gallnut (BG) in Galla Chinensis (GC). The HPLC fingerprints from 15 batches of HG and 15 batches of BG were performed, and chemometric analysis including similarity analysis (SA), hierarchical clustering analysis (HCA), principal component analysis (PCA), and orthogonal partial least squares discrimination analysis (OPLS-DA) was also set up. The results showed that the similarity of all batch samples was more than 0.9. In fingerprint analysis, 8 distinct common peaks were detected, among which gallic acid (GA), 1,3,6-tri-O-galloyl-β-D-glucose (TGG), and 1,2,3,4,6-O-galloyl-D-glucose (PGG) were identified by comparing with the standard compounds. Meanwhile, samples were clearly grouped into two classifications corresponding to HG and BG. This study demonstrated that HPLC fingerprints coupled with chemometric analysis could be applied to discriminate HG and BG and evaluate the qualities of HG and BG rapidly, which provided a certain experimental basis for the selection of GC raw materials and subsequent use.


Introduction
Galla Chinensis (GC), also known as wubeizi in Chinese, is a gall caused by the aphids on the Rhus leaves of the Anacardiaceae family (mainly Rhus chinensis Mill, Rhus potaninii Maxim, and Rhus punjabensis).Tese three original plants are mainly distributed in China, Japan, Korea, and other regions at an altitude range of 350-2700 m [1].In the complex process of GC formation, the presence of the gall aphid and these three host plants are indispensable, which is the main reason why relatively few countries produce GC.It was frst recorded as a medicine in the ancient Chinese medicine book Bencao Shiyi which was written in the Tang Dynasty and then was recorded for the treatment of lung defciency, lung heat, phlegm cough, diarrhea, night sweat, bloody stool, hemorrhoid, and traumatic hemorrhage in Compendium of Materia Medica, an ancient Chinese medical book that was written by Li Shizhen in the Ming Dynasty [1,2].GC contains about 50-70% hydrolysable tannins which have a variety of biological effects, including antioxidant, antibacterial, anticancer, and antiviral, and is widely used in medicine, commercial inks, leather tanning, and chemicals, of which in leather tanning, GC has a softening and toughening efect [3][4][5][6][7][8].GC is mainly distributed in most parts of China (except Xinjiang and Qinghai) and other East Asian countries, of which China accounts for more than 95% of the world's total production and has long-term been exported to Europe, America, Japan, and other developed countries [9].
Based on shape, GC can be categorized into two groups: horned gallnut (HG) and bellied gallnut (BG), also known as jiaobei and dubei in Chinese.Although the output of HG is higher than that of BG, the quality of BG is better than that of HG [9,10].In the measurement of wall thickness, it was found that the wall thickness of BG was signifcantly greater than that of HG, giving BG a frmer appearance, which is probably the main reason why traditional BG is superior to HG [10].Te GC formed on the leaves of Rhus chinensis Mill is called HG, and the acquisition time is mostly concentrated in September to October, while the GC formed on the leaves of Rhus potaninii Maxim and Rhus punjabensis J. L. Stewart ex Brandis is called BG, and the acquisition time is centered around May to June [1].According to Chinese Pharmacopoeia (2020 edition), HG is rhomboid with irregular obtuse-angled branches and more prominent pilose, while BG is oblong or fusiform without bulge or branches.According to statistics, GC can be produced in many provinces of China, of which Guizhou, Sichuan, Hubei, Hunan, Shanxi, and Yunnan account for more than 90% of the country's total production [11,12].Te HG is mainly distributed in the south of the Changjiang River, accounting for about 75% of China's total production, while the BG is mainly distributed in the north of the Changjiang River, accounting for about 20% of the total production [13].Although HG and BG are very diferent in appearance and formation process, HG and BG are collectively known as GC, and there is no clear classifcation in use.Traditional Chinese medicine generally possesses complex compositions, and the formation and content of internal chemical composition are infuenced by a variety of factors.Te choice of raw materials also afects the composition of the extract and may change its chemical properties and biological activity [14].Terefore, this study will investigate the diferences and connections of HG and BG in their internal chemical composition, providing a certain experimental basis for the subsequent use and the selection of raw materials of GC.At present, there are fewer studies on the chemical fngerprinting of GC, most commonly on the gall parasitic process of GC, while the simultaneous study of HG and BG chemical fngerprints has not been reported [15,16].
Chromatography fngerprint is a comprehensive and quantifable analysis method that can elucidate the complexity and relevance of components [17].Te most commonly and conveniently used detection methods for fngerprint establishment are chromatography and spectrometry, among which high-performance liquid chromatography (HPLC) has the advantages of high separation, fast analysis speed, good selectivity, and wide application range [18][19][20].HPLC is based on the principle that substances reach equilibrium through multiple partitioning between two liquid phases and uses the diference in partition coefcients of the components in the two phases for separation, preparation, and collection.It is widely used for the analysis and separation of various components.With the development of computer software, chemometric analysis is becoming more and more important in the discrimination of diferent Chinese medicine [21,22].Chemometric analysis can rapidly and accurately diferentiate samples that contain virtually identical compounds by processing HPLC fngerprint data.Terefore, the method could be chosen to study the diferences and connections of chemical composition between the HG and BG, which could develop the medicinal value and expand its applications, providing a wider and more efective choice of raw materials.

