Insecticidal amino acids in legume seeds

Biochemical Systematics. 1973. Vol. 1. pp. 63 to 67. Pergamon Press. Printed in England.

Insecticidal Amino Acids in Legume Seeds S. S. REHR Department of Entomology. Cornell University. Ithaca. NY 14850. USA

E. ARTHUR BELL* Department of Botany. University of Texas. Austin. TX 78712. USA

..

DANIELH. JANZENt

t

Department of Biology. University of Chicago. Chicago. IL 60637. USA

and PAUL P. FEENY Department of Entomology. Cornell University. Ithaca. NY 14850. USA

Key Word Index-Prodenia eridania; Leguminosae; seed toxins; plant-insect interactions; albizzine; canavanine; amino a,cids. Abstract-When incorporated into artificial diets. powdered seeds of several Central American legume species were found to be toxic or repellent to larvae of the southern armyworm. Similar results were obtained from bioassay experiments with certain unusual amino acid components of these seeds. suggesting that these compounds may contribute to the known resistance of many legume seeds to natural attack by insects.

Introduction In 1918 Bridwell suggested that the presence of natural insecticides. such as alkaloids. protected the seeds of certain legumes in the Hawaiian Islands from attack by introduced beetles of the family Bruchidae. 1 Howe and Currie 2 have also implicated seed toxins as the cause of failure of bruchid larvae to develop in some legume seeds. and Applebaum 3 has shown experimentally that seed toxins and nutritional imbalance may result in failure of bruchid larvae, to develop in some economically important legumes. In Central America. legume seeds are attacked by a variety of bruchid species. as well

as by the larvae of a few species of Lepidoptera. Certain legume species. however. are conspicuously free from such attack.4 and it was our objective in the present study to explore the possibility that such seeds may be prote-cted against insects by the high concentrations of unusual amino acids known to occur in some at least of these. 5 The toxicity both of ground seed powders and of pure compounds extracted from seeds was tested by incorporating them into an artificial diet suitable for rearing larvae of the southern armyworm. Prodenia eridania (Noctuidae). Caterpillars of this species are polyphagous with a highly effective system for

'Present address: Department of Botany. King's College. 68 Half Moon Lane. London. England. tPresent address: Department of Zoology. University of Michigan. Ann Arbor. MI 48104. USA. 1 Bridwell. J. C. (1918) Proc. Hawaiian Entomol. Soc. 3, 465. 2Howe. R. W. and Currie. J. E. (1964) Bull. Entomol. Res. 55,437. 3Applebaum. S. W. (1964) J. Insect Physiol. 10, 783; Applebaum. S. W .. Gestetner. B. and' Birk. T. (1965) J. Insect Physiol. 11, 611. 4Janzen. D. H. (1969) Evolution 23, 1. 5Fellows. L. E. and Bell. E. A. (1970) Phytochemistry 9, 2389: Dardenne. G. A .. Bell. E, A.. Nulu. J. R. and Cone. C. (1972) Phytochemistry11 , 791 ; Dardenne. G. A .. Casimir. J .. Bell. E, A. and Nulu. J. R. (1972) Phytochemistry 11, 787: Bell. E. A. and Janzen. D. H. (1971) Nature 229, 136: Bell. E. A. (1972) in Phytochemical Ecology (Harborne. J. B .. ed,). pp. 163-177. Academic Press. London.

(Received 21 August 1972.

Accepted 21 September 1972) 63

s. S.

64

REHR. E. ARTHUR BELL. DANIEL H. JANZEN AND PAUL P. FEENY

TABLE 1. A COMPARISON BETWEEN PREDATION ON LEGUME SEEDS AND THEIR CHEMICAL CONSTITUENTS Species

Entada polystach ya Gleditsia triacanthos Gymnocladus dioicus Albizi8 julibrissin

Predation on seeds' Two Bruchidae spp. at least two Lepidoptera spp. One Lepidoptera sp. larvae on green seeds Heavy (bruchid preda!ors unidentified) None None

Enterolobium cyclocarpum Dioclea megacarpa Guilandina crista

None Caryedes brasiliensis§ (Bruchidae) None

Hymenaea courbaril (C.R.)

