Introduction

Azadirachtin (C35H44O16) is a high-value secondary metabolite obtained from Azadirachta indica (Neem), commercially used as a broad-spectrum biopesticide. Limitations like climatic and geographical variability in the natural resource, contamination with toxic metabolites during seed collection, and uneconomical chemical synthesis with the conventional method of seed extraction has led to the development of plant tissue culture technology as an alternative route for continuous supply of azadirachtin to meet the increasing demand of azadirachtin [1]. In literature, in vitro production of azadirachtin has been widely reported using callus/cell culture of A. indica [2–7]. Prakash and Srivastava [8] have reported the azadirachtin yield enhancement up to 12.2 % (w/w) in continuous cultivation (with cell retention) of A. indica plant cells. However, the major constraint with cell cultures is the gradual decrease (deteriorations) in the biosynthetic capacity of cell suspension cultures resulting in low yields of secondary metabolites during subcultures. In this context, hairy root culture has been established in literature by using transformation of plant cell with a natural vector system, Agrobacterium rhizogenes, as a promising system for in vitro secondary metabolite production due to their inherent genetic/biochemical stability with fast growth in hormone-free economical medium. High yields of various secondary metabolites have been reported in hairy root cultures of several plant species [9]. However, the major constraint in their commercial exploitation as a production platform has been its scale-up limitation. As a result, identification of a suitable scalable bioreactor configuration to facilitate mass cultivation of hairy roots for large-scale production of plant metabolites has become the need of the hour. The filamentous growth of hairy roots presents limitation for scale-up. Due to branching, the roots form an interlocked matrix that exhibit resistance to nutrient and oxygen transfer. It has been invariably observed in literature that hairy root shake flask results are unable to get reproduced at bioreactor level. The hairy root morphology is quite plastic as the roots eventually respond to the local environmental changes. Changes in morphology, including changes in the density and length of the root hairs, directly affect the secondary metabolite production from hairy roots [10]. Hence, adaptation to the reactor environment is species dependent (culture behavior and characteristics). This is also evident from the fact that hairy root lines of different species have been reported to require different bioreactor cultivation conditions for successful scale-up, i.e., some roots grow better if completely submerged in the growth medium (liquid-phase reactors) while other demonstrate profuse growth if medium droplets are sprinkled on it (gas-phase reactors) [11, 12]. Thus, exploitation of hairy root culture as a source of bioactive chemicals largely depends on the selection/development of a suitable bioreactor system, where several physical and chemical parameters (nutrient availability, nutrient uptake, oxygen transfer, mixing, and shear sensitivity) specific to the culture characteristics and requirements must be taken into account. In view of the above factors, present work was, thus, undertaken to establish a suitable bioreactor configuration for mass propagation of hairy root cultures of A. indica for healthy growth and enhanced azadirachtin production.

Materials and Methods

Hairy Root Culture of A. indica

A fast growing high-yielding hairy root line (Az-35) of A. indica was used to carry out the present study. The hairy root culture was developed as per the protocol reported earlier by the authors [13]. The hairy root inoculum used for the experiments (3 g l−1 dry weight (DW)) was prepared by growing the hairy roots on solid medium (MM2) of the following composition: Murashige and Skoog's (MS) medium [14] major salts + MS medium minor salts + Gamborg's (B5) medium [15] vitamins + 3 % (w/v) sucrose + 0.75 % (w/v) agar for 30 days. The temperature was maintained at 25 °C and the initial pH of the medium was set at 5.8. The hairy roots were subcultured in fresh MM2 medium after every 30 days of cultivation period.

Establishment of Growth and Production Kinetics in Shake Flask

Growth and azadirachtin production kinetics in A. indica hairy root culture in liquid medium was established in 250-ml Erlenmeyer flasks rotating on a gyratory shaker at 60 rpm containing 38 ml of the liquid MM2 medium [13]. Temperature was maintained at 25 °C and the initial pH of the medium was maintained at 5.8. The experiment was conducted under 16/8 h L/D illumination conditions, and the flasks (containing roots) were harvested at an interval of 5 up to 40 days for generating the profiles for biomass production (in grams per liter, on DW basis) and azadirachtin accumulation (content) in hairy roots (in milligrams per gram) along with its overall production (in milligrams per liter) with substrate (residual sucrose; in grams per liter) utilization. The experiment was done in duplicate and average values were reported.

