Biotechnol. Prog. 1990, 6, 243-248
243
Indole Alkaloid Formation by Catharanthus roseus Cells in a Biofilm Reactor F. Kargi,* B. Ganapathi, and K. Maricic Biotechnology Engineering Laboratory, Department of Chemical Engineering, Washington University, St. Louis, Missouri 63130
Catharanthus roseus cells producing indole alkaloids were grown in the form of a biofilm. Production medium was circulated through the reactor parallel to the upper surface of the horizontal biofilm. Sugar consumption and indole alkaloid formation were followed to compare the performance of cultures with different biofilm thicknesses. Dissolved oxygen concentration gradients within the biofilms were determined at the end of each run. RNA and protein content of the cells in the upper and lower layers of the biofilms were compared. Results obtained in the biofilm experiments were compared to those obtained with suspension cultures. At optimized biofilm thicknesses, the biofilm reactor was more effective than suspension cultures in maximizing indole alkaloid titers. This is thought to be due to better cell-cell contact within the biofilm and nutrient concentration gradients, which resulted in low growth rates.
Introduction A wide variety of chemicals are produced by higher plants. Products of interest include specialty chemicals, pharmaceuticals, flavor and fragrance agents, agrochemicals, and oils (Stafford et al., 1986). The most valuable plant cell products are pharmaceuticals. Approximately 25 ?' 4 of all prescription medicines are derived from plants (Stafford et al., 1986). The estimated annual United States market for plant-produced pharmaceuticals is about 9 billion dollars (Sahai and Knuth, 1985). Presently, most plant products are extracted from whole plants. Several problems associated with extraction of chemicals from intact plants are seasonal variations, pests, diseases, and inconsistent product quality and yield. A viable alternative is the use of large-scale plant cell cultures for production of pharmaceuticals and chemicals. However, there are several difficulties inherently associated with the production of secondary metabolites by plant cell cultures. Slow growth and product formation rates, low product concentrations, and shear sensitivity of plant cells due to large aggregate formation are some of the problems. Some degree of cellular differentiation also seems to be important for maximizing product formation. Finally, product recovery is difficult since the plant secondary metabolites are often intracellular compounds that must be separated from the cells following production. According to recent economic evaluations, indole alkaloid titers and product formation rates in plant cell culture need to be improved significantly to make plant cell culture a viable alternative to extraction from intact plants (Drapeau et al., 1987). Various cultivation methods have been developed for plant cells. Several recent review articles discuss advantages and disadvantages of these methods (Sahai and Knuth, 1985; Kargi and Rosenberg, 1987). Slow growth imposed by nutrient limitations, cytodifferentiation, intercellular organization, and certain hormones, such as kinetin, are necessary for high levels of secondary metabolite formation (Lindsey and Yeoman, 1983; Morris, 1986). Nutrient limitations and desired hormone concentrations can be achieved in suspension cultures. However, control 8756-7938/90/3006-0243$02.50/0
of aggregate size and intercellular organization may be difficult to achieve in suspension cultures. Immobilized systems have the ability to impart spatial organization of cells, as well as to allow for the establishment of nutrient concentration gradients within the support matrix. However, control of microenvironmental conditions and achieving adequate levels of cell-to-cell contact are potential problems in immobilized cell systems. B i o f i i cultures of plant cells have significant advantages over other cultivation methods (Kargi and Rosenberg, 1987). High levels of cell-to-cell contact can be achieved; nutrient limitations and the degree of cell differentiation can be controlled by controlling the biofilm thickness. Moreover, high cell concentrations can be achieved in biofilm cultures, and shear effects on cells within the biofilm would be reduced. In a previous study, high levels of indole alkaloid formation were obtained from Catharanthus roseus cells grown on Ca-alginate beads in the form of a b i o f i i and aggregates in a packed column reactor (Kargi, 1988). In this study, plant cells were grown in the form of a horizontal film on an agar-nutrient support. Production medium was circulated through the reactor parallel to the upper surface of the biofilm. Sugar consumption and alkaloid formation were determined and compared with results obtained in suspension cultures. Dissolved oxygen concentration gradients within the b i o f i s were measured. RNA and protein content of the cells in the upper and lower layers of the biofilms were determined as an indication of relative growth rates at different positions in the biofilms.
