The Expression Profiles of the Salvia miltiorrhiza 3-Hydroxy-3-methylglutaryl-coenzyme A Reductase 4 Gene and Its Influence on the Biosynthesis of Tanshinones

Salvia miltiorrhiza is a medicinal plant that synthesises biologically-active tanshinones with numerous therapeutic properties. An important rate-limiting enzyme in the biosynthesis of their precursors is 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGR). This study presents the organ-specific expression profile of the S. miltiorrhiza HMGR4 gene and its sensitivity to potential regulators, viz. gibberellic acid (GA3), indole-3-acetic acid (IAA) and salicylic acid (SA). In addition, it demonstrates the importance of the HMGR4 gene, the hormone used, the plant organ, and the culture environment for the biosynthesis of tanshinones. HMGR4 overexpression was found to significantly boost the accumulation of dihydrotanshinone I (DHTI), cryptotanshinone (CT), tanshinone I (TI) and tanshinone IIA (TIIA) in roots by 0.44 to 5.39 mg/g dry weight (DW), as well as TIIA in stems and leaves. S. miltiorrhiza roots cultivated in soil demonstrated higher concentrations of the examined metabolites than those grown in vitro. GA3 caused a considerable increase in the quantity of CT (by 794.2 µg/g DW) and TIIA (by 88.1 µg/g DW) in roots. In turn, IAA significantly inhibited the biosynthesis of the studied tanshinones in root material.


Organ-Specific Expression of S. miltiorrhiza HMGR4 Gene
Real-time qPCR results showed that HMGR4 gene was expressed in all analysed S. miltiorrhiza organs, but with different intensities. The leaves and stems demonstrated higher levels of the HMGR4 transcript than the reference, with R = 1.14 ± 0.08 and R = 1.05 ± 0.01, respectively; in roots, the level was lower than in the reference with R = 0.95 ± 0.07.
Due to their level of transcript and high availability of material for research, it was decided to use the leaves to study the effect of hormones on S. miltiorrhiza HMGR4 activity.

Potential Regulators of S. miltiorrhiza HMGR4 Gene Expression
Sequence analysis of the S. miltiorrhiza HMGR4 promoter using the PlantPan 2.0 tool showed the existence of 5369 potential TFBSs and 365 interacting TFs previously detected in the Arabidopsis thaliana model plant. The similarity score between the TFBSs found in HMGR4 promoter and those identified in A. thaliana was set to 0.7-1.0. Of all the TFs detected, a large group was able to respond to hormonal agents; many of these were sensitive to GA 3 , IAA and SA (Table 1, Tables S1-S3). It is worth emphasising that these TFs also had potential binding sites in the HMGR4 proximal promoter region (Tables S1-S3), where most functional TFBSs are believed to be located [38,39]. Therefore, it was decided to investigate the importance of these hormones on the expression of S. miltiorrhiza HMGR4. The hormones used in the experiment changed the expression of HMGR4 in treated leaves compared to control (leaves not incubated with hormones) (Figure 1). At the beginning of each study, a lower HMGR4 transcript level was observed in the test materials than in corresponding control (R < 1). Treatment with GA 3 or IAA or SA for 12 and 24 h resulted in the stimulation of HMGR4 expression against untreated samples (R > 1). It is worth noting that the exposure of leaves to GA 3 resulted in an approximately 2.86-fold increase in HMGR4 expression between 12 and 24 h, and the stimulation effect was also maintained at 48 h. From 48 h, the level of HMGR4 transcript in the hormone-treated samples decreased compared to the control (R < 1). In the final part of the testing (72 h for SA and 96 h for GA 3 and IAA) the level of HMGR4 mRNA increased again in leaves incubated with hormones compared to untreated samples (R > 1). Based on the obtained results, GA 3 and IAA were selected for experiments determining the tanshinone content.

Impact of pRI201-AN-HMGR4 Transformation on S. miltiorrhiza HMGR4 Gene Expression
Higher levels of the HMGR4 transcript were observed in all S. miltiorrhiza organs taken from plants transformed with the pRI201-AN-HMGR4 construct compared to control (R > 1). The R values for stems, roots and leaves, were 1.28 ± 0.19, 1.25 ± 0.14 and 1.10 ± 0.14, respectively.

