Introduction

Hypericum perforatum L. (St. John’s wort) is an herbaceous perennial plant with five-petalled bright yellow flowers that has for centuries been utilized in traditional medicine for the treatment of wounds, neurological disorders and depression (Nathan 2001). A wide range of biologically active secondary metabolites, belonging to different chemical groups, are produced in several types of secretory structures, such as multicellular dark or translucent glands and secretory canals present on leaves, stems, flower petals and stamens (Zobayed et al. 2006; Ciccarelli et al. 2001). Hypericins and hyperforins are considered the main contributors to the anticancer, antidepressive and antimicrobial activities of Hypericum herbal extracts (Kubin et al. 2005; Beerhues 2006; Fiebich et al. 2011), and a broad spectrum of flavonoids and phenolic acids contributes to its potent antioxidant activity (Butterweck 2003; Silva et al. 2005).

Cultivated for medical purposes, the plants require approximately 2 years to mature; unfortunately, breeding for increased secondary metabolite accumulation sometimes leads to a loss of biological activity (McCoy and Camper 2002). Also, commercial Hypericum preparations show inconsistency in their contents of bioactive compounds as a result of different field environments (Southwell and Campbell 1991; Büter et al. 1998; Çirak et al. 2014). Consequently, natural populations are still highly exploited and endangered by human harvesting.

The use of in vitro culture systems is an alternative strategy for the production of satisfactory amounts of plant material, as well as to obtain less variable preparations (Bourgaud et al. 2001; Zhou and Wu 2006; Karuppusamy 2009). Large-scale production of adventitious roots and secondary metabolites in bioreactors has recently been employed (Cui et al. 2010; Wu et al. 2014). In addition, genetic transformation by Agrobacterium rhizogenes results in hairy roots that can be subsequently propagated in vitro in growth regulator-free medium at low cost. Moreover, it has been reported that plants regenerated from transformed roots can produce secondary metabolites in greater quantities than untransformed ones (Chaudhuri et al. 2006; Gangopadhyay et al. 2010; Majumdar et al. 2011). Because of these convenient aspects, efficient A. rhizogenes-mediated transformation protocols have been established in many medicinal plants (reviewed by Roychowdhury et al. 2013), including H. perforatum (Di Guardo et al. 2003; Vinterhalter et al. 2006; Bertoli et al. 2008; Koperdáková et al. 2009, 2011).

In a previous report, transgenic hairy roots were successfully induced by inoculating in vitro-grown shoots of H. perforatum with A. rhizogenes agropine-type strain A4M70GUS (Vinterhalter et al. 2006). Spontaneous direct bud regeneration was achieved by culturing transformed roots on hormone-free medium. Since the main biologically active compounds of H. perforatum, hypericins and hyperforins, are predominantly found in the aerial parts of the plant (Karppinen and Hohtola 2008), improvement of shoot regeneration from established transgenic hairy root cultures and their growth was an objective of the present study.

A preliminary experiment showed that sucrose significantly affected the shoot regeneration potential of H. perforatum hairy root clone 14 (Vinterhalter et al. 2006); thus, we focused further on the effects of sucrose on the shoot morphogenesis potential of transgenic hairy root clones and enhanced production of the aerial parts of plants. Molecular characterization of tested clones was also performed to provide further understanding of the processes involved in morphogenesis in transgenic H. perforatum.

Materials and methods

Hairy root cultures and growth conditions

Hairy root clones used in this study were obtained by transformation of plants grown in vitro, initially established from the local (border area around Belgrade, Serbia) H. perforatum plant population, with Agrobacterium rhizogenes strain A4M70GUS (Vinterhalter et al. 2006). A4M70GUS carries the non-disarmed plasmid pRiA4 in which the uidA sequence (coding for the β-glucuronidase enzyme, GUS), under the 70S promoter (enhancer-doubled 35S CaMV promoter) and followed by nos polyadenylation sequence, was integrated between the rolC and rolD genes in the TL-DNA (Tepfer and Casse-Delbart 1987). The presence of the uidA gene was previously detected by PCR in clones 2, 14, 15, 27 and 29, obtained from individual transformation events (Vinterhalter et al. 2006). Root explants of all above mentioned clones were cultured in Petri plates (Ø 90 mm; Spektar, Čačak, Serbia) containing 25 ml of basal medium (BM) with WPM macrosalts (Woody Plant Medium; Lloyd and McCown 1981), MS microsalts and vitamins (Murashige and Skoog 1962), 100 mg l−1 myo-inositol and 0.64 % (w/v) agar (Torlak Institute of Virology, Vaccines and Sera, Belgrade, Serbia). The pH of all media was adjusted to 5.8 before autoclaving at 114 °C and 80 kPa for 25 min. The cultures were maintained at 25 ± 2 °C, under cool white fluorescent light with a photosynthetic photon flux density of 55 μmol m−2 s−1 and 16-h light/8-h dark photoperiod.

