Introduction

Lignans constitute a large group of secondary metabolites synthesized by many plants. These compounds are usually formed from two phenylpropanoid units and manifest considerable biological activity (Smollny et al., 1998). Podophyllotoxin (PTOX), a lignan with antiviral and antineoplastic activities, is used today primarily as a precursor for the semi-synthesis of established cancer therapeutics such as etoposide, teniposide and etopophos (Gordaliza et al., 2004). To date, extracting PTOX and 6-methoxy podophyllotoxin (MPTOX) from the subterrestrial parts of Podophyllum species is still preferred to their chemical synthesis due to low product yields and the complexity of the enantioselective reactions (Botta et al., 2001). Plants from the genus Linum (the family Linaceae), with their ability to synthesize and accumulate lignans in callus and hairy root cultures, provide an additional source of PTOX and MPTOX (Farkya and Bisaria, 2008).

Biotechnological methods are being increasingly studied for their application as alternative ways of obtaining these valuable metabolites (Ramachandra and Ravishankar, 2002), although only a few have so far been produced at an industrial level owing to certain limitations (Chattopadhyay et al., 2002; Zhong, 2002). As suspension cultures are not genetically stable (Bais et al., 2001), new methods have been sought to improve lignan production. Hairy roots are transformed roots derived from the infection of wounded plant tissue by the soil bacterium, Agrobacterium rhizogenes. Such hairy roots often grow as fast as plant cell cultures or faster, and do not require hormones in the growth medium (Giri and Narasu, 2000). The greatest advantage of using hairy roots is that their biosynthetic capacity for metabolite production is similar to that of the mother plants or greater. Most of the research on the production of PTOX and related lignans has used undifferentiated cell cultures derived from non-transformed explants of naturally growing plants or aseptically growing seedlings (Smollny et al., 1998; Farkya et al., 2004; Kuhlmann et al., 2002; Nikolay and Ionkova, 2005). The in vitro production of PTOX in Podophyllum hexandrum has been reported at the bioreactor level under batch, fed-batch and continuous modes with optimized medium conditions only in non-transformed undifferentiated cell suspension cultures (Chattopadhyay et al., 2002, 2003a, b, c). Transformed calli of P. hexandrum obtained by infecting its embryos with A. rhizogenes, in the presence of phytohormones, showed a threefold increase in PTOX content compared with non-transformed cell cultures (Giri et al., 2001).

In this study, hairy root cultures were established by infecting derived leaf segments of L. album with A. rhizogenes strains LBA9402, A4, AR15834 and A. tumefaciens strain C58C1 (pRiA4). PTOX and MPTOX production in two lines of hairy roots was compared. The presence of PTOX and MPTOX in the samples was verified by ESI/MS in positive ion mode. In order to confirm MPTOX and determine its enantiomer, nuclear magnetic resonance (NMR) spectroscopy and circular dichroism (CD) spectroscopy were used, respectively.

Experimental

Plant material

Linum album seeds were collected from Sohanak, (35° 48′ N, 51° 32′ E and altitude of 1,900 m) Iran. The plant material was kindly identified by Dr. Shahrokh Kazempour (Tarbiat Modares University, Tehran, Iran) and a voucher specimen was deposited at the herbarium of the Biological Sciences Faculty, Tarbiat Modares University, Tehran, Iran. The seeds were surface-sterilized with 70 % (v/v) ethanol for 70 s, and 2 % (v/v) sodium hypochlorite for 20 min, and then were washed with sterile distilled water more than three times. Sterile seeds were treated with GA3 (500 mg L−1) overnight, then cultured on hormone-free MS medium supplemented with sucrose (3 %) (Murashige and Skoog, 1962) and incubated at 25 °C with a 16-h photoperiod. Sterile grown seedlings were used to initiate cultures.

Agrobacterium culture

Agrobacterium rhizogenes strains LBA9402, A4, AR15834 and A. tumefaciens C58C1 (pRiA4) were used for infections. The bacteria from glycerol stock (stored at −80 °C) were cultured on LB solid medium (Park and Facchini, 2000) containing 50 mg L−1 (w/v) rifampicin for screening and selecting the desired colony. Bacterial suspensions were incubated in liquid LB medium for 2 days at 28 °C.