Experimental Section Materials and Methods. Culture. C. roseu cells were obtained from Dr. H. Blanch of the University of California, Berkeley. The cells were maintained in callus and suspension cultures under fluorescence light. The growth medium used in suspension and callus culture was Murashige and Skoog medium supplemented with 0.5 mg/L (2,4-dichlorophenoxy)aceticacid (Dixon, 1985). The
0 1990 American Chemical Society and American Institute of Chemical Engineers
244
Biotechnol. Prog., 1990,Vol. 6, No. 4
Air
Nutrient Reservoir
Figure 1. Schematic of the biofilm reactor used.
medium was adjusted to pH 5.8. The cells were maintained at 26 OC under continuous lighting. Production Media. The production medium used in the biofilm and suspension culture experiments was the NB5 medium, which consisted of Gamborg's B5 basal medium supplemented with 1.0 mg/L l-naphthaleneacetic acid and 0.1 mg/L kinetin (Morris, 1986). The media was adjusted to pH 5.8 with 1 N NaOH. Biofilm Experimental Setup. A schematic of the biofilm reactor configuration is depicted in Figure 1. The reactor was operated as a batch loop system. The biofilm was located on top of a nutrient-agar layer in a rectangular glass inner tube. The surface dimensions of this inner tube were 17.5 cm X 3.5 cm. This inner tube was maintained aseptically inside a glass reactor equipped with ports for venting. Liquid nutrient media was aerated and agitated in a nutrient reservoir. T h e air was humidified by being bubbled through sterile water before entering the reservoir to avoid evaporation losses and was filtered at the extrance and exit of the nutrient reservoir and the biofilm reactor. The liquid nutrient medium was circulated continuously between the reservoir and the biofilm reactor and flowed in the form of a liquid film parallel to the upper surface of the biofilm. Five different biofilm experiments were conducted with initial biofilm dry weights of 0.5, 0.75, 1.0, 1.25, and 1.5 g. These cell weights corresponded to biofilm thicknesses of 2, 3,4, 5, and 6 mm, respectively. Biofilm dry weights a t the start of experiments were determined by filtering samples from the suspension culture flasks and measuring the dry weight of biomass after filtration. The initial nutrient volumes for the 0.5-, 0.75-, and 1.5-g biofilm runs were 475 mL. The initial nutrient volumes for the 1.0- and 1.25-g biofilm runs were 1000 mL. The nutrient medium was circulated a t rates varying between 4 and 12 mL/ min. Residence times in the inner tube were in the range of 4-9 min. Biofilm Experimental Procedure. Prior to the startup of an experiment, 60-70 mL of agar-NB5 nutrient media was poured into t h e inner tube. Cells were cultivated in suspension culture for 7 days up to a certain dry weight and were centrifuged aseptically. Centrifuged cells were resuspended in a smaller volume of media, and this thick suspension of cells having the desired dry weight was then aseptically transferred on top of the agar-NB5 layer, to form a biofilm. The cells were allowed to form a solid film for 2 days under aseptic conditions before being placed in the reactor. The biofilm experiments were conducted for 16-18-day periods. Media samples were taken aseptically once every 2 days and were analyzed for total reducing sugars and alkaloid concentration. At the end of each run, the biofilm was removed from the reactor and the dissolved oxygen profile in the biofilm was determined. Cell samples from the top and bottom of each