Rhinochenus spp. (Curculionidae) after pod resin lowered None

Cassia grandis

H ymenaea courbaril (P. R.) Griffonia simplicifolia

Potentially toxic secondary chemicals in seedst No uncommon free amino acids Unidentified uncommon amino acid 5-Hydroxypipecolic acid Il-Hydroxy-y-methylglutamic acidt 2-3% Albizziine 2-3% S-(Il-Carboxyethyl) -cysteine Albizziine 5-10% Canavanine y-Methylglutamic acid y-Methyleneglutamic acid y-Ethylideneglutamic acid No uncommon amino acids. no alkaloids No uncommon amino acids. no alkaloids 6-10% 5- Hydroxytryptophan II

• Janzen. D. H .. unpublished data. tBeli. E. A.. unpublished data. Dardenne. G. A. et al. (1972) Phytochemistry 11. 787. §Janzen. D. H' (1971)* Am. Nat. 105. 97. II Fellows. L. E. and Bell. E. A. (1970) Phytochemistry 9.2389.

detoxifying both insecticides 6 and. presumably. also secondary chemicals encountered in their wide array of foodplants.7 We have found that several uncommon free amino acids (nonprotein amino acids of sporadic occurrence in plants) which occur in high concentrations in the seeds of certain legumes are indeed insecticidal or repellent to the otherwise rather resistant larvae of this species.

Results Seed Species Attacked in the Field Seeds of Cassia grandis are heavily preyed

upon by insects.s and extracts of these seeds have yielded no uncommon free amino acids (Table 1). This species apparently compensates for its losses by producing such large numbers of seeds that its bruchid predators are satiated before all seeds are killed. 8 . 9 As expected. diets of 5% C. grandis seed powder proved acceptable to southern armyworms and produced negligible mortality (Table 2). Seeds of Entada polystachya. which are known to be eaten while green by larvae of one species of Lepidoptera. contain an as yet unidentified free amino acid (Table 1). A 5% E. polystachya

TABLE 2. EFFECTS OF ARTIFICIAL DIETS CONTAINING 5% SEED POWDER ON PRODENIA ERIDANIA LARVAE REARED ON THEM INITIAL NUMBER- OF LARVAE ON EACH DIET WAS 20 (X ± S.E.) Species Cassia grandis Entada polystachya

Mean initial wt(mg) 15,6 ± 1·0 11·2 ± 0·4

Mean pupal wt (mg) 310,9 ± 11·3 192·2 ± 9·3

GlfJditsia triacanthos

13·2 ± 0,8

316·2 ± 10·7

Gymnocladus dioicus

14'0±0'7

219·2 ±

7·5

Albizia julibrissin

11-1±0'5

254,5 ±

7·9

Enterolobium cyclocarpum

13·7 ± 0,8

201·5 ±

7·9

Dioclea megacarpa Guilandma crista

15·7 ± 1·2 11'3± 0-4

290,5 ± 259·3 ±

9·9 4,6

Hymenaea courbaril (C.R.) Hymenaea courbaril (P.R.) Griffonia simplicifolia

18,4 ± 1·4 18'0± 1·3 11'8± 0·4

281,9 ± 272-5 ±

9·3 8·1

Pupae 20 normal 15 normal 3 inc' 10 normal 8 inc. 4 normal 11 inc. 14 inc. 11 normal 3 hpbk.· 19 normal 20 normal 20 normal 20 normal 0

Mean adult wt (mg) 109·8± 7·4 142,2 ± 12·2 102'5±

7·9

129·7 ±

8·3

85'4± 4·2 134,3 ±

3·5

Adults 18 normal 17 normal 1 bt. wg.· 9 normal 3 st. wg.· 3 normal 6 sl. wg. 2 bt. wg.· 2 sl. wg. 11 normal 18 normal 15 normal 5 bt. wg. 20 normal 20 normal ·0

·Abbreviations. inc.-first three ventral abdominal segments soft and light in color but not highly vulnerable to puncture; hpbk.-pupa compressed to give a humpbacked appearance; bl. wg.-bent or twisted wings; sl. wg.-stunted wings.

6Krieger. R. I. (1970) Ph.D. Thesis. Cornell University: Krieger. R. I. and Wilkinson. C. F. (1969) Biochem. Pharmacol. 18, 1403. 'Krieger. R. I.. Feeny. P. P. and Wilkinson. C. F. (1971) Science 172, 579: Gordon H. T. (1961) Ann. Rev

Entomol. 6, 27. 8Janzen. D. H. (1973) Ecology in press. 9Janzen. D. H. (1969) Evolution 23, 1.