Materials and Reagents.
A total of 15 batches of HG (HG210601-HG2106015) and 15 batches of BG (BG210601-BG2106015), named HG1-HG15 and BG1-BG15, respectively, were collected from Jiangsu Province, China.According to their morphological characteristics, all voucher specimens were taxonomically identifed by Associate Professor Jie Zhang (Teaching and Research Department of Chinese Medicine Identifcation, Yunnan University of Chinese Medicine).Te morphology can be found in Figure 1.

HPLC Instrumentation and Chromatographic Conditions.
HPLC fngerprint analysis was operated on an Agilent 1200 high-performance liquid chromatograph (Agilent, Santa Clara, CA, USA), and all samples were separated on an Agilent Zorbax C18 column (150 × 4.6 mm, 5 μm, Agilent, Santa Clara, CA, USA).Te mobile phase was composed of 0.1% phosphoric acid-water (A) and methanol (B).Te condition used for the gradient program was developed as follows: 0-6 min,

Preparation of the Sample and Standard Solutions
Preparation of Standard Solutions.Te reference compounds were accurately weighed and dissolved in methanol to prepare the stock solutions.Teir concentrations were as follows: GA (1.12 mg/mL), TGG (0.88 mg/mL), and PGG (1.78 mg/mL).All the solutions were stored in the refrigerator at 4 °C before use.
Preparation of Sample Solutions.15 batches of each category were cracked, and then the gall wasps in the shell were removed, powdered, and passed through an eighty-mesh sieve.0.5 g of powder that was accurately weighed was refuxed with 20 mL of water at 90 °C for 2.5 hours.Te extraction was fltrated and the residue was washed with water.Te fltrate was collected and transferred to a rotary evaporator to recover the solution water at 50 °C.Methanol was added to dissolve and transferred to a 50 ml volumetric fask to volume.After fltrating through a 0.45 μm membrane, 10 μL of the obtained solution was injected into the HPLC system for analysis.

Validation of Analytical Method
3.1.1.Precision, Stability, and Repeatability.According to the guidelines of the methodology, the system precision, stability, and repeatability were determined to verify the feasibility of the HPLC fngerprint method.
In order to test the precision of the method, the same sample was continuously measured six times.Te stability experiment was carried out by measuring the same sample in 0, 4, 8, 12, 16, 20, and 24 h, respectively.To confrm the repeatability of the method, six replicate samples were prepared by using the same sample preparation procedure.Te relative standard deviations (RSDs) of the relative retention time and relative peak area were calculated separately.Te results are shown in Table 1, indicating that the analytical method was repeatable, the sample was stable, and the instrument had good precision.

