Effect of Coumarin and Related Compounds on Blister Beetle Feeding in Sweetclover

The ashgray blister beetle. EpiCtluta [abricii (LeConte); th e striped blister beetle. E. lilt/ala (F.) ; and the margined bli ster beetle, E. p estijer« Werner. collected in the field from plants of sweetclover low in coumarin content, fed preferentially in the Iaboratory on low-coumarin leaves when offered a choice between leaves from highand low-coumarin plants. The black bli ster beetle. E. peTl1IS)'/va nica (Dc Geer) , not known to feed 011 sweetclover foli age, also fed preferenti all y in th e laboratory on lea ves from low-coumarin plants, although th e rate of Sweetclover iMelilotus 5pp.) is one of a wide range of plants fed upon by blister beetles, Ep icauta spp . No serious damage to sweetclover was observed until low-coumarin strains became available following the interspecific transfer of the gene conditioning low-coumarin content from a wild form to the cultivated white-flowered species. M. alba Desr, (Smith 1943). Preferential feeding by blister beetles on low-coumarin sweetclover plants was observed in I Coleoptera: Melold ae. # Cooperati ve Investigations of th e Plant Scien ce and Entomology Research Dhislons, Agrtc. Res. Serv., USDA, and the Nebraska Agricult ura l Experiment Station . Linc oln 68505. Published as Paper no. 3259. Journal Seri es. Nebraska Agr icultural Experimen t Station. Re sear ch reported was cond ucted under Project 1227. The technical assistance of R alph D. Kulm Is gr atefull y acknowl edged. Recei ved for publicat ion Feb. 25. 1972. • Plant Science Research Division, Agri c. Res. Serv.. USDA. Lincoln , Nebr. 68£.03. • Department 01 Agr onomy. Uulversltv of Nebr aska. Lin coln 68503. • Entomology Research Division , Agr ic. Res. Sen'., USDA. Li ncoln , Nebr. 68503. feeding was mum slower than for the other species. By using appropriate genotypes of sweetclover and controlling the type of light to which lite plants were exposed , it was possible to offer th e beetles choices among leaves differing greatly in th eir conten t of specifiG coumarin-related compounds. Both cis-o-hydl"oxycinnamic acid (cis-o-HCA) glucoside and coumarin (the lactone form of cis-o-HCA) were strong feeding deterrents: transa -H CA glucoside had no significant deterring effect, the Great Plains region (Howe and Gorz 1960). but quantitati ve data on feeding differences were not obtained, nor were the species of Epicau ta identified. Although the term "coumar in" is used in the preceding paragraph. in reality. there is little if apy coum arin in the intact sweetclover plant (Haski~s and Gorz 1961). However, the term "coumarin" Wl~ be retained in some part s of thi s paper {or convent" ence . The plant constituents previously assayed as coumarin are the glucosides of cisand trans-o-~y­ droxycinnamic acid (o-HCA) _ Available evidence 1!ldicates that lrans-o-HCA glucoside is synthesized m very young leaves and is then pa rti ally converted to the corresponding cis isomer by the action of '(JV light (Has kins et al. 196·1). In plants grown 1D growth chambers equipped with cool-white fluorescent lamps (ca. 10,000 lux), most of the total o-H~A is present as the trans-glucoside. whereas simIlar chamber-grown plants exposed to sunlight for a few hours contain a . preponderance of cis-o-HCA gluDecembe~r 1972 GORZ ET AL.: BLISTER BEEll_E FEEDING 1633 coside. In those plants having a high content of cis-o-HeA glucoside as well as high activity of fJglucosidase (the enzyme which effects hydrolysis of cis-o-HCA glucoside), tissue disruption (such as that caused by blister beetIe feeding) results in rapid hydrolysis of the cis-o-HCA glucoside to yield coumarink acid. which lactonizes spontaneously to form coumarin (Kosuge and COlin 1961). However, no free coumarin formation would be expected during feeding on leaves containing only trems-a-HCA glucoside because the fJ-glucosidase in sweetclover does not hydrolyze the t.l·411S glucoside. Thus, by selecting proper g'enolypes and lighting conditions, leaves were obtained for use in feeding trials in which the effect of each glucoside could be assessed and the added effect of coumarin could be measured. The objectives of this study were to (I) measure the relative amount of feeding by various Epicauta species among lines of sweetclover differing in content of o-HeA glucosides and the presence or absence of .p-glucosidase activity. and (2) determine the deterring action of coumarin and the glucosides of cisand tml/.l-o-HCA on heetle feeding. l\fATI~RIALS AND METHoDs.-In 1970 and 1971, collections were made of the ashgray blister beetle, E. fab"icii (LeConte); margined blister beetle, E. pestifet"a Werner; striped blister beetle, E. vittata (F.); and black blister bee lie, E. pennsylvanica (De Geer) • The 1st 3 species were collected from low-coumarin sweeldover plants {variety 'Denta') and tbe black blister beetles were found on pigweed, Amaranthus spp. .Neur-isogenic lines of biennial white-blossomed sweetclover (M. alba) of the euCuBB, ClICubb, cut'UB11.• and.Ct,cubb genotypes were used. The derivation of these genotypes was described by Gilchrist et al. (19iO). In commenting on the review of Whittaker and Feeny (1971), lones (1971) stressed the impotUlJlce of wlthin-spedes comparisons to establish the role of secondary metabolites in the resist.ance of plants to insects and diseases. We submit that such intraspecific comparisons call best be made through use of near-Isogenlc lines of the type used in .the present study. Sweetclover plants of the CnCu genot)'pe are high ill content of the glucosides of cisand tt'/17u-o-HCA; c'u'u plants are low in content (AkesclIl et a1. 1963). Similarly, preparations of BB ~lanls (Ire high in ~-glucosidase activity; bb llreparahons :Ippear to he inactive