biofilm were obtained with the aid of a sterile loop and were analyzed for RNA and protein content. Suspension Culture Experimental Procedure. Shake flask suspension culture experiments were conducted with the same cell concentration as that of biofilm experiments t o compare the performance of biofilm cultures with suspension culture. Three suspension culture experiments were conducted with initial cell concentrations of 1.02, 1.64, and 2.92 g/L. The experiments were conducted for a 16-day period. Cell growth, sugar consumption, and alkaloid concentration were monitored throughout the course of the experiments. Analytical Methods. D r y Weight. The dry weight of the cells was estimated by filtering the cells through a 0.45-y glass-fiberfilter paper. The preweighed filter paper was wetted with distilled water, and the cells were filtered onto the paper. The cells were then washed with distilled water. After filtration, the filter paper was dried overnight in a vacuum oven at 60 "C and was reweighed. Cell Viability. The viability (respiration activity) of the cells was determined a t the end of each biofilm run. The TTC assay was used, which was based on the cells ability to reduce tetrazolium salts to water-insoluble red formazan (Towill and Mazur, 1975). The TTC assay was used only qualitatively for the purposes of this research. The red color produced in a given sample was compared visually to the color formed by viable cells from a suspension culture. Total Reducing Sugar Concentration. The first step was enzymatic hydrolysis of the sucrose in the media to glucose and fructose. This was accomplished by mixing the sample with an invertase solution and heating for 30 min at 60 "C. Samples from the hydrolysis treatment were then assayed by the dinitrosalicylic acid (DNS) method (Miller, 1959). Alkaloid Concentration. Both extracellular and intracellular indole alkaloid concentrations were determined. The cells were extracted for 24 h with 95% ethanol to remove intracellular alkaloids. Following centrifugation, the ethanol was separated from the cells and evaporated. The residue was redissolved in 0.1 N HCl. The alkaloids were then separated from hydrophobic compounds at low pH followed by separation from hydrophilic compounds at high pH. The organic solvent used was ethyl acetate. Following the extraction procedure, the ethyl acetate was evaporated, and the residue was redissolved in methanol. The total alkaloid concentration was determined by absorbance measurement a t 280 nm and subsequent comparison to a calibration curve prepared by using the alkaloid ajmalicine (Payne, 1986; Shuler, 1986). Each measurement was done three times, and average values were reported. Deviations from the average were less than 10%. Protein Content. The cells were first collected on a glass-fiber filter and were washed twice with boiling 70 % aqueous ethanol. The cells were then dried with acetone and were transferred to a solution of 1 N NaOH. The solution was heated at 85 "C for 3 h. Following filtration, the filtrate was assayed for protein content by the Lowry method (Lowry, 1951). RNA Content. Assay of the cells for RNA consisted of multiple extractions followed by spectrophotometric determination at 260 nm (Dixon, 1985). A calibration curve was developed with yeast RNA. Dissolved Oxygen Profile. A t the end of each biofilm run, the dissolved oxygen (DO) profile in each biofilm was determined with a dissolved oxygen microelectrode (Diamond Electrotech, MI). The polarographic
245
Biotechnol. Prog., 1990,Vol. 6, No. 4
f
51
I
*
0
Time (days)
2
4
6
8
10 12
14
1 6 18 2 0
Time (days)
Figure 2. Sucrose consumption profde+qfor b i o f i i s with different thicknesses: (0) 0.5 g (2 mm); (0) 0.75 g (3 mm); ( 0 ) 1.0 g (4 mm); ( A ) 1.25 g (5 mm); ( 0 )1.5 g (6 mm).
Figure 3. Extracellular media alkaloid content for biofilms with 0.5 g (2 mm); (0) 0.75 g (3 mm); ( 0 ) different thicknesses: (0) 1.0 g (4 mm); (A) 1.25 g (5 mm); ( 0 )1.5 g (6 mm).
electrode was calibrated a t 100 % dissolved oxygen saturation. Current measurements were converted to percent dissolved oxygen saturation.