Chromatography Fingerprint Analysis of Samples.
Te HPLC fngerprints from 15 batches of HG and 15 batches of BG were established by Similarity Evaluation System for Chromatography Fingerprint of Traditional Chinese Medicine (version 2012A, Chinese Pharmacopoeia Commission, Beijing, China).Peaks that were existed in all sample chromatograms with reasonable heights and good resolutions were assigned as "common peak."Te time window was set to 0.1 s, and the calibration method was multipoint calibration.Te reference chromatogram fngerprint was generated by using the average method.As shown in Figures 2 and 3, there were 8 distinct common peaks in the HPLC fngerprints, three of which (peaks 2, 6, and 7) were identifed as GA, TGG, and PGG, respectively, by comparing retention times with the standard compounds.
Te similarity of chromatographic fngerprint data is indicated by the correlation coefcient.Te value of the correlation coefcient is close to 1.0, indicating that the diferent samples there have high similarity.On the contrary, a low correlation coefcient indicates a poor mathematical International Journal of Analytical Chemistry quality for identifying the relationship between diferent samples.As shown in Table 2, the similarity of the samples between HG and BG was larger than 0.983 and 0.961, respectively, indicating that HG and BG had good similarity and shared similar chemical components.Te results confrmed that the fngerprints established in this study were reliable in assessing the quality of HG and BG.However, similarity analysis (SA) could not give more information for the connections and diferences between HG and BG.Terefore, in order to analyse the inner quality of HG and BG more precisely, it is necessary to perform subsequent chemometric analysis, including hierarchical clustering analysis (HCA), principal component analysis (PCA), and orthogonal partial least squares discrimination analysis (OPLS-DA).

Hierarchical Cluster Analysis of Samples (HCA).
To assess the connections and diferences between HG and BG, the HCA was performed by SIMCA-P14.1 (Umetrics, Umea, Sweden), and the Euclidean distance was used to measure the closeness between the samples.HCA is a multivariate analysis method that displays complex raw data in a visual form and provides classifcation information for test samples [23].In this study, the relative peak areas of 8 common peaks from the 15 batches of HG and 15 batches of BG were used to form a matrix.As shown in Figure 4, the tested samples could be approximately categorized into two groups.Te left of the dendrogram consisted of HG samples, whereas the BG samples were distributed on the right.From the results of  8) S14 ( 8) S13 ( 8) S11 ( 8) S10 ( 8) S9 ( 8) S8 ( 8) S7 ( 8) S6 ( 8) S5 ( 8) S4 ( 8) S3 ( 8) S2 ( 8) S1 ( 8) R ( 8) S15 ( 8) S14 ( 8) S13 ( 8 4 International Journal of Analytical Chemistry cluster analysis, there were obvious diferences between HG and BG, which might be related to the diferent formation processes of HG and BG.

Principal Component Analysis of Samples (PCA).
PCA was required to assemble the original variables into a new set to further explore the homogeneity and quality of HG and BG.PCA is a useful approach to efciently reduce the original high-dimension data into low-dimension data without much information loss and can be used to investigate the interrelationships between multiple variables [24].Te relative peak areas of 8 common peaks from 15 batches of HG and 15 batches of BG were imported into the SIMCA-P software to obtain the scores and loadings of multivariate analysis.As shown in the score plot of PCA (Figure 5(a)), the distribution distances of the samples represented the similarities and diferences between these   International Journal of Analytical Chemistry samples.Obviously, all samples were clearly grouped into two classifcations corresponding to HG and BG, which refected the diferences in the chemical profles between the HG and BG.In conclusion, PCA could be used to discriminate the HG and BG.Te result of PCA was similar to that of HCA.

