Transient Effects of Overexpressing Anthranilate Synthase r and ß
Catharanthus roseus produces two economically valuable anticancer drugs, vinblastine and vincristine. These drugs are members of the terpenoid indole alkaloids and accumulate in small quantities within the plant; thus these two drugs are expensive to produce. Metabolic engineering efforts have focused on increasing the alkaloids in this pathway through various means such as elicitation, precursor feeding, and gene overexpression. Recently we successfully expressed Arabidopsis genes encoding a feedback-insensitive anthranilate synthase R subunit under the control of the glucocorticoid-inducible promoter system and the anthranilate synthase ß subunit under the control of a constitutive promoter in C. roseus hairy roots. In this work we look at the transient behaviors of terpenoid indole alkaloids over a 72 h induction period in late exponential growth phase cultures. Upon induction, the tryptophan, tryptamine, and ajmalicine pools accumulated over 72 h. In contrast, the lochnericine, ho¨rhammericine, and tabersonine pools decreased and leveled out over the 72 h induction period. Visible changes within the individual compounds usually took from 4 to 12 h.
Introduction
Catharanthus roseus produces several economically valuable terpenoid indole alkaloids, including the anti- cancer compounds vinblastine and vincristine. These alkaloids, which are produced in small amounts within the plant (1), are used in combination with other drugs for treatment of cancers such as lymphomas and leuke- mia (2). Because of their economic and pharmacological value, researchers have focused on characterizing and engineering the complex network of the terpenoid indole alkaloid pathways to increase the production of vinblas- tine and vincristine. Previous attempts to engineer the terpenoid indole alkaloid pathways have focused on media optimization (3), elicitation (4, 5), precursor feeding (4, 6, 7), and the expression of a few cloned genes within cell suspension or hairy root cultures of C. roseus (8- 11).
The first committed step in the indole pathway is the conversion of tryptophan to tryptamine, whereas the first committed step in the terpenoid pathway is the conversion of geraniol to 10-hydroxygeraniol (Figure 1). 10- Hydroxygeraniol is converted to secologanin via multiple steps. Secologanin and tryptamine are combined to form strictosidine. Strictosidine is converted to a wide range of terpenoid indole alkaloids. The formation of trypto- phan is brought about by a series of reactions that start with the conversion of chorismate to anthranilate by anthranilate synthase (AS). The AS enzyme is a tetramer composed of two R subunits (ASR) and two ß subunits (ASß). The R subunit catalyzes the aromatization of chorismate while the ß subunit donates an amino group from glutamine (11).
The expression of a feedback-resistant ASR subunit from Arabidopsis (12) under the control of a glucocorti- coid-inducible promoter system (13) in C. roseus hairy roots for 72 h in late exponential growth phase sig- nificantly increased the amounts of tryptophan and tryptamine within the roots but had little effect on the terpenoid indole alkaloids (11). Similar expression of the ASR subunit under the control of a glucocorticoid- inducible promoter system along with the constitutive expression of the ASß subunit from Arabidopsis (14) in
C. roseus hairy roots caused a significant increase in the amount of tryptophan and tryptamine within the roots and caused a decrease in lochnericine, ho¨rhammericine, and tabersonine (15). To determine whether terpenoid indole alkaloid levels may be transiently affected by AS overexpression, we measured metabolite levels in late exponential growth phase C. roseus hairy roots over- expressing a feedback-resistant ASR subunit and an ASß subunit from Arabidopsis for time periods ranging from 4 to 72 h.
Figure 1. Biosynthesis of terpenoid indole alkaloids in Catharanthus roseus. DXS, 1-deoxy-D-xylulose 5-phosphate synthase; DXR, 1-deoxy-D-xylulose 5-phosphate reductoisomerase; GPPS, geranyl pyrophosphate synthase; G10H, geraniol 10-hydroxylase; AS, anthranilate synthase; TDC, tryptophan decarboxylase; STR, strictosidine synthase; SGD, strictosidine ß-D-glucosidase; T16H, tabersonine 16-hydroxylase; D4H, desacetoxyvindoline 4-hydroxylase; DAT, deacetylvindoline acetyltransferase.
Materials and Methods
Chemicals, Culture Conditions, and Hairy Root Lines. Dexamethasone (Sigma) was dissolved in ethanol and added to the culture media. Serpentine (Sigma), ajmalicine (Sigma), catharanthine (Qventas), and taber- sonine (a gift from Dr. Hamada, Okayama University, Japan) were dissolved in methanol and used as HPLC standards. The culture media for the hairy roots consisted of a filter-sterilized (0.22 ym filter) solution of 30 g/L sucrose (Sigma), half-strength Gamborg’s B5 salts (Sigma), and full-strength Gamborg’s vitamins (Sigma) with the pH adjusted to 5.7. Hairy root cultures were initiated by placing five root tips in 50 mL of the culture media in a 250-mL Erlenmeyer flask. These cultures were grown in the dark at 26 °C at 100 rpm. Roots were subcultured at 21 day intervals, by cutting off 5 root tips and transfer- ring them to 50 mL of fresh media. The generation of ASAB-1 has been previously described (15). This hairy root line expresses a feedback-resistant ASR subunit from Arabidopsis under the control of a glucocorticoid-induc- ible promoter and an ASß subunit from Arabidopsis under the control of the constitutive CaMV 35S pro- moter.