Influence of HMGR4 Overexpression on the Biosynthesis of Tanshinones in S. miltiorrhiza
The conducted studies indicate that S. miltiorrhiza HMGR4 overexpression had a significant influence on the quantity of measured tanshinones.
Clear differences in the levels of individual tanshinones were observed between roots with HMGR4 overexpression and those without. More specifically, these values were as follows (the first value indicating soil conditions and the second in vitro): 2.43-or 3.62-fold (5.39 or 2.40 mg/g DW) for CT, 2.19-or 2.47-fold (0.59 or 0.44 mg/g DW) for DHTI, 1.86-or 2.21-fold (0.71 or 0.65 mg/g DW) for TI, 1.51-or 1.82-fold (1.88 or 0.55 mg/g DW) for TIIA, and reflected the place of cultivation.
Moreover, overexpression of HMGR4 gene induced the quantity of TIIA to about 50 µg/g DW in in vitro and in soil-grown stems and leaves ( Figure 2D).  Bars are medians with first and third quartile. ** significant difference at p < 0.01 compared to control; * significant difference at 0.01 < p < 0.05 compared to control; DW, dry weight; OX, overexpression.

Organ-Dependent Accumulation of Tanshinones in S. miltiorrhiza
Roots appeared to be the main site of accumulation of all studied metabolites in S. miltiorrhiza. All of the examined roots were found to contain all tested tanshinones (Figure 2A-D). CT was present at the highest levels (0.91-9.17 mg/g DW), while lower amounts were found for TIIA (0.67-5.61 mg/g DW), TI (0.54-1.53 mg/g DW) and DHI (0.30-1.08 mg/g DW) (Figure 2A-D). The quantity of the identified metabolites was highest in soil-grown roots overexpressing HMGR4.
Some tanshinones were detected in stems and leaves with median values ranging from 50 to 73.5 µg/g DW (Figure 2A,B,D,E). The most common tanshinone present in the tested stems and leaves was TIIA. The TIIA content was typically 104.5-fold higher (by 5.55 mg/g DW) in roots than in stems or leaves in the soil-grown plants, and 23.4-fold higher (by 1.16 mg/g DW) in in vitro roots ( Figure 2D). No tanshinones were detected in flowers (Figure 2A-E). Hence, apart from slight changes in TIIA level in stems and leaves, HMGR4 overexpression did not appear to significantly change the organ-specific pattern of accumulation of the compounds in S. miltiorrhiza.

Impact of Growth Environment on the Biosynthesis of Tanshinones in S. miltiorrhiza
The soil environment favoured a significantly higher production of all tested tanshinones in the root material compared to in vitro conditions (p < 0.01). This was true both in the group of roots with and without HMGR4 overexpression, and the differences were from 1.45-to 4.62-fold (0.34-5.86 mg/g DW) and from 1.63-to 5.58-fold (0.19-3.06 mg/g DW), respectively, depending on the type of metabolite. The quantities of individual tanshinones varied considerably between the soil-grown roots and those grown in vitro. More specifically, these differences were as follows (first value = HMGR4 overexpression; the second value = without): 4.62-or 5.58-fold (4.39 or 3.06 mg/g DW) for TIIA, 2.77-or 4.14-fold (5.86 or 2.87 mg/g DW) for CT, 1.45-or 1.63-fold (0.34 or 0.19 mg/g DW) for DHTI, 1.28or 1.52-fold (0.34 or 0.28 mg/g DW) for TI. It is worth noting that the in vitro roots with HMGR4 overexpression demonstrated 1.51-fold higher DHTI (0.25 mg/g DW) and 1.45-fold higher TI (0.37 mg/g DW) than the soil-grown roots without overexpression.
In leaf material, the content of TIIA was significantly higher in soil than in in vitro conditions (p = 0.0000), amounting to 50 or 56.8 µg/g DW, depending on HMGR4 overexpression status.
DHTI and CT were detected in stems grown in vitro but not in stems grown in soil (Figure 2A,B). Median levels were 66.8 or 65.9 µg/g DW for DHTI, and 72 or 67.3 µg/g DW for CT, depending on the presence or absence of HMGR4 overexpression.

