Sparge gas composition affects biomass and ajmalicine production from immobilized cell cultures of Catharanthus roseus

Enzyme and Microbial Technology 37 (2005) 424–434

Sparge gas composition affects biomass and ajmalicine production from immobilized cell cultures of Catharanthus roseus Carolyn W.T. Lee-Parsons a,∗ , Michael L. Shuler b a

342 Snell Engineering Center, Chemical Engineering Department, 360 Huntington Avenue, Northeastern University, Boston, MA 02115-5000, USA b 120 Olin Hall, School of Chemical and Biomolecular Engineering, Cornell University, Ithaca, NY 14853-5201, USA Received 16 April 2004; received in revised form 23 February 2005; accepted 25 February 2005

Abstract Despite their low solubility in aqueous medium, dissolved gases play important roles in the cultivation and successful scale-up of plant cell cultures. In this paper, the effects of O2 and CO2 on growth and secondary metabolism were investigated using the production of ajmalicine from Catharanthus roseus cultures. The effects of gas composition were investigated using shear-protected alginate-immobilized cells (diameter < 2 mm) cultured in bubble columns sparged with premixed gases, including nine combinations of O2 and CO2 . A wider range of concentrations (10–95% O2 , 0.03–10% CO2 by mole) was studied to explore potential benefits or drawbacks. Sparge gas composition significantly altered growth and ajmalicine production. Low and high O2 concentrations (10, 90, 95% O2 ) were either inhibitory or toxic to growth and ajmalicine production. The effects of CO2 depended on O2 concentration. At lower O2 concentrations (21% O2 ), increasing the CO2 concentration decreased both growth and specific ajmalicine production. At higher O2 concentrations (78.4% O2 ), increasing the CO2 concentration decreased growth while specific ajmalicine production was not affected. In these studies, extracellular ajmalicine concentration was maximized with a gas mixture of 50% O2 + 0.03% CO2 . © 2005 Elsevier Inc. All rights reserved. Keywords: Catharanthus roseus cell cultures; Gas composition; Oxygen; Carbon dioxide; Secondary metabolism; Ajmalicine; Alginate immobilization; Amberlite® XAD-7 resin; Bubble column reactors

1. Introduction Dissolved gases such as O2 , CO2 , and ethylene have generally not been optimized for growth and secondary metabolism in plant cell cultures. However, literature indicates that dissolved gases do impact growth and secondary metabolism and ultimately impact the success of scale-up [1–16]. Although gases have low solubility in aqueous medium, their role or effects are not negligible with respect to growth and secondary metabolism. Abbreviations: 2,4-D, 2,4-dichlorophenoxy-acetic acid; FW, fresh weight; HPLC, high performance liquid chromatography; IAA, indole-3acetic acid; kL a, mass transfer coefficient; LB, Luria-Bertani media; MS, Murashige-Skoog; n-HS, n-heptanesulfonic acid; SD, standard deviation; TLC, thin layer chromatography; W/v, weight per volume ratio; V/v, volume to volume ratio; Vvm, volume of gas per volume of culture per min ∗ Corresponding author. Tel.: +1 617 373 2989; fax: +1 617 373 2209. E-mail address: [email protected] (C.W.T. Lee-Parsons). 0141-0229/$ – see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.enzmictec.2005.02.016

For instance, molecular oxygen plays a key role in the cell’s bioenergetics as the final electron acceptor in respiration. In addition, molecular oxygen may also be involved in cytochrome P450-mediated oxygenation reactions in the biosynthesis of secondary metabolites such as terpenoid indole alkaloids [17] and taxanes [18]. Since different enzymes are involved, growth and secondary metabolism may be optimized at different O2 concentrations [10,19–21]. Carbon dioxide and ethylene are other gases produced in plant cell cultures that can act as essential nutrients or hormones, altering growth-related and secondary metabolic activities. For instance, CO2 stripping is associated with high aeration rates in bioreactors and has been implicated in lowering growth rates, lengthening lag periods, and causing browning [5–7,12,13,22]. Although the mechanism is not fully understood, maintaining a minimum CO2 concentration in the cultures (i.e. 20 mbar in Catharanthus roseus cultures) is essential for promoting growth and culture viability [23–25].

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

The endogenous plant hormone ethylene is involved in many physiological responses in the whole plant [26]. In plant cell cultures, ethylene either stimulated [27–30], inhibited [31], or had no effect [32] on the production of various secondary metabolites. Carbon dioxide and ethylene can also interact and modulate each other’s effects [3,26,29,33]. This paper investigates the effects of a wide range of O2 and CO2 concentrations on growth and secondary metabolism in C. roseus cultures, using the production of ajmalicine as a model system. Potential benefits or toxicities associated with operating outside the range achieved with air or air supplemented with CO2 (typical sparge gas compositions) could then be explored. This gas effect study was designed to (1) investigate a wider range of O2 and CO2 concentrations in the gas phase (i.e. 10–95% O2 and 0.03–10% CO2 ) than that previously reported in the literature, (2) vary O2 and CO2 independently, and (3) alter gas composition while minimizing effects due to shear. The effects of gas composition were investigated using shear-protected alginate-immobilized cells cultured in bubble columns sparged with premixed gases. A Box-Wilson statistical design [34] was used to prescribe the minimum number and the specific combinations of O2 and CO2 in the study. Ethylene was not included in this study since previous experiments indicated that ethylene inhibited ajmalicine production above 1 ppm [31]. The effect of oxygen transport on dissolved oxygen profile within immobilized cells (spherical beads of less than 2 mm in diameter) was also analyzed. This present study illustrates the impact of a wide range of dissolved gas compositions on growth and secondary metabolism and suggests that gas composition, like media composition, should be optimized for growth and secondary metabolism.

425

ics, 80 g/l sucrose, and 10 ␮M IAA [36]. The production medium was autoclaved at 121 ◦ C for 15 min. The cultures were incubated in the dark at 25 ◦ C on a gyratory shaker at 120 rpm. 2.3. Cell immobilization procedure After being primed in production medium for 10 days, cells were immobilized in alginate under sterile conditions, as previously described in Lee and Shuler [35]. Primed cell suspensions were sieved for cell aggregates of less than 500 ␮m. The large cell aggregates were removed to prevent clogging the pipette tip used in the extrusion of alginate beads. The sieved fraction was then suction-filtered and mixed with 2% (w/v) sodium alginate solution (medium viscosity; Sigma Chemical Co., St. Louis, MO) to achieve an initial cell concentration of 25% by cell weight. The cellalginate mixture was extruded through a 1000 ␮l pipette tip with concentric forced air flow. The cell-alginate droplet size was controlled by altering the air flow rate. Alginateimmobilized cells in the form of spherical beads were dropped and hardened in MS production medium plus 50 mM CaCl2 ·H2 O for 30–40 min. After hardening, the final cell concentration in the beads was approximately 33% by cell weight. The hardened beads (less than 2 mm in diameter) were rinsed in production medium to remove the excess calcium and suction-filtered; the cell-alginate beads, containing 5 g FW cells, were cultured in 100 ml of MS production medium supplemented with 5 mM CaCl2 ·H2 O in a 500 ml Erlenmeyer flask. The MS production medium promoted fresh weight accumulation in immobilized cells. The growth and ajmalicine production profiles of immobilized cells in shake flasks were shown in Lee and Shuler [35].