Transient Study. ASAB-1 hairy root cultures were treated with 0.2 yM dexamethasone (induced cultures) or with an equal amount of ethanol (uninduced cultures) as a negative control on day 18 of the growth cycle. ASAB-1 transitions into stationary phase around day 23 of the growth cycle (data not shown). Three induced and three uninduced cultures were randomly harvested at 4, 8, 12, 24, 36, 48, 60, and 72 h after being fed with dexamethasone or ethanol.
Alkaloid Extraction. The fresh weight of the har- vested hairy root cultures was measured after the cultures were blotted to remove excess media. The cultures were then frozen at -80 °C. The dry weight was measured after lyophilization. Approximately 50 mg dry weight of the ground tissues was extracted with 10 mL of MeOH in a sonicating bath for 1 h at 15 °C. The extracts were then clarified by centrifugation at 1,300g for 15 min at 15 °C. The supernatant was removed and the cellular debris was re-extracted in the same manner. The supernatants from both extracts were combined, concentrated to 2 mL using a vortex evaporator, and passed through a 0.22 ym nylon filter (13 mm). The media was not tested for alkaloid production because it has been previously shown that the terpenoid indole alkaloids were not excreted into the media (3).
HPLC Analysis. Ten microliters of the alkaloid ex- tract was injected onto a Phenomenex (Torrance, CA) Luna 5A C18(2) HPLC column (250 mm × 4.6 mm) under two different solvent systems. The HPLC system used was a Thermo Separations Products (San Jose, CA) Spectrasystem HPLC consisting of a P4000 pump, an AS3000 autosampler, and an UV2000 detector. To detect tryptophan and tryptamine, a protocol was adapted from
the literature with UV detection at 218 nm (11, 16). For the first 12 min, a 15:85 mixture of MeCN:100 mM phosphoric acid (pH 2) was maintained at a flow rate of 1 mL/min. The column was then washed with an 85:15 mixture of the same chemicals for 15 min and re- equilibrated. Another solvent system was used to detect the indole alkaloids ajmalicine, serpentine, catharan- thine, ho¨rhammericine, lochnericine, and tabersonine. Ajmalicine (retention time of 32.2 min), serpentine (16.9 min), and catharanthine (30.6 min) were detected at 254 nm and were quantified by comparison with authentic standard curves. Tabersonine (45.1 min), ho¨rhammeri- cine (22.9 min), and lochnericine (34.6 min) were detected at 329 nm by retention time of standards and were quantified on the basis of a tabersonine standard curve as previously reported (11, 17). For the first 5 min, the mobile phase consisted of a 30:70 mixture of 50% MeOH/ 50% MeCN:5mM (NH4)2HPO4 (pH 7.3) at a flow rate of 1 mL/min. Over the next 10 min, the mobile phase was linearly ramped to a 64:36 mixture, where it was maintained for 15 min at a flow rate of 1 mL/min. Over the next 5 min, the flow rate was linearly increased to 1.4 mL/min. During the next 5 min, the mobile phase ratio was increased to 80:20 where it was maintained for 15 min at a flow rate of 1.4 mL/min. The ratio was then brought back to 30:70 at 1 mL/min and reequilibrated (11). Statistical Analysis. Data were analyzed using the Student’s t-test.
Results and Discussion
Tracking the dynamic changes in terpenoid indole alkaloids over a period of time after gene induction can shed light on the metabolic fluxes within these pathways. In this study we use a previously generated C. roseus hairy root line (15) that carries an Arabidopsis gene encoding a feedback-resistant ASR subunit (12) expressed under the control of a glucocorticoid-inducible promoter (13) and an Arabidopsis gene encoding an ASß subunit (14) expressed under the control of the constitutive CaMV 35S promoter. This hairy root line thus has constitutively high levels of ASß expression but has high levels of expression of the feedback-resistant ASR subunit only when grown in the presence of a glucocorticoid, such as dexamethasone. This hairy root line therefore permits analysis of the specific effects of altering ASR activity in a background of high ASß activity.