Effect of GA 3 and IAA on the Biosynthesis of Tanshinones in S. miltiorrhiza
The addition of GA 3 to S. miltiorrhiza in vitro root culture significantly increased CT, TIIA and total tanshinone levels in comparison to untreated roots (p = 0.0000, p = 0.0404, p = 0.0404, respectively) ( Figure 2B,D). The observed increases were 1.24-fold (0.79 mg/g DW) for CT and 1.07-fold (88.1 µg/g DW) for TIIA. Treatment had no effect on DHTI and significantly decreased the amount of TI by 1.29-fold (0.27 mg/g DW) (p = 0.0000) (Figure 2A,C).
In vitro cultivation of stems grown in the presence of GA 3 showed a significant 1.15-fold (9.9 µg/g DW) rise in DHTI (p = 0.0000) and a significant 1.04-fold (1.9 µg/g DW) reduction in TIIA (p = 0.0235) compared to untreated controls (Figure 2A,D). However, in vitro cultivation of leaves with GA 3 resulted in a significant 1.02-fold (1.2 µg/g DW) increase in TIIA compared to control (p = 0.0000) ( Figure 2D).
The use of IAA resulted in a significant decrease in the content of all tested tanshinones in in vitro root culture compared to untreated roots (p = 0.0000) ( IAA treatment only appeared to have a slight influence on the quantity of tanshinones in stems and leaves: a significant 1.05-fold (3.2 µg/g DW) rise in DHTI and a significant 1.05-fold (2.3 µg/g DW) fall in TIIA were observed in stems compared to control (p = 0.0000) (Figure 2A,D), while a significant 1.01-fold (0.7 µg/g DW) increase in TIIA was noted in leaves relative to control (p = 0.0001) ( Figure 2D).

