Journal of Chemical Ecology, Vol. 21, No.

4, 1995

ANTITERMITIC ACTIVITY OF Lonchocarpus castilloi FLAVONOIDS AND HEARTWOOD EXTRACTS

R. REYES-CHILPA, *'l N. VIVEROS-RODRIGUEZ, z F. GOMEZ-GARIBAY, 2 and D. A L A V E Z - S O L A N O2

tlnstituto de Ecologia A. C. Apartado Postal 63 Xalapa, Veracruz, 91000, M~xico 71nstituto de Quhnica, Universidad National Aut6noma de MFxico Ciudad Universitaria Coyoacdn, 04510 MFxico, D.F.
(Received September 23, 1994; accepted December 30, 1994)

Abstract--The heartwood of the tropical tree Lonchocarpus castilloi Standley (Leguminosae) is highly resistant to attack by the drywood termites Cryptotermes brevis (Walker); nevertheless successive extraction with hexane, diethyl ether, acetone, methanol, and water reduced its resistance to these organisms. Antitermitic properties of the extracts were bioassayed using impregnated filter paper disks. Although the five extracts reduced both feeding and survival of C. brevis, no significant differences among them were detected. Choice feeding tests showed that termites avoided eating the paper treated with the extracts. Two flavonoid compounds isolated from the heartwood, eastillen D and castillen E, impregnated into filter paper showed concentration-dependent feeding deterrent activity, but were not toxic to C. brevis. Key Words--Feeding deterrents, termites, Cryptotermes brevis, lsoptera, Lonchocarpus castilloi, Leguminosae, heartwood, flavonoids, antitermitic activity, castillen D, castillen E.

INTRODUCTION

Lonchocarpus castilloi Standley (Leguminosae) is a tree indigenous to the tropical subhumid forest in southern Mexico, Central America, and northern South America. It can reach 30 m in height and 40 cm diameter at breast height (Ortega-Escalona et al., 1991). In Mexico it is known as "machiche." Locally,
*To whom correspondence should be addressed.

455
0098-0331/95/0400-0455507.50/0 1995 PlenumPublishingConporation

456

REYES-CHILPA ET AL,

the timber is used in bridges, housing, rural wagons, and for manufacturing veneer, plywood, staves, and parquet. It is also recommended for railroad crossties, stakes, fenceposts, and post cross-arms (Echenique-Manrique, 1970). Lonchocarpus castilloi heartwood has been reported to be very durable in accelerated field tests (B~ircenas-Pazos, 1995); it is also highly resistant to several wood-rotting fungi (Reyes-Chilpa et al., 1987) and to the drywood termite C~ptotermes brevis (Walker) (P6rez-Morales et al., 1984). Current evidence indicates that natural resistance to fungi (Laks, 1988; D6on et al., 1980; Hart and Shrimpton, 1979) and termites (Scheffrahn, 1991; McDaniel, 1992) is primarily determined by the concentration and type of secondary compounds found in the heartwood. In this context, it has been shown that the natural decay resistance of Lonchocarpus castilloi heartwood is also chemically determined. Extraction of the heartwood with a series of organic solvents reduced its decay resistance (Reyes-Chilpa et al., 1987), and the extracts showed antifungal activity in vitro (Reyes-Chilpa, 1986). Chromatographic separation of the hexane extract afforded four 2-methoxyfuran auronols named castillens A, B, C, and D, as well as one a-hydroxy-dihydro-chalcone, castillen E (G6mez-Garibay et al., 1990). These flavonoid compounds (Figure 1), inhibited the mycelial growth of the fungus Lenzites trabea Pets. ex Ft. The most active compound was the castillen E, followed by castillen D (G6mez-Garibay et al., 1990). We now have examined the effect of L. castilloi heartwood extracts and of

II u , ~

6'

s'

0 1 .- Castillen A: R1 = Rz - H 2.- Castillen D: R1 + Rz = OCHzO

0Me 3.- Castillen B: R1 = R - H 2 4.- Castillen C: Rt + Rz = OCHzO

o oHLO"
S , ~ o 0 S.- Castillen
E

H ~"

FIG. l. Flavonoids from Lonchocarpus castilloi heartwood. Castillen A (1), Castillen D (2), Castillen B (3), Castillen C (4), and Castillen E (5).

ANTITERMIFIC ACTIVITY OF

L. castilloi

457

castillens D (2) and E (5) on the feeding and survival of C. brevis. We also report the isolation of two additional compounds, pongaglabrone and 4-hydroxybenzofuran, from the heartwood.