INSECTICIDAL AMINO ACIDS IN LEGUME SEEDS

65

seed powder diet resulted in low mortality, but some deformity. for larvae reared on it (Table 2). In deformed pupae the first three ventral abdominal segments were soft and light in color. but not highly vulnerable to puncturing with subsequent loss of hemolymph. 10 However. such deformed pupae generally either died or emerged as adults with such bent or twisted wings that mating was impossible.

L-glutamic acid, the analagous protein amino acid, was acceptable to the larvae and supported growth, in spite of the resulting low pH of the diet (Table 3). A similar response to an uncommon amino acid has been reported; L-DOPA (3A-dihydroxyphenylalanine) is toxic to southern armyworms at low levels (0,25%), but repellent at the high concentrations found in certain legume seeds. 12 NH2

NH2

I

I

•

H2N-CO-NH-CH2-CH-COOH

HOOC-CH-CH-CH-COOH

I

(I)

I

(II)

Me OH p.Hydroxy-y-methylglutamic acid

Seeds of Gleditsia triacanthos suffer severe damage from bruchids in the field (Table 1). Unexpectedly a diet of 5% G. triacanthos seed powder produced moderate mortality and high incidence of pupal deformity among southern armyworms reared on it (Table 2). Seed Species Relatively Free from Insect Attack in the Field Seeds of Gymnoc/adus dioicus are reported to be immune to bruchid damage (Table 1). and heavy mortality as well as adult deformity resulted among armyworms reared on a 5% G. dioicus seed powder diet (Table 2). A diet of 4% p-hydroxy-y-methylglutamic acid (I), the concentration found in G. dioicus seeds,11 repelled armyworms; that is. no feeding on the diet was observed. and 100% mortality resulted within three days (Table 3). In contrast a diet of 4%

Albizziine

Both species that contain albizziine (II) proved toxic to P. eridania larvae. but the 5% Albizia julibrissin seed powder diet produced greater mortality than the 5% Enterolobium cyc!ocarpum diet (Table 2). All pupae were deformed, and no normal adults emerged from larvae reared on the A. julibrissin diet. Jacobson 13 also reported insecticidal properties in powdered seeds both of Albizia species and of £ cyc/ocarpum. There was not sufficient albizziine available to rear armyworms through an entire life cycle; however, a 3% albizziine diet was acceptable to larvae and supported life, but little growth, for 1 week (Table 3). A diet of 3·2% S-(p-carboxyethyl)-cysteine (III), a second uncommon free amino acid in A. julibrissin seeds,14 produced a moderate incidence of deformed adults and moderate mortality in P. eridania larvae reared on it (Table 4). The heavy mortality produced by the A. julibrissin diet can perhaps be attributed

TABLE 3. EFFECTS OF PURE AMINO ACIDS IN ARTIFICIAL DIETS ON PRODENIA ERIDANIA LARVAE FED THEM. INITIAL NUMBER OF LARVAE ON EACH DIET WAS 20 (X: S.E.)

Compound 4% p-Hydroxy-y-methylglutamic acid 4% L -Glutamic acid 3% Albizziine 5% Canavanine sulfate 5% Arginine hydrochloride 8% 5-Hydroxytryptophan

Mean initial wt(mg) 11-1 ± 0·3 11'0± 0·3 14·1 ± 0·6 13'3: 0'5 13·1 ± 0·5 12·0± 0·6

Mean larval wt at 3 days (mg)

% mortality at 7 days

18'5 ± 0·8 27·7 ± 2·9

0

13'1: 0·6

5 lOOt

Mean days to death

Mean wt on death (mg)

2-7 ± 0·2

5'0± 0'5 3·5 5'0

4'0± 0·1

5·9 ± 0·3

2·9± 0·1

5·2± 0·3

25

lOOt

pH of diet"

4·5

"Jumbo pHydrion paper, pH 3,0-5·5. The pH of 5·0% seed powder diets was 6·0. tAfter 3 days.

10See also Rehr, S. S.. Janzen, D. H. and Feeny, P. P., in preparation. 11 Dardenne, G. A., Beli, E. A,Nulu, J. R. and Cone, C. (1972) Phytochemistry 11. 791. 12Rehr, S. S.. Janzen. D. H. and Feeny, P. P., in preparation. 13Jacobson. M. (1958) Insecticides from Plants, USDA Agric. Handbook 154. 14Bell, E. A, unpublished data.

