The difference in coumarin content between the roots of HBZ and QBZ in seven periods
The chromatographic peaks of eight coumarin components were well separated under the set chromatographic conditions (Fig. 1). Tables S1 ~ S5 display the parameters of linearity, precision, stability, repeatability, and recovery test. Figure 2 and Table S6 present the content changes of eight coumarin components in two varieties of A. dahurica at seven development stages. Figure 2 reveals that the eight coumarin components of the two types of A. dahurica displayed dynamic changes in the seven periods with an overall increasing trend. The mass fraction of imperatorin should be more than 0.08% based on the dried product as a quality control component specified in the 2020 edition of Chinese Pharmacopoeia. In our experiment, all 14 measured samples conformed to the regulations, and the imperatorin content was the highest of any period compared to other coumarin components. The eight coumarins can be roughly divided into three categories by comparing HBZ with QBZ from the perspective of the composition. In the first category, the overall change of imperatorin, phellopterin, and oxypeucedanin content exhibited an “S” pattern. They were lower in HBZ than in QBZ in the early developmental stage (late March and April). However, HBZ was higher than QBZ and had been in the lead composition in the late growth stage. In the second category, the overall change of isoimperatorin, bergapten, byakangelicol, and byakangelicin content exhibited an “M” pattern. They were more in HBZ than in QBZ in late May and late June, and HBZ had been in a dominant position until late July. In the third category, HBZ contained more oxypeucedanin hydrate than QBZ in seven periods. From the perspective of the development period, the total content of eight coumarins in HBZ peaked in late June, ranked second in the middle of July, and peaked in the middle of July in QBZ.
Principal component analysis (PCA) of eight coumarins in HBZ and QBZ
PCA of eight coumarin components of two varieties of A. dahurica was performed using SMICA software. Two principal components were extracted. The contribution rates of the characteristic values were 55.4% and 21.0%, respectively. The cumulative contribution rate reached 76.4%, indicating that the two principal components can represent the internal quality of A. dahurica. Figure 2B presents the biplot of eight coumarin components of HBZ and QBZ. PC1 primarily reflected the information on isoimperatorin, byakangelicol, bergapten, imperatorin, byakangelicin, and phellopterin. PC2 mainly reported the information on oxypeucedanin and oxypeucedanin hydrate. The quality difference between the two varieties of A. dahurica was evaluated using the PCA comprehensive score of two principal components. The comprehensive score was calculated as the sum of the product of the principal component factor scores and the contribution rate of the eigenvalues (Table 1). The comprehensive score of H6 was the highest, followed by H5, and Q6 ranked third.
Table 1
The comprehensive quality score of A. dahurica samples.
Sample | PC1 score | PC2 score | Comprehensive score | Rank |
H1 | -1.06717 | 0.43248 | -0.50039 | 11 |
H2 | -1.8607 | 0.15784 | -0.99768 | 14 |
H3 | 0.78783 | -2.07 | 0.001758 | 8 |
H4 | 0.49112 | 1.00693 | 0.483536 | 4 |
H5 | 1.18612 | -0.56896 | 0.537629 | 2 |
H6 | 1.56886 | 0.02822 | 0.875075 | 1 |
H7 | 0.51752 | 0.66186 | 0.425697 | 5 |
Q1 | -1.32177 | -0.58144 | -0.85436 | 13 |
Q2 | -0.86973 | -0.86773 | -0.66405 | 12 |
Q3 | -0.06719 | -1.59218 | -0.37158 | 10 |
Q4 | 0.17609 | 1.24777 | 0.359586 | 6 |
Q5 | -0.53504 | 0.68804 | -0.15192 | 9 |
Q6 | 0.69912 | 0.66314 | 0.526572 | 3 |
Q7 | 0.29492 | 0.79404 | 0.330134 | 7 |
Characterization of coumarins in A. dahurica using DESI-MSI
Figure 3 illustrates the mass spectrum of the coumarin components in A. dahurica roots obtained in positive ion mode. The ions of m/z 317.1014, 287.0910, 271.0960, 301.1067, and 271.0970 were separately identified as byakangelicol [M + H]+, oxypeucedanin [M + H]+, imperatorin [M + H]+, phellopterin [M + H]+, and isoimperatorin [M + H]+, respectively. The errors are all less than 5 ppm. Table S7 summarizes the peak assignment of these identified ions. The fragment of oxypeucedanin hydrate bound to ions was not detected because it reacted with ethanol during mass spectrometry analysis and existed as a hydrogen loss peak, while byakangelicin underwent water loss during mass spectrometric detection, and its m/z was consistent with that of byakangelicol so that byakangelicin was discarded. Bergapten has extremely strong sublimation properties, causing component shifts during the detection process, which is not conducive to subsequent DESI-MSI detection.