In vitro evaluation of hairy root regeneration potential

Morphological measurements were done on 35-day-old roots grown on BM supplemented with varied concentrations of sucrose (0.5, 1, 2, 4, 6 and 8 %). Each treatment consisted of three to six Petri plates with ten root apical explants, each 15 mm long. The final length of the root explants, the number of lateral roots and regenerated shoots per root explant were scored. The effects of different sucrose concentrations and the genotype on the regeneration potential of transgenic roots were evaluated using standard two-way analysis of variance (ANOVA). Data were subjected to square root transformation and percentage data to angular transformation prior to statistical analysis and inversely transformed for presentation. Differences between the corresponding means were separated using Fisher’s LSD test at P ≤ 0.05. Statistical analyses were performed using SAS software (SAS Institute, 2002; SAS/STAT, ver. 9.00. SAS Institute Inc., Cary, NC, USA).

PCR

For PCR analysis, genomic DNA from 5-week-old in vitro-grown transformed and non-transformed control plants was isolated according to the method of Zhou et al. (1994). The pairs of primers used for amplification of the rolA, -B, -C and -D genes from the TL-DNA region are shown in Table 1.

Table 1 Primers used for real-time RT-PCR analyses of rol genes and GAPDH reference gene with the expected amplification product sizes
Full size table

PCR reactions were set up using GeneAmp® Gold PCR Reagent Kit (Applied Biosystems Co., Foster City, CA, USA) according to the manufacturer’s instructions. All reactions were carried out for 36 cycles of denaturation at 95 °C for 1 min, primer annealing at 60 °C for 1 min and extension at 72 °C for 1 min, after initial denaturation at 95 °C for 5 min. The PCR products were separated on a 1.2 % agarose gel and visualized under UV light after ethidium bromide staining.

Real-time RT-PCR

Total RNA was extracted from 5-week-old shoots spontaneously regenerated on transgenic hairy roots and grown in vitro on BM with 2 % sucrose. Extraction was performed according to the method of Gasic et al. (2004), including treatment with DNAseI (Thermo Scientific, Waltham, MA, USA). Reverse transcription was performed on 1 μg of total RNA, using the GeneAmp® Gold RNA PCR Core Kit (Applied Biosystems Co.) with oligo-dT primers, according to the manufacturer’s instructions. cDNA corresponding to 200 ng of RNA was used for setting up the real-time PCR reactions using Maxima™ SYBR Green/ROX Master Mix (Thermo Scientific) and an ABI PRISM 7000 Sequence Detection System (Applied Biosystems Co.). To amplify the rolA, -B, -C and -D genes, the specific primers shown in Table 1 were used. PCR reaction conditions included initial denaturation (95 °C for 5 min), followed by 40 cycles of denaturation at 95 °C for 30 s, annealing at 60 °C for 30 s and extension at 72 °C for 30 s. Additional melting curve analyses were performed by cooling the reactions to 60 °C and then increasing the temperature to 95 °C with a slope of 0.1 °C s−1, while measuring the fluorescence continuously. Constitutively expressed glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH, GenBank accession No. GU014528) served as the internal reference gene (Table 1). All reactions were carried out in triplicate and repeated twice. The results were analyzed using 7000 System SDS Software (Applied Biosystems Co.).