Plant transformation

Two procedures were tested for initiation of L. album hairy roots: the syringe needle and the leaf disk methods (Lin et al., 2003). In the first procedure, separate syringe needles were contaminated with Agrobacteria colonies strains LBA9402, A4, AR15834 and C58C1 (pRiA4) and used to infect 4-week-old L. album seedlings by pricking the leaves and stems, and then placed on solid MS medium as described above. Also, seedlings pricked with a sterile syringe needle, without infection, were used as the control. After incubation in the dark at 25 °C for 2 days, the infected seedlings were placed on solid MS medium containing 200 mg L−1 cefotaxime antibiotic and incubated at 25 °C.

In the leaf disk method, leaves of L. album seedlings were excised and cut in half transversally. The leaf disks were immersed in the A. rhizogenes suspension for 10 min, blotted to remove excess bacteria, and placed on solid MS medium as described above. Uninfected sterile leaf explants were grown on the MS medium culture as the control. The obtained hairy roots were cultured in MS medium by serial sub-culturing every 2 weeks.

PCR analysis

DNA was extracted from non-transformed root, obtained from seedlings, and transformed hairy roots using the acetyl trimethyl ammonium bromide (CTAB) method (Khan et al., 2007). The transgenic nature of hairy roots was confirmed by PCR rolC, Ags (agropine synthase) and VirD specific primers. Sequences of these primers are shown in Table 1. Amplification conditions for rolC, Ags and VirD were 1 cycle at 94 °C for 2 min followed by 35 cycles of amplification (45 s at 94 °C, 40 s at 55 °C and 45 s at 72 °C) and 1 cycle at 72 °C for 10 min. PCR products were analyzed by electrophoresis in a 1 % (w/v) agarose–ethidium bromide gel. Meanwhile, the Ri-plasmid of A. rhizogenes was used as a positive control (Wang et al., 2006).

Table 1 Sequences of used primers
Full size table

Growth kinetics and biomass measurements

The growth rate was measured as hairy root growth kinetics at the first of the experiment. The experimental conditions were as follows. Hairy root lines with 2.4 g of root mass (on fresh weight basis) were transferred to a 250-ml Erlenmeyer flask containing 50 ml of hormone-free MS liquid medium at 25 °C for period of 4 weeks. Flasks were harvested with three replications each week (up to day 28 after inoculation). The growth of hairy root lines was measured every week. Hairy roots were dried by blotting paper to remove all external moisture and weighed to determine the final fresh weight (FW) before being frozen for lignan extraction.

Extraction of lignans

Powdered hairy roots (2 g) were extracted by sonication in methanol (80 %) during 1 h. Dichloromethane and water (1:1 v/v) were added to obtain a partition of compounds between two layers. The dichloromethane fractions were collected, dried and dissolved in 500 μl of HPLC grade methanol and then injected into the HPLC (Van Uden et al., 1989) (Philips, UV/Vis detector; Pu 41110). The presence of PTOX and MPTOX in the hairy roots was determined by ESI/MS in positive ion mode and 0.05 % formic acid was used.

MPTOX purification and confirmation

Analytical separation was carried out to confirm MPTOX and determine its enantiomer. Separations were performed on a C18-S5ODS3 (250 × 4.6 mm) RP column with a flow rate of 1.0 ml min−1. A fraction containing MPTOX was collected, lyophilized and dissolved in DMSO and subjected to 500 MHz NMR spectroscopy. A NMR spectrum was recorded on Bruker DRX500 Avance running XWIN-NMR 2.6 software using 500 μl samples. The chemical shifts (delta) are reported in parts per million (ppm). Also, FT-RI was carried out with a Bruker tensor 27 spectrometer.

The CD spectrum was obtained with a JASCO J-715 spectropolarimeter using cuvettes with optical lengths of 1 cm and detection at 250–350 nm. The final data were obtained by subtracting the buffer contribution (DMSO) from the original spectrum.

Optical rotation ([α]D values) of MPTOX was measured on a Perkin Elmer 241 automatic polarimeter (Überlingen, Germany) at 20 °C. The specific rotation ([α]D) at the sodium D line (λ 589 Å) was calculated as follows:

[α]D = (100 × α)/(l × c), α = observed angle of rotation, l = length of the path of light through the solution in decimeters (dm), c = concentration of the solution in g/100 ml.

