Rhizobium rhizogenes -mediated transformation of Rhodiola rosea leaf explants

: Rhodiola rosea L. is an endangered medicinal plant distributed in mountains and in high latitude regions. For its conservation, sustainable methods for the obtaining of its bioactive compounds must be developed. This work hypothesized that leaf, stem and rhizome explants of R. rosea from different geographical origins respond differently to inoculation with Rhizobium rhizogenes agropine strain ATCC43057. The objective was to generate R. rosea hairy roots (HRs) containing rol -genes. These HRs could be cultivated under axenic conditions for the extraction of the medical compounds rosavinoids and salidroside. Hereby, production of bioactive compounds could be improved per plant biomass. Thirteen R. rosea accessions of Alpine, Scandinavian, Nordic Gene Bank (NGB) and Russian origins were compared for their explant survival and HR formation. Significant differences were observed among plants from different geographical origins, where the NGB leaf explants exhibited up to 70% of HR formation and the Russian accessions did not exhibit HRs at all. Moreover, maintaining explants in light conditions after R. rhizogenes inoculation resulted in higher explant survival and HR formation rate (35%) when compared with explants kept in darkness (9%). Taken together, an efficient HR formation in roseroot by inoculation of R. rhizogenes following culturing in light was reported as a required step. This work represents a stepping-stone to R. rosea HR cultivation in bioreactors as well as regenerating whole plants. Hence, it is initiating a novel route towards high-throughput production of bioactive compounds as well preventing depletion of natural roseroot populations. following with . The effect of different from thirteen accessions of this from the and the and explants of rosea from different geographical differently with


INTRODUCTION
The medicinal plant Rhodiola rosea L. (roseroot, golden root or arctic root) has multiple uses in traditional and modern medicine, such as treating mental and physical fatigue, stress-induced depression, anxiety and high-altitude sickness (Anghelescu et al. 2018). Its wide application range is based on its health-promoting and adaptogenic properties, including mental and physical stimulation (avoiding fatigue and stress) (Zhang et al. 2016;Bangratz et al. 2018). The plant's pharmacological activity is based on the phenylpropenoid compound rosavin and its derivatives, which are only present in R. rosea, and salidroside, found in most Rhodiola species (Peschel et al. 2018).
Increasing demand of R. rosea-derived products has led to over-exploitation of plants from their natural habitats, which has placed R. rosea as an endangered species in several countries and posing the imminent risk of low quality and adulteration of roseroot related products (Booker et al. 2016). Since cultivation of this plant is challenging and costly, new

Inoculation with Rhizobium rhizogenes
The experiment targeted comparison of various R. rosea plant accessions as well as plant organs in respect to hairy root formation. Leaves, stems and rhizomes of R. rosea were inoculated with R. rhizogenes agropine strain ATCC43057 containing plasmid pRiA4 (Slightom et al. 1985;Jouanin et al. 1987). The inoculation process was based on Hegelund et al. (2017) with minor modifications. Bacteria were grown in malt, yeast and agar (MYA) medium (Tepfer and Casse-Delbart 1987).
Following sterilization, the base of the R. rosea leaves was cut off, in order to create fresh wounds for inoculation and these pieces were immersed into either inoculation (R. rhizogenes in MYA OD 600 = 0.5) or control (MYA) solution for 30 min. Subsequently, explants were dried on filter paper, transferred to cocultivation media [half strength Murashige and Skoog medium (Murashige and Skoog 1962) with vitamins (Duchefa Biochemie), 15 mg·L -1 acetosyringone (Sigma-Aldrich, St. Louis, MI, USA), 4·gL -1 gelrite (Duchefa Biochemie), pH 6] and incubated in darkness at room temperature for 48 h. The explants were then washed in sterile water containing 10 mg·L -1 timentin (ticarcillin/clavulanate, 15:1 mixture, Duchefa Biochemie), dried on filter paper and transferred to R-medium which consisted of half strength MS with 100 mg·L -1 timentin and 0.5 mmol·L -1 arginine (Sigma-Aldrich).