On the basis of a previous feeding study of C. roseus hairy roots that showed feeding tryptophan in the late exponential growth phase significantly increased the accumulation of tabersonine and serpentine (6), hairy root cultures in the late exponential growth phase were harvested 4, 8, 12, 24, 36, 48, 60, and 72 h after addition of 0.2 yM dexamethasone (induced cultures) or an equal volume of ethanol (uninduced cultures). As predicted from previously reported results (11, 15), tryptophan and tryptamine pools increase over the 72 h period in the induced cultures, whereas these pools remain at low and constant levels in the uninduced cultures (Figure 2). The concentration of tryptophan increases rapidly in the first 48 h from a basal level of 80 yg/g dry weight (DW) to 2240 yg/g DW, a 28-fold increase. The tryptophan con- centration is still increasing after 48 h, though at a slower rate, reaching a level of 2660 yg/g DW by 72 h, a 33-fold increase. The most rapid increase in the concentration of tryptamine occurs in the first 24 h, reaching a concentration of 375 yg/g DW, a 3-fold increase over the basal levels. The concentration of tryptamine continues to increase to a level of 700 yg/g DW at 72 h, a 5.4-fold increase over basal levels. The increases in tryptophan and tryptamine pools suggest that bottlenecks occur in reactions downstream of these compounds or that the supply of secologanin is limiting (i.e., the flux through the terpenoid pathway is limiting), both of which could cause the buildup of these pools of metabolites.
The terpenoid indole alkaloids are difficult to analyze since many of these compounds have not been well characterized. We quantified the levels of six terpenoid indole alkaloids to evaluate the metabolic effects of ASRß overexpression. These alkaloids are ajmalicine and ser- pentine from the corynanthe family, catharanthine from the iboga family, and tabersonine, ho¨rhammericine, and lochnericine from the aspidosperma family. Overexpres- sion of ASRß has little effect on the serpentine and catharanthine pools over the 72 h induction period (Figure 2). The ajmalicine pool in the induced cultures increases over time with significant increases occurring in the first 60 h. At this time point the ajmalicine concentration reached 920 yg/g DW, about twice that of the uninduced culture (Figure 2).
Ho¨rhammericine, lochnericine, and tabersonine con- centrations in the induced cultures decrease over time (Figure 2). The ho¨rhammericine concentration in the induced cultures does not vary significantly from the uninduced cultures in the first 12 h after induction. Between 12 and 36 h, there is a 50% decrease in ho¨rhammericine levels, from 650 to 320 yg/g DW, in the induced cultures when compared to the basal level. After 36 h ho¨rhammericine levels remain relatively constant around 300 yg/g DW. Similarly, the concentration of lochnericine in the induced cultures does not vary significantly in the first 12 h after induction. Between 12 and 48 h, there is a 2.3-fold decrease in lochnericine levels in the induced cultures, from 700 to 300 yg/g DW, when compared to the basal level. After 48 h there is still a slow decrease in lochnericine to 185 yg/g DW in the induced cultures. The tabersonine concentration in the induced cultures starts to drop after 4 h from 700 to 91 yg/g DW at 48 h, a 7-fold decrease from basal levels. After 48 h the tabersonine concentration remains approxi-
mately constant.
The accumulation of the corynanthe family of alkaloids consisting of ajmalicine and serpentine shows a trend similar to that of ajmalicine alone; the induced cultures show an increase over the first 60 h to a level of 2345 yg/g DW, a 35% increase over basal levels (Figure 3). The pools of the aspidosperma alkaloids consisting of loch- nericine, ho¨rhammericine, and tabersonine decrease 64% from 4 to 48 h, 2020 to 740 yg/g DW, in the induced cultures before leveling off. The total alkaloid pools remain relatively constant, with the induced cultures showing approximately the same or slightly lower levels of total alkaloids than the uninduced cultures.
Given the increase in tryptophan and tryptamine levels, the increase in the corynanthe family of alkaloids is not surprising. In contrast, the decrease in aspido- sperma family alkaloids is less expected but might be explained by a couple of possibilities. First, these alka- loids could be converted to other alkaloids in the pathway that are uncharacterized. Their expression could be regulated by an increase in certain metabolites. For example, the increase in ajmalicine might regulate the production of the aspidosperma alkaloids in order to maintain a relatively constant pool of terpenoid indole alkaloids. These alkaloids could also be experiencing degradation by rapid catabolism within the cell caused by factors such as metabolic stress within the tissue. Such metabolic stress could explain the browning of the C. roseus hairy roots seen upon induction of ASRß and could be caused by the large pools of tryptophan that might trigger the stress response. This observation is consistent with results from another transgenic hairy root line in our lab that increases expression of 1-deoxy- D-xylulose-5-phosphate synthase (DXS) under the control of the glucocorticoid-inducible promoter. When DXS is overexpressed within C. roseus hairy roots, we observe browning plus a decrease in the aspidosperma alkaloids (data not shown). For the DXS line, the browning/stress response may be caused by a drain of the central metabolism metabolites pyruvate and glyceraldyhyde-3- phosphate.
It is also important to note that the induced pools of any of the compounds measured do not start to diverge from the uninduced samples for 4-12 h after induction. This is a good reminder that changes in metabolism do not happen immediately within the hairy roots and may not stabilize for a couple of days. This is consistent with a number of studies of inducible promoter systems used within plants (18-21).
The results of this experiment along with other experi- ments can aid in the development of metabolic flux maps of the terpenoid indole alkaloids within C. roseus. The development of such a map would be useful in determin- ing the potential effects of changes to the network and thus could provide insight into the metabolic engineering of the terpenoid indole alkaloid pathway.