Analysis of lignans

PTOX and MPTOX quantification was analyzed using HPLC as described previously by Yousefzadi et al. (2010b). The stationary phase was a C18-ODS3, 5 μm (250–4.6 mm) column. The elution solvent was composed of water and acetonitrile with a gradient system. This system with acetonitrile (A%) and water (B%) was used as follows: (0 min, flow rate 0.8): A (40), B (33); (10 min, flow rate 1): A (67), B (33); (17 min, flow rate 1): A (40), B (60); (21 min, flow rate 0.8): A (40), B (60). A spectrophotometric UV detection was used (290 nm). The calibration was performed with standard PTOX (RT = 9.5 ± 0.1 min) and MPTOX (RT = 11.5 ± 0.1 min) (Fig. 1). Every sample was assayed with three replications. The lignan production was determined using a calibration curve compared with standards and a co-chromatogram of the standards and samples.

Fig. 1
figure 1

6-Methoxy podophyllotoxin (a) and podophyllotoxin (b) calibration curve using HPLC. Each value is the mean of three replications

Full size image

Statistical analysis

All data were analyzed using MSTAT-C software. Duncan’s multiple range tests were used to show statistical differences between treatment methods and controls. P ≤ 0.05 or ≤0.01 was considered significantly different.

Results and Discussion

L. album hairy root induction

Although all four Agrobacterium strains applied in this work are among the most virulent in common usage (Lin et al., 2003), despite several attempts using different bacterial strains and infection sites, no hairy roots were obtained using the leaf disk infection technique. Infection by syringe needle was more successful, with strains LBA9402 and C58C1 (pRiA4) showing their rooting ability, the former more readily. LBA9402 hairy roots of L. album were induced after 4 weeks of infection and had a higher transformed percentage (20 %), while C58C1 (pRiA4) hairy roots were induced after 5 weeks with a lower percentage of transformation (10 %) (data are not shown). Some transgenic roots turned brown and aged considerably faster than others. These brown roots were discarded and only those showing a good growth capacity were maintained for further characterization. Although various explant sources can be used to produce hairy roots in vitro, in some cases specific tissues perform better than others (Mishra et al., 2011). Regarding L. album, young stems produced more primary cultures than other tissues. Efficiency of transformation differed among the Agrobacterium strains, as reported by other authors (Giri et al., 2001; Mishra et al., 2011).

A time course experiment (up to 4 weeks at defined time intervals) was performed to monitor root growth and biomass production and also to estimate lignan content. The growth kinetics of L. album hairy roots showed three different phases: the lag phase (first week) when the specific growth rate is essentially slow, the exponential growth phase (second and third weeks) when the specific growth rate is constant and the stationary phase (fourth week) with the emergence of at least one limiting parameter. The highest hairy root growth was recorded with strain LBA9402 (12.455 g FW) after 4 weeks of culture (Fig. 2). Hairy roots induced by C58C1 (pRiA4) showed comparatively slow growth (11.083 g FW). However, transformed hairy roots induced by both strains exhibited negative geotropism. The transformed roots with A. rhizogenes have an altered phenotype, showing profuse lateral growth, lack of geotropism and fast growth in the culture (Mishra et al., 2011).

Fig. 2
figure 2

Growth pattern of hairy roots induced by A. rhizogenes strains LBA9402 and A. tumefactions C58C1 (pRiA4). Data represent average values from four separate experiments ± SD

Full size image

Confirmation of transformation

In order to assess the genetic status of the hairy roots, we used a PCR-based analysis that targeted rolC, Ags (opine synthase) and virD genes. The rolC and Ags genes are diagnostic for T-DNA integration into the host genome. The virD gene, located outside the T-DNA, is diagnostic for the presence of any remaining Agrobacteria in the root tissue. PCR analysis revealed the presence of the rolC and Ags genes in hairy roots, indicating that the T-DNA was retained in hairy roots (Fig. 3). Confirmation of contamination was also done by PCR analysis using virD (483 bp) gene-specific primers. It was not found in the hairy root lines, thus indicating the absence of contamination with Agrobacterium strains LBA9402 and C58C1. Also, rolC, Ags and virD genes were not detected in the non-transformed root.