Hairy root induction
The effect of four genotypes, i.e. Alpine, Scandinavian, NGB and Russian, separated in two growth locations, i.e. outdoor and greenhouse, in the responsiveness to R. rhizogenes A4 was tested. Additionally, the light inducing effect was investigated by keeping explants either in darkness or in light (225 µmol·m -2 ·s -1 ) after cocultivation. Following inoculation, the explants were kept at room temperature and regularly monitored for first HR appearance (days), explants with HR, HR formation rate (%) and contamination (%), i.e. fungi and/or bacterial growth. As a WT strain of R. rhizogenes, i.e. no antibiotic selection marker was used in the inoculation. Explants of Kalanchoë blossfeldiana 'Molly' , highly responsive to this bacterial strain (Christensen et al. 2008), were used as a positive control. The HR clusters were separated from the explant once 2 cm growth was achieved, followed by maintenance in R-medium and sub-cultured to fresh medium every three weeks.

DNA extraction and polymerase chain reaction (PCR)
Hairy roots were harvested after approximately 9 weeks and minimum 3 series of subculturing to ensure the establishment of axenic putatively transformed lines. The DNA was extracted from the hairy roots with the DNA isolation kit from TaKaRa-Clontech (TaKaRa Bio Inc., Shiba, Japan) according to the manufacturer' s instructions. A NanoDrop (ThermoFischer, Waltham, MA, US) was used to measure DNA concentration and purity. Specific primer sets for fragments of rolB, aux1, virD2 and Rractin (control) ( Table 1) were used. Polymerase chain reaction products were amplified in a DNA thermal cycler (MyCycler, Biorad, Hercules, CA, USA) with the following program: 95 °C for 10 min, 40 cycles of [95 °C for 30 s, 57 °C for rolB and aux1 / 52 °C for virD2 and Rractin for 30 s, 72 °C for 30 s] and 72 °C for 7 min. For separation of the amplified PCR products, these were mixed with GelRed (Biotium, Hayward, CA, USA) and subjected to TAE 1.5% agarose gel electrophoresis at 100 V for 55 min. Finally, the products were visualized under UV-light.

Statistical analysis
All the statistic studies were performed with the GraphPad Prism 7.03 program, where significance was assessed through t-test (p < 0.05). The transformation overview is presented in Table 2. Each individual plant provided explants to at least two independent transformation experiments on which the number of explants varied according to availability, ranging from 20 to 80.

RESULTS AND DISCUSSION
In the current study, hairy roots were successfully induced from leaves of R. rosea ( Fig. 1) of diverse geographical origin by R. rhizogenes strain A4. In addition, different culture conditions were evaluated in terms of viability of surface sterilized explants and HR formation rate. This represents an initial step towards the obtaining of a transformed R. rosea plant.

Cultivating R. rosea in greenhouse decreased in vitro contamination
Leaf explants from experiments TE1-TE3 were collected from plants kept outdoor (Taastrup, Denmark) (TE1 and TE2 from plant NGB1 and TE3 from plant NGB2), which caused a high contamination rate (100% in TE2 and TE3) and the discard of the plant material. Such difficulties related to the surface sterilization of R. rosea plants from collected wild material were also reported by Khapilina et al. (2016) and Tasheva and Kosturkova (2010), who tested several sterilization methods on different explant tissues, and only 3 out of 14 combinations resulted in successful decontamination. Moreover, it is also likely that some of the contamination encountered could be due to overgrowth of R. rhizogenes, and additional washes with 10 mg·L -1 timentin and several rounds of subculture were conducted to inhibit R. rhizogenes (data not shown).
In order to avoid severe contamination due to an uncontrolled environment, the subsequent transformation experiments were performed using material from plants kept in a greenhouse environment for a week prior to inoculation. This resulted in much lower contamination rates for the greenhouse plant material (TE4-12) with an average contamination of 9% versus outdoor sourced (TE1-3) with an average of 86% (Table 2).