Discussion
This work analyses the expression profiles of the S. miltiorrhiza HMGR4 gene and its influence on the biosynthesis of tanshinones.
Previous studies have shown that S. miltiorrhiza HMGR genes are expressed in the roots, stems and leaves, but with different intensities in each organ. HMGR showed the strongest activity in roots, and weaker in stems and leaves [17]. The level of the HMGR2 transcript was about four-fold higher in leaves than in stems, and about two-fold higher in stems than in roots [15]. HMGR3 was vigorously expressed in stems and root steles, and to a much greater degree than in root cortices and leaves [16]. HMGR4 activity was the highest in flowers, lower in stems and leaves, and lowest in root steles and root cortices [16]. In the present study, a higher level of HMGR4 mRNA was noted in leaves and stems than in the control; however, this was not observed in roots. Previous transcriptomic analyses have indicated that within the S. miltiorrhiza root, the strongest expression of HMGR4 occurred in xylem [40].
The S. miltiorrhiza HMGR4 gene showed a biphasic response to GA 3 treatment. After initial stimulation of its expression relative to the control at 12, 24 and 48 h, it then decreased and subsequently increased at 96 h ( Figure 1). It has been found that 2.89 µM GA 3 has a similar influence on the S. miltiorrhiza HMGR2 gene; however, in this case, HMGR2 expression increased compared to control at 12 h, followed by a fall and a second increase at 72 and 96 h [41]. Elsewhere, stimulation with 400 µM GA 3 resulted in an initial rise in Malus domestica HMGR1 transcripts against control until four hours, followed by a decrease at six h [42]. We hypothesise that stimulation of S. miltiorrhiza HMGR4 gene expression by GA 3 and subsequent enzyme production could activate the next stages of the MVA pathway and the production of mediators necessary for the biosynthesis of endogenous gibberellins, such as ent-kaurene [43]. The newly-produced endogenous GA 3 could stimulate the HMGR4 transcription which decreased as a result of metabolising the exogenous hormone. However, this hypothesis needs to be verified by monitoring endogenous GA 3 levels during the course of an experiment.
The impact of IAA on S. miltiorrhiza HMGR4 expression was very similar to that induced by GA 3 (Figure 1). Although the effect of IAA on plant HMGR genes has not been widely studied, we have noticed some similarities in our results with previous research. IAA at a final concentration of 100 µM first raised the level of M. domestica HMGR4 transcripts relative to control, and then lowered them [42]. The biphasic effect, which we observed in our experiment, may result from the stimulation of various TFs, some of which increase expression of the gene, while others reduce it.
The use of SA caused a rise in HMGR4 expression at 12, 24 and 72 h and a fall at 48 and 96 h in relation to untreated material (Figure 1). A similar effect was observed for 10 mM SA against HMGR3 in Ginkgo biloba leaves; however, in contrast to our present findings, the level of HMGR3 mRNA rose against control values in the final phase of the study (96 and 120 h) [44]. Elsewhere, SA treatment was found to result in continually elevated HMGR transcript levels versus untreated controls in S. miltiorrhiza hairy roots throughout the experiment [20] and Salvia przewalskii hairy roots [45]. A maximum three-fold increase in HMGR expression was noted after 36 h of stimulation [20], and an eight-fold rise after six days [45].
The present study is the first investigation of the role of HMGR4 in the biosynthesis of tanshinones in S. miltiorrhiza. Overexpression of this gene resulted in a significant increase in DHTI, CT, TI and TIIA content: by 1.51-to 2.43-fold (0.59-5.39 mg/g DW) in soil-grown roots, and by 1.82-to 3.62-fold (0.44-2.40 mg/g DW) in in vitro roots (Figure 2A-D). Of all tanshinones tested, CT showed the highest rise relative to control: 2.43-fold (5.39 mg/g DW) for roots grown in soil and 3.62-fold (2.40 mg/g DW) for in vitro roots. The results are in agreement with data received for other S. miltiorrhiza HMGR enzymes. Kai et al. reported that overexpression of the HMGR gene led to an increase in CT, TI, TIIA quantity in hairy root culture ranging from 1.17-to 3.19-fold (0.844-1.515 mg/g DW) compared to control [46]. As in our research, CT showed the highest rise in all seven transgenic lines tested. In another study, HMGR2 overexpression significantly enhanced the amount of DHTI, CT, TI, TIIA by 1.23-to 2.46-fold (0.99-3.16 mg/L) at day 40 of root culture relative to control [15]. In the experiment, CT demonstrated the greatest increase, i.e., by 2.46-fold (3.16 mg/L).
Our results indicate that tanshinone accumulation in S. miltiorrhiza was organ-dependent, with roots as the primary storage place for DHTI, CT, TI, TIIA ( Figure 2E). Li et al. specif-ically indicate the root periderm of S. miltiorrhiza as the main site of accumulation of all tested tanshinones, viz. DHTI, CT, TI, TIIA, Tanshinone IIB, Dehydrotanshinone IIA, Dashenxinkun B, Trijuganone A, Trijuganone C; the inner layer of the roots and the outer part of stems contained much smaller amounts [47]. Subsequent research also pointed to S. miltiorrhiza root periderm as the main storage place for TIIA, although traces were also detected in root phloem [40]. In addition, transcriptomic analyses of the MVA and MEP pathway genes and other enzymes leading to the production of tanshinones indicated that the strongest expression of most of the tested genes (AACT1 to AACT6, HMGS2, HMGR1, HMGR2, MK, PMK, MDC1, MDC2, IPI1, GGPPS3, DXS2, DXS4, DXR, MCT, CMK, MDS, HDS, HDR1 to HDR3, CPS1, CPS5, KSL1, KSL7, KSL8, CYP76AH1) occurred in the periderm of S. miltiorrhiza roots [40]. Hence, the root periderm layer appears to be not only the main storage site, but also the main place of biosynthesis of tanshinones. The examined stems turned out to be a better source of the metabolites than leaves, but their content was quite low (several dozen µg/g DW) ( Figure 2E). These results are in line with previously-performed studies [48]. Organ-specific accumulation and production of tanshinones may result from the existence of various mechanisms regulating the activity of enzymes involved in the biosynthesis of these compounds [49].
Our findings indicate that soil cultivation favoured 1.28-to 5.58-fold (0.19-5.86 mg/g DW) higher production of DHTI, CT, TI and TIIA in roots and 1.12-fold (6.2 µg/g DW) greater production of TIIA in leaves compared to in vitro conditions. This may be due to the community of microorganisms naturally present in the rhizosphere, phyllosphere and endosphere; it is possible that these may affect the biosynthesis of metabolites [50,51]. According to Yan et al., the endophytic bacteria Pseudomonas brassicacearum subsp. neoauraniaca raised the activity of HMGR and DXS enzymes by 2.1-and 4.2-fold, respectively, in S. miltiorrhiza hairy root culture. This resulted in a significant increase in the content of all tanshinones tested, with particular gains found for DHTI (19.2-fold) CT (11.3-fold) and total tanshinones (3.7-fold) compared to controls [52]. In addition, the polysaccharide fraction isolated from rhizobacterium Bacillus cereus stimulated the accumulation of tanshinones in S. miltiorrhiza root culture by about seven-fold (1.59 vs. 0.19 mg/g DW) compared to control [53]. Another potential reason for the lower in vitro yields of tanshinones may be changes occurring in the morphology, anatomy and physiology of plants during in vitro cultivation [54,55].
Additionally, our findings provide further information about the influence of hormones on the biosynthesis of tanshinones in S. miltiorrhiza. GA 3 stimulated CT and TIIA production, but had no significant effect on DHTI content and decreased TI in in vitro root culture compared to untreated controls (Figure 2A-D). We hypothesize that the presence of GA 3 may strongly induce the expression of some key enzyme/-s involved in the terminal stage of CT biosynthesis. This could be the reason for the higher TIIA content which arises from CT; however, as GA 3 may not have a similar effect on DHTI production, the resulting TI does not rise, and may even fall [14]. GA 3 has been found to increase DHTI, CT, TI and TIIA levels in most GRAS3-overexpressing S. miltiorrhiza hairy root culture lines and in untransformed controls [56]; however, these results cannot be directly compared to ours, as the experiment used a 34.6-fold higher concentration of the hormone (100 µM) and a much shorter incubation time with GA 3 , of only six days. The second hormone used, IAA, significantly reduced the accumulation of CT by 34.06-fold (3.21 mg/g DW), TIIA by 11.49-fold (1.11 mg/g DW), DHTI by 8.84-fold (0.66 mg/g DW) and TI by 5.05-fold (0.96 mg/g DW) in an in vitro root culture versus control (Figure 2A-D). Reduced CT, TI and TIIA synthesis was also observed in S. miltiorrhiza hairy roots treated with 5.71 µM IAA: 1.61-fold decrease (82 µg/g DW) for TI, 1.50-fold decrease (125 µg/g DW) for CT, and 1.24-fold decrease (23 µg/g DW) for TIIA, compared to control [57].

