*2.1. Dry Cell Weight in Response to A. flavus Elicitation*

Different concentrations of *A. flavus* medium elicitor were added to four-day-old or six-day-old suspensions of *C. roseus* CMCs. In the four-day-old suspensions of CMCs, we observed that the dry cell weight was slightly higher than that of the check group (CK) after *A. flavus* medium elicitor treatment for 24 h (Figure 2a). In the six-day-old suspensions of CMCs, we observed no significant difference in dry cell weight after the addition of the *A. flavus* medium elicitor (Figure 2b). Besides, the concentration of the *A. flavus* mycelium elicitor had no significant effect on the dry cell weight of *C. roseus* CMC suspension cultures (Figure 2c,d). For six-day-old suspensions of CMCs, the dry cell weight was slightly higher than that of the CK after 24-h treatment of 15 mg/L *A. flavus* mycelium elicitor (Figure 2d). Based on the results of Figure 2, the *A. flavus* elicitor (different concentrations: 5, 15, or 25 mg/L) had no negative effect on the growth of *C. roseus* CMCs in the selected concentration.

**Figure 2.** The effect of the *A. flavus* elicitor on the dry cell weight of *C. roseus* cambial meristematic cell (CMC) suspension cultures. The effect of the *A. flavus* medium elicitor on dry cell weight of four-day-old suspensions of *C. roseus* CMCs (**a**) or six-day-old suspensions of *C. roseus* CMCs (**b**). The effect of the *A. flavus* mycelium elicitor on the dry cell weight of four-day-old suspensions of *C. roseus* CMCs (**c**) or six-day-old suspensions of *C. roseus* CMCs (**d**). *C. roseus* CMC cultures treated by sterile water were labeled as the check group (CK). Data are given as the means ± SD (*n* = 3).

#### *2.2. HPLC-MS/MS Analysis of Alkaloids*

The *C. roseus* CMC cultures were harvested after suspension culture for six days. Then, the separation of the compounds of *C. roseus* CMCs was performed by reversed-phase high-performance liquid chromatography (RP-HPLC) with photodiode array detection, as well as by electrospray ionization tandem mass spectrometry (ESI-MS/MS) in positive mode (Figure S1). The MS spectra of the identified compounds are displayed in Figure S1c.

Compound **1**, identified by MS and chromatographic behavior comparison with that of the authentic standard, was ajmalicine (Rt 26.9 min. +MS: 353 [M + H]+; +MS2: 353, 321, 284, 252, 222, 210, 178, 144, 143, 117) (Figure S1).

Compound **2**, equally identified by MS and chromatographic behavior comparison with that of the authentic standard, was catharanthine (Rt 31.3 min. +MS: 337 [M + H]+; +MS2: 337, 248, 219, 204, 173, 165, 144, 143, 133, 128, 127, 93, 91, 77) (Figure S1).

Another compound (**3**) with [M + H]+ at 457 putatively corresponded to vindoline. Compound **3** (Rt 37.1 min) with [M + H]+ at 337 was noticed, and its UV spectrum (UV: 240, 294 nm) was similar to that of catharanthine; therefore, it was the catharanthine isomer. Further comparison showed that its MS spectrum (+MS: 457 [M + H]+; +MS2: 457, 439, 397, 379, 347, 337, 295, 258, 232, 222, 188, 173, 162, 157, 145, 134, 122) (Figure S1c) was in line with that of the vindoline authentic standard.

#### *2.3. TIA Content in Response to Elicitation*

To obtain the optimal *A. flavus* elicitor treatment condition, different concentrations of the *A. flavus* medium elicitor and mycelium elicitor were tested for their inducing effects in *C. roseus* CMC suspension cultures.

The effect of the *A. flavus* elicitor on TIA content in *C. roseus* CMCs is shown in Figures 3 and 4. As shown in these figures, *A. flavus* increased the TIA content in *C. roseus* CMCs. For different alkaloids, the induction condition for the highest yield was different. The content of catharanthine reached its maximum (3.39 mg/L), which was 3.6-times as high as that of the CK, after 24-h treatment with

25 mg/L *A. flavus* medium elicitor in four-day-old suspensions of CMCs (Figure 3b). The content of vindoline reached as high as 8.79 mg/L, which was 1.45-times as high as that of the CK, after 48-h treatment with 25 mg/L *A. flavus* mycelium elicitor in six-day-old suspensions of CMCs (Figure 4a). As for ajmaline, the content reached 9.84 mg/L, which was 3.4-times as high as that of CK, after 24-h treatment with 25 mg/L *A. flavus* mycelium elicitor in six-day-old suspensions of CMCs (Figure 4c).