2. Materials and methods 2.4. Bubble column reactors 2.1. Maintenance of C. roseus cell suspension cultures The C. roseus cell suspension cultures were maintained as previously described in Lee and Shuler [35]. Every 9 days, 3.0 g FW of suction-filtered cells was transferred to 100 ml of fresh growth medium in a 500 ml Erlenmeyer flask. The growth medium [36] consisted of MS minimal organics [37] (GIBCO Life Technologies Inc., Grand Island, NY), 30 g/l sucrose, and 5 ␮M 2,4-D. The growth medium was autoclaved at 121 ◦ C for 15 min. The cultures were incubated in the dark at 25 ◦ C on a gyratory shaker at 120 rpm. 2.2. Priming C. roseus cell suspension cultures for alkaloid production After 9 days in growth medium, the cultures were suctionfiltered and 5.0 g FW of cells was transferred to 50 ml of MS production medium in 250 ml Erlenmeyer flasks for 10 days, as previously described in Lee and Shuler [35]. MS production medium consisted of MS minimal organ-

The effects of gas composition on growth and ajmalicine production were investigated using immobilized cells cultured in bubble column reactors (Fig. 1). The bubble column reactors were fritted cylindrical filter funnels (Kontes, Vineland, NJ), which were modified to include glass ports for dissolved oxygen probes (New Brunswick Scientific, Edison, NJ). The bubble column dimensions were 45 mm × 310 mm (diameter versus height). Humidified gas was forced through the fritted glass disk to disperse the gas as bubbles and to provide mixing. The mass transfer coefficient (kL a) of each bubble column was determined using the unsteady state method, i.e. after displacing dissolved O2 with N2 , air was sparged into the bubble column and the dissolved O2 concentration was monitored with time. The mass transfer coefficient of the bubble columns was greater than 30 h at 1.5 vvm of air (i.e. 150 ml/min). The variation in growth and ajmalicine production between three replicate bubble columns was determined and is presented in Fig. 2 and in Section 3.1.

426

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434 Table 1 Sparge gas compositions prescribed by the Box-Wilson statistical designa Axial combinations

Factorial combinations

Center point combination

% O2

% CO2

% O2

% CO2

% O2

% CO2

10 50 50 90b

5 0.03 10 5

21.6 21.6 78.4 78.4

1.5 8.6 1.5 8.6

50

5

a Gas compositions are expressed in terms of % by mole with N as the 2 inert gas. b The gas composition, 95% O + 5% CO , replaced the statistical design 2 2 composition of 90% O2 + 5% CO2 in the first experiment due to availability. Later experiments with 90% O2 + 5% CO2 yielded similar results as 95% O2 + 5% CO2 .

Fig. 1. The bubble column reactor used for studying the effect of gas composition on growth and ajmalicine production in immobilized C. roseus cells.

2.5. Investigation of gas effects Cell-alginate beads, containing 5 g FW cells, were initially cultured in 500 ml Erlenmeyer flasks containing 100 ml of MS production medium supplemented with 5 mM CaCl2 ·H2 O. After 5–6 days (i.e. after lag phase), the immobilized cells were transferred to bubble column reactors in a laminar flow hood (i.e. the contents of one 500 ml flask were transferred to one bubble column). Bubble columns were then sparged with humidified, premixed gases (Matheson Gas Products, Twinsburg, OH) at 150 ml/min. Filtersterilized gases were humidified by bubbling the gas through a fritted filter tube immersed in sterilized, distilled water (300 ml of water in a 500 ml filter flask). Amberlite® XAD-7 resin (Sigma) was added at the time of bubble column inoculation and then exchanged aseptically every 3–4 days. XAD-7 resin enhanced ajmalicine production, potentially by reducing feedback inhibition, and promoted high extracellular recovery of ajmalicine [38]. Resin was added on day 5 or day 6 since ajmalicine production was maximized and ajmalicine recovery was nearly complete with resin addition on day 5 [38]. XAD-7 resin (1 g) was enclosed in Miracloth (Calbiochem, La Jolla, CA), forming a resin bag. Two resin bags were tied to nylon lines and suspended from a short copper wire (not immersed in the media) for easy sampling (see Fig. 1). In addition, 1–2 ml of 20% (v/v) antifoam C (Sigma) was added to control foaming. A Box-Wilson statistical design was used to prescribe the minimum number and the compositions of O2 and CO2 combinations in the study [34]. The concentration range of inter-

est was 10–90% O2 and 0.03–10% CO2 . The upper CO2 limit (i.e. 10%) was chosen based on previous experiments in shake flask cultures with limited gas exchange [31]. The Box-Wilson statistical design consisted of nine combinations of O2 and CO2 with two replicates at the center point to measure scatter (50% O2 and 5% CO2 as prescribed by the Box-Wilson statistical design; see Table 1). Inherent to the Box-Wilson design, replicates are not required at every gas composition. To study all the gas compositions prescribed by the BoxWilson design in one experiment, 10 bubble column reactors with dissolved oxygen probes would be required. Due to limitations in equipment and the scale of immobilization required, the gas composition study was divided into two separate blocks. The axial combinations of O2 and CO2 were run in parallel in one block and the factorial combinations along with the center point replicates were run in another block. Replicates in shake flasks or air-sparged bubble column reactors were also included in each block for determining trends with respect to these replicates since ajmalicine production generally varied from one subculture to another and hence between the two experimental blocks. In the first block involving axial combinations, 95% O2 + 5% CO2 was used in place of 90% O2 + 5% CO2 due to availability of gas mixture. However, cultures sparged with 90% O2 + 5% CO2 were studied in later experiments with similar results to that observed at 95% O2 + 5% CO2 . Certain runs terminated early due to gas exhaustion or culture browning. Therefore, two additional experiments were conducted, including repeats at gas compositions where browning or early termination had occurred. Although the actual ajmalicine concentrations varied with subculture, the trends on growth and ajmalicine production observed in the first two blocks were similar to those of the two additional experiments performed later. 2.6. Fresh and dry weight determination Fresh weight was determined using the entire contents of the bubble column reactor on its final day. The actual fresh weight was determined by subtracting the weight of the al-