Orthogonal Partial Least Squares Discriminant Analysis of Samples (OPLS-DA)
. OPLS-DA is a supervised model in which the users give the identity of each group sample in order to attain maximum variance of the groups in the hyperspace.It divides the systematic variation in the X matrix into two distinct parts: Y-predictive block and Y-    6 International Journal of Analytical Chemistry orthogonal (Y-uncorrelated) block to improve the quality of the model [25].Tree main parameters R 2 X, R 2 Y, and Q 2 are often used to assess the performance of the model.If the values of R 2 X and R 2 Y are close to 1.0, this suggests that the method is very suitable.Generally, the value of Q 2 greater than 0.5 is acceptable, and the diference between R 2 and Q 2 values should be less than 0. To further confrm whether there were diferences between HG and BG and identify the characteristic components which have a signifcant infuence on the chemical profling of them, OPLS-DA was established with SIMCA-P14.1.Te results are shown in Figure 5(b).Te values of R 2 Y and Q 2 were 0.996 and 0.965, respectively, indicating that the model had good stability and predictability.Te OPLS-DA score scatter plot showed that all samples could be classifed into two groups, corresponding to HG and BG.Te OPLS-DA model could efciently diferentiate between HG and BG.Te results were consistent with HCA and PCA, suggesting that these three methods could be used to distinguish between HG and BG.In addition, in order to further test the validity of the model and the diferences between the samples, a permutation test of 200 iterations of the established OPLS-DA model was performed.As shown in Figure 5(c), the R 2 Y and Q 2 Y values of the original OPLS-DA models were still signifcantly higher than the corresponding values of the permuted models.In addition, the R 2 Y intercept and Q 2 Y intercept for the established OPLS-DA models were less than 0.3 and 0.05, respectively.Tese results showed the validity of the model.
Te variable importance plot (VIP) produced by OPLS-DA gave a good reference for identifying variables that had a signifcant impact on the classifcation.Te VIP value is usually used to explain the contribution of a variable to the model, and the variable that VIP value is greater than 1 is considered to be important [26].As shown in Figure 5(d), four components with VIP >1 including chromatography peaks 7 (PGG), 8, 2 (GA), and 4 were selected as potential chemical markers, which could be used to diferentiate between HG and BG.Peaks 8 and 2 could not be identifed because of experimental limitations, which also was the shortcoming of this experiment, and subsequent studies will be combined with mass spectrometry to help identify more compounds in HG and BG.PGG has shown strong biological and pharmacological activities in antiviral, anticancer, anti-infammatory, antimicrobial, and antidiabetic [27].Tere are diverse scientifc reports on the biological and pharmacological activities of GA, focused on antioxidant, antimicrobial, anti-infammatory, anticancer, cardioprotective, gastroprotective, and neuroprotective efects [28].Our group has also related research on PGG and GA in the treatment of diabetes and found that PGG and GA could inhibit islet β-cell apoptosis in a high glucose state [29][30][31].

Method Validation of Quantitative Analysis
(1) Linearity, Limits of Detection (LOD), and Limit of Quantifcation (LOQ).A series of calibration curves were constructed between the peak areas (Y) and the concentrations of reference substances (X) in the investigated ranges.Te limits of detection (LOD) were calculated with S/ N � 3, and the limit of quantifcation (LOQ) was calculated with S/N � 10.Te correlation coefcient values were more than 0.9993 at a series of gradient concentrations, which indicated that there was a satisfactory correlation between the concentrations and peak areas of the three compounds at a relatively wide range of concentrations.Te results are shown in Table 3.
(2) Precision.Te precision check was performed by continuously injecting 6 times according to the same chromatography conditions.Te RSD values for the peak areas of GA, TGG, and PGG were 1.06%, 2.28%, and 1.34%, respectively, indicating that the instrument had good precision.
(3) Stability.Sample stability test was performed by running the same sample at 0, 4, 8, 12, 16, 20, and 24 h.Te RSD values for the peak areas of GA, TGG, and PGG were 3.18%, 1.11%, and 1.24%, respectively, indicating that the test solution was stable within 24 h.
(4) Repeatability.To confrm the repeatability of the method, six replicate samples were prepared by using the same sample preparation procedure.Te RSD values for the peak areas of GA, TGG, and PGG were 2.81%, 1.49%, and 3.78%, respectively, indicating that the method was reproducible.
(5) Recovery.In the recovery test, an appropriate amount of sample was weighed and spiked with a known amount of each standard compound.Ten the sample was treated and analysed as previously described.Te results showed that the average recovery of GA, TGG, and PGG was 100.13%, 98.04%, and 99.41%, and the RSDs were 1.39%, 1.82%, and 1.09%, respectively, which showed that the experimental method was suitable.