**Figure 3.** The effect of the *A. flavus* medium elicitor on the concentrations of TIAs. The effect of the *A. flavus* medium elicitor on the concentrations of vindoline (**a**), catharanthine (**b**), ajmaline (**c**), and total alkaloids (**d**) in four-day-old suspensions of *C. roseus* CMCs. The effect of *A. flavus* medium elicitor on the concentrations of vindoline €, catharanthine (**f**), ajmaline (**g**), and total alkaloids (**h**) in six-day-old suspensions of *C. roseus* CMCs. *C. roseus* CMC cultures treated by sterile water were labeled as the check group (CK). Data are given as the means ± SD (*n* = 3). \* *p* < 0.05, \*\* *p* < 0.01, compared to the CK group.

According to the timing when the contents of vindoline, catharanthine, and ajmaline were all relatively high (compared with those of the CK and other experimental groups), the optimal condition was confirmed. The optimal condition for the *A. flavus* elicitor treatment in *C. roseus* CMCs was as follows: after suspension culture of *C. roseus* CMCs for six days, 25 mg/L *A. flavus* mycelium elicitor were added, and the CMC suspensions were further cultured another 48 h. Although the content of total alkaloids was slightly lower than that of the CK under this condition, the contents of vindoline, catharanthine, and ajmaline reached 8.79, 2.81, and 8.95 mg/L, respectively, which were 1.45-, 3.29-, and 2.14-times as high as those of the CK, respectively (Figure 4e–g).

**Figure 4.** The effects of the *A. flavus* mycelium elicitor on the concentrations of TIAs. The effects of the *A. flavus* mycelium elicitor on the concentrations of vindoline (**a**), catharanthine (**b**), ajmaline (**c**), and total alkaloids (**d**) in four-day-old suspensions of *C. roseus* CMCs. The effect of the *A. flavus* mycelium elicitor on the concentrations of vindoline (**e**), catharanthine (**f**), ajmaline (**g**), and total alkaloids (**h**) in six-day-old suspensions of *C. roseus* CMCs. *C. roseus* CMCs treated by sterile water were labeled as the check group (CK). Data are given as the means ± SD (*n* = 3). \* *p* < 0.05, \*\* *p* < 0.01, compared with the CK group.

#### *2.4. Functional Annotation and Functional Classification of Unigenes*

To identify the key genes in the TIA biosynthetic pathway in *C. roseus* CMCs, the transcriptomes of *C. roseus* DDCs (callus) treated with an equal amount of sterile water, *C. roseus* CMCs (the check treated with an equal amount of sterile water, CK), and *C. roseus* CMCs under optimal *A. flavus* elicitor treatment condition (the experimental group, EG) were determined.

In total, approximately 170.7 million Illumina raw data were generated from the three different samples (Table S1). After filtering the raw data, approximately 59.7, 50.7, and 57.8 million clean reads remained for the callus, CK, and EG transcriptomes, respectively (Table S1). All clean reads were subsequently subjected to de novo assembly with the Trinity program, producing 121,532 transcripts and 105,552 unigenes (Table S2).

Functional annotations of unigenes in the seven largest databases are shown in Table S3. A total of 83,742 unigenes (79.33%) were annotated based on the information available from seven public protein databases including the NCBI non-redundant protein sequences (NR), NCBI non-redundant nucleotide sequences (NT), Swiss-Prot, Protein family (Pfam), Gene Ontology (GO), euKaryotic Ortholog Groups (KOG), and the Kyoto Encyclopedia of Genes and Genomes (KEGG) using the Basic Local Alignment Search Tool (BLAST) with an *E*-value cut-off of 1 e−<sup>10</sup> (Table S3). A total of 79,711 unigenes (75.51% of the total assembled unigenes) had a match in the NR database, and 59,272 (56.15%), 63,496 (60.15%), 61,591 (58.35%), 61,829 (58.57%), 26,674 (25.27), and 34,367 (32.55%) unigenes showed significant similarity to sequences in the NT, Swiss-Prot, Pfam, GO, KOG, and KEGG databases, respectively (Table S3).

All unigenes were subjected to a search against the GO database to classify unigene functions based on the NR annotation. Of the 105,552 assembled unigenes, 61,829 unigenes were successfully assigned to one or more GO terms, and these unigenes were classified into three main GO categories and 56 groups (Figure S2). Within the "biological process" domain, the most evident matches were the terms "cellular process" (37,633), "metabolic process" (35,132), and "single-organism process" (28,250). In the "cellular component" domain, the terms "cell" (20,343) and "cell part" (20,343) were most frequently assigned. For the "molecular function" domain, the assignments were mostly enriched in the terms "binding" (37,680) and "catalytic activity" (31,249).