In vitro light exposure positively influenced hairy root formation
To investigate the optimal growth conditions to induce HR formation, the presence or absence of light on the HR regeneration phase were investigated. This study found that the inoculated explants produced more HR (35 ± 4 %) under constant light, while the explants kept in darkness conditions had lower HR formation (9 ± 1%) (Fig. 2 a). This dynamic change is described in the dark TE1-6 and light-conditioned T7-11 (Table 2). In more details, TE4 using leaves from NGB3 exhibited the first hairy root (HR) 27 days after inoculation and the percentage of explants forming HRs reached 19%. TE5 was also performed on plant material from NGB3 as a repetition. However, the plant utilized in the experiment exhibited yellowish and weak leaves and the explants did not develop hairy roots. Hence, lack of consistency among independent inoculation experiments were generally encountered.
As reflected in the literature (Flem-Bonhomme et al. 2004;Grech-Baran et al. 2014), plant explants transformed with R. rhizogenes are typically kept in darkness for root formation, emulating roots naturally growing inside soil where there is no light. However, R. rosea is a plant that grows in northern regions, where light intensity is high and days are long during summer ). Moreover, periodic light was proven to have a positive effect on the performance of the tissue and the root generation in in vitro cultures of Glycine max and Agastache foeniculum by modulating organogenesis and growth (Nourozi et al. 2016;Chen et al. 2018). Therefore, in the present work a novel strategy was pursued to compare the effect of darkness and light on explant survival.
In a pilot experiment, 24 explants from a R. rosea plant of Russian origin were placed in darkness and 24 explants of the same plant were subjected to a light intensity of approximately 225 µmol·m -2 ·s -1 and a light period of 16 h. After 17 days, the explants kept in darkness started exhibiting necrosis, while explants kept in light still maintained their green color (data not shown). This showed that R. rosea explants kept in light displayed longer viability than those kept in darkness. Leaf morphology was variable among plants from different geographical origins (Fig. 3) and they also responded differently to inoculation  ( Fig. 2 b). Hence, the interplay between plant origin, genotype and leaf morphology is an important factor to take into consideration.
Overall, statistical analysis revealed that HR formation rate was significantly (p < 0.001) higher when the explants were kept in light prior to inoculation (Fig. 2a). This is supported by Siegień et al. (2013), who tested the effect of light on shoot regeneration and root genesis on explants and sterile plantlets of Linum usitatissimum. In that study, explants exhibited 30% higher shoot regeneration when cultured in light than when cultured in darkness, and root formation was similar in both conditions. Light has also proven to have a positive effect on Artemisia annua hairy root cultures obtained by inoculation with R. rhizogenes. In that study, hairy root cultures were exposed to five different light intensities. The lowest hairy root growth was observed in darkness and hairy root growth was greater as light intensity increased (Liu et al. 2002). However, another study performed on Stevia rebaudiana reported that hairy root organogenesis should be induced in darkness, while the subsequent growth of HR-cultures performed well only under continuous light of approximately 40 μmol photons·m -2 ·s -1 (Pandey et al. 2016). From these contradictory results, it can be deducted that the effect of light on organogenesis and performance of in vitro plant cultures differs significantly among plant species.

Leaves were the most responsive tissue to transformation
Moreover, in TE5, stem explants were also used in order to test a different plant organ. However, the explants exhibited necrosis 2 weeks after inoculation (data not shown). Similarly, rhizomes were assessed as source of explants for inoculation, however no response was observed in terms of HR formation, and severe contamination developed and all the plant material died fast (data not shown).

Geographical origin strongly affected the transformation rate
Several attempts succeeded in generating micropropagated vegetative material of R. rosea species (Tasheva and Kosturkova, 2010;. However, when conducting Rhizobium-mediated inoculation of R. rosea leaves and calli, Tasheva and Kosturkova (2012 b) observed that most of the explants were dead, infected or exhibited necrosis within 2-4 weeks after inoculation. In the current study, HR formation was observed on leaf explants originating from the Nordic Gene Bank, Alpine, and Scandinavian regions, leading to the first successful report of HR formation in this species.
Moreover, the HR formation rate in NGB and Alpine plants in inoculations where explants were kept in light conditions was significantly higher (p < 0.001 for NGB and p < 0.05 for Alpine) than in explants of Scandinavian and Russian origin (Fig. 2b). Rhodiola rosea is a highly variable species, and major morphological differences are observed not only among plants from different countries but also among accessions from the same country (NGB 2005;Serebryanaya and Shipunov 2009). Likewise, morphological differences among the plants from different geographical origin (Nordic Gene Bank, Scandinavian, Russian, Alpine) were observed in the present study (Fig. 2b). This high variability can affect the way R. rosea responds to inoculation with R. rhizogenes. This is in accordance with results revealed in this work, where plants grown at the same conditions, but with different geographical origins, responded differently to inoculation with R. rhizogenes.