For absolute quantification, standards were made from cDNA fragments of each rol gene amplified using conventional PCR with specific primers (Table 1) and the pRiA4M70GUS plasmid DNA as a template, and subsequently purified from agarose gels (GeneJET Gel Extraction Kit, Thermo Scientific). After spectrophotometric quantification, the number of recovered DNA molecules was calculated (http://endmemo.com/bio/dnacopynum.php) and serial dilutions of the DNAs were made (from 108–10 copies per μl). Overall efficiencies (E) of PCR were calculated from the slopes of the standard curves for serial dilutions. Results are presented as the mean transcript number per 1 ng of total RNA with the standard error (SE) of the mean. The statistical significance of the differences between the means was determined by ANOVA and Fisher’s LSD test (P ≤ 0.05).

Flow cytometry

Shoots from in vitro-grown regenerants and young leaves of field-grown plants were collected and used for analysis. The DNA content was determined using flow cytometry on isolated nuclei stained with propidium iodide (PI), according to Doležel (1991). Briefly, leaf tissue of the studied H. perforatum lines was chopped together with leaves of Trifolium repens L. cv. Milo (2C = 2.07; Arumuganathan and Earle 1991) in LB01 buffer. Nuclei were passed through a 10-µm nylon-mesh filter and stained with 50 µg ml−1 PI and 50 µg ml−1 ribonuclease. Measurements were done on a PAS flow cytometer (Partec, Münster, Germany) using a linear scale. Samples were analyzed using an argon laser tuned to 488 nm, with emissions measured through an RG 590 long-pass optical filter. Seven thousand nuclei per sample were measured with three repetitions for each line. Flomax software (Partec) was used for calculating the positions of the G0/G1 peaks.

Results and discussion

Impact of sucrose concentration on shoot regeneration potential of hairy roots

The morphological evaluation of the hairy root clones in response to 0.5–8 % sucrose indicated that both genotype and sucrose concentration affected regeneration processes (Fig. 1). ANOVA confirmed that factor interaction significantly affected not only root elongation and branching, but also shoot regeneration potential (Table 2).

Fig. 1
figure 1

Shoot regeneration on transgenic Hypericum perforatum hairy roots of clones 14 and 29 on BM supplemented with various concentrations of sucrose (0.5–8 %). Bars represent 1 cm

Full size image
Table 2 ANOVA table of F and P values on the effect of genotype and sucrose concentration for morphological parameters of H. perforatum hairy roots
Full size table

The measurement of root elongation showed that clones responded differently to various sucrose concentrations in media (Fig. 2a). Roots of clones 2 and 29 elongated intensively even at the lowest sucrose concentration of 0.5 % with an almost tripling of the initial length. These two clones reached the highest root lengths of 68.3 and 61.0 mm, respectively, on 2 % sucrose. All other clones had length increases of about 5 mm at the lowest sucrose level and reached a maximum on 4 % sucrose (63.5, 58.6 and 67.7 mm for clones 14, 15 and 27, respectively).

Fig. 2
figure 2

Hypericum perforatum hairy root clones performance on BM with various concentrations of sucrose (0.5–8 %). Main root length (a), percentage of roots with developed lateral roots (b), and number of lateral roots per root explant (c) were scored after 35 days of treatment. Results are expressed as mean ± SE (n = 30–60). Values with the same letters are not significantly different at the P ≤ 0.05 level according to Fisher’s LSD test

Full size image

A similar response to sucrose was observed for lateral root differentiation (Fig. 2b, c). For clones 2 and 29, about 80 % of hairy roots developed more than six lateral roots per explant on 0.5 % sucrose. Conversely, for the other three clones, only 20 % of roots tended to form just one lateral root per explant. For these clones, the highest lateral root differentiation was obtained on medium with 4 % sucrose, but with significantly lower frequencies (60–70 %) and with a maximum of only six roots per explant. The highest number of lateral roots (almost 14) was scored for clone 2 grown on 4 % sucrose. On concentrations higher than 4 %, all clones showed a decline in the average number of lateral roots. Similar results showing that sucrose concentration affects the growth of in vitro-maintained hairy roots, with optimal responses on between 3 and 5 % sucrose, were obtained by Lourenço et al. (2002) for Centaurea calcitrapa and Yu et al. (1996) for Solanum aviculare.