Spatial localization of coumarin compounds in HBZ and QBZ
Figure 4 depicts the spatial distribution of five coumarins in the root sections of HBZ and QBZ under different development stages. The spatial localization of five coumarin components in HBZ and QBZ roots is consistent. The five coumarin components of the two varieties of A. dahurica were distributed in the roots’ outer layer, namely, the cortex in late March. The five coumarin components of the two A. dahurica were distributed in the cortex and phloem from late April to late July and the coumarin components in the outer cortex had a significantly lower ionic strength than those in the phloem. In addition, it can be seen from the figure that the relative ionic strength of the five coumarins in HBZ is stronger, indicating that the coumarins in HBZ are higher, which is consistent with the content determination results by HPLC.
Growth and Development of the root secretory tract
HBZ and QBZ did not differ significantly regarding secretory tract development. The primary structural development of the A. dahurica root did not result in any secretory canal formation. Therefore, the secretory canals in the A. dahurica root should be classified as secondary structures, primarily formed where the cambium's periphery was connected with the secondary phloem (Fig. 5AB). Secretory canals are more common in the cortex and secondary phloem, their distribution is round, oval, and roughly umbrella-shaped. The image illustrates that the size of secretory canals from the secondary phloem to the outside gradually grows. The number is increasing with the growth and development of the A. dahurica root (Fig. 5C ~ F). After the secondary structure formation in the A. dahurica root, the secretory canal initially formed in the secondary phloem. The secretory canal continued to grow outward with the secondary phloem development. During this process, new epithelial cells were constantly embedded into the mature secretory canal epithelial cells, resulting in the gradual expansion of the secretory canal from the secondary phloem to its periphery. Moreover, the expansion of the secondary phloem and the constant formation of new secretory canals contribute to the increase in the number of secretory canals.
Area and quantity statistics of the root secretory tract
The root cross-section radius, xylem radius, 1/16 of the cortex area, the average area of the secretory canal, and the number of secretory canals increased as the root of A. dahurica grew, but the above indicators exhibited differences between the two varieties (Table S8, Fig. 6). Specifically, the data measured in HBZ were greater than those measured in QBZ, indicating that the root growth rate of HBZ was generally superior to that of QBZ during the growth and development of A. dahurica root. The distribution coefficient of secretory canals was obtained by multiplying the cortex area, the size of secretory canals, and the number of secretory canals in 1/16 of the root cross-sections. The distribution coefficient of secretory canals of the two species exhibited an upward trend with root development, and the amplification was more pronounced as the A. dahurica root grew. However, the two varieties of A. dahurica still have differences. The distribution coefficient and amplification of the secretory tract of HBZ were greater than those of QBZ, presenting that the cortex area, secretory tract size, and secretory tract growth rate of HBZ were, to a certain extent, faster than those of QBZ.
Correlation analysis
Figure 7 reveals the correlation between coumarins and distribution coefficient of root secretory tract. The QBZ secretory tract distribution coefficient and oxyimperatorin and imperatorin did not correlate significantly. However, hydrated oxyimperatorin, byakangelicin, bergapten, phellopterin, isoimperatorin, and the total coumarins correlated positively with the distribution coefficients of the secretory tracts in two varieties of A. dahurica. There may be differences between the two germplasms in the accumulation of oxyimperatorin and imperatorin, as a quality control component specified in the Chinese Pharmacopoeia, merits further investigation. The study on the localization of coumarin components and root secretory tract indicates that the spatial distribution of coumarin components exhibits a pattern consistent with the localization of the root secretory tract.