66

S. S. REHR. E. ARTHUR BELL. DANIEL H. JANZEN AND PAUL P. FEENY

to the synergistic effects of albizziine and S - (~-carboxyethyl) -cysteine. NH2

I

HOOC-CH 2-CH 2-S-CH 2 -CH-COOH (III)

A 5% Guilandina crista seed powder diet produced no mortality. but 25% of the adults of P. eridania larvae reared on it were deformed (Table 2); however. pure chemicals present in the seeds (Table 1) were not tested.

S -(p-Carboxyethyl) -cysteine

Seeds of Dioclea megacarpa are attacked orly by the highly host-specific bruchid. Caryedes brasiliensis. 1 5 and are known to contain 5-10% canavanine (IV). Southern armyworm larvae were able to tolerate the relatively low canavanine concentration in a 5% D. megacarpa seed powder diet (Table 2); however. at a concentration of canavanine sulfate equivalent to that found in whole D. megacarpa seeds. the diet proved repellent and produced 100% larval mortality (Table 3). In contrast a 5% diet of the analagous normal amino acid. arginine hydrochloride. sustained for 1 week 75% of the twenty larvae fed on it (Table 3). NH2

I

H2N-C(=NH)-NH-O-CHz-CH2-CH-COOH (IV)

Canavanine

Seeds of Hymenaea courbaril are free from insect attack in Puerto Rico. but in Costa Rica several species of weevils of the genus Rhinochenus oviposit on the pods after their resin content is lowered; That a 5% H. courbaril seed powder diet was entirely innocuous to southern armyworm larvae (Table 2) supports the hypothesis 15 that the pod wall resins· provide mechanical protection for the seeds within. as an alternative strategy to seed toxins.· ..

(V)

5- Hydroxytryptophan

Seed Species with No Available Field Data on Insect Damage Field data on degree of herbivore damage to seeds are unfortunately not available for the African legume. Griffonia simplicifolia. This species produced 100% larval mortality when seed powder was incorporated into an artificial diet at a concentration of 5% (Table 2). Leaves of this species are reputedly used to kililice. 16 and the seeds are rich in 5-hydroxytrytophan (V).17 A diet of· 0·5% 5-HTP. simulating the concentration of that chemical in the seed powder diet. produced some mortality and adult deformity in southern armyworms (Table 4). A diet of 8% 5-HTP, an average level found in whole G. simplicifolia seeds. repelled armyworms and produced 100% larval mortality within three days (Table 3). Discussion Since the armyworm is phytophagous. whole legume seeds were ground and presented to the larvae as a constituent of an artificial diet. A major limitation in the use of such a diet for screeni ngpotentially toxic seeds- was that. in order to maintain sufficient water content for theTarvae.the seed powder concentration in the

TABLE 4. EFFECTS OF PURE AMINO ACIDS IN ARTIFICIAL DIETS ON PRODENIA ERIDANIA LARVAE REARED ON THEM. INITIAL NUMBER OF LARVAE ON EACH DIET WAS 20 (X ± S.E.)

Compound a'2% S-(~-Carboxyethyl)-cysteine 0·5% 5-Hydroxytryptophan

Mean initial wt (mg) ·10'8 ± 0·3

Mean pupal wt (mg) 246·9 ±8'0

11·3± 0'3

205'7± 11·2

Pupae 16·nOrmal . 3 abnorm.* 15 normal 4 abnorm.

Mean adult wt(mg) 120·8 ± 9·3

"Abbreviations. abnorm.-various abnormalities· in sclerotization of pupal case: sl. wg.-stunted wings.

15Janzen. D. H. (1971) Am. Nat. 105, 97. 16lrvine. F. R. (1961) Woody Plants of Ghana. p. 308. Oxford UniversityPress. Oxford. 17Fellows. L. E. and Bell. E. A. (1970) Phytochemistry 9,2389.

93'7 ± 8:0

Adults 7 normal 5 st. wg.* 13 normal 2 Sl.wg.