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

ginate in the cell-alginate beads on day 0 from the weight of the rinsed and suction-filtered beads on the sample day. The dry weight of the immobilized cells was determined after freeze-drying the frozen cell-alginate beads for 1–2 days at <500 mTorr and −40 ◦ C using a Flexi-Dry Microprocessor Manifold Lyophilizer (FTS Systems Inc., Stone Ridge, NY). 2.7. Extraction of alkaloids from resin or medium After freeze-drying, each Amberlite® XAD-7 resin bag was extracted twice with 25 ml of HPLC grade methanol for 3 h while being agitated on a 120 rpm gyratory shaker [36]. Methanol extracts were pooled from the two extraction steps and evaporated to dryness with a Buchi Rotary Vaporizer (Buchi, Switzerland). Indole alkaloids were extracted from culture medium with methylene chloride using two-phase solvent–solvent extraction [36]. Negligible levels of ajmalicine were found in the medium in which resin had been added. 2.8. Quantification of alkaloids in crude extracts by TLC Experimental data shown in Figs. 3 and 4 were obtained from samples prepared using the following pretreatment method and then quantified by TLC. Initially, dried crude extracts from resin samples were redissolved in 20 ml of 10% methanol and passed through Sep-pak C18 liquid chromatography cartridges (Waters Associates, Milford, MA). Alkaloids of interest adsorbed onto the Sep-pak C18 packing and were then eluted with 100% methanol, subsequently evaporated, and redissolved in known volumes of methanol. Indole alkaloids in resin samples were separated by TLC using a solvent system consisting of chloroform: acetone: diethylamine (20:4:1 by volume) [36]. Spotted resin samples were flanked by two consecutive ajmalicine concentrations. The sample concentration was estimated by comparing the intensity and size of the spot with the surrounding standards. The error bar was calculated by dividing the difference between the flanking ajmalicine standard concentrations by two. At the higher ajmalicine concentrations (>20 mg/l), the ajmalicine concentrations determined by TLC were 30–40% higher than the concentrations determined by HPLC. However, the trends were similar between the results obtained by TLC and by HPLC. 2.9. Quantification of alkaloids in crude extracts by HPLC Experimental data shown in Figs. 5 and 6 were obtained using the following pretreatment method (adapted from [39]) and quantified using HPLC. Dried crude extracts from resin samples were redissolved in 20 ml of an aqueous solution containing 10% methanol plus 50 mM n-heptanesulfonic acid (n-HS, Sigma). The solubilized extract was then passed through a Sep-pak C18 liquid chromatography cartridges (Waters Associates). Alkaloids of interest adsorbed onto the

427

Sep-pak C18 packing and were subsequently eluted with 100% methanol plus 50 mM n-HS, evaporated, and redissolved in known volumes of 10% methanol (aqueous) plus 50 mM n-HS. The final sample was filtered through a Gelman 0.2 ␮m filter (Fisher Scientific) before HPLC analysis. The Sep-pak C18 step was intended to remove compounds which would strongly adhere to the HPLC’s C18 packing and which would interfere with separation of chemical species or reproducibility of retention times. Alkaloids were separated using a Waters ␮Bondapak C18 stainless steel (3.9 cm × 30 cm) column with Waters C18 precolumn. Elution was performed at 1 ml/min with a SpectraPhysics Pump (SP8800) for 35 min with an isocratic method adapted from Morris et al. using 65% methanol: 35% water + 50 mM n-HS [39]. Ajmalicine eluted at 23.7 min. Alkaloids were detected by UV absorbance at 254 nm with a Spectra-Physics UV–vis Detector (SP8490). The maximum standard deviation in ajmalicine concentration between replicate flasks analyzed using this protocol was 10%. The column was washed with 100% methanol + 50 mM n-HS every 10–15 samples for 1–2 h and then re-equilibrated with 65% methanol + 50 mM n-HS for 1–2 h. After 150 injections, column characteristics had changed and ajmalicine and serpentine could not be separated. Hence, a similar sample preparation method and HPLC protocol using a different solvent system (acetonitrile, water, and trifluoroacetic acid) was the preferred method [35] and was used for the analysis of Fig. 2. The maximum standard deviation in ajmalicine concentrations between replicate flasks was 14% using the sample preparation and HPLC protocol described in Lee and Shuler [35]. 2.10. Calculating the dissolved oxygen profile in immobilized cells To calculate the dissolved oxygen profile [S] within the cell-alginate beads, saturation kinetics for oxygen uptake rate was assumed and the differential mass balance of oxygen within the bead at steady state (Eq. (1)) was solved numerically using the fourth-order Runge-Kutta method (Mathcad 2001 Professional, MathSoft, Inc.):
 d2 [S] 2 d[S] (q max ρbead )[S] De + = (1) 2 dr r dr Ks + [S] The maximum specific oxygen uptake rate (qmax) was measured using suspension cultures during exponential growth phase (4.69 × 10−4 mmol O2 /g FW min). The saturation constant (Ks ) was estimated at 10% of saturation with air (2.53 × 10−5 mmol O2 /cm3 ) [19]. The effective diffusivity of oxygen in alginate (De ) was estimated as 9.72 × 10−4 cm2 /min [19]. The dissolved oxygen concentration profile was calculated at different cell densities within the bead (ρbead = g cell FW/cm3 of bead volume) from the initial cell inoculum density (ρbead = 0.33 g FW/cm3 ) to the maximum cell density reached (which depended on the sam-

428

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

pling day and the sparge gas composition). The oxygen profile was calculated from the center of the bead to the bead diameter of 0.1 cm and the dissolved oxygen concentration on the surface of bead was estimated at 90% of saturation with the sparge gas (i.e. for air, [Ssurface ] = 2.28 × 10−4 mmol O2 /cm3 ).