Establishment of S. miltiorrhiza Culture and Treatments
S. miltiorrhiza plants were cultivated from seeds provided by the Garden of Medicinal Plants of the Medical University of Lodz. To establish in vitro plant cultures, the seeds were surface sterilised utilising 70% ethanol for 1 min and subsequent 1% sodium hypochlorite solution for 5 min, and then rinsed three times with sterile distilled water for 5 min. The seeds were thereafter transferred aseptically onto Murashige and Skoog (MS) basal medium [58] with 3% sucrose (Chempur, PiekaryŚląskie, Poland) and 0.65% agar (Sigma-Aldrich, Saint Louis, MO, USA) and a final pH of 5.7. Germination was carried out in the dark at 26 ± 2 • C. After germination, aerial parts of S. miltiorrhiza were grown in solid MS medium at 26 ± 2 • C under 16/8 h (light/dark) photoperiod at a cool fluorescent light with intensity of 40 µmol m −2 s −1 . Roots were cultivated in the dark at 26 ± 2 • C in Gamborg B5 liquid medium [59] agitated at 70 rpm. Subcultures were carried out every five weeks.
Five-week-old leaves, stems and roots, grown as described above, were used to study organ-specific expression of the HMGR4 gene.
The effect of hormones on HMGR4 activity was determined in five-week-old leaves.

Preparation of pRI201-AN-HMGR4 Overexpression Construct
The S. miltiorrhiza HMGR4 coding sequence (1653 bp) was synthesised on the basis of JN831103.1 sequence and inserted into a pUC57 vector (Gene Universal Inc., Newark, DE, USA). The correctness of the insert was determined by double-strand Sanger sequencing. Afterwards, the HMGR4 insert was excised from pUC57 and inserted into a pRI201-AN binary expression vector (Takara Bio Inc., Kusatsu, Japan) at NdeI/SalI sites of MCS1 (Eurofins Genomics, Ebersberg, Germany). HMGR4 gene overexpression was driven by the strong and constitutive promoter of Cauliflower Mosaic Virus 35S (CaMV), which facilitates high levels of RNA transcription in a wide variety of plants. Analysis of the HMGR4 sequence and flanking regions was performed by double-strand Sanger sequencing. A map of the prepared pRI201-AN-HMGR4 construct is presented in Figure 3.

