Index | Search | Home | Table of Contents

Ratnayake, S., J.K. Rupprecht, W.M. Potter, and J.L. McLaughlin. 1993. Evaluation of the pawpaw tree, Asimina triloba (Annonaceae), as a commercial source of the pesticidal annonaceous acetogenins. p. 644-648. In: J. Janick and J.E. Simon (eds.), New crops. Wiley, New York.

Evaluation of the Pawpaw Tree, Asimina triloba (Annonaceae), as a Commercial Source of the Pesticidal Annonaceous Acetogenins*

Sunil Ratnayake, J. Kent Rupprecht, William M. Potter, and Jerry L. McLaughlin

  4. Table 1
  5. Table 2
  6. Fig. 1
  7. Fig. 2
  8. Fig. 3
  9. Fig. 4

The potent antitumor, pesticidal and/or insect antifeedant properties of the Annonaceous acetogenins have been previously reported (Rupprecht et al. 1986; Hui et al. 1989; Alkofahi et al. 1989) and patented (Mikolajczak et al. 1988, 1989). A recent compilation reviews the chemistry and biological actions, known to date, of this new class of diversely bioactive botanical compounds (Rupprecht et al. 1990). The Annonaceae is almost exclusively a tropical plant family and encompasses over 2000 species (Heywood 1978). However, the pawpaw (Asimina triloba Dunal) is a temperate representative and is an abundant native of eastern North America. Work on the seeds and stem bark of the pawpaw initially revealed pesticidal actions of the acetogenins and focused on asimicin (1), a major bioactive component of a complex mixture of these compounds (Rupprecht et al. 1986; Alkofahi et al. 1989).

Asimicin (1) (Fig. 1) was originally isolated from a 90% aqueous methanol partition fraction (Fig. 2) (F005) from the 95% ethanol extract of the stem bark of pawpaw (Mikolajczak et al. 1988). F005 is identical to the F020 as described in the initial patent of Mikolajczak et al. (1988). It was isolated via bioactivity-directed fractionation using a simple test involving lethality to brine shrimp larvae (Artemia salina Leach) (BST). Promising pesticidal activities paralleled the BST throughout the fractionation (Alkofahi et al. 1989) and included significant activity against blowfly larvae (Colliphora vicina Meig), two-spotted spider mite (Tetranychus urticae Koch), melon aphid (Aphis gossyphii Glover), mosquito larvae (Aedes aegypti Linnaeus), Mexican bean beetle (Epilachna varivestis Mulsant), striped cucumber beetle (Acalymma vittatum F.), and a free-living nematode [Caenorhabditis elegans (Maupas) Dougherty]. Subsequent testing of F005 has shown promising activities against a host of additional pests, especially those of horticultural concern.

Asimicin (1) contains eight chiral centers (256 possible stereoisomers) and, thus, is not a good candidate for commercial chemical synthesis. However, plant extracts, such as F005, containing a mixture of the bioactive acetogenins can be quickly and inexpensively prepared. As is true with the pyrethrins, such botanical mixtures often increase the pesticidal spectrum and are less likely to induce pest resistance after repeated application. Therefore, a plant extract, such as F005, would be a logical means for incorporation of the acetogenins into a pesticidal product. F005, obtained from the stem bark of pawpaw and suspended with 2% aqueous Tween 80 in the concentration range of 0.05 to 0.5% (500 to 5000 ppm), provides effective pesticidal action and plant protection in laboratory and field tests. Fig. 2 illustrates the protective effect of F005 against foliar damage on bush beans infested with bean leaf beetles (Cerotoma trifurcata Forster).

Asimina triloba is quite common as an understory shrub and small tree throughout the eastern United States, and the edible fruits, called "pawpaw" (Fig. 3), have generated considerable interest for commercial production (Callaway 1990). Over 18 million trees (3 to 13 cm diam) are growing in Indiana (Spencer et al. 1990), and other eastern states have similarly high populations. Thus, an evaluation of the relative biological potencies and percentage yields of standardized extracts from the various plant parts is needed. The development of such new, environmentally compatible, biologically degradable, botanically-derived, pesticides has become more desirable (Arnason et al. 1989).


The plant material was collected in Indiana during July 1988 and October 1990 with the exception of the seeds which were collected in November 1989 from plantings at the University of Maryland, established in 1982 by R. Neal Peterson. The Indiana collections came from a single clone at the Horticultural Research Farm, Purdue University, and were identified by George R. Parker, Professor of Forestry. The unripe fruits were frozen and freeze-dried, while the seeds were air dried at room temperature. The other plant materials were dried in an oven at 40°C, and all of the materials were ground in a Wiley mill (unless otherwise stated, 4 mm mesh size was used in all cases). The woody collections were chipped and shredded (Mantis, Chipmate) prior to drying.