3. Results 3.1. Growth and ajmalicine production in replicate bubble column reactors Growth and ajmalicine production between replicate bubble columns were monitored to determine the variability between bubble columns. Immobilized cells were cultured in shake flasks initially and then transferred on day 6 to bubble column reactors sparged with air. Bags containing Amberlite® XAD-7 resin were added on day 6 and exchanged every 3–4 days thereafter to adsorb the ajmalicine produced by the immobilized cells (see Section 2.5). Biomass and ajmalicine production were similar between the three bubble columns. Biomass increased six-fold in this production medium, yielding an average final fresh weight concentration on day 24 of 300 g/l and a standard deviation of 8 g/l (i.e. 3%). The extracellular ajmalicine profiles (i.e. ajmalicine adsorbed onto resin) of these cultures were also similar (Fig. 2). After 24 days, the average extracellular ajmalicine accumulation was 42 mg/l and the standard deviation was 6 mg/l (i.e. 14%). The standard deviations in growth and ajmalicine production in this experiment represented the typical variation between replicate bubble columns (using the analytical method described in Section 2.9).

Fig. 2. Ajmalicine production in immobilized C. roseus cells from replicate bubble column reactors sparged with air. Immobilized cells (inoculum density of 50 g FW/l) were cultured in shake flasks initially and transferred to bubble column reactors on day 6. Two resin bags, each containing 1 g of Amberlite® XAD-7, were added on day 6. Resin was sampled and exchanged with fresh resin every 3–4 days. Extracellular ajmalicine concentrations from three replicate bubble column reactors are shown.

3.2. Effects of O2 and CO2 composition on growth and ajmalicine production The effects of O2 and CO2 composition on growth and ajmalicine production were investigated with immobilized cells cultured in bubble column reactors and with a BoxWilson statistical design to prescribe nine sparge gas combinations. Due to limitations in equipment and the scale of immobilization required, the nine sparge gas combinations were divided into two separate experimental blocks: one consisting of the axial gas combinations and the other consisting of the factorial and center point gas combinations (see Table 1). This gas effects study was conducted similarly to the previous experiment, i.e. resin bags were added beginning on day 6 and exchanged every 3–4 days thereafter. 3.2.1. Axial combinations of O2 and CO2 The sparge gas composition significantly altered the growth and ajmalicine production of immobilized C. roseus cultures. Replicates in shake flasks were cultured in parallel as a measure of scatter and as a basis of comparison to the second experimental block consisting of factorial and center point gas compositions. Plasmolysis, low viability, and graying/browning were observed in the culture sparged with 95% O2 + 5% CO2 . The culture sparged with 50% O2 + 10% CO2 was terminated early on day 13 because the sparge gas supply was exhausted. No adverse effects at this composition were observed during the first 13 days. However, when this gas composition was repeated in a later experiment for a period of 24 days, browning was observed in cultures sparged with 50% O2 + 10% CO2 after day 18. The culture sparged with 10% O2 + 5% CO2 contained small starch grains but exhibited high viability with fluorescein diacetate [40]. High viability and the presence of large starch grains were observed in shake flasks and with 50% O2 + 0.03% CO2 . Ajmalicine production was completely inhibited in immobilized cultures sparged with 10% O2 + 5% CO2 and 95% O2 + 5% CO2 throughout the culture period (Fig. 3). In contrast, ajmalicine production was promoted in shake flasks and with 50% O2 + 0.03% CO2 (Fig. 3). Similarly, growth was also promoted in shake flasks and with 50% O2 + 0.03% CO2 but limited with 10% O2 + 5% CO2 and 95% O2 + 5% CO2 (Fig. 4). The specific ajmalicine concentration in shake flask cultures and with 50% O2 + 0.03% CO2 was at least 4–11 times higher than cultures sparged with 10% O2 + 5% CO2 or 95% O2 + 5% CO2 (Fig. 4). As a result, the combined higher fresh weight concentration and higher specific ajmalicine content resulted in higher extracellular ajmalicine concentrations in shake flask cultures and with 50% O2 + 0.03% CO2 . The sparge gas composition, 50% O2 + 0.03% CO2 , resulted in the highest extracellular ajmalicine accumulation, 57 mg/l. However, this ajmalicine concentration did not represent the maximum concentration attainable since production had not leveled off by the last sampling day

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

Fig. 3. Ajmalicine production in immobilized C. roseus cells from bubble column reactors sparged with axial combinations of O2 and CO2 . Immobilized cells (inoculum density of 50 g FW/l) were cultured in shake flasks initially and transferred to bubble column reactors on day 6. Two resin bags, each containing 1 g of Amberlite® XAD-7, were added on day 6 and exchanged every 3–4 days. Several runs were terminated earlier than day 20 due to gas supply exhaustion (i.e. 50% O2 + 10% CO2 and 50% O2 + 0.03% CO2 ). Browning was observed in cultures sparged with 95% O2 + 5% CO2 . Determination of error bars is described in the Section 2.8.

429

and in a bubble column sparged with air were conducted in parallel with the center point and factorial gas compositions. In addition to the replicated center point composition, the shake flask and air-sparged bubble column also served as a measure of scatter since previous experiments showed that immobilized cells cultured in shake flasks yielded similar biomass and ajmalicine production to that of air-sparged bubble columns. On a given day, the extracellular ajmalicine concentration decreased with increases in CO2 concentration at either 21.6% O2 (Fig. 5A) or 78.4% O2 (Fig. 5B). The effect of CO2 concentration on extracellular ajmalicine production appeared to be dependent on the O2 concentration. For example, on day 21, the ajmalicine production from cultures sparged with 21.6% O2 + 8.6% CO2 was 24 mg/l compared

in either shake flasks or with 50% O2 + 0.03% CO2 (Fig. 3). 3.2.2. Factorial and center point combinations of O2 and CO2 The effects of the factorial and center point gas compositions on growth and ajmalicine production were investigated next. As a basis of comparison with the previous block consisting of axial gas combinations, cultures in a shake flask

Fig. 4. Growth and specific ajmalicine production in immobilized C. roseus cells from bubble column reactors sparged with axial combinations of O2 and CO2 . The specific extracellular ajmalicine production (
) and the final fresh weight concentration (bar graph; day of sampling shown in parenthesis on the x-axis) are shown. The inoculum density was 50 g FW/l. Error bars for growth represent a standard deviation of 3%. Determination of error bars for the specific ajmalicine concentration is described in Section 2.8.

Fig. 5. Ajmalicine production in immobilized C. roseus cells from bubble column reactors sparged with factorial and center point combinations of O2 and CO2 . Immobilized cells (inoculum density of 50 g FW/l) were cultured in shake flasks initially and transferred to bubble column reactors on day 6. Two resin bags, each containing 1 g of Amberlite® XAD-7, were added on day 6 and exchanged every 3–4 days. Several runs were terminated earlier than day 33 due to gas supply exhaustion (i.e. 78.4 O2 + 1.5% CO2 ) or due to culture browning (i.e. 21% O2 + 1.5% CO2 and 78.4% O2 + 8.6% CO2 ). Error bars for growth and ajmalicine concentration were determined using the shake flask and the bubble column reactor as replicates. The error bars for growth and ajmalicine concentration represent a standard deviation of 3 and 10%, respectively.