METHODS AND MATERIALS

Lonchocarpus castilloi was collected at the "Lacandona" tropical rain forest, Chiapas, Mrxico (B~rcenas-Pazos et al., 1980). Wood samples and voucher specimens are deposited at the Ecology Institute's Herbarium (Xal) in Xalapa, Mexico. Termite Collection. Cry.ptotermes brevis termites were obtained from old beams in Xalapa City. The beams were sawed and stored under controlled conditions (23C, 75% relative humidity) until employment, and bioassays also were conducted under these conditions. Beam pieces were split on a paper; termites were taken up with entomological tweezers. Only active pseudergates above the third instar were used for bioassays. Selection of active specimens was done as follows. A group of 40 pseudergates was placed inside a 50-ml beaker, and that beaker was placed perpendicular to another. The first 30 termites that passed to the second beaker were chosen (Prrez-Morales et al., 1984). Wood Block Bioassays. Sapwood and heartwood blocks (2 x 1 z 0.5 cm, smallest dimension being parallel to the fiber) were cut from a board. The blocks were carefully sand-papered, oven dried (105C, 24 hr), and weighed. Heartwood blocks were successively extracted in a Soxhlet apparatus with hexane, diethyl ether, acetone, methanol, and water. After each extraction step (10 hr), an equal number of blocks were removed. Extractive yields were calculated as the percentage of oven-dried weight loss (Table 1). Sapwood and heartwood (extracted and unextracted) blocks were tested against C. brevis. Each block was placed in a glass vial with 30 pseudergates. The vials were capped with filter paper. Each treatment was replicated five times. Mortality was recorded periodically, and the wood consumption was determined at the end of the test period (seven weeks) as the percentage of ovendried weight loss. Filter Paper Bioassays. Lonchocarpus castilloi heartwood shavings (500 g) were extracted under reflux with the same sequence of solvents described previously. The extracts were concentrated in a rotary evaporator and freezedried when the solvent was water. Solutions (2.5% w/v) of each extract were prepared in order to impregnate filter paper disks. This concentration was chosen since 2% was the mean extractive yield from blocks. Hexane and diethyl ether extracts were dissolved in a mixture of hexane-ethyl acetate 1 : 1 (HE), and acetone and methanol extracts were dissolved with a mixture of acetone-meth-

458

REYES-CHILPA ET AL.

anol 70:30 (AM). Each solution was poured into a Petri dish, and filter paper disks (Whatman No. 1, diameter = 7 cm) were dipped for 30 sec and then air dried. Impregnated disks were then placed in a hood for 12 hr. Control disks were treated with the solvent mixtures only. Two bioassays were run with the impregnated filter paper: (1) For force feeding, a disk was placed into a Petri dish and 15 termites were added. (2) For the choice test, one disk of each treatment, including the AM control, was sectioned into six equal slices. A slice of each treatment was then fixed to the bottom of a Petri dish (Figure 2) with a bit of "Blue Tac" plaster (Williams, 1980) and 15 termites were added. Five replicates per test were run. Mortality was recorded periodically. Paper consumption was quantified at the end of the test (36 days). Isolation of Compounds. Heartwood shavings (650 g) were extracted at room temperature with hexane and then with methylene chloride. After concentration, the methylene chloride extract was chromatographed over silica gel. The compounds were obtained by preparative TLC and identified by ~H NMR, ~3C NMR, IR, UV, and MS. Castillens D and E were tested using impregnated filter paper and following the force-feeding protocol. The solutions tested were 0.025, 0.05, 0.1, and 0.15% w/v with acetone as solvent. Control paper disks were treated with the solvent only.

FIG. 2. Choice test. Feeding of termites on filter paper impregnated with hexane (hx), diethyl ether (et), acetone (ac), methanol (me), or water (ag) extracts of Lonchocarpus castilloi heartwood. Control disks (c) treated only with a mixture of acetone-methanol (7:3).