Transformation, Selection, Regeneration and Treatments of S. miltiorrhiza Culture
Agrobacterium tumefaciens (Rhizobium radiobacter) GV2260 (C58C1Rif R with pGV2260) competent cells were transformed with the pRI201-AN-HMGR4 construct or the empty pRI201-AN vector using the freeze/thaw method [62]. The transformed bacteria were firstly grown for 84 h at 26 • C on solid selective YEB medium containing 50 mg/L kanamycin, 100 mg/L carbenicillin and 30 mg/L rifampicin (Chem-Impex International, Wood Dale, IL, USA) and then on liquid selective YEB medium with shaking at 140 rpm until OD 600 reached 0.4-0.8. To confirm the transformation, plasmid DNA was isolated by alkaline lysis and extracted with a phenol/chloroform/isoamyl alcohol mixture [63]; this was then subjected to PCR amplification using GoTaq Hot Start Green Master Mix (Promega, Madison, WI, USA) and Kanamycin primers ( Table 2). The PCR reactions were carried out in an MJ Mini Personal Thermal Cycler (Bio-Rad, Hercules, CA, USA) with the following parameters: initial denaturation (95 • C, 5 min), denaturation (95 • C, 45 s), primer annealing (60 • C, 30 s), extension (72 • C, 30 s), final extension (72 • C, 5 min). In total, 40 PCR cycles were conducted. The obtained products were separated by 2% agarose gel electrophoresis.  To induce virulence, bacterial cultures with confirmed transformation were collected by centrifugation and resuspended to OD 600 = 0.1 in sterile induction medium, i.e., liquid MS medium supplemented with 100 µM acetosyringone (Sigma-Aldrich, Saint Louis, MO, USA), and then agitated on a rotary shaker at 140 rpm for five hours at 26 • C [64].
Three-month-old leaves of S. miltiorrhiza grown in pots were surface sterilised using the same protocol described earlier for the seeds; however, 0.8% sodium hypochlorite solution was applied. Preparation, infection of leaves and co-cultivation were performed according to Dandekar and Fisk with some modifications [64]. The composition of the induction medium was as mentioned above. Co-cultivation solid MS medium was supplemented with 1 mg/L 6-benzylaminopurine (BAP), 0.2 mg/L 1-naphthaleneacetic acid (NAA) and 100 µM acetosyringone (Sigma-Aldrich, Saint Louis, MO, USA). Overall regeneration frequency, non-transgenic regeneration under selection and non-transgenic controls were included in the research. After 72 h of incubation, leaf discs were transferred every two weeks onto fresh A. tumefaciens (R. radiobacter) killing medium, i.e., solid MS medium with 1 mg/L BAP, 0.2 mg/L NAA and 250 mg/L cefuroxime. After another six weeks, the obtained calluses were moved onto solid MS medium supplemented with 0.5 mg/L BAP, 0.2 mg/L IAA and 250 mg/L cefuroxime. In the following weeks, cefuroxime was gradually phased out and the selection antibiotic kanamycin (Biological Industries, Kibbutz Beit-Haemek, Israel) was introduced (10-50 mg/L). The aerial parts of the S. miltiorrhiza transformants and controls were cultivated in solid MS medium at 26 ± 2 • C under 16/8 h (light/dark) photoperiod using a cool fluorescent light with intensity of 40 µmol m −2 s −1 and their roots in the dark in liquid Gamborg B5 medium agitated at 70 rpm. Subcultures were carried out every five weeks. Additionally, in order to compare the influence of different growth environments on the biosynthesis of tanshinones, the transformants and control were transferred from in vitro cultures to pots containing sterile composite soil. The plants were covered with a transparent glass jar for three weeks and grown at 26 ± 2 • C under natural light. Figure 4 shows S. miltiorrhiza cultures at various stages of the experiment.  Table 2). The concentration and purity of the DNA were assessed based on A 260 /A 280 and A 260 /A 230 ratios using a Nanophotometer P300 (Implen, Munich, Germany). PCR reaction parameters were as mentioned above. The obtained products were separated via 2% agarose gel electrophoresis.
The effect of the pRI201-AN-HMGR4 construct or the empty pRI201-AN vector (control) on S. miltiorrhiza HMGR4 expression was investigated in five-week-old leaves, stems and roots.
The importance of the HMGR4 gene for tanshinone biosynthesis was assessed in five-week-old S. miltiorrhiza roots, leaves, and stems growing in soil and in vitro and overexpressing HMGR4 relative to plant material that did not overexpress HMGR4.
The role of the growth environment for the tanshinone content was evaluated in five-week-old S. miltiorrhiza roots, leaves, and stems with and without HMGR4 overexpression growing in soil in relation to the plant material grown in in vitro conditions. The effect of GA 3 (1 mg/L, 2.89 µM) or IAA (0.5 mg/L, 2.85 µM) on the production of tanshinones was estimated in five-week-old roots, leaves, and stems of S. miltiorrhiza overexpressing HMGR4, grown in vitro and treated with the hormones against untreated plant material.
The role of plant organ for the accumulation of tanshinones was assessed in fiveweek-old S. miltiorrhiza roots, leaves, and stems with and without HMGR4 overexpression growing in soil and in in vitro conditions.

