*2.2. Identification of Metabolites*

The M and NM growth for six months were used for metabolome analysis. Overlapping analysis of the total ion chromatography (TIC) in different quality control (QC) samples showed that the retention time and peak intensities were consistent (Figure S1), which indicated that the instrument had a good stability and could thus be used for subsequent analysis. Principal component analysis (PCA) was performed on the NM, M, and mixed samples. Principal component 1 (PC1) and principal component 2 (PC2) were 54.7% and 26.3%, respectively (Figure 2A). The metabolite profiles of *A. roxburghii* were then subjected to orthogonal partial least squares discriminant analysis (OPLS-DA). The result showed that the R2X, R2Y, and Q2 were 0.862, 1.000, and 0.995, respectively (Figure 2B), which indicated the model of OPLS-DA was stable and reliable. The score plots of PCA and OPLS-DA exhibited an obvious separation between the M and NM, and each formed a cluster. These suggested that mycorrhizal fungus AR2 affected the metabolism in *A. roxburghii*.

**Figure 2.** PCA and OPLS-DA scores plots derived from ultra-performance liquid chromatography-electrospray ionization-tandem mass spectrometry (UPLC-ESI-MS/MS) profiling of non-mycorrhizal *A. roxburghii* (NM) and mycorrhizal *A. roxburghii* (M) growth for six months. (**A**) PCA scores plot of the two samples (NM and M growth for six months) and the quality control sample (mix, the same volume of sample extract from the NM and M growth for six months was prepared by mixing); the x-axis represents the PC1 and the y-axis represents PC2. (**B**) OPLS-DA scores plot of the putatively annotated metabolites from NM and M growth for six months. The x-axis represents the score value of main components in the orthogonal signal correction process and the differences between the groups can be seen from the direction of the x-axis; the y-axis represents the scores of orthogonal components in the orthogonal signal correction process and the differences within the groups can be seen from the direction of the y-axis.

A total of 709 metabolites with known structures were identified in M and NM under quality validation, each of which was analyzed using three biological replicates. Detailed information about the identified metabolites, including the compounds, classes, molecular weights, ionization models, Kyoto encyclopedia of genes and genomes (KEGG) pathways, and quantities for each of the three periods is shown in Table S1. Flavonoid (20.9%), organic acids, and derivatives (15.4%), amino acids and derivatives (12.8%), lipid (9.6%), and phenylpropanoid (8.7%) accounted for a large proportion of these 709 metabolites (Figure 3A).

**Figure 3.** Component analysis of the putatively annotated metabolites and pathway enrichment analysis of the DAMs. (**A**) Component analysis of the putatively annotated metabolites from non- mycorrhizal *A. roxburghii* (NM) and mycorrhizal *A. roxburghii* (M) growth for six months. The percentage of the top six metabolites are shown in the graph. The percentage after each compound represents the percentage of the number of DAMs of a certain class of compounds in the total DAMs. (**B**) Pathway (top 20) enrichment analysis of the DAMs between the NM and M growth for six months. The x-axis represents the corresponding rich factor of each pathway. The y-axis represents the name of pathway. The color of the dot is *p*-value, and the closer it is to 0, the more significant the enrichment is. The size of the point represents the number of DAMs enriched in the corresponding pathway. The rich factor is the ratio of the number of metabolites in the corresponding pathway to the total number of metabolites detected and annotated in the pathway. The higher the value of rich factor is, the higher the enrichment degree is.
