14 Natural Products Containing Phenylalanine as Potential Bioherbicides Mikhail M. Bobylev, Ludmila I. Bobyleva, and Gary A. Strobel CONTENTS 14.1 Introduction 14.2 Maculosin-I and -II 14.3 Maculosin-I and -II Analogs: Biological Activity 14.4 Role of Phenylalanine Acknowledgments References ABSTRACT Our previous study of maculosin (I, cyclo-TyrPro), a host specific toxin pro- duced by Alternaria alternata on spotted knapweed (Centaurea maculosa), showed that a number of phenylalanine analogs of I possess similar activity, the unsubstituted analog (cyclo-PhePro, II) being the most active. Interestingly, II appeared to be active against a wide variety of plants. We suggested that protected phenylalanine and not a diketopiper- azine is an active moiety of II and other analogs. To prove this idea we synthesized and tested two compounds: (1) where the proline carbonyl was cut off of the pyrrolidine ring, and (2) where in addition the pyrrolidine ring itself was cut to form diethylamine. Both compounds produced the same symptoms on spotted knapweed plants as II. KEYWORDS: natural products; phenylalanine; herbicides; maculosin; spotted knapweed 14.1 Introduction Natural products comprise a voluminous source of new and strikingly diverse bioactive compounds for pharmaceutical and agrochemical development. In medicine they have already been heavily used for many years and form a very substantial part of prescribed drugs — antibiotics being one very good example. However, this is not the case in plant protection. Very few natural products or their derivatives are currently used as agrochem- icals; pyrethroids are probably the only example of true success. There are three major rea- sons for this drastic dearth of natural products in the plant protection area: complex structure, insufficient stability, and, sometimes, high toxicity. © 1999 by CRC Press LLC Indeed, many of the natural bioactive compounds have very complex structure and, therefore, are very difficult to synthesize in the lab and very expensive to produce indus- trially. This obstacle is very easy to overcome with the virtually unlimited resources of the pharmaceutical industry; people are ready to pay large sums of money for a cure. In con- trast, in crop protection there is a very strict, simple, and low limit on spending. It should always be lower than the cost of the saved crops. This sole consideration makes most of the natural products prohibitively expensive. Insufficient stability also is a very serious problem. Many active natural products cannot withstand harsh conditions of field application because sunlight and oxygen break them down before they produce the desired biological effect. Again, pyrethroides are a very good example. Their natural prototypes were extremely unstable and had to be heavily modified for agricultural application. As a result, some of the pyrethroides bear very little resemblance to the initial molecule and could hardly be considered as natural products or their derivatives. The third reason is toxicity. Contrary to widespread expectations, natural products are often very toxic and might have a devastating effect if spread over large areas of field or pasture land. Probably the most familiar example is nicotine. It is an excellent insecticide, but its application is limited because of its high mammalian toxicity. Consequently, the right candidate for a natural biopesticide must not only possess relevant biological activity, have a simple structure, be stable, but have a low mammalian toxicity as well. Maculosin (I) is one of the very few compounds completely satisfying these requirements. 14.2 Maculosin-I and -II Maculosin {(I), (3S-cis)-hexahydro-3-[(4-hydroxyphenyl)methyl]pyrrolo[1,2-a]pyrazine- 1,4-dione} is a host specific fungal toxin produced by Alternaria alternata on spotted knap- weed (Centaurea maculosa). 1 It was discovered in the course of a systematic search for bio- active natural products for weed control among weed pathogens, a novel approach developed by professor Gary Strobel at Montana State University. 2 Initially the authors suggested the name maculosins for the entire series of related dipeptides isolated from Alternaria alternata. Only two compounds in the series were phytotoxic and they were assigned individual names of maculosin-1 (I) and maculosin-2 (II). However, the less active maculosin-2 was not mentioned after that and maculosin-1 became known simply as maculosin. Since the present work reveals some interesting properties of II, we return to the authors’ original terminology and address the compound as maculosin-2. Maculosin-1 possesses a truly remarkable combination of useful properties. First, and most important, it is highly toxic to the target species. In primary tests maculosin-1 produced necrotic lesions on detached and punctured spotted knapweed leaves at the concentration as low as 10 –5 mole/l. Second, its structure is very simple. It is just a combination of two © 1999 by CRC Press LLC common amino acids — proline and tyrosine. Third, in maculosin-1 these amino acids form a cycle and the whole structure becomes very stable. This stability is a general quality of cyclic dipeptides 3 which might be considered as a terminal product of metabolism. They form very easily as a result of metabolism or degradation of proteins, but once formed they usually resist further metabolism or degradation. This process of formation and accumula- tion of cyclic peptides takes place during cooking or even storage of any protein-containing food, and we consume these compounds daily throughout our lives without any adverse effect. This fact reveals the fourth important quality of maculosin-1 — its potential (although not proven) safety. 