430

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

Fig. 6. Growth and specific ajmalicine production in immobilized C. roseus cells from bubble column reactors sparged with factorial and center point combinations of O2 and CO2 . The specific extracellular ajmalicine production (
) and the final fresh weight concentration (bar graph; day of sampling shown in parenthesis on the x-axis) are shown. The inoculum density was 50 g FW/l. Error bars for growth and ajmalicine concentration were determined using the shake flask and the bubble column reactor as replicates. The error bars for growth and ajmalicine concentration represent a standard deviation of 3 and 10%, respectively.

to 46 mg/l from cultures sparged with 78.4% O2 + 8.6% CO2 . Increasing the O2 concentration appeared to counteract the suppressing effects of CO2 concentration on extracellular ajmalicine production. These trends were also observed in later experiments. Increasing the CO2 concentration at 78.4% O2 appeared to depress ajmalicine concentration by depressing growth. For example, at 78.4% O2 (Fig. 6), the specific ajmalicine production with either 1.5% CO2 or 8.6% CO2 was similar (i.e. 0.39 mg of ajmalicine/g FW). But the fresh weight concentration observed with 1.5% CO2 (on day 29) was approximately double that observed with 8.6% CO2 (on day 27; Fig. 6). As a result, the overall ajmalicine concentration of the culture sparged with 1.5% CO2 was nearly double of that with 8.6% CO2 . Although the comparison was made on different sampling days, limited growth in cultures sparged with 78.4% O2 + 8.6% CO2 would have been expected between day 27 and 29 since the doubling time under these conditions was over 14 days (Fig. 6). By day 27, only residual fructose remained in the medium of cultures sparged with 78.4% O2 + 8.6% CO2 . At 21% O2 , an increase in the CO2 concentration (from 0.03 to 8.6%) depressed both biomass concentration and specific ajmalicine production when compared on day 33 (Fig. 6). For instance, on day 33, the fresh weight concentration of the culture sparged with air (21.6% O2 + 0.03% CO2 ) was 36% higher than that of the culture sparged with 21.6% O2 + 8.6% CO2 . On day 33, the specific ajmalicine production of the culture sparged with air was also double that of the specific ajmalicine production of the culture sparged with 21.6% O2 + 8.6% CO2 (Fig. 6). Compared with 21.6% O2 + 8.6% CO2 , the combination of a higher biomass concentration and a higher specific ajmalicine concentration in the culture sparged with air resulted in a nearly

three-fold increase in the overall ajmalicine concentration (Fig. 5A). The center point composition was replicated in order to assess the scatter of the data. However, browning occurred in one of the replicates (i.e. Fig. 6, 50% O2 + 5% CO2 (2)) by day 15. Browning was also observed at the center point composition in one of the replicates in a later experiment. The variability in ajmalicine production between these two replicates was not consistent with the low variability observed between replicate bubble columns sparged with air (i.e. S.D. = 14%) or between the shake flask and the air-sparged bubble column in this experiment (i.e. S.D. = 10%). Browning was observed in three bubble columns, two of which contained malfunctioning dissolved carbon dioxide probes, i.e. one replicate at the center point and one with 21.6% O2 + 1.5% CO2 . The rupture of the probe’s membrane and the leakage of electrolyte into the culture may have contributed to browning. Based on these and other experiments, browning was consistently observed in cultures sparged with a combination of high O2 and high CO2 concentration, i.e. at 50% O2 + 10% CO2 , 78.4% O2 + 8.6% CO2 , and 95% O2 + 5% CO2 ; browning was observed 50% of the time with 50% O2 + 5% CO2 . Culture browning was usually observed between day 13 and 20 and associated with lowered cell viability and either partial or complete suppression of ajmalicine production. Browning was not a result of detectable biological contamination (i.e. when medium was plated on LB agar plates). Various axial and factorial gas compositions were later rerun in two additional experiments. The trends observed with growth and ajmalicine production in the presence of increasing CO2 concentrations were reproducible although the actual ajmalicine concentrations varied from one experiment to another. 3.2.3. Correlating ajmalicine production to sparge gas O2 and CO2 concentrations Using the data from the block consisting of the factorial, center point, and air compositions, the coefficients to the mathematical correlation proposed by the Box-Wilson statistical design was determined using the least squares method and is shown in Table 2. The correlation is: z = b0 + b1 x + b2 y + b11 x2 + b22 y2 + b12 xy where z is the extracellular ajmalicine concentration in mg/L, x the % O2 in the sparge gas, and y the % CO2 in the sparge gas. The ajmalicine levels between the two blocks differed Table 2 Coefficients of the correlation: z = b0 + b1 x + b2 y + b11 x2 + b22 y2 + b12 xy b0 b1 b2 b11 b22 b12

−17 5.6 −29 −0.052 2.7 −0.012

Where z = ajmalicine concentration at day 21 (mg/L), x = % O2 in the sparge gas and y = % CO2 in the sparge gas.

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

431

by the correlation coefficients. The values and signs of b1 and b2 indicated that ajmalicine production was increased with an increase in the O2 concentration but more strongly decreased by an increase in the CO2 concentration. But as the CO2 concentration increased to 10%, the negative effects of the linear term, b2 , was counteracted by the interactive effects of CO2 , represented by b22 , leading to a minimization of ajmalicine production at 5.6% CO2 . The interaction between O2 and CO2 was minor as indicated by the value of b12 . From the correlation, we learned that the interaction between O2 and CO2 was minimal and that a CO2 concentration of 5.6% minimized ajmalicine production. 3.3. Dissolved oxygen profile in the cell-alginate beads Fig. 7. Extracellular ajmalicine production (on day 21) predicted by the Box-Wilson correlation as a function of the % O2 and % CO2 in the sparge gas.