Content Determination of Samples.
Te established HPLC method was used to determine the contents of GA, TGG, and PGG in 15 batches of HG and 15 batches of BG.As shown in Figure 6 and Table 4, found that the content of TGG was the lowest in HG and BG compared with GA and PGG.GA and PGG were used as potential compounds to diferentiate HG and BG.Te GA content in BG was signifcantly higher than HG (P < 0.05), while the contents of TGG and PGG in BG were signifcantly lower than HG (P < 0.001).If only the diference in GA content between HG and BG were analysed, this result was consistent with the traditional belief that "bellied gallnut is superior and horned gallnut is inferior."However, the evaluation of the quality of International Journal of Analytical Chemistry Based on the results of OPLS-DA, PGG was identifed as an important chemical marker for HG and BG, and its content was higher in HG than in BG.Terefore, the quality of HG and BG should be evaluated from multiple aspects, perspectives, and components rather than a single component, which provided some reference for the selection of GC raw materials.

Conclusions
Tis study is diferent from previous studies, which only focused on the HPLC fngerprint of GC or content determination of some compounds in HG or BG [8,32].In this study, we analysed the connections and diferences between HG and BG by analyzing diferent batches of HG and BG by SA, HCA, PCA, OPLS-DA, and content determination of samples, which provided a certain experimental basis for the selection of GC raw materials and subsequent use.Te results indicated that fngerprint combined with chemometric analysis was a powerful and practical method to objectively and rapidly diferentiate between HG and BG.However, this paper only analysed the connections and diferences between HG and BG from the perspective of chemical fngerprint analysis.In the follow-up, our group will further investigate HG and BG in more depth from metabolomics, transcriptomics, rat models, and clinical treatments so as to more comprehensively observe the diferences between HG and BG.

Figure 2 :
Figure 2: HPLC fngerprints of 15 batches of HG (a) and 15 batches of BG (b).Te X-axis represents the retention time (t/min) and the Yaxis represents the response value (mAU).

Figure 3 :
Figure 3: Representative chromatograms of HG (a) and BG (b), and HPLC chromatogram of mixed reference solution (c).Peaks 2, 6, and 7 were GA, TGG, and PGG, respectively.Te X-axis represents the retention time (t/min).Te Y-axis represents the response value (mAU).Te peaks 1-8 were the main common peaks.

3 .
In this research, OPLS-DA was conducted to obtain three important parameters: the diagram of the score scatter plot, the diagram of the OPLS-DA model replacement verifcation, and the variable importance plot (VIP).Te results are shown in Figures 5(b)-5(d).

Table 1 :
Te relative standard deviation (RSD) results of methodology validation.

Table 2 :
Te similarity values for 15 batches of HG and 15 batches of BG.
Figure6: Te contents of GA, TGG, and PGG in HG and BG (mg/g).Compared with the HG, the contents of GA, TGG, and PGG had signifcant diference ( * p < 0.05).

Table 4 :
Te contents of GA, TGG, and PGG in HG and BG (n � 3) (mg/g).International Journal of Analytical Chemistry traditional Chinese medicine is not comprehensive if it is based only on a single component.