Although it has been reported that hairy root clones of H. perforatum spontaneously regenerate shoots on hormone-free medium supplemented with sucrose at concentrations of 2 % (Vinterhalter et al. 2006) or 3 % (Di Guardo et al. 2003), our results show that this process is considerably affected by sucrose. Root segments of clones 2, 27 and 29 regenerated the maximum percentage of shoots at lower sucrose concentrations (2 % for clones 2 and 27 and 1 % for clone 29). In contrast, for the most intensive regeneration in clones 14 and 15, more sucrose (4 %) was required. The highest number of root explants with initiated shoots was recorded for clone 14 on 4 % sucrose (more than 80 %) (Fig. 3a), which also had a remarkably high number of regenerated shoots (6.5 shoots per root explant) (Fig. 3b).

Fig. 3
figure 3

Shoot regeneration potential of Hypericum perforatum hairy roots grown on BM with various concentrations of sucrose (0.5–8 %). Percentage of roots with regenerated shoots (a) and the mean number of regenerated shoots per root explant (b) were scored after 35 days of treatment. Results are expressed as mean ± SE (n = 30–60); values with the same letters are not significantly different at the P ≤ 0.05 level according to Fisher’s LSD test

Full size image

Plant cells and tissues grown in vitro due to a lack of autotrophic ability require the addition of an external carbon source into the growth medium. The most commonly used carbon supply is sucrose, which at an optimal concentration is sufficient for the energy requirements of basic cell division and differentiation. Moreover, it has been well-documented that sucrose is one of the factors that controls the induction and growth of shoots and that it can, at physiologically relevant concentrations, significantly promote regeneration processes in plants grown in vitro (Gibson 2000; Pavlović et al. 2010; Ilyas et al. 2013; Rathore et al. 2013), especially in non-autotrophic systems such as hairy roots (Yu et al. 1996; Lourenço et al. 2002).

In addition to acting as the main energy source, sugars also act as potent signaling molecules in plants (Hanson and Smeekens 2009). Since Wu et al. (2005) found that exogenous sucrose in the growth medium fully restored the elongation of the Arabidopsis growth-arrested stip mutants by reactivating their primary shoot and root meristems, there have been indications that sucrose has a signaling function regarding the sequence of events in the cell cycle and, in combination with auxins and cytokinins, regulates plant cell proliferation and development (Francis and Halford 2006; Hartig and Beck 2006). Some of the mechanisms by which sugar regulates morphogenesis have already been elucidated. Earlier reports found that sucrose can accelerate growth by activating D-type cyclin (CYCD) expression, which promotes the G1 to S cell cycle phase transition (Riou-Khamlichi et al. 2000; Gaudin et al. 2000). CYCD4 expression during the formation of lateral root primordia in Arabidopsis was also found to be induced by the addition of sucrose in a suspension culture (De Veylder et al. 1999). More recent studies have confirmed the regulatory role of sugar on CycD gene induction and expression (Kwon and Wang 2011; Rosa et al. 2013). In addition, Skylar et al. (2011) demonstrated that sugar signals also indirectly activate the transcription of key cell components for the G2 to M phase transition and thus affect meristem proliferation, size and structure. Therefore, differential reaction of an individual clone to a particular sucrose concentration could be explained by its individual genetic and physiological makeup, since sugar signals are linked in a signaling network to plant hormones, resulting in a specific morphogenetic pathway.

Although the morphogenetic responses of the analyzed hairy root clones were different, we observed a strong influence of the sucrose concentration on this process. Low, suboptimal sucrose concentrations (below 2 %) probably could not provide enough energy and acted only as building blocks, so hairy roots developed few secondary roots and failed to regenerate shoot buds. However, when present in high, supraoptimal concentrations (6–8 %), sucrose had negative effects, possibly due to higher osmotic pressure, which has been shown to be deleterious to root primordium induction (Cui et al. 2010).