ANTITERMITIC ACTIVITY OF L. castilloi

459

RESULTS

Extracted Wood. Cryptotermes brevis pseudergates ate over six times more of Lonchocarpus castilloi fully extracted heartwood blocks than of the sound heartwood (Table 1). The most efficient solvent that rendered heartwood more susceptible was methanol and then water. Heartwood blocks successively extracted with hexane, diethyl ether, and acetone were eaten equally when compared with sound heartwood. Termites consumed 1.7 times more of extracted heartwood blocks than sapwood. Sapwood seemed also to be eaten more than sound heartwood, but the difference was not significant when examined by A N O V A (Table 1); nevertheless, it was significant (P = 0.001) by Student's t test. At the end of the test period, termite survival was similar among all treatments (Table 1). Effects of Extracts. All the heartwood extracts incorporated into filter paper reduced both the paper consumption and the survival of termites (Table 2). No significant differences among the extracts were detected. On average, there was a fivefold reduction in consumption when the treatments were compared with the control; there was also a twofold reduction in termite survival. The above data suggested that all the extracts contained antitermitic compounds. In the choice feeding test, pseudergates consumed only the control and avoided all the filter paper slices treated with the heartwood extracts (Figure 2). Survival in this test was similar to that depicted for controls in the force-feeding experiment (see Table 2). Compounds Isolated. T L C of the hexane extract of L. castilloi heartwood showed the presence of castillens A, B, C, D, and E as compared with authentic

TABLE 1. FEEDING AND SURVIVAL OF

C~_ptotermes brevis PSEUDERGATES

ON L.

castilloi WOOD AFTER47-DAY EXPOSURE"

Extractive yields (% w/w) Consumption (% weight loss) 1.77a 0.52a 2.79 0.36 2.28 3.84 0.76 0.61 a 0.87a 0.40a 2.17b 3.16b

Treatment Sapwood Heartwood Extracted heartwood Hexane Ethe~ Acetone Methanol WaterJ'

Survivors 9.0a 7.8a 6.0a 6.2a 5.8a 6.6a 11.5a

"Means (N = 5") using 30 termites per replicate. Numbers followed by same letter are not significantly different at P > 0.05 (Duncan's NMRT"). hN = 4.

460

REYES-CHILPA ET AL.

TABLE 2. FEEDING AND SURVIVAL OF C~.ptotermes brevis PSEUDERGATES ON PAPER IMPREGNATED WITH HEARTWOOD EXTRACTS OF L. castilloi AFTER 36-DAY EXPOSURE"

Treatment Control HA Control AM Hexane Ether Acetone Methanol Water

Consumption (mg) 13.7a 14. la 1.8b 3.2b 2.6b 4.5b 3. Ib

Survivors 10.4a 9.8a 4,4b 4.4b 3.8b 4.4b 4.2b

"Means (N = 5) using 15 termites per replicate. Numbers followed by same letter are not significantly different at P > 0,05 (Duncan's NMRT).

samples. The methylene chloride extract rendered castillen D (25 mg), castillen E (49 mg), pongaglabrone (6 mg), and 4-hydroxy-benzofuran (13 mg) by column chomatography and preparative TLC. Castillens D and E were then crystallized, and the spectral data from the purified compounds agreed with those previously reported (G6mez-Garibay et al., 1990). Castillen D. Colorless solid, mp 121C. 13C NMR (CDCI3, 50 MHz): 197.0 [C = 0], 166.5 [C-6], 162.2 [C-la], 147.1 [C-3'], 146.4 [C-4'], 145.6 [C-o~], 126.4 [C-I'], 123.8 [C-4], 119.7 [C-6'], 115.2 [C-3a], 113.3 [C-7], 110.9 [C-5'], 110.1 [C-2], 107.7 [C-2'], 107.2 [C-/3], 103.7 [C-5], 100.7 [OCH20], 52.5 [OCH3] , 41.0 [Ar-CH2]. Castillen E. Colorless solid, crystallyzed from diethyl ether, mp 72-73 C. 13C NMR (CDCI3, 50 MHz): 204.5 [C = 0], 159.4 [C-4'], 144.8 [C-o~], 136 [C-2'], 130.8 [C-6'], 129.4 [C-3], 128.8 [C-5], 128.4 [C-2], 126.9 [C-6], 125.5 [C-4], 118.1 [C-l], 112.3 [C-I'], 111.1 [C-3'], 105.1 [C-/3], 104.5 [C-5'], 77.8 [C-~'l, 42.9 [C-/3']. Pongaglabrone (Khanna and Seshadri, 1963) and 4-hydroxy-benzofuran were also obtained as solids and identified by their spectral data. Effects of Castillens D and E. Consumption by termites of filter paper impregnated with castillen D decreased as the concentration of this compound increased (Table 3). The highest concentration tested (0.15 % w/v) caused a 3.7fold reduction in the amount of paper eaten as compared to the control. Survival of termites was not affected by any concentration of castillen D tested. Similar results were obtained with castillen E (Table 4). In this case, the 0.15% treatment caused a 2.6-fold reduction in paper consumption. Survival of termites was not affected at any concentration tested. Since small amounts of castillen C, and a mixture of castillens A and B