Scale-Up of A. indica Hairy Root Culture in Liquid-Phase Reactors

Batch cultivation of A. indica hairy roots was carried out in different liquid-phase reactor configurations in order to select the most suitable bioreactor system for mass propagation (scale-up) of A. indica hairy roots. As per the observed and deduced culture requirements, the conventional bioreactor designs were modified and a comparison was made on the basis of azadirachtin production (in milligrams per liter) obtained after 25 days of batch cultivation period (as deduced from the shake flask kinetics).

A working volume of 1 l liquid MM2 medium was used in the reactor studies. Temperature and pH of the medium were controlled by ADI 1030 bio-controller (Applikon Dependable Instruments, The Netherlands). Temperature was maintained at 25 °C by circulating cold water through the glass jacket around the reactor vessel. The pH of the medium was controlled at 5.8 by the automatic addition of 0.1 M aq. NaOH/0.1 M aq. HCl. Evaporation of the medium was minimized by directly passing cold water (≤10 °C) from the chilled water circulation unit through the exit gas condenser. The inlet sterile aeration rate was kept at 0.2 volume of air per volume of reactor per minute which was manipulated, if required, to maintain the dissolved oxygen above 50 % saturation throughout the cultivation period by manually adjusting the aeration rate. Experiments were conducted under 16/8 h L/D illumination conditions and roots were harvested after 25 days for the estimation of root biomass (in grams per liter, DW) and azadirachtin accumulation (in milligrams per gram) in the hairy roots.

Batch Cultivation of A. indica Hairy Roots in Conventional Bioreactors

Batch cultivation of A. indica hairy roots was carried out in a conventional 3 l stirred tank reactor (STR) (Applikon Dependable Instruments, The Netherlands) and a 3 l bubble column reactor (BCR) (Applikon Dependable Instruments, The Netherlands) configurations successfully used by the authors for A. indica plant cell cultivation in the past [5, 16] to investigate the possibility for scale-up of the A. indica hairy root cultivation.

Batch Cultivation of A. indica Hairy Roots in Modified Bubble Column Reactor (with Polypropylene Mesh and Polyurethane Foam as Root Support)

The hairy roots are self immobilizing (tendency to club together and form a matrix) in nature [17], which also is an essential feature during growth. Hence, to facilitate growth in bioreactor by allowing self-immobilization of the hairy roots, a smaller and uniform height-to-diameter ratio of the vessel was proposed to be used in the study as bubble column reactor which could provide slow and less bubble dispersion with low shear environment. A 3-l stirred tank reactor vessel (Applikon Dependable Instruments, The Netherlands) was, hence, operated like a bubble column reactor with the impeller and baffles removed and a sintered air sparger fitted at the bottom (Fig. 1). A glass fiber mesh fitted inside a circular polypropylene basket type casing (of 9 cm outer diameter and 8 mm thickness) (Tarsons Product Pvt. Ltd., Kolkata, India) was also incorporated in the reactor vessel as an attachment for root support. Due to the light weight, the polypropylene mesh support remained floated on the surface of the medium which prevented the roots from being in complete submerged state.

Fig. 1
figure 1

Modified bubble column reactor with polypropylene mesh basket as root support

Full size image

In a separate modification made inside a bubble column reactor, a polyurethane foam (PUF) disc (13-cm diameter and 10-mm thickness) was used as a root support (Fig. 2). The root support not only prevented the hairy roots from being in completely submerged state but was also more porous and adsorbent material as compared to polypropylene mesh, to facilitate better oxygen and mass transfer of nutrients from the medium to the roots (inoculated on the support). The PUF disc could also ensure unhindered growth of the hairy roots unlike in basket type enclosure around the actively growing mass of the hairy roots used earlier, which could limit the growth of the culture. Moreover, PUF was more economical and easily available root support material.