significantly since they were conducted at different times and thus could not be used collectively to determine the coefficients of the correlation. Using the data shown in Fig. 5, the extracellular ajmalicine concentration at day 21 was used to determine the coefficients of the equation. Since all the cultures were cultivated for at least 21 days, no extrapolation was necessary. The ajmalicine concentrations predicted by the mathematical correlation were plotted as a function of the % O2 and % CO2 in the sparge gas (Fig. 7); the surface plot resembled a saddle. According to this surface plot, the optimum O2 concentration for ajmalicine production was only slightly dependent on the CO2 concentration in the sparge gas. For the CO2 concentration range of 0.03–10%, the optimum O2 concentration for ajmalicine production ranged from 54–55% O2 . For the O2 concentration range of 10–90%, ajmalicine production was minimized at a CO2 concentration of 5.6%. The applicability of this correlation was supported by the data from the axial gas combinations. Note that only the factorial, center point, and air combinations were used in fitting the correlation. The inhibition of ajmalicine production was observed experimentally at 10% O2 + 5% CO2 and 95% O2 + 5% CO2 . At these gas phase concentrations, the correlation predicted negative values for ajmalicine concentration. Although the correlation did not predict the actual numerical value, it did predict the severe inhibition of ajmalicine production at 10% O2 + 5% CO2 and 95% O2 + 5% CO2 . For the O2 range studied, the gas composition predicted by the mathematical correlation which would yield the maximum ajmalicine concentration, ∼55% O2 + 0.03% CO2 , was similar to that observed experimentally, 50% O2 + 0.03% CO2 , in the block consisting of the axial gas combinations. Based on our experimental results, increasing the CO2 concentration at a given O2 concentration appeared to inhibit ajmalicine production (Fig. 5). The inhibitory effect of increasing CO2 concentration on ajmalicine production was more pronounced at 21.6% O2 than at 78.4% O2 , as captured

The dissolved O2 profile within the cell-alginate beads was calculated for cultures sparged with either 10% O2 or 21% O2 at different cell densities, ranging from the inoculum density to the maximum cell density achieved experimentally with 10% O2 or 21% O2 cultures (i.e. 120 and 300 g FW/l). The specific O2 uptake rate for C. roseus was experimentally determined using cell suspensions during exponential growth (qmax = 4.69 × 10−4 mmol O2 /g FW min). Initially, this specific O2 uptake rate was assumed to be constant and repre-

Fig. 8. The dissolved oxygen profile within spherical immobilized beads sparged with (A) 10% O2 and (B) 21% O2 at different cell culture densities. The dissolved oxygen profile was calculated using an oxygen uptake rate of 4.69 × 10−4 mmol O2 /g FW min. The inoculum cell density was 50 g FW/l. The maximum cell density observed in cultures sparged with 10% O2 + 5% CO2 and with 21% O2 + 0.03% CO2 was 120 and 300 g FW/l, respectively.

432

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

sentative of the rate in immobilized cells and was used for calculating the dissolved O2 profile within the cell-alginate beads assuming saturation kinetics (see Section 2.10). The dissolved O2 profile within the cell-alginate beads (bead radius of 0.1 cm) as a function of distance from the center and cell density is shown in Fig. 8. The predicted dissolved O2 concentration would have reached zero at the center of the bead at cell densities higher than the inoculum density (50 g/l) with either 10% O2 or 21% O2 in the sparge gas. However, experimentally, cell densities of 120 and 300 g FW/l were achieved in cultures sparged with 10% O2 + 5% CO2 and 21% O2 + 0.03% CO2 , respectively. High cell viability was observed in these cultures. These experimental observations suggest that the dissolved oxygen was not exhausted in the cell-alginate beads but that the actual specific O2 uptake rates of the immobilized cells was less than that of the suspensions. Since growth rates in the immobilized cultures were reduced compared to suspensions (i.e. 25–50% of suspensions), the actual specific O2 uptake rates of immobilized cells may have been much lower than that of suspensions. Using a spe-

Fig. 9. The dissolved oxygen profile within spherical immobilized beads sparged with (A) 10% O2 and (B) 21% O2 at different cell culture densities. The dissolved oxygen profile was calculated using an oxygen uptake rate of 1.41 × 10−4 mmol O2 /g FW min. The inoculum cell density was 50 g FW/l. The maximum cell density observed in cultures sparged with 10% O2 + 5% CO2 and with 21% O2 + 0.03% CO2 was 120 and 300 g FW/l, respectively.

cific O2 uptake rate which was 30% that of suspensions (i.e. 1.41 × 10−4 mmol O2 /g FW min), the dissolved O2 profile within the cell-alginate beads was re-calculated for cultures sparged with either 10% O2 or 21% O2 , as shown in Fig. 9. Using this lower specific oxygen uptake rate, the dissolved oxygen was not depleted at the center of the bead with 10% O2 in the sparge gas, even at the culture’s highest cell density of 120 g/l. With 21% O2 in the sparge gas, the dissolved oxygen becomes limiting at the center of the bead only when the maximum cell density of 300 g FW/l was reached; this maximum cell density was reached after day 24 (i.e. Fig. 2 and in Ref. [35]) and remained at this density through day 33 (i.e. Fig. 6). Hence, this specific O2 uptake rate may be more representative of that in immobilized cells.

4. Discussion The O2 and CO2 concentration in the sparge gas significantly altered growth and ajmalicine production from immobilized C. roseus cells. Benefits and drawbacks of operating at O2 and CO2 concentrations outside that of air composition were also observed. For example, ajmalicine production was enhanced with O2 concentrations higher than that of air (i.e. 50% O2 + 0.03% CO2 ) but O2 concentrations of 90% and 95% O2 appeared to be toxic to the cells. Lower oxygen concentrations (i.e. 10% O2 + 5% CO2 ) resulted in high viability but inhibited growth and ajmalicine production. High carbon dioxide concentrations appeared to inhibit growth or specific ajmalicine production. At higher O2 concentrations (i.e. 78.4% O2 ), increasing CO2 concentration appeared to affect ajmalicine production by decreasing growth while the specific ajmalicine production remained the same. At lower O2 concentrations (i.e. 21–21.6% O2 ), increasing CO2 concentration decreased both growth and specific ajmalicine production. The effects of O2 and CO2 on ajmalicine production and growth in this study were compared to that in the literature with C. roseus and other cell culture systems. With C. roseus suspensions, Schlatmann et al. [41] reported that a dissolved oxygen concentration greater than 43% of saturation with air promoted high ajmalicine productivities while dissolved oxygen below 29% of saturation with air resulted in low ajmalicine productivities. In our study, ajmalicine production was also reduced with low O2 (10%) in the sparge gas. In contrast, Mirjalili and Linden [3] observed that paclitaxel production was initiated earlier from Taxus cuspidata suspensions with reduced O2 (10%). Schlatmann et al. [32] observed a negligible effect of CO2 level on ajmalicine production with C. roseus suspensions although the removal of CO2 from the recirculated gas enhanced the secretion of ajmalicine. In Schlatmann et al. [2], reduced ajmalicine production was attributed to higher levels of dissolved gas metabolites, which included CO2 . With respect to growth, Van Gulik et al. [42] observed no difference in growth of C. roseus suspension sparged with 0.3 and 2%