14.3 Maculosin-I and -II Analogs: Biological Activity For this reason, 3 years ago we started a systematic investigation of maculosin-1 and its analogs with the initial goal to explore their potential as knapweed control agents and to determine primary structure–activity relationships. We synthesized a series of 17 maculosin-1 analogs (III–VIII) carrying different substituents on the aromatic ring and tested them on whole knapweed plants in the greenhouse. 4 We found that neither macu- losin-1 (I) nor any other analog with the free hydroxyl group (IIIa,b,c) were active against whole and intact knapweed plants. We also found that the elimination of the free hydroxyl group by any means — protection (IVa-f), substitution (Va-e), or complete removal (II, VI–VIII) — restores the activity. The activity greatly depended on the size of the substitu- ent or the protecting group; the smaller the substituent, the higher the activity. The most active compound appeared to be the one without substituents, maculosin-2 (II). At the con- centration of 6 × 10 –2 mole/l, which is approximately equivalent to 1.5%, it induced sweep- ing necrosis on spotted knapweed leaves and destroyed up to two thirds of the foliage © 1999 by CRC Press LLC within 1 week after application. This result seemed to be of special importance because the compound had already been described as a phytotoxin and shown to be toxic to another plant. 5 Therefore, we might expect that maculosin-2, in contrast to uniquely selective mac- ulosin-1, possesses a broad spectrum herbicide activity. We tested maculosin-2 first on more closely related to spotted knapweed plants, like yel- low star thistle and Canada thistle, Russian knapweed, rush skeleton weed, dandelion, and sunflower. Later we tested it on a wide variety of totally unrelated weeds, like hound’s tongue, lambsquarters, plantain, sulphur cinquefoil, white top, field bindweed, wild buck- wheat, common mallow, leafy spurge, and hollyhock. The results of these tests are given in the Table 14.1. A more detailed report of the study is being prepared for publication in Plant TABLE 14.1 Broad Spectrum Phytotoxicity of Maculosin-2 The Highest Observed Toxicity Sweeping necrosis, more than half of the leaf surface is damaged; symptoms appear within several hours after application and fully develop within a day or two Weeds Canada thistle (Cirsium arvense) Hound’s tongue (Cynoglossum officinale) High Toxicity Sweeping necrosis, more than half of the leaf surface is damaged; symptoms appear in 2 to 3 three days after application and fully develop within a week Weeds Yellow starthistle (Centaurea solstitialis) Russian knapweed (Centaurea repens) Moderate Toxicity 2 to 3 mm necrotic spots all over leaf surface; symptoms appear in 5 to 7 days after application and fully develop within 10 days Weeds Crops Rush skeletonweed (Chondrilla juncea) Potato Dandelion (Taraxacum officinale) Tomato Broad-leaved plantain (Plantago major) Sunflower Lambsquarters (Chenopodium album) Redroot pigweed (Amaranthus retroflexus) Ornamentals Whitetop or Hoary cress (Cardaria draba) Hollyhock (Alcea rosea) Common mallow (Malva neglecta) Low Toxicity Slightly “burned” tips of the leaves Weeds Crops Sulfur (erect) cinquefoil (Potentilla recta) Wheat Leafy spurge (Euphorbia esula) Barley No Toxicity Weeds Crops Field bindweed (Convolvulus arvensis) Beans (Phaseolus vulgaris) Wild buckwheat (Polygonum convolvulus) © 1999 by CRC Press LLC Science. 6 As we expected, maculosin-2 appeared to be toxic to all of the tested weeds except field bindweed and wild buckwheat. This toxicity, apparently, did not depend on the plant family or genus, but on a quite unexpected quality — hairiness of the leaves. Plants with hairy leaves developed much stronger symptoms within shorter periods of time than those without or with less hair. We suppose that the presence of plant hair may somehow improve absorption of maculosin-2 by the leaves and thus facilitate its phytotoxic action. Among other plants with hairy leaves, Canada thistle and hound’s tongue appeared to be the most sensitive, even more sensitive than spotted knapweed. First, necrotic spots develop on these two plants within 2 to 3 h after application, and in 24 h most of the leaves, except for one or two youngest, are completely desiccated. In this case, the level of phyto- toxic action was almost sufficient for practical application and was approaching that of the commercial biopesticide “Scythe”. 