Hairy root phenotype was confirmed by PCR
In the current study, an agropine R. rhizogenes strain was used, hereby the transferred (T) DNA is part of a root inducing (Ri) plasmid, which harbors two distinct regions. The T L -DNA contains the rol-genes rolA, rolB, rolC and rolD, among others (Tepfer 2017;Otten 2018;Desmet et al. 2019), while the T R -DNA carries two genes involved in auxin synthesis (aux1 and aux2) (Camilleri and Jouanin 1991) and a rolB homolog, called rolB Tr (Bouchez and Camilleri 1990). T L and T R regions of the Ri-plasmid can be integrated in the plant DNA independently, but the hairy root development is mainly attributed to the presence of the rol-genes from the T L region (Halder and Jha 2016). This system is often preferred to plant/cell callus and suspension cultures as it has a strong potential of mimicking the multienzyme biosynthetic potential of the parent plants with a relative low-cost production and without significant loss of metabolic activity (Banerjee et al. 2012;Häkkinen et al. 2016).
In order to confirm successful transformation, hairy roots derived from different accessions (NGB, Alpine and Scandinavian) were selected for PCR analysis. The rolB fragment represents the T L -DNA integration into the plant genome, and its presence was detected in all the putatively transformed samples, and it was absent in the nontransformed sample (Fig. 4). Hence, R. rosea -although being in the same family (Crassulaceae) as Kalanchoë -seems not to form adventitious roots to a similar extend (Christensen et al. 2008). Additionally, nontransformed adventitious "hairy-looking" roots can develop from non-inoculated tissue. This process is highly dependent on the plant species being transformed and up to 50% adventitious root formation has been observed on non-inoculated leaf explant in comparison to inoculated (Christensen et al. 2008). Hence, this issue needs to be taken in to consideration when assessing the hairy-root formation. An aux1 fragment was chosen as representative for T R -DNA and was only present in the NGB transformed sample, indicating the combined T L +/T R + insertion. The independent integration of T L and T R -DNA into the plant genome has a ratio favoring T L insertion alone instead of both T L and T R (Roychowdhury et al. 2015;Halder and Jha 2016), which was also the case Note: 1. NGB transformed, 2. Alpine transformed 1, 3. Alpine transformed 2, 4. Scandinavian transformed, 5. Scandinavian nontransformed, 6. Water, 7. R. rhizogenes A4 Ri-plasmid. Rractin observed in this study. Moreover, a virD fragment, which is a part of the Ri-plasmid not integrated in the plant genome, was only present in the plasmid (positive control). Hence, absence of contamination by R. rhizogenes in the samples was demonstrated, verifying that the lines were true transformants. Rractin was chosen as a reference gene for R. rosea; however, only 2 of the 5 samples showed presence of this gene fragment (Fig. 4), likely due to the un-sequenced status of R. rosea and its high genetic variability (György et al. 2012). Although, further growth of the obtained HRs, both in Erlenmeyer flasks and bioreactors, was not achieved (data not shown), this work represents a stepping-stone to R. rosea HR cultivation as well as regeneration of whole plants. Hence, a novel route towards high-throughput production of bioactive compounds in HRs as well preventing depletion of natural roseroot populations is being outlined.

CONCLUSION
This study reports, for the first time, effective transformation of R. rosea with R. rhizogenes and the obtaining of viable hairy root cultures. The results indicated that light has a positive effect on survival of leaf explants of R. rosea after bacterial inoculation and resulted in higher HR formation rates. Additionally, differences in response to inoculation of R. rosea plants from different geographical origin, genotype and morphology were observed. Overall, the plant material provided by the Nordic Gene Bank was the most responsive to R. rhizogenes inoculation. Therefore, further studies on superior HR lines from this source should be conducted towards the obtaining of transformed R. rosea plants.