Fig. 2
figure 2

Modified bubble column reactor with polyurethane foam (PUF) disc as root support

Full size image

Biopesticidal Efficacy in the Hairy Roots of A. indica

Azadirachtin has been the main component in A. indica which has been responsible for the antifeedancy and toxic effects in insects. The desert locust (Schistocerca gregaria (Forskal)) is known to be highly sensitive to azadirachtin as an antifeedant. The antifeedant effects are correlated with the sensory response of chemoreceptors on the insect mouthparts [18]. Azadirachtin as a feeding deterrent is known to stimulate the specific “deterrent” cells in the chemoreceptors and block the firing of “sugar” receptor cells, which stimulates feeding [19–22]. Hence, the biopesticidal efficacy in the hairy roots of A. indica was evaluated on the basis of its antifeedant effect on the desert locust S. gregaria. Samples tested were as follows: (a) root hair powder (b) terpenoid extract from the hairy root (c) standard azadirachtin (35 % pure, Indian Agricultural Research Institute, New Delhi, India). The insects were derived from a permanent laboratory colony of the desert locust, S. gregaria (Forskal) (Family, Acrididae; Order, Orthoptera). Forth day of final instar (V) nymphs were prestarved overnight (8–10 h) before the bioassay. No choice feeding test on graded concentrations of the test samples dissolved in pure ethanol was carried out. The formulated sample was applied to 1.0 g of cabbage leaf discs offered in a cage (8″ × 8″ × 8″) under natural light at 25 °C for 4 h (during 11 a.m. till 3 p.m.) to the prestarved insect (Fig. 3). Each of the feeding tests was undertaken with ten replications. Antifeedant activity was determined as the antifeedant index (AI) calculated as per the protocol suggested by Mordue (Luntz) A.J. et al. [23].

$$ \begin{array}{l}\mathrm{A}.\mathrm{I}=\left[\mathrm{C}-\mathrm{T}/\mathrm{C}\right]\times 100\hfill \\ {}\mathrm{C}:\mathrm{Amount}\;\mathrm{of}\;\mathrm{food}\;\mathrm{consumed}\;\mathrm{in}\;\mathrm{the}\;\mathrm{control}\;\mathrm{sample}\hfill \\ {}\mathrm{T}:\mathrm{Amount}\;\mathrm{of}\;\mathrm{food}\;\mathrm{consumed}\;\mathrm{in}\;\mathrm{the}\;\mathrm{test}\;\mathrm{sample}\;\hfill \end{array} $$
Fig. 3
figure 3

A prestarved insect (Schistocerca gregaria) feeding on cabbage leaf discs

Full size image

Analytical Methods

The harvested roots were blotted on a filter paper to remove excess water and weighed for fresh weight estimation and then dried at 35 °C until a constant weight was obtained, which was accounted as the DW of the roots. Extraction and HPLC analysis of azadirachtin present in the hairy root biomass was determined as per the protocol defined elsewhere [24]. Residual sucrose in the medium was estimated by the phenol–sulfuric acid method [25]. The amount of phenolics released in the medium in response to the stress on hairy roots was measured by the protocol suggested by Yuan et al. [26].

Results and Discussion

Growth and Azadirachtin Production Kinetics of Az-35 Hairy Root Line in Shake Flask

As has been reported by the authors [13], the growth and product profile in shake flask hairy root cultivation exhibited a specific growth rate of 0.07 day−1, where the maximum root biomass obtained was 13.3 g l−1 DW and the azadirachtin concentration was 44 mg l−1 in 25 days of cultivation period with depletion of residual sucrose in the medium to 0.42 g l−1 from 30 g l−1. Hence, the hairy root cultivation period for bioreactor studies was selected as 25 days in order to achieve maximum azadirachtin production by in vitro cultivation of A. indica hairy roots.

Scale-Up of A. indica Hairy Root Cultivation in Liquid-Phase Reactors

The hairy roots of A. indica failed to grow in the conventional stirred tank reactor configuration. Even after 25 days of batch cultivation period, no hairy root growth was observed (evident from the fact that no increase in the biomass was observed from the initial value of 3.0 g l−1). High amount of residual sugar (25.6 g l−1) was found to be present in the medium even after 25 days of the cultivation period on account of no hairy root growth. The high amount of shear stress on the roots (by the moving parts of the reactor) was evident from the large release of phenolics (150 mg l−1) in the medium (as suggested by Yuan et al. [26]), which can be related to the complete inhibition of growth and low secondary metabolite accumulation in hairy roots observed in the present study. Similarly, no significant increase (from 3.0 to 3.2 g l−1 DW after 25 days) in the hairy root biomass was observed in the bubble column reactor also. Significant amount of phenolics (133 mg l−1) were released in the medium with high amount of residual sugar (24 g l−1) remaining in the medium even after 25 days.

The results suggested that the conventional reactor configurations (STR and BCR) were not suitable for the scale-up of A. indica hairy root cultivation.