Interestingly, shoots regenerated from hairy roots did not show the phenotypic changes typical of so-called “hairy root syndrome” (Fig. 4a). This is quite unusual for A. rhizogenes transformed plants, since the expression of rol genes alters the normal phenotypic appearance in many plant species, including H. perforatum (Di Guardo et al. 2003; Koperdáková et al. 2009). All transgenic Hypericum plants obtained by A4M70GUS transformation in the present study and regenerated on media with sucrose concentrations up to 4 % were phenotypically normal, with articulated apical growth, opposite, simple oval leaves, and characteristic black glands on the leaf edges (Fig. 4a, b). However, when applied at concentrations of 4 % and higher, sucrose had an adverse effect on the growth of both shoots and roots (Fig. 4c, d). Roots became dark brown, with red tips (Fig. 4c) and the dark glands were less prominent on most of the leaves (Fig. 4d).

Fig. 4
figure 4

Shoot regeneration on transgenic Hypericum perforatum hairy roots of clone 15. Regenerated shoots on 2 % sucrose (a) with normal phenotypic appearance and black glands on leaf edges indicated by an arrow (b) and hairy roots grown on 8 % sucrose with no regenerated shoots and red root apices (c). Shoots regenerated on 6 % sucrose had fewer black glands on leaf edges (d). Bars represent 1 cm

Full size image

Genetic characterization of H. perforatum regenerated shoots

Although hairy root-derived shoots displayed a lack of morphological alterations typical of the hairy root phenotype, the presence of the rolA, -B, -C and -D genes was demonstrated by PCR in all clones (Fig. 5). No specific amplifications of rol transgenes were observed in control plants. Furthermore, Southern hybridization confirmed the stable integration of the 4.3 kb EcoRI fragment, containing the rolA, -B and -C genes, in all five clones (data not shown). These genes are essential and sufficient for induction and development of hairy roots (Jouanin et al. 1987).

Fig. 5
figure 5

PCR analysis of rolA, -B, -C and -D genes in A. rhizogenes-mediated transformed Hypericum perforatum in vitro-grown plants. L 100 bp DNA molecular weight marker, NTC non-transformed control, A.r. positive control (DNA from the plasmid used for A. rhizogenes-mediated transformation); B PCR blank. Numbers refer to individual H. perforatum transformed lines (2, 14, 15, 27 and 29)

Full size image

In addition, flow cytometric measurements revealed that all transgenic as well as non-transformed control plants were diploid with an average nuclear DNA content of 0.670 ± 0.002 pg per nucleus, with no significant differences between the analyzed lines. Since flow cytometry is a reliable tool used for the detection of variants among regenerated transgenic plants (Ochatt et al. 2013), the genomic stability found in the present study confirmed that it was not affected either by transformation or by the in vitro regeneration process, indicating trueness-to-type of the reported transgenic plants.

According to their morphogenetic process, all analyzed H. perforatum transgenic clones were grouped into two categories. Clones 2 and 29 possessed high lateral root differentiation potential at lower sucrose concentrations (below 2 %), but with less effective shoot bud proliferation. On the other hand, clones 14, 15 and 27 had lower lateral roots, but higher shoot differentiation potential on sucrose-enriched media (2–4 %).

We assumed that the observed differences in the morphogenic potential of the tested clones could also be result of the transformation process itself. Alternations in the development and metabolism of transformed plants could be attributed to unique transformation events, such as the number of inserted copies and their position in the plant genome, levels of expression, methylation and the chromatin structure at the insertion site (Kooter et al. 1999; Vaucheret and Fagard 2001).

To determine the mRNA levels of the rolA, -B, -C and -D genes, we used the real-time RT-PCR approach. Three of five analyzed clones were selected according to their morphogenetic performance (clones 2, 14 and 27). Real-time RT-PCR confirmed the expression of the rolA, -B and -C genes, while the expression of the rolD gene was not detected, although its presence was confirmed by PCR, as mentioned above. Similar loss of expression for rolD has been reported by others (Slightom et al. 1986; Chriqui et al. 1996; Alpizar et al. 2008; Koperdáková et al. 2009; Komarovská et al. 2010). For the other three analyzed rol genes, considerable differences were detected in the absolute amounts of the transcripts between clones (Table 3). The highest levels of expression for all three analyzed rol genes were detected in clone 2. About 13 times more rolA and rolB and 4 times more rolC transcripts were detected in the shoots of clone 2 than in clones 14 and 27, which had similar levels of expression. Differences in the relative proportions of the expression of each transgene were also documented by Koperdáková et al. (2009), with decreased amounts of the rolC transcript compared to the rolA and rolB transcripts. In both cases, an effect of the differential position on the expression of each transgene, despite having been linked in the same inserted T-DNA, could be the explanation (Beltrán et al. 2009).