The dried plant materials (100 g) were then repeatedly extracted (Fig. 4) with 95% ethanol (5 x 750 ml). The combined extract was evaporated under reduced pressure (below 40°C) to provide a syrupy residue (F001). F001 was transferred to a 250 ml separatory funnel with a mixture of 100 ml of water and 100 ml of chloroform. After removal of the chloroform layer, the water layer was then extracted seven additional times with 100 ml portions of chloroform, and the combined layers were reduced under rotary vacuum (<40°C) to obtain the chloroform residue (F003). F003 was further partitioned between hexane and 10% water in methanol (1:1, 4 x 100 ml), and the solvents were evaporated under reduced pressure to afford the hexane solubles (F006) and the 90% aqueous methanol solubles (F005). The yield of F005 in grams from 100 g of plant material then directly corresponded to the percentage yields from the various plant parts.

The brine shrimp lethality bioassay (BST) (Meyer et al. 1982) was performed with the aqueous methanol soluble fractions (F005) as follows: 20 mg of each fraction was dissolved in 2 ml of methanol and 5, 50, and 500 µl amounts were transferred to 2 dram vials to correspond to 10, 100 and 1000 ppm concentrations; three vials were prepared for each concentration. These vials were dried overnight to permit evaporation of the methanol, and control vials were prepared using 500 µl of methanol alone. Ten brine shrimp larvae, taken 48 h after initiation of hatching in 3.8% aqueous sea salt brine (Instant Ocean, Metaframe) were added to each vial, and the final volume of each vial was adjusted to 5 ml using the artificial sea water. After 24 h, survivors were counted, and LC50 values with 95% confidence intervals were computed using a Finney's probit analysis program adapted to an IBM personal computer (discs of the program are available from J.L. McLaughlin).


To access the pesticidal potential of the available biomass of A. triloba, various plant parts were collected and extracted in a standardized way (Fig. 4), and the resulting 90% aqueous methanol fractions (F005) were bioassayed in the BST. The percentage of F005 obtained from each plant part was also calculated, and this yield was correlated with the BST data to determine which plant part would yield the greatest quantity of an extract that could be suitably active for incorporation into a commercial product.

Table 1 shows the brine shrimp lethality of F005 for the various plant parts extracted and tested. The twigs, unripe fruit, seeds, root wood, and all the stem bark samples were the most active plant parts with LC50 values ranging from 0.042 to 0.104 ppm. These differences in activity, however, were not significant since the BST 95% confidence intervals all overlap. The BST activity of the root bark is very close to the above plant parts and is probably equivalent in activity even though there is a slight difference in the 95% confidence interval overlap between the root bark and the unripe fruit and root wood (D 0.01 ppm). The whole above-ground plant (leaves, stem, wood, and stem bark) and stems consisting of wood and bark were approximately 1/8 to 1/2 as active as the above plant parts. The BST of the wood and of the leaves were between 40 to 100 times and 500 to 1,200 times, respectively, less active than the most active plant parts. Furthermore, there were no significant differences in the bioactivities of the stem bark collected in July 1988 and October 1990, indicating that collections of bark made during the summer and fall are biologically equivalent and that the pesticidal constituents are stable over some period of time when the dried bark is stored.

The percentages of 90% aqueous methanol extractables (F005) of each plant part tested were also determined (Table 1). The unripe fruit, leaves, and seeds gave the highest percentages of F005. They would appear to be suitable sources of F005; however, the F005 from the leaves is the least active of the plant parts tested, and therefore, the leaves are a poor candidate for use. The extracts of the unripe fruit and seeds are extremely bioactive and also give a very high yield of F005; unfortunately, these plant parts represent the least available forms of biomass and are therefore unsuitable for commercial development. The root bark and stem bark afforded 1.96 to 2.72% yields into the 90% aqueous methanol extractables, and both plant parts possess potent brine shrimp activity. However, uprooting of the trees would destroy the stands and would require replanting on a commercial scale. When the stem bark was ground more finely (2 mm versus 4 mm sieves) the percentage of F005 increased by approximately 20%. The amount of 90% aqueous methanol extractables (F005) of the stem wood and bark, whole (above-ground) plant, root wood, and stem wood were 1.61, 1.60, 1.27, and 0.90%, respectively. The plant part which represents the best balance of yield and biological activity to its availability as a source of biomass is either the stem bark or the stem wood and stem bark in combination. Stripping of the bark is quite labor intensive; thus, mechanically chipping and shredding of the stem would be most practical.

To assess further the pesticidal potential of the combined stem wood and stem bark biomass, stems of various diameters were collected, chipped and shredded, dried, and pulverized, and standardized extracts were prepared; the BST was made on the F005 extract from stems of the different diameters. The percentage of F005 obtained and the brine shrimp activity were both inversely proportional to the stem diameter (Table 2). The activity of the smallest stems, 64 mm or less in diameter, was significantly greater than the other diameters tested.