For further analysis, the unigenes were mapped onto the KEGG database for categorization of gene function and identification of biochemical pathways. A total of 34,367 unigenes were annotated and assigned to 5 main KEGG metabolic pathways, 19 sub-branches, and 300 KEGG pathways. Among them, the most common sub-branch was "carbohydrate metabolism" (2904), followed by "translation" (2711) and "folding, sorting and degradation" (2575) (Figure S3). In addition, there were seven unigene matches in "indole alkaloid biosynthesis" (ko00901), 43 in "monoterpenoid biosynthesis" (ko00902), and 314 in "terpenoid backbone biosynthesis" (ko00900) (Table S4).

As shown in Table S5, the detected unigenes contained transcription factors of the orphans, AP2-EREBP, and SET families.

#### *2.5. High and Differential Expression Analysis of Unigenes*

The correlation of gene transcription levels between samples could reflect differences in gene expression patterns. The closer the correlation coefficient is to one, the higher the degree of similarity of the gene expression pattern between samples. On the other hand, the closer the correlation coefficient is to zero, the bigger the difference in gene expression pattern between samples. As shown in Figure S4a, the square of the Pearson correlation coefficient (R2) was 0.539, indicating that there was a significant difference between the callus and CK groups in the level of gene expression. Thus, the gene expression pattern of the *C. roseus* CMCs was very different from that of the *C. roseus* DDCs. Besides, R2 was 0.86 between the EG and CK groups (Figure S4b), showing that *A. flavus* elicitor treatment could indeed cause differences in gene expression levels.

Differentially-expressed genes (DEGs) (|log2 (fold change)| ≥ 1 and *q*-value ≤ 0.005) were defined as unigenes that were significantly enriched or depleted in one sample relative to the other. A volcano plot was constructed to illustrate the distribution of DEGs in the callus vs. CK and EG vs. CK groups (Figure S5). The results of the differential expression analysis indicated that the expression levels of some genes in the callus group were significantly up- or down-regulated when compared with the CK. Further, compared with the CK, the expression levels of a few genes in the EG group were significantly up- or down-regulated, but the significant degree was lower than that of the callus vs. CK. The number of common differential genes among the two comparative combinations was 133, while the number of unique differential genes was 4825 in callus vs. CK and 139 in EG vs. CK (Figure S6).

To further understand the biological functions of DEGs, they were annotated with the GO and KEGG pathways. Compared with the CK, there were 2035 up-regulated DEGs and 1859 down-regulated DEGs in the callus group and 232 up-regulated DEGs and 114 down-regulated DEGs in EG (Figure S7). Partial results of KEGG pathways analysis of DEGs, which were associated with TIA biosynthesis and induced by the *A. flavus* elicitor, are shown in Figure S8. Compared with the *C. roseus* DDCs (the callus group), some genes related to the biosynthesis of TIAs were up-regulated in the *C. roseus* CMCs (the CK group). Combining the results of differential expression analysis, GO enrichment, and KEGG pathway analysis of DEGs in callus vs. CK and EG vs. CK, the DEGs associated with TIA biosynthesis and induced by the *A. flavus* elicitor were screened, and their sequences were aligned with the genes in the NCBI database via the BLAST tool. The genes identified were *D4H*, 10-hydroxylase geraniol (*G10H*), geraniol synthase (*GES*), iridoid synthase (*IRS*), loganic acid O-methyltransferase (*LAMT*), *SGD*, *STR*, tryptophan decarboxylase (*TDC*), and *ORCA3*, of which *ORCA3* was a transcription factor gene.

#### *2.6. qRT-PCR*

Under the optimal *A. flavus* elicitor treatment condition, the transcription levels of *D4H*, *G10H*, *GES*, *IRS*, *LAMT*, *SGD*, *STR*, *TDC*, and *ORCA3* were much higher in EG than in CK. Specifically, their expression were 4.49-, 1.75-, 1.71-, 1.42-, 3.12-, 2.33-, 2.87-, 2.51-, and 5.97-times as high as those of CK, respectively (Figure 5).

**Figure 5.** The effects of the *A. flavus* elicitor on the expression of TIA biosynthesis key genes in *C. roseus* CMCs under the optimal *A. flavus* elicitor treatment condition. *C. roseus* CMCs cultured under the optimal *A. flavus* elicitor treatment condition were labeled as the experimental group (EG). *C. roseus* CMCs treated by sterile water were labeled as the check group (CK). Data are given as the means ± SD (*n* = 3). \* *p* < 0.05, \*\* *p* < 0.01, compared with the CK group.

#### **3. Discussion**

Since the low content of the pharmacological TIAs in *C. roseus* [18] and commercial production has used a semi-synthetic route to couple vindoline and catharanthine [3], developing methods to enhance the TIA production of *C. roseus* cell cultures has become a focus for domestic and foreign scholars [19]. According to our previous research, undifferentiated *C. roseus* CMCs were capable of maintaining not only good cellular morphology, but also stable and high production of alkaloid metabolites [25]. Therefore, *C. roseus* CMCs were used as plant cell materials for the investigation of the biosynthesis of TIAs.