Table 3 Real-time RT-PCR analysis of rolA, rolB and rolC expression in shoots of 5-week-old in vitro-grown transgenic Hypericum perforatum clones 2, 14 and 27, and non-transformed controls (NTC)
Full size table

Taken altogether, our results show that clone 2, when grown on regular BM with 2 % sucrose, had higher expression of rol genes; this was reflected in the increased elongation of root explants and intensive lateral root development. Two other clones had significantly lower levels of rol gene transcripts, which correlated with shorter roots and fewer lateral roots. In accordance with these findings, shoot development was more favored in clones 14 and 27 than in clone 2. We believe that more rol transcripts direct morphogenesis to root versus shoot development. There have already been some attempts to correlate the phenotypic appearance with the number of rol gene insertions in Hypericum. Komarovská et al. (2010) claimed that the number of rol genes integrated into clones transformed with the agropine ATCC 15834 strain influenced almost all analyzed quantitative characteristics, as well as the total content of hypericins. Plants with a higher transgene copy number resembled the controls, while those with one or two integrations expressed the features typical of hairy root syndrome more intensively. These authors believed that post-transcriptional silencing, caused by a high number of copies, was responsible for the similarities between clones with more rol copies and controls.

The behavior of hairy roots grown on elevated concentrations of sucrose (4 %) emphasized the role of rol genes in directing transformed plant development. An increase in the sucrose levels for clone 2 caused similar or even more intensive rooting, while the number of regenerated shoots significantly declined. For clones 14 and 27, sucrose promoted rooting, supporting the hypothesis of the inductive effects of sucrose on rol gene expression. Namely, it is known that sucrose can modulate the expression of the rolC promoter (Yokoyama et al. 1994; Nillson et al. 1996). Furthermore, it is thought that rolB has a crucial role for the induction of rhizogenesis, and that both sucrose and auxins, as regulators of its promoter, are necessary. Only cells with sufficient levels of these compounds will be competent to serve as targets for A. rhizogenes infection and will induce the formation of adventitious roots (Nilsson and Olsson 1997). At the same time, sucrose enhances the sensitivity of cells to auxins (Lazzeri et al. 1988), thereby stimulating rolB promoter activity. Elevated expression of the rolB and rolC unequivocally leads to intensified rhizogenesis as opposed to shoot development.

Conclusions

Variations in shoot morphogenesis potential promoted by sucrose were observed for H. perforatum hairy root clones. Different sucrose concentrations in the growth media affected shoot development, with clones 2 and 29 more effective on lower sucrose concentrations, while the other three clones, 14, 15 and 27, intensified shoot development only on elevated sucrose concentrations (4 %). For all clones, concentrations above 4 % (6–8 %) had deleterious effects on both root and shoot development. Thus, in further investigation, selection of clones with intensified regeneration of morphologically normal shoots on the media with regular or slightly elevated sucrose concentration will be undertaken. We assumed the observed variations in morphogenesis potential could be attributed to different levels of expression of the integrated A. rhizogenes genes rolA, -B and -C. Clones with higher levels of rol genes expression had more intensive root development, which was expected according to their original role in the promotion of root formation. So far, H. perforatum hairy root cultures have not been subjected to phytochemical analysis of their secondary metabolite contents. Therefore, the established transgenic H. perforatum cultures represent a well-defined and powerful tool for further biotechnological and metabolomic investigations. The examination of the secondary metabolite content in regenerated shoots, as well as the influence of sucrose on this, will be the focus of subsequent studies.

Author contribution statement

B. V. and D. V. designed research; B. V. performed sucrose treatments and did statistical analyses; S. Z-K. performed Southern analysis and contributed in writing of the manuscript; N. M. performed PCR analyses and contributed in writing of the manuscript; B. B. performed flow cytometry; J. S. performed real-time RT-PCR analyses, wrote the manuscript and prepared figures.