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

(v/v) CO2 . On the other hand, Archambault [43] observed higher specific growth rates and higher biomass concentrations in immobilized C. roseus cultures sparged with air + 2% CO2 than in cultures sparged with air or air + 5% CO2 . In our study, increasing CO2 levels (i.e. from 1.5 to 10% CO2 ) had an inhibitory effect on either growth, specific ajmalicine production, or both. Mirjalili and Linden [3] also reported an inhibitory effect with high CO2 concentration (10%) combined with 2 ppm ethylene on paclitaxel production from T. cuspidata suspensions. Similarly, Linden et al. [1] observed an inhibition in artemisinin production from Artemisia annua suspensions with high CO2 concentration (i.e. 2.8–12.8% CO2 ). The changes in biomass concentration or ajmalicine concentration with gas composition in our immobilized C. roseus cultures were not mediated by differences in ammonium, nitrate, or sugar uptake rates (data not shown). Other researchers have observed differences in uptake of sugars, phosphate, calcium, and nitrogen with gas composition [3,22,43,44]. In our study, browning was observed in cultures sparged with high concentrations of both O2 and CO2 . Other researchers report that ethylene or low CO2 concentrations are associated with browning [5–7,12–13,22]. Kobayashi et al. [28,29] linked culture browning to increased levels of ethylene in the late exponential and stationary growth phases of Thalictrum minus suspensions. They observed that enriching air with 2% CO2 prevented browning and enhanced berberine production. Schlatmann et al. [4] observed that browning disappeared when a significant fraction of the exhaust gas was recirculated back to the bioreactor. Their results suggested that the removal of other volatile gas components by aeration was responsible for browning and for the reduced secondary metabolism with scale-up. Differences in the response to gas composition between different plant cell culture systems suggest that the effects are specific to the biosynthetic pathway of the secondary metabolite and to the metabolic state of the cells. Relevant to ajmalicine production, gas composition effects on ajmalicine production may be partly attributed to the interaction of O2 and CO2 with the enzymes involved in the biosynthesis of terpenoid indole alkaloids. The enzyme geraniol 10hydroxylase (G10H) catalyzes the first committed step in the biosynthesis of secologanin, the terpene precursor to the terpenoid indole alkaloids. In some cases, G10H has been shown to be the bottleneck in terpenoid indole alkaloid production from C. roseus cultures [45]. Schlatmann et al. [41] proposed that the gas effects were mediated through G10H which requires O2 for the reaction and which is partly but reversibly inhibited by CO2 . This explanation would be consistent with our observations and suggestive of a mechanism. In summary, O2 and CO2 are slightly soluble nutrients that can significantly impact both growth and secondary metabolism. Gas composition must not only meet the oxygen demand of a culture but may require optimization for enhancing secondary metabolism.

433

Acknowledgements The authors would like to thank Chris Hanscom, Shana Chacko Mueller, and Paula Miller for their assistance with the sample preparations and their good company. The authors would also like to thank James Falabella for his assistance with MathCad. This work was sponsored by National Science Foundation (#BCS-9117252). References [1] Linden JC, Haigh JR, Mirjalili N, Phisaphalong M. Gas concentration effects on secondary metabolite production by plant cell cultures. Adv Biochem Eng Biotechnol 2001;72:27–62. [2] Schlatmann JE, ten Hoopen HJG, Heijnen JJ. A simple structured model for maintenance, biomass formation, and ajmalicine production by non-dividing Catharanthus roseus cells. Biotechnol Bioeng 1999;66(3):147–57. [3] Mirjalili N, Linden JC. Gas phase composition effects on suspension cultures of Taxus cuspidata. Biotechnol Bioeng 1995;48(2):123– 32. [4] Schlatmann JE, Nuutila AM, van Gulik WM, ten Hoopen HJG, Verpoorte R, Heijnen JJ. Scaleup of ajmalicine production by plant cell cultures of Catharanthus roseus. Biotechnol Bioeng 1993;41:253– 62. [5] Leckie F, Scragg AH, Cliffe KC. An investigation into the role of initial kL a on the growth and alkaloid accumulation by cultures of Catharanthus roseus. Biotechnol Bioeng 1991;37:364–70. [6] Leckie F, Scragg AH, Cliffe KC. Effect of bioreactor design and agitator speed on the growth and alkaloid accumulation by cultures of Catharanthus roseus. Enzyme Microb Technol 1991;13:296–305. [7] Leckie F, Scragg AH, Cliffe KC. Effect of impeller design and speed on the large-scale cultivation of suspension cultures of Catharanthus roseus. Enzyme Microb Technol 1991;13:801–10. [8] Reinhard E, Kreis W, Barthlen U, Helmbold U. Semicontinuous cultivation of Digitalis lanata cells: production of ␤-methyldigoxin in a 300-L airlift bioreactor. Biotechnol Bioeng 1989;34:502–8. [9] Scragg AH, Morris P, Allan EJ, Bond P, Fowler MW. Effect of scale-up on serpentine formation by Catharanthus roseus suspension cultures. Enzyme Microb Technol 1987;9:619–24. [10] Spieler H, Alfermann AW, Reinhard E. Biotransformation of ␤methyldigitoxin by cell cultures of Digitalis lanata in airlift and stirred tank reactors. Appl Microbiol Biotechnol 1985;23:1–4. [11] Ulbrich B, Wiesner W, Arens H. Large scale production of rosmarinic acid from plant cell cultures of Coleus blumei Benth. In: Neumann KH, Barz W, editors. Primary and secondary metabolism of plant cell cultures. Berlin, Germany: Springer Verlag; 1985. p. 293–303. [12] Pareilleux A, Chaubet N. Mass cultivation of Medicago sativa growing on lactose: kinetic aspects. Eur J Appl Microbiol Biotechnol 1981;11:222–5. [13] Smart NJ, Fowler MW. Effect of aeration on large-scale cultures of plant cells. Biotechnol Lett 1981;3(4):171–6. [14] Wagner F, Vogelmann H. Cultivation of plant tissue cultures in bioreactors and formation of secondary products. In: Barz W, Reinhard E, Zenk MH, editors. Plant tissue culture and its biotechnological application. Berlin, Germany: Springer Verlag; 1977. p. 245–52. [15] Zenk MH, El-Shagi H, Arens H, Stockigt J, Weiler EW, Deus B. Formation of the indole alkaloids serpentine and ajmalicine in cell suspension cultures of Catharanthus roseus. In: Barz W, Reinhard E, Zenk MH, editors. Plant tissue culture and its biotechnological application. Berlin, Germany: Springer Verlag; 1977. p. 27–43. [16] Kato A, Shimizu Y, Nagai S. Effect of initial kL a on the growth of tobacco cells in batch culture. J Ferment Technol 1975;53:744–51.