Fig. 3
figure 3

Electrophoretogram showing the presence of rolC (a), Ags (b) genes and absence of VirD (c) gene (534, 347, 483 bp bands, respectively) in transformed hairy roots; M molecular weight marker, PC positive control, R1 and R2: transformed root lines LBA9402 and C58C1, respectively, R−: the non-transformed root

Full size image

Confirmation of MPTOX and PTOX

Identification of MPTOX in transformed root cultures

Lignan (dimmers of monolignols) formation and lignin (polymers of monolignols) biosynthesis are catalyzed by different enzymes, and a consequence is that natural lignans are normally enantiomerically pure because they arise from stereochemically controlled coupling (Dewik, 2002). For example, the lignan pinoresinol, and its derived matairesinol, was found only as their (-)-isomers in suspension-cultured Pinus taeda (Eberhardt et al. 1993). Data acquired using a polarimeter (optical rotation = [α]D −60°) and CD spectrum confirmed the presence of the (-)-isomer of MPTOX (Fig. 4).

Fig. 4
figure 4

CD spectrum of MPTOX. Detected solvents: methanol of chromatographic grade, detected temperature: 25 °C. UV detected wavelength at 250–350 nm

Full size image

Moreover, this less hydrophilic compound was identified by NMR and IR spectroscopy. Significant 1H-NMR data are shown in the following order: chemical shift (δ) expressed in ppm, multiplicity (s, singlet; d, doublet; m, multiplet and ABq, AB quartet).

1H-NMR (500 MHz, (CD3)2 SO): δ = 6.39 (2H, s), 6.29 (1H, s), 5.97 (2H, d), 5.04 (1H, d), 4.93 (1H, m), 4.45 (2 H, ABq), 4.08 (1H, dd), 3.95 (3H, s, OCH3), 3.62 (6H, s, (OCH3)2), 3.6 (3H, s, OCH3), 3.0 (1H, dd). Melting point: 80–83. IR: 3449.20, 2924.41, 2862.45, 1775.34, 1634.25, 1465.59, 1247.14, 1124.43, 1050.77.

MS analyses of MPTOX

For identification, liquid chromatography–electrospray–mass spectrometry/mass spectrometry (LC–ESI–MS/MS) analysis was carried out. Several reports have been cited in the literature demonstrating the applicability of MS and MS/MS coupled with liquid chromatography in the analyses of PTOX and its derivatives (Li et al., 2007; Lin et al., 2008). In this work, [M + H]+ and [M + Na]+ ions of 6-MPTOX were determined. According to Fig. 5, the observed mass at 467 Da corresponds to the [M + Na]+ ion of 6-MPTOX and the observed mass of 445 Da is related to the [M + H]+ ion of 6-MPTOX. The other peaks at mass of 427 and 343 Da correspond to fragments illustrated in Fig. 6.

Fig. 5
figure 5

Selected molecular ion (445 m/z) in TIC (up) and mass spectrum of 6-MPTOX (down)

Full size image
Fig. 6
figure 6

The fragmentation of MPTOX (according to Li et al. 2007)

Full size image

Confirmation of PTOX

The presence of PTOX in the hairy roots was verified by ESI/MS in positive ion mode, as previously described (Yousefzadi et al., 2010a).

Quantification of lignans

Among the examined samples, hairy roots induced by strain LBA9402 were found to be the best in terms of biomass as well as lignan production. The maximum concentrations of PTOX (105 μg g−1 DW), and 6-MPTOX (48 mg g−1 DW) were recorded in hairy roots induced by strain LBA9402 (Fig. 7). In the roots of mother plants which were collected from its natural habitat, concentrations of PTOX and 6-MPTOX were 14.11 and 1.13 mg g−1 DW, respectively. When compared to the roots of mother plant, the hairy root cultures were found to contain higher concentrations of lignans. This confirms that the insertion sites of Ri-T-DNA interfere with the secondary metabolite biosynthetic pathway of lignans. Similar results have been reported by several groups (Mishra et al., 2011; Verma et al., 2007).

Fig. 7
figure 7

PTOX and MPTOX production in a 4-week period. PTOX: podophyllotoxin. MPTOX: 6-methoxy podophyllotoxin. Data represent average values from three separate experiments ± SD

Full size image

Conclusion

In general, the amount of anti-cancer lignans (PTOX and MPTOX) in plants is considerably less than in transformed hairy roots. In addition, the hairy roots exhibited more rapid growth than plants. This study may open possibilities for bioreactor-based production of the pharmaceutically important metabolites PTOX and 6-MPTOX.