Table I. Final Alkaloid Content for Biofilm Runs biofilm initial final % % cell alkaloid dry run wt, g concn, g/L content, % intracellular extracellular 1 0.50 1.05 0.044 87.5 12.5 2 0.75 0.073 45.5 54.5 1.58 3 1.00 1.00 0.233 15.2 84.8 4 1.25 1.24 0.270 18.7 81.3 5 1.50 3.16 0.178 66.7 33.3
Results and Discussion Biofilm Experiments. Five different biofilm experiments were conducted with 0.5-, 0.75,1.0-, 1.25, and 1.5-g biofilms, respectively. The extent of growth during each run was less than 10% of the initial dry weight. Therefore, biofilm thickness was assumed to be constant throughout in each experiment. Sugar consumption curves for the biofilm runs are shown in Figure 2. In each case, sucrose consumption was less than 20 % , which is an indication of slow growth. The media alkaloid contents (extracellular) of each biofilm run are shown in Figure 3. It is apparent that there is significant variation among these runs. This is a result of differences in the total production of both alkaloids, i.e., from variations in the percentage of the total alkaloids released into the media (extracellular) and perhaps due to the inherent instability of the source culture. The percentage of extracellular alkaloids seemed to be related to the final viability of the cells. The experiments having the highest extracellular content had few viable cells. It seems that intracellular alkaloids were released into the media due to hostile conditions, leading to the loss of cell viability. Table I summarizes the final alkaloid content (intracellular and extracellular) for the biofilm runs. Alkaloid production increased with increasing film thickness, and thicker films tended to produce more extracellular alkaloids up to a film thickness of 5 mm. This is thought to be due to nutrient gradients within the biofilms. Figure 4 shows the variation in the total final alkaloid content with the thickness of the biofilm. The biofilm thickness appeared to be roughly linear with the bio-
mass content. Increases in biofilm thickness caused favorable nutrient gradients that may have resulted in the higher production of alkaloids up to a certain point. This increase in alkaloid production is most likely a result of low nutrient levels, slow growth, and primitive levels of differentiation caused by the nutrient gradients. Final alkaloid concentration (titer) reached an optimum level with the 5-mm (1.25-g) biofilm. Alkaloid formation decreased with a further increase in biofilm thickness to 6 mm (1.5 g). Apparently, at this biofilm thickness, nutrient gradients became too severe and prevented the cells from maximizing secondary metabolite biosynthesis. Dissolved oxygen measurements at different depths within the biofilms at the end of experiments showed that steep nutrient gradients were present. Figure 5 depicts a typical dissolved oxygen gradient for the 2-mm biofilm (0.5-g biomass). In almost all cases, the dissolved oxygen content decreased from 100% saturation at the top of the biofilm to only 5 % a t the bottom. However, the DO concentration profile was steeper with larger biofilm thicknesses. Higher alkaloid contents obtained with steeper DO profiles indicated that DO limitations may have stimulated indole alkaloid formation. Figure 6 depicts variation of the total alkaloid content of the culture with the slope of the DO profile, indicating an optimal level of DO gradient or limitation to maximize alkaloid titer. Very
Biotechnol. Prog., 1990,Vol. 6, No. 4
246 0.301
-m
I
4 rl
al
3
n
m \ E
x rl
0.20
m E
o\o
c .ri
u U
u ffl
2
0.10
u d
.rl
0
rl
a x
4
4
rl
a u H
I
0
2
4
)O
6
Biofilm Thickness (mm)
D.O.