INSECTICIDAL AMINO ACIDS IN LEGUME SEEDS

diet could not be raised above 5%. Concentrations of potentially toxic secondary chemicals were therefore much diluted and unless the chemicals were extremely toxic, the tolerant armyworms may have developed normally with low mortality. This problem was overcome by feeding diets containing pure chemicals, identified in seeds free from insect attack, at concentrations similar to those found in such seeds. We were thus able to demonstrate that canavanine (IV) and p-hydroxy-y-methylglutamic acid (I) were repellent to southern armyworm larvae, while equivalent concentrations of normal amino acid analogs were acceptable to the larvae, in spite of the resulting low pH of such diets. 5-Hydroxytryptophan (V) proved toxic to larvae at low levels and repellent at those high levels found in Griffonia simp/icifo/ia seeds. A similar variability in toxicity has been demonstrated with L-DOPA a constituent of Mucuna seeds. 18 Other explanations have been proposed for the presence in legume seeds of high concentrations of unusual free amino acids. 19 The high percentage of nitrogen in certain uncommon amino acids suggests that they may function to store nitrogen reserves. However. one may argue that one of the normal amino acids, for which the plant must simultaneously contain enzymes for synthesis and metabolism. could easily play the same role. The presence of an unusual amino acid confers on the plant an additional adaptive advantage. Many of these compounds are known to be toxic to man and his domestic animals. but it is unlikely that the initial selective pressures favoring synthesis of high concentrations of toxic compounds were exerted by such organisms. More likely the pressure of insect and rodent herbivores and of

67

fungi and microorganisms was a more significant force in selecting for seed species containing these toxins.

Experimental Source of seeds. Legume seeds were collected (D.H.J.) at the following sites: Hymenaea courbaril (C.R.). north of Canas. Guanacaste Province. Costa Rica. 9 February 1969; Hymenaea courbaril (P.R.). near Rio Piedras. San Juan. Puerto Rico. March 1969; Cassia grandis. south of Canas. Guanacaste Province. Costa Rica. 12 July 1969; Dioc!ea megacarpa. north of Bagaces. Guanacaste Province. Costa Rica. 20 June 1969; Enterolobium cyc!ocarpum. Guanacaste Province. Costa Rica. 10 June 1969; Gymnocladus dioicus. northern Indiana. June 1969; Gleditsia triacanthos. Hyde Park. Chicago. Illinois. November 1969; Guilandina crista. Playa Coco. Guanacaste Province. Costa Rica. 11 September 1970; Entada polystachya. Puntarenas. Puntarenas Province. Costa Rica. 4 February 1970. Albizia julibrissin seeds were collected (EAB.) in Austin. Texas. Griffonia simplicifolia seeds were obtained from Prof. G. Lawson. University of Ghana. Legon. Ghana. Chemicals. L - DO PA. L -5-hydroxytryptophan. and L-canavanine sulfate were purchased from Nutritional Biochemicals. Cleveland. Ohio. S-(p-carboxyethyl)cysteine 20 and 2(S).3(S).4(R)- p-hydroxy-y-methylglutamic acid 21 were isolated from seeds. Albizziine was a gift from Prof. L. Fowden. University College. London. Arginine hydrochloride and L-glutamic acid were obtained from Calbiochem. Los Angeles. California. Bioassay experiments. Eggs of Prodenia eridania 24 were initially obtained from the Niagara Chemical Division. FMC Corporation. Middleport. N.Y. The artificial diet used for bioassay was a version of the one developed by Feeny.22 modified to incorporate 5% seed powder. Twenty early fourth instar larvae were reared on each diet in individual plastic pots under constant conditions (16/8 hr light/dark photoperiod. 21' day. 19' night. 30-40% relative humidity). An excess of diet was always present. The pots were half-filled with moistened vermiculite immediately prior to pupation. Pupae were weighed and their sex determined 23 as soon as the pupal case hardened. Adults were weighed after emerging. 24

Acknowledgement-This research was supported by Hatch Grant NYC-139413.

18Rehr. S. S.. Janzen. D. H. and Feeny. P. P.. in preparation. 19Bell. E. A. (1971) in Chemotaxonomy of the Leguminosae (Harborne. J. B.. Boulter. D. and Turner. B. L.. eds.). pp. 179-206. Academic Press. London. 2°Bell. E. A.. unpublished data. 21 Dardenne. G. A.. Casimir. J .. Bell. E. A. and Nulu. J. R. (1972) Phytochemistry 11. 787. 22Feeny. P. P. (1968) J. Insect Physiol. 14. 805; Rehr. S. S.. Feeny. P. P. and Janzen. D. H.. J. Anim. Ecol.. in press. 23Butt. B. A. and Cantu. E. (1962) U.S.D.A. Agric. Res. Servo ARS-33-75. 24Specimens of the insects used in these experiments have been deposited in the Cornell University Insect Collection. Lot No.1 023. Sublot No.9.