Batch Cultivation of A. indica Hairy Roots in Modified Bubble Column Reactor (with Polypropylene Mesh Support)

A significant reduction in the mechanical and hydrodynamic stress on the roots was observed. This was evident from the reduced amount of phenolics released (694.3 μg l−1) in the medium as compared to the amount released in the medium (150 and 133 mg l−1) when the hairy roots were grown under complete submerged state in the conventional stirred tank reactor and bubble column reactor designs, respectively, in which they failed to grow. An improved biomass of 9.5 g l−1 DW with an azadirachtin accumulation (content) of 2.1 mg g−1 was obtained in the present reactor configuration after 25 days of the hairy root cultivation period. As a result, an overall azadirachtin production of 20.23 mg l−1 (equivalent to a volumetric productivity of 0.81 mg l−1 day−1) was achieved. Although, hairy roots were able to grow in the present bioreactor set up, but the final biomass obtained was still 0.7 times less than that obtained in the shake flask studies (with 13.3 g l−1 of biomass production in 25 days of cultivation period). A residual sucrose concentration of 18.3 g l−1 still remained in the medium after 25 days. This decrease in biomass and incomplete utilization of medium nutrients was presumably due to the small size of the mesh support which might have led to volume and mass transfer limitations as the size of the tissue matrix increased with growth.

Batch Cultivation of A. indica Hairy Roots in Modified Bubble Column Reactor (with PUF Support)

The incorporation of a PUF disc as a support for the hairy roots inoculated inside the bubble column reactor facilitated a biomass production of 9.2 g l−1 DW and azadirachtin accumulation in hairy roots of 3.1 mg g−1 in 25 days of cultivation period. This resulted in an azadirachtin production of 28.52 mg l−1 (equivalent to a volumetric productivity of 1.14 mg l−1day−1) in the present bioreactor configuration. Higher azadirachtin accumulation (∼1.5 times) in the hairy roots could be achieved as opposed to 2.1 mg g−1 in the modified bubble column reactor (with polypropylene mesh support). Increase in the secondary metabolite production by fungi (heterogeneously growing culture) was observed when polyurethane foam was used for immobilization due to the increase in the culture exposure to oxygen-rich atmosphere [27]. It has also been reported in literature that inclusion of PUF in air-sparged reactors reduces the entrapment of gases by the hairy roots which might result in enhanced biomass and secondary metabolite production [28]. Polyurethane foam has been used in air-lift reactors to support the root tissues during the hairy root cultivation of Armoracia rusticana [29] and Duboisia leichhardtii [30]. Kondo et al. [31] have reported successful immobilization of hairy roots by using a reticulate polyurethane foam sheet fixed on the inner wall of a rotating drum bioreactor. The inherent advantages of using PUF like easier availability, scalability, and cost effectiveness favored its use for the scale-up of A. indica hairy root cultivation. Moreover, the modified liquid-phase bioreactor design could prove to be more favorable in comparison to the gas-phase reactor design reported earlier by the authors [32] in terms of low maintenance and operational cost during scale-up of A. indica hairy root cultivation.

Even though the overall volumetric productivity of azadirachtin obtained in the present bioreactor set up (1.14 mg l−1day−1) was better than that obtained in the previous reactor design of modified bubble column reactor (with polypropylene mesh), but still it could not match up to the productivity obtained in the shake flask studies (1.76 mg l−1day−1). The roots present on the top of the root mass were found to be dry and dead. This could be related to the increase in the size of the root mass with growth due to which the roots present on the matrix surface lost contact with the liquid medium and eventually died of nutrient starvation. Hence, this non-reflection of growth (with 13.3 g l−1 of biomass production) and overall azadirachtin production (44 mg l−1) of the shake flask study at the bioreactor level (with only 9.2 g l−1 of biomass and 28.52 mg l−1 of azadirachtin production) was presumably due to the uneven distribution of nutrients and oxygen in high-density culture of hairy roots.