14.4 Role of Phenylalanine These results, as well as proving the idea of a broad spectrum herbicide activity of macu- losin-2, lead us to the suggestion that there is a much broader phenomenon than the phy- totoxicity of a certain compound, or even a group of compounds, to a certain plant or group of plants. Indeed, maculosins are not the only cyclic dipeptides with phytotoxic properties. Pyriculamide (IIIa), a 3-nitro derivative of maculosin, was described by Russian scientists as being somewhat phytotoxic to rice. 7 Two other nitrated dipeptides — thaxtomins A and B (IXa,b) — were found to be responsible for producing the symptoms of potato scab. 8 Recently, the same two products were shown to be phytotoxic to a wide variety of seed- lings. 9 All these compounds have one component in common, namely, phenylalanine (tyrosine should naturally be considered as a substituted phenylalanine), and it is quite logical to assume that phenylalanine is responsible for their phytotoxic action. This assumption is supported by the fact that there are two other Alternaria alternata phytotoxins (Xa,b) — AK toxin I and AK toxin II — that comprise esters of phenylalanine. 10,11 We suppose that phenylalanine may be toxic to higher plants and that this toxicity reveals itself when a properly protected molecule of phenylalanine reaches the target. In that sense there is no difference between maculosins and other phytotoxic cyclic peptides on one hand and esters like AK I and II on the other; all are just protected phenylalanine. To check this idea, we used something similar to a disconnection approach. We synthesized a compound where proline carbonyl was cut off of the pyrrolidine (XI), and another one © 1999 by CRC Press LLC (XII) where in addition the pyrrolidine ring itself was cut to form diethylamine. Essentially, the two compounds are still very similar to maculosin-2 both in shape and size, but are no longer cyclic dipeptides. Instead, they both are just phenylalanine, protected with amido and formyl groups. As we expected, both compounds produced the same symptoms on spotted knapweed plants as did maculosin-2. Although far from being conclusive evidence, this experiment proved to us that we are on the right track and that a special study should be done to investigate the phytotoxicity of protected phenylalanine. ACKNOWLEDGMENTS: The authors thank the Montana Noxious Weed Trust Fund, the Montana Agricultural Experimental Station, and Beim Foundation for their financial support. References 1. Stierle, A., Cardellina, J.H., and Strobel, G.A., Maculosin, a host-specific phytotoxin for spotted knapweed from Alternaria alternata. Proc. Natl. Acad. Sci., 85, 8008-8013, 1988. 2. Strobel, G., Sugawara, F., and Clardy, J., Phytotoxins from plant pathogens of weedy plants. In Allelochemicals: Role in Agriculture and Forestry, American Chemical Society, Washington, D.C., 1987, 516-523. 3. Prasad, C., Bioactive cyclic dipeptides. Peptides, 16(1), 151-164, 1995. 4. Bobylev, M.M., Bobyleva, L.I., and Strobel, G.A., Synthesis and bioactivity of analogs of maculosin, a host specific phytotoxin produced by Alternaria alternata on spotted knapweed (Centaurea maculosa). J. Ag. Food Chem., 44(12), 3960-3964, 1996. 5. Chen, Y., Studies on the metabolic products of Rosellinia necatrix. I. Isolation and characteriza- tion of several physiologically active neutral substances. Bull. Agr. Chem. Soc. Japan, 24, 372- 381, 1960. 6. Bobylev, M.M., Bobyleva, L.I., and Strobel, G.A., Maculosin-2 as broad spectrum bioherbicide. Plant Science. (Being prepared for publication.) 7. Sviridov, S.I. and Ermolinskiy, B.S., Secondary metabolites of Pyricularia oryzae. Khimiya Pr- irodnykh Soedineniy, 7(4), 811-818, 1990. 8. King, R.R., Lawrence, C.H., Clark, M.C., and Calhoun, L.A., Isolation and characterization of phytotoxins assotiated with Streptomyces scabies. J. Chem. Soc. Chem. Commun., 13, 849-50, 1989. 9. Leiner, R.H., Fry, B.A., Carling, D.A., and Loria, R., Probable involvment of thaxtomin A in pathogenicity of streptomyces scabies on seedlings. Phytopathology, 86(7), 709-713, 1996. 10. Nakashima, T., Ueno, T., and Fukami, H., Structure elucidation of AK toxins, host specific phytotoxic metabolites produced by Alternaria kikuchiana Tanaka. Tetrahedron Lett., 23, 4469-4472, 1982. 11. Nakashima, T., Ueno, T., Fukami, H., Taga, T., Masuda, H., Osaki, K., Otani, H., Kohmoto, K., and Nishimura, S., Isolation and structures of AK toxins I and II, host specific phytotoxic metabolites produced by Alternaria alternata Japanese pear pathotype. Agric. Biol. Chem., 49, 807-15, 1985. © 1999 by CRC Press LLC . completely satisfying these requirements. 14. 2 Maculosin-I and -II Maculosin {(I), (3S-cis)-hexahydro- 3-[ (4-hydroxyphenyl)methyl]pyrrolo[1,2-a]pyrazine- 1,4-dione} is a host specific fungal toxin. 14 Natural Products Containing Phenylalanine as Potential Bioherbicides Mikhail M. Bobylev, Ludmila I. Bobyleva, and Gary A. Strobel CONTENTS 14. 1 Introduction 14. 2 Maculosin-I and -II 14. 3. analog (cyclo-PhePro, II) being the most active. Interestingly, II appeared to be active against a wide variety of plants. We suggested that protected phenylalanine and not a diketopiper- azine is an active