ANTITERMITIC ACTIVITY OF

L. castilloi

461

TABLE 3. FEEDING AND SURVIVAL OF

C~ptotermes brevis

PSEUDERGATES ON PAPER

IMPREGNATED WITH CASTILLEN D AFTER 2 8 - D A Y EXPOSURE"

Treatment Control 0.025 % 0.050% 0.100% 0.150%

Consumption (rag)
9.4a 7.4ac 4. lbc 3. lb 2.5b

Survivors
14.8a 13,8a 13.2a 13.6a 14.4a

"Means (N = 5) using 15 termites per replicate. Numbers followed by the same letter are not significantly different at P > 0.05 (Duncan's NMRT),

TABLE 4. FEEDING AND SURVIVAL OF

Cryptotermes brevis PSEUDERGATES ON PAPER

IMPREGNATED WITH CASTILLEN E AFTER 2 8 - D A Y EXPOSURE"

Treatment Control 0.025% 0.050% O, 100% 0.150%

Consumption (rag)
14.6a 13.3a 16.2a 7.0b 5.5b

Survivors
11.8a 12.0a 13.2a 10.6a 11.0a

"Means (N = 5) using 15 termites per replicate. Numbers lbllowed by same letter are not significantly different at P > 0.05 (Duncan's NMRT),

were available, the substances were also tested at tl,e lower concentration (0.025 % w/v) following the force-feeding protocol. Neither paper consumption nor survival of termites was significantly different from controls.
DISCUSSION

The heartwood of Lonchocarpus castilloi contains antitermitic compounds to Cry.ptotermes brevis pseudergates. These compounds were extracted with four organic solvents and water as indicated by the force-feeding test with impregnated filter paper (Table 2). At the concentration tested, all the extracts were equally effective. Similar results have been reported for other tropical hardwoods (Carter and de Camargo, 1983; Carter et al., 1975). Feeding-deterrent activity, but not toxicity, was detected for the previously known fungistactic flavonoids (G6mez-Garibay et al., 1990), the castillens D and E (Tables 3 and 4). These

462

REYES-CHILPA ET AL.

compounds could be removed by hexane and other nonpolar solvents; nevertheless, reduction of the heartwood termite resistance was only noted after extraction with methanol and water (Table 1). The above results suggests that other compounds could be directly involved in L. castiUoi termite resistance or could play a synergistic role. Besides castillens D and E, other flavonoids have been previously tested against termites (Scheffrahn, 1991; Supriana, 1983). Several flavonoids were able to repell C. brevis attack when impregnated into a highly susceptible wood (Wolcott, 1953). Most of them showed poor activity, but taxifolin and pomiferin were considered outstanding. For instance, treatment with 0.05, 0.10, or 0.2% taxifolin solutions for 10 min was enough to protect susceptible wood for two years (Wolcott, 1953). Rotenone, a well-known insecticidal isoflavonoid, which can be isolated from the roots of several legume species, has also been found to be toxic and to deter the feeding of subterranean termites Coptotermes f o r mosanus Shiraki (K~e~kov~i et al., 1988). Afzelin also has been reported to be toxic to Heterotermes indicola Wasmann (Supriana, 1983). At present, however, flavonoids seem to be more involved in the resistance of wood to fungi than to termites. Antifungal properties of monomeric and polymeric wood flavonoids are documented (D6on et al., 1980; Hart and Hillis, 1972). On the other hand, antitermitic compounds identified from wood have been found to be predominantly terpenoids (from monoterpenes to triterpenes), aliphatic acids, naphthoand anthroquinones, simple quinones, stilbenes and, naphthopyrans (Scheffrahn, 1991). Therefore, in the case of L. castilloi heartwood resistance to C. brevis, it seems probable that besides the feeding deterrents castillen D and E, other active compounds, probably of a nonflavonoid nature, remain to be identified.

Acknowledgments--R.R.C. and N.V.R. thank Dr. Raymundo D~valos Sotelo for his support while carrying on this work.

REFERENCES
BARCENAS-PAZOS, G.M. 1995. Caracterizaci6n tecnol6gica de veinte especies maderables de Ia

Selva Lacandona. Madera y Bosques (Mexico) I ( 1):9-38. BARCENAS-PAZOS, G.M., REJON,R.R., and ECHENIQUE-MANRIQUE, 1980. Especies Maderables R. de la Selva Lacandona. La Madera y su Uso en la Construcci6n No. 6. INIREB, Xalapa, M6xico, 18 pp. CARTER,F.L., and OE CAMARGO,C.R.R. 1983. Testing antitermitic properties of Brazilian woods and their extracts. Wood Fib. Sci. 15(4):350-357, CARTER, F.L., BEAL, R.H.. and BULTMAN.J.D. 1975. Extraction of antitermitic substances from 23 tropical hardwoods. Wood Sci. 8(1):406-410. Dt~ON, G., CHADENSON,M., and HAUTEVILLE,M. 1980. Influence des extraits naturels du bois sur sa r6sistance a la pourriture. Bois For. Trop. 191:75-90.