In previous reports, fungal elicitors could increase the accumulation of secondary metabolites of interest by activating specific secondary metabolic pathways [20]. Tonk et al. showed that not only the callus growth rate, but also the alkaloid content in the embryos of *C. roseus* could be effectively increased after being treated with a low-dose *A. flavus* fungal elicitor [24]. However, there has been no

research on the effect of *A. flavus* fungal elicitor on the yields of TIAs in *C. roseus* CMCs. Generally, the optimal time for adding the fungal elicitor to cell cultures is at the log phase of the culture cycle, when the cell cultures are most sensitive to elicitation. Our previous research showed that the log phase of the *C. roseus* CMC culture cycle lasts from 4–9 days [25]. Besides, the results of our preliminary experiment in this research showed that the *C. roseus* CMC cultures turned brown after three days of treatment or when the suspension cultures of *C. roseus* CMCs were treated with the fungal elicitor after a nine-day culture. Thus, we finally added three different concentrations of the *A. flavus* medium elicitor or the *A. flavus* mycelium elicitor to four-day-old or six-day-old suspensions of *C. roseus* CMCs. After inducing treatment, we found that the *A. flavus* mycelium elicitor could promote the TIA production without a negative effect on the growth of *C. roseus* CMCs in the selected concentration. According to the results determined by HPLC and HPLC-MS, the optimal condition of the *A. flavus* elicitor was as follows: after suspension culture of *C. roseus* CMCs for six days, 25 mg/L *A. flavus* mycelium elicitor were added, and the CMC suspensions were further cultured for another 48 h. Under this condition, the contents of vindoline, catharanthine, and ajmaline were 1.45-, 3.29-, and 2.14-times as high as those of the CK group, respectively (Figure 4e–g).

The TIA biosynthetic pathway in *C. roseus* is complex and highly regulated [2,6–8]. Hao et al. comprehensively analyzed the differential gene expression profiles of 24 h of continuous light, 24 h of dark treatment, 4 h of MeJA treatment under continuous light conditions, and 4 h of MeJA treatment under dark conditions in *Artemisia annua* seedlings using Illumina transcriptome sequencing. As a result, some TFs in the light signaling pathway were identified that can respond to MeJA [26]. Thus, transcriptome analysis is an effective method of investigating the inducing mechanism. To explore the inducing mechanism at the gene expression level, we analyzed the transcriptome data of *C. roseus* DDCs treated with sterile water, *C. roseus* CMCs treated with sterile water, and *C. roseus* CMCs under the optimal *A. flavus* elicitor treatment condition. This transcriptome analysis showed that *D4H*, *G10H*, *GES*, *IRS*, *LAMT*, *SGD*, *STR*, *TDC*, and *ORCA3* were involved in the regulation of this induction process. The functions of the above genes in the TIA biosynthetic pathway are shown in Figure 1. Among the above genes, *ORCA3* is the core transcription factor gene in the TIA biosynthetic pathway [6]. Then, we further analyzed the transcription levels of the above genes by qRT-PCR. The results of qRT-PCR showed that, under the optimal *A. flavus* elicitor treatment condition, the transcription levels of *D4H*, *G10H*, *GES*, *IRS*, *LAMT*, *SGD*, *STR*, *TDC*, and *ORCA3* were much higher in EG than in CK (Figure 5). These results indicated that the increasing accumulations of vindoline, catharanthine, and ajmaline in *C. roseus* CMCs were correlated with the increasing expression of the above genes. As previous reports showed, the expression of the transcription factor *ORCA3* gene could be induced by elicitors (such as MeJA and jasmonic acid), and elicitors could increase the expression of *D4H*, *G10H*, *STR*, *GES*, *SGD*, and *TDC* [6,17,25,27,28]. Since overexpression of *ORCA3* increases the expression of key genes such as *STR* in the TIA biosynthetic pathway [6], it is extremely meaningful to explore whether the up-regulation of TIA biosynthesis-related genes is related to the increasing transcription level of *ORCA3* under the *A. flavus* mycelium elicitor treatment.

In conclusion, the *A. flavus* mycelium elicitor could promote the TIA production of *C. roseus* CMCs. Under the optimal *A. flavus* elicitor treatment condition, the contents of vindoline, catharanthine, and ajmaline were 1.45-, 3.29-, and 2.14-times as high as those of the CK group, respectively. Transcriptome analysis and qRT-PCR experiment revealed that *D4H*, *G10H*, *GES*, *IRS*, *LAMT*, *SGD*, *STR*, *TDC*, and *ORCA3* were involved in the regulation of this induction process. Furthermore, the up-regulation of TIA biosynthesis-related genes was related to the inducting effect of the *A. flavus* mycelium elicitor, which in turn promoted the accumulation of vindoline, catharanthine, and ajmaline in *C. roseus* CMCs.