434

C.W.T. Lee-Parsons, M.L. Shuler / Enzyme and Microbial Technology 37 (2005) 424–434

[17] Meijer AH, Verpoorte R, Hoge JHC. Regulation of enzymes and genes involved in terpenoid indole alkaloid biosynthesis in Catharanthus roseus. J Plant Res 1993;3:145–64. [18] Jennewein S, Rithner CD, Williams RM, Croteau RB. Taxol biosynthesis: taxane 13␣-hydroxylase is a cytochrome P450-dependent monooxygenase. Proc Natl Acad Sci USA 2001;98(24):13595–600. [19] Payne GF, Bringi V, Prince C, Shuler ML. Plant cell and tissue culture in liquid systems. Munich, Germany: Hanser Publishers; 1992. p. 346. [20] Breuling M, Alfermann AW, Reinhard E. Cultivation of cell cultures of Berberis wilsonae in 20-L airlift bioreactors. Plant Cell Rep 1985;4:220–3. [21] Tabata M, Fujita Y. Production of shikonin by plant cell cultures. In: Zaitlin M, Day P, Hollaender A, editors. Biotechnology in plant science. Orlando, FL: Academic Press; 1985. p. 207–18. [22] Smart NJ, Fowler MW. Mass cultivation of Catharanthus roseus cells using a nonmechanicaly agitated bioreactor. Appl Biochem Biotechnol 1984;9:209–16. [23] DiIorio AA, Cheetham RD, Weathers PJ. Carbon dioxide improves the growth of hairy roots cultured on solid medium and in nutrient mists. Appl Microbiol Biotechnol 1992;37(4):463–7. [24] Ducos JP, Feron G, Pareilleux A. Growth and activities of enzymes of primary metabolism in batch cultures of Catharanthus roseus cell suspension under different pCO2 conditions. Plant Cell Tissue Organ Cult 1988;13(3):167–77. [25] Ducos JP, Pareilleux A. Effect of aeration rate and influence of pCO2 in large-scale cultures of Catharanthus roseus cells. Appl Microbiol Biotechnol 1986;25:101–5. [26] McKeon TA, Yang SF. Biosynthesis and metabolism of ethylene. In: Davies PJ, editor. Plant hormones and their role in plant growth and development. 2nd ed. The Netherlands: Kluwer Academic Publishers; 1987. p. 94–112. [27] Kim D-I, Pedersen H, Chin C-K. Cultivation of Thalictrum rugosum cell suspension in an improved airlift bioreactor: stimulatory effect of carbon dioxide and ethylene on alkaloid production. Biotechnol Bioeng 1991;38:331–9. [28] Kobayashi Y, Masakazu H, Fukui H, Tabata M. The role of ethylene in berberine production by Thalictrum minus cell suspension cultures. Phytochemistry 1991;30:3605–9. [29] Kobayashi Y, Fukui H, Tabata M. Effect of carbon dioxide and ethylene on berberine production and cell browning in Thalictrum minus cell cultures. Plant Cell Rep 1991;9:496–9. [30] Cho GH, Kim DI, Pedersen H, Chin C-K. Ethephon enhancement of secondary metabolite synthesis in plant cell cultures. Biotechnol Prog 1988;4:184–8. [31] Lee CWT, Shuler ML. Different shake flask closures alter gas phase composition and ajmalicine production in Catharanthus roseus cell suspensions. Biotechnol Technol 1991;5:173–8.

[32] Schlatmann JE, Fonck E, ten Hoopen HJG, Heijnen JJ. The negligible role of carbon dioxide and ethylene in ajmalicine production by Catharanthus roseus cell suspensions. Plant Cell Rep 1994;14:157–60. [33] Burg SP, Burg EA. Molecular requirements for the biological activity of ethylene. Plant Physiol 1967;42:144–52. [34] Perry RH, Chilton CH. Chemical engineer’s handbook. 5th ed. New York: McGraw-Hill; 1973. pp. (2–74)-(2–75). [35] Lee CWT, Shuler ML. The effect of inoculum density and conditioned medium on the production of ajmalicine and catharanthine from immobilized Catharanthus roseus cells. Biotechnol Bioeng 2000;67(1):61–71. [36] Asada M, Shuler ML. Stimulation of ajmalicine production and excretion from Catharanthus roseus: effects of adsorption in situ, elicitors, and alginate immobilization. Appl Microbiol Biotechnol 1989;30:475–81. [37] Murashige T, Skoog F. A revised medium for rapid growth and bioassay with tobacco tissue cultures. Physiol Plant 1962;15:473– 97. [38] Lee-Parsons CWT, Shuler ML. The effect of ajmalicine spiking and resin addition timing on the production of indole alkaloids from Catharanthus roseus cell cultures. Biotechnol Bioeng 2002;79(4):408–15. [39] Morris P, Scragg AH, Smart NJ, Stafford AJ. Secondary product formation by cell suspension cultures. In: Dixon RA, editor. Plant cell culture: a practical approach. Oxford, England: IRL Press Limited; 1985. p. 142–67. [40] Widholm JM. The use of fluorescein diacetate and phenosafranine for determining viability of cultured plant cells. Stain Technol 1972;47(4):189–94. [41] Schlatmann JE, Vinke JL, ten Hoopen HJG, Heijnen JJ. Relation between dissolved oxygen concentration and ajmalicine production rate in high-density cultures of Catharanthus roseus. Biotechnol Bioeng 1995;45(5):435–9. [42] Van Gulik WM, Nuutila AM, Vinke KL, ten Hoopen HJG, Heijnen JJ. Effects of carbon dioxide, air flow rate, and inoculation density on the batch growth of Catharanthus roseus cell suspensions in stirred fermentors. Biotechnol Prog 1994;10(3):335–9. [43] Archambault J. Large-scale (20 L) culture of surface-immobilized Catharanthus roseus cells. Enzyme Microb Technol 1991;13: 882–92. [44] Taticek RA, Moo-Young M, Legge RL. Effect of bioreactor configuration on substrate uptake by cell suspension cultures of the plant Eschscholtzia californica. Eur J Appl Microbiol Biotechnol 1990;33:280–6. [45] Moreno PRH, van der Heijden R, Verpoorte R. Effect of terpenoid precursor feeding and elicitation on formation of indole alkaloids in cell suspension cultures of Catharanthus roseus. Plant Cell Rep 1993;12(12):702–5.