ANTITERMITIC ACTIVITY OF L. castitloi

463

ECHENIQUE-MANRIQUE. R. 1970. Descripcidn, Caracteffsticas y Usos de 25 Maderas Tropicales Mexicanas. Serie Maderas de M~ico. Biblioteca del Constructor No, 5. CANAIC, Mexico. 237 pp. GOMEZ-GARIBAY, F., REYES-CHILPA, R., QUIJANO, L., PARDO-CALDERON, J.S., and RfoS-CASTILLO, T. 1990. Methoxy furan auranols with fungistatic activity from Lonchocarpus castilloi. Phytochemistry 29(2):459-463. HART, J H . , and HILLIS, W.E. 1972. Inhibition of wood-rotting fungi by ellagitannins in the heartwood of Quercus albo. Phytopathology 62:620-626. HART, J . H , and SHRIMPTON, D.M. 1979. Role of stilbenes in resistance of wood to decay. Phytopathologr 69:1138-1143. KHANNA R.N., and SESHADm, T.R. 1963. Pongaglabrone, a new component of the seed of Pongamia glabra its constitution and synthesis. Tetrahedron 19:219-225. KREC~KOVA, J., KRE('EK, J.. and HARMATHA, J. 1988. Feeding deterrent activity of certain plant substances against subterranean termite Coptotermes formosanus (Rhinotermitidae; Isoptera), in F. Sehnal, A. Zabza, and D.L. Denlinger, (eds.). Endocrinological Frontiers in Physiological Insect Ecology. Wroclaw Technical University Press, Wroclaw, p. 105-107. LAKS, P.E. 1988. Wood preservative as trees do it. Am. Wood Pres. Assoc. 1988:147-155. MCDANIEL, C.A. 1992. Major antitermitic components of the heartwood of southern catalpa. J. Chem. EcoL 18(31:359-639. ORTEGA-ESCALONA. F., CASTILLO-MORALES, 1., and CARMONA-VALDOVINOS, T. 1991. Anatom/a de la madera de veintiseis especies de la Selva Lacandona, Chiapas. La Madera y su Uso, Vol. 26. lnstituto de Ecologia A.C.-Universidad Aut6noma Metropolitana, Xalapa, M~xico, 200 pp. PI~REz-MORALES, V.. SEGURA-WARNHOLTZ, G. and BRUNHUBER-MENENDEZ, J.L. 1984. Resistencia natural de 18 especies de maderas mexicanas contra el ataque de termitas de madera seca, Cryptotermes brevis (Walker), (Kalotermitidae, Is6ptera). Memorias del Congreso Mexicano de Patolog/a Forestal, September 26-29, 1984. Durango, M6xico. REYES-CmLPA, R. 1986. Los metabolitos secundarios en la resistencia natural de {a madera a la pudrici6n: Nuevos flavonoides fungist~iticos aislados de l_zmchocarpus castilloi. MSc thesis. INIREB, Xalapa, M6xico. 95 pp. REYES-CHILPA, R.. PI~REZ-MORALES, V. and DEL ANGEL-BLANCO, S. 1987. Influencia de los extractiros en la resistencia natural de seis maderas tropicales al hongo de pudrici6n morena Lenz.ites trabea. Bi6tica (M6xico) 12(1):7-13. SCHEFFRAHN, R . H 1991. Allelochemical resistance of wood to termites. Sociobiology 19(1]:257281. SUPRIANA. N. 1983. Preliminary tests on the effect of naturally occurring chemicals on termites. IRG/WP/1181. Fourteenth Annual Meeting of the International Research Group on Wood Preservation. Australia, May 1983. II pp. WILLIAMS, R.M.C. 1980. A visit to Mexico to advise on the termite research programe at INIREB (LACITEMAI Xalapa, Veracruz. November 7-29, 1980. Report CVR/81/5. Manuscript of limited distribution. Centre for Overaseas Pest Research, London, 17 pp. WOLCOTr, G.N. 1953. Stilbene and comparable materials for dry wood termite control. J. Eeon. Entomol. 46(21:374-375.