Asimicin (1) and several other acetogenin compounds play the major role in the pesticidal activity. Asimicin alone gives BST LC50 values at 0.03 ppm (Rupprecht et al. 1986), whereas its stereoisomer, bullatacin gives LC50 values at 0.00159 ppm (Hui et al. 1989). In addition to asimicin, bullatacin, bullatacinone, and trilobacin are all potent acetogenins recently isolated from F005 of pawpaw (Zhao et al. 1992). The variations of observed toxicity are probably due to quantitative variations of these and additional toxic acetogenins as distributed in different parts of the plant. The amounts of these compounds may also vary seasonally, but this variable has not been studied.

It is probable that the toxicities are higher in the twigs, unripe fruit, seeds, and bark because these are the plant parts which can be most easily damaged by herbivores, environmental pests and pathogens, and, therefore, for the purpose of defense, higher concentrations of the protective acetogenins may accumulate in these parts. The root (both bark and wood) may accumulate these compounds to avoid attack by nematodes and other soil pathogens.

In conclusion, the bioactivity of the A. triloba tree is concentrated mainly in the twigs, unripe fruits, seeds, root, and bark, but stems, especially those with the smallest diameters, are also significantly toxic. Therefore, the smaller aerial parts, minus the leaves, could be used as a biomass for the pesticidal extract. Finer milling would give more solubles while maintaining the level of the bioactivity. The stem bark collected in July showed more activity which, however, was not significantly different from the values of F005 obtained from the batch collected in October from the same clone. It is possible that further screenings of wild populations of pawpaw could identify some clones with acceptably high levels of bioactivity in the leaves. The brine shrimp toxicity bioassay can be conveniently used as a rapid quality control measure of the total bioactivity of mixtures of the acetogenin compounds. Thus, the commercially used extract could be easily standardized as to bioactivity in furnishing a consistently bioactive product. A recent study with asimicin and F005 in the guinea pig maximization test showed only weak skin sensitization and suggested that few or no dermatologic problems are expected as a result of the pesticide use and human dermal contact (Avalos et al. 1992). Furthermore, in the Ames test, F005 was negative in nine of ten determinations with five histidine mutants of Salmonella typhimurium (Loeffler), with and without hepatic enzyme activation; such results are superior to many common substances, such as caffeine, and suggest that F005 is not a serious mutagen (P.E. Kirby pers. commun.). Mode of action studies in three separate laboratories have recently determined that F005, asimicin, and bullatacin are superb inhibitors (at subnanomolar concentrations) of Complex I in mitochondrial electron transport systems from several organisms; this is very similar to the site of action of rotenone, an established and approved botanical pesticide (J.T. Arnason; R. Hollingworth, C.E. Snipes pers. commun.).


*This work was supported by NIH/NCI RO1 grant no. CA 30909.
Table 1. Brine shrimp lethality and percentage yields of F005 (90% aqueous methanol fraction) from various plant parts of Asimina triloba.

Plant part extracted LC50 (ppm)z Yield of F005 (%)
Twigsy 0.042 (0.02-0.09) 1.78
Unripe fruitx 0.060 (0.03-0.08) 5.11
Root woodw 0.060 (0.03-0.08) 1.27
Seedv 0.065 (0.03-0.10) 4.03
Stem barkx,u 0.077 (0.04-0.12) 2.71
Stem barky,u 0.102 (0.06-0.15) 1.96
Stem barky 0.104 (0.03-0.20) 2.52
Root barkw 0.135 (0.09-0.21) 2.72
Whole stem: wood, bark and leavesy 0.202 (0.14-0.31) 1.60
Stem woody 4.86 (0.37-11.93) 0.90
Leavesy 53.6 (33-82) 4.36
zWith 95% confidence intervals in parenthesis.
yCollected in October 1990.
xCollected in July 1988.
wCollected in November 1990.
vCollected in November 1989.
uGround to a mesh size of 2 mm, all other samples were ground to a 4 mm mesh size.

Table 2. Variation of the percentage F005 (90% aqueous methanol solubles and brine shrimp lethality with differing stem diameters.

Stem diameter of extracted material (mm) Percentage of F005 LC50 (ppm)z
251-5001.36 0.36 (0.23-0.55)
189-2501.45 0.22 (0.13-0.34)
65-188 1.76 0.22 (0.13-0.34)
0-64 (twigs) 1.78 0.04 (0.01-0.09)
zWith 95% confidence intervals in parentheses.

Fig. 1. Structure of asimicin.

Fig. 2. Scheme for extraction and partitioning of Asimina triloba.
Fig. 3. Bush beans (cv. Blue Lake). The row on the left has a natural infestation with bean leaf beetles (Cerotoma trifurcata Forster). The row on the right was sprayed three times during a ten day period with 0.5% of F005 (F020), suspended with 2% Tween 80, in water. F005 (F020) is prepared from the bark of the pawpaw tree (Asimina triloba). Note the protective effect of the Annonaceous acetogenins in a field test against garden insects.

Fig. 4. Fruits and seeds of pawpaw (Asimina triloba).

Last update May 16, 1997 aw