Biopesticidal Efficacy in the Hairy Roots of A. indica

In order to investigate the antifeedant activity (AI) present in the hairy roots of A. indica, a bioassay was carried out on S. gregaria as described under Materials and methods. Dose-response curve was obtained by plotting the percentage antifeedancy achieved versus concentration of the hairy root sample taken on a semi-log plot. The ED50 value (the concentration giving 50 % antifeedancy) was calculated from the dose-response curve for comparison. As can be observed in Fig. 4a, increasing concentration of the hairy root powder in ethanol resulted in increased antifeedant activity. Highest antifeedant activity (AI 97 %) was observed in the concentration range tested when the applied concentration on the leaf disc (feed for prestarved S. gregaria) reached 2 % (w/v). The calculated ED50 value from the dose-response curve was found to be 0.45 % (w/v) for the hairy root powder. Similarly, as demonstrated in Fig. 4b, the crude form of azadirachtin obtained by the solvent extraction of the hairy root powder also demonstrated high antifeedant activity (AI 83.5 %) even at concentrations as low as 0.05 % w/v. A lower value of ED50 (0.004 % w/v) was obtained from the respective dose-response curve as opposed to the ED50 of 0.45 % (w/v) for the ethanolic solution of the hairy root powder as a whole. The results obtained in the present study, thus, established the possibility of using the hairy root powder as a biopesticide. Dose-response curve (between the percent antifeedancy (AI) and concentration (in parts per million) on semi-log scale) was also plotted for graded concentrations of pure azadirachtin (in parts per million) (Fig. 5) for comparison of the results obtained in the present study with the available standard azadirachtin. The ED50 value for the pure azadirachtin sample tested was found to be as low as 0.0087 ppm. As can be observed in Fig. 6, a linear relationship (with regression coefficient (R2) ≥0.9) was obtained between the percent antifeedancy (in terms of AI) and logarithm (loge = ln) concentration (in grams per liter) of the samples (hairy root powder, hairy root crude solvent extract, and pure azadirachtin) except for deviations at very low and high concentration levels studied for pure azadirachtin sample. The regression lines for both the hairy root samples (powder and the crude solvent extract) exhibited gradients (slope ∼9.1) similar to those of pure azadirachtin (slope ∼9.2). This demonstrated similar antifeedancy response of the insect S. gregaria towards the hairy root samples (powder and extract) and the pure azadirachtin probably due to the presence of similar bioactive compound, as suggested by Allan et al. [33]. Smaller would be the value of ED50 if higher is the yield (w/w) of the bioactive compound (azadirachtin) present in the sample. This could explain the reason for the decreasing trend of the ED50 values observed among the three samples (ED50 for the hairy root powder (0.45 % w/v), >ED50 for the hairy root solvent extract (0.004 % w/v), and >ED50 for pure azadirachtin (0.0087 ppm), respectively.

Fig. 4
figure 4

a Antifeedant activity () in the hairy root powder observed on S. gregaria. b Antifeedant activity () in the crude azadirachtin extract of the hairy root powder

Full size image
Fig. 5
figure 5

Antifeedant activity () present in pure azadirachtin sample

Full size image
Fig. 6
figure 6

A plot between the antifeedant index versus logarithmic concentration of the samples (hairy root (HR) powder (), hairy root (HR) solvent extract (), and pure azadirachtin ())

Full size image

Conclusions

Incorporation of a root support facilitated scale-up of A. indica hairy roots up to 3 l bioreactor level. Among the two bubble column reactor configurations studied, the highest volumetric productivity of azadirachtin (1.14 mg l−1day−1) was achieved (equivalent to a biomass production of 9.2 g l−1 and azadirachtin accumulation of 3.1 mg g−1 in the hairy roots in 25 days) in modified bubble column reactor with PUF as root support. Moreover, its simpler and economical design (for easy operation and maintenance during scale-up) could prove to be more favorable. Hence, modified bubble column reactor (with PUF disc as root support) was recommended for the scale-up of the hairy root cultivation of A. indica. Satdive et al. [34] have reported a maximum azadirachtin yield up to 0.14 % DW from A. indica hairy root cultivation in shake flasks. In the present investigation, azadirachtin yield in the hairy roots has been found to be 0.31 % DW (∼twofold higher), when hairy root cultivation was scaled-up to 3 l bioreactor level. However, the volumetric productivity of azadirachtin obtained during the scale-up of the hairy root culture in bioreactor was found to be 35 % less than that obtained in the shake flask studies. This was presumably due to larger mass transfer limitations encountered during scale-up (in bioreactors) as evident from the increased growth rate of the culture in the presence of optimum inlet gas flow rate and pure oxygen [35]. Thus, the biomass and azadirachtin production in the selected bioreactor configuration can be improved further by optimizing the various operating conditions like gas-phase composition, air-flow rate, etc., at the bioreactor level to further curb the mass transfer limitations during scale-up. Nevertheless, the selection of the modified bubble column reactor with PUF as root support facilitated the scale-up of the A. indica hairy root culture in liquid to bioreactor level which has not been reported earlier. The presence of antifeedant activity in the hairy roots of A. indica and its solvent extract was demonstrated by conducting a bioassay on the insect S. gregaria.

High amount of antifeedant activity (maximum up to AI 97 %) observed in the A. indica hairy root powder, and its crude azadirachtin extract (maximum up to AI 83.5 %) demonstrated and supported the prospective application of in vitro hairy root cultivation for the biopesticide production.