Figure 4. Variation of total alkaloid content with the thickness of the biofilm. 1
4.00
5.00
6.00
7.00
Gradient Within the Biofilm (mg/l.mm)
Figure 6. Variation of total alkaloid content of cells with the slope of the DO profile within the biofilm of various thicknesses.
both diffusion and convection within the biofilm reactor. To estimate the importance of convection within the biofilm reactor, the effective diffusivity of oxygen was estimated with the following equation
C .rl
u
U
u (0
c m
sr
0
u ffl U 04
0.0
1.0
2.0
3.0
4.0
5.0
Distance From Top Surface of Biofilm (mm)
Figure 5. Typical dissolved oxygen (DO) profile within the biofilm of thickness of 2 mm.
steep DO profiles within the biofilm resulted in low alkaloid titers, possibly as a result of severe DO limitations. It was expected that the oxygen gradient would be influenced by
where De is the effective diffusivity of oxygen within the b i o f i i (cmZ/s); (ds/dy)j,,o is the initial slope of DO profile (mg 02/cm4);80,is the oxygen consumption rate (mg O,/ (cm2.(s-gof cells)); and X is the cell dry weight (g). With use of this correlation, the effective diffusivities and were calculated and found to vary between 2.4 X 7X cm2/s, yielding an average value of De = 4 X lo4 cmz/s. High diffusivity values obtained from t h e experiments indicated the presence of convective diffusion of DO within the biofilm, since diffusivities for pure diffusion of oxygen in such a medium would be on the order of 10-5 cm2/s. The RNA and protein contents at the top and bottom of the biofilms were determined. These results are shown in Table 11. RNA and protein results are presented as percentages of the total cell dry weight (mg of RNA or protein/100 mg of cells). Fast-growing cells have been shown to have higher RNA content than slow-growingcells. In most cases, the cells in the lower layer of the biofilm had a lower RNA content than the cells in the upper layers. This indicates that the cells at the bottom of the film were growing more slowly than those a t the top. Apparently, these differences, caused by nutrient gradients, resulted in favorable cell differentiation and increased levels of alkaloid production. In general, the protein results indicate the same trend. In most runs, the protein content was lower in the bottom layer of the biofilm. Again, this indicates slower growth at the bottom of the biofilm, which
247
Biotechnol. Prog., 1990, Vol. 6, No. 4
Table 11. RNA and Protein Contents of Biofilm Cells % RNA, % protein, biofilm mg/100 mg mg/100 mg dry position run wt, g in biofilm of cells of cells 1 0.50 top 1.23 12.67 bottom 1.04 12.67 2 0.75 top 0.72 bottom 0.52 3 1.00 top 0.35 14.09 bottom 0.29 12.79 14.20 4 1.25 top bottom 1.08 10.16 5 1.50 top 0.96 13.45 bottom 0.90 13.29
probably resulted in primitive levels of differentiation and, hence, increased levels of alkaloid production. Suspension Culture Experiments. Three different suspension culture experiments were conducted with initial cell concentrations of 1.02,1.64, and 2.92 g/L, respectively. In each suspension culture experiment, sugar consumption was between 80% and 90%, which was an indication of the rapid growth. Growth in suspension cultures was more significant than that in biofilm cultures, ranging from 30% to 40% of the initial dry weight. Time courses of the total alkaloid contents in suspension culture experiments (intracellular and extracellular) are depicted in Figure 7. The majority of the alkaloids in the suspension culture runs were intracellular. The percentages of intracellular alkaloids were 77.6%, 97.6%, and 98.7% for suspension culture experiments with 1.02, 1.64, and 2.92 g/L initial cell concentrations, respectively. Comparison of Biofilm and Suspension Culture Experiments. Table I11 is a comparison of the final alkaloid yields of the biofilm experiments with those of the suspension culture. Each biofilm experiment was compared to a suspension culture experiment having approximately the same initial cell concentration. At low biofilm thicknesses ( L I 3 mm), alkaloid titers in suspension culture were higher than those of the biofilm cultures. However, at high biofiim thicknesses ( L 1 4 mm), biofilm culture resulted in higher alkaloid titers than that of t h e suspension culture. Above certain biofilm thicknesses, the biofilm culture performance was superior to that of the suspension culture in terms of the total alkaloid titers. This is most likely a result of the low nutrient concentrations within the biofilm, resulting in slow growth. This difference in growth is especially apparent when the sugar utilization within the biofilm reactor (