RNA Isolation, Reverse Transcription and Quantitative Real-Time PCR
Total RNA was isolated in accordance with the protocol given in NucleoSpin RNA Plant and Fungi kit (Macherey-Nagel, Duren, Germany). Plant material was ground under liquid nitrogen to a fine powder using mortar and pestle. The samples were digested by RNase-free rDNase (Macherey-Nagel, Duren, Germany) to assure removal of genomic DNA. Isolated RNA was stored at −80 • C. The concentration and purity of the RNA were evaluated using Nanophotometer P300 (Implen, Munich, Germany). The obtained A 260 /A 280 ratios were within the range of 1.9-2.1 and A 260 /A 230 ratios were~2.
The reverse transcription reactions were carried out using Maxima H Minus Reverse Transcriptase, Oligo(dT)18 Primer, dNTP Mix, RiboLock RNase Inhibitor, and nuclease-free water (Thermo Scientific, Waltham, MA, USA) according to the manufacturer's protocol. The quantity of RNA was adjusted to achieve the same final RNA concentration in a given experiment. No reverse transcriptase and no template controls were applied. The prepared cDNA was stored at −20 • C.

Quantitative Analysis of Tanshinones
Roots, stems and leaves of S. miltiorrhiza were freeze-dried in a lyophiliser Alpha 1-2 LD (Martin Christ, Osterode am Harz, Germany) under 0.1 mbar pressure and ground with a pestle and mortar to a fine powder. The obtained powder (50 mg) was extracted with methanol (2 mL) under ultrasonic treatment for one hour at room temperature. The mixture was then centrifuged at 14,000× g for 5 min and then the supernatant was filtered through a 0.2 µm organic membrane filter (Millipore, Burlington, MA, USA) [67].

Conclusions
The most important observation from the conducted research concerns the important role played by HMGR4 in the biosynthesis of tanshinones, which is reflected in the content of DHTI, CT, TI, TIIA in the roots and TIIA in the stems and leaves with gene overexpression.
Other conclusions: GA 3 , IAA and SA regulated the expression of the S. miltiorrhiza HMGR4 gene, confirming the results of the in silico promoter analysis.
The soil environment promoted a higher accumulation of all tested metabolites in roots and TIIA in leaves compared to in vitro conditions. However, it is worth noting that the amounts of DHTI and TI in in vitro roots with HMGR4 overexpression were higher than in soil-grown roots without overexpression.
Apart from the positive effect on the appearance of TIIA in the studied stems and leaves of S. miltiorrhiza, HMGR4 overexpression did not change the characteristic organdependent pattern of tanshinone accumulation, i.e., the main source was the root, with trace amounts observed in stems and leaves. GA 3 increased CT and TIIA production in roots, while IAA reduced the biosynthesis of all tested metabolites.
The greatest efficiency of tanshinone biosynthesis was found to result from a combination of three traits, namely HMGR4 gene overexpression, root organ, and cultivation in soil conditions. Future research could investigate the mechanisms controlling S. miltiorrhiza HMGR4 gene expression. TFs regulating HMGR4 expression could be isolated using the yeast-one hybrid (Y1H) system and then functionally characterised [68]. The role of specific TFBSs in the response of HMGR4 to abiotic or biotic factors could be verified by its mutagenesis [69]. TF networks that play a key role in the regulation of HMGR4 gene expression could be explored through transcriptomic RNA sequencing and weighted gene co-expression network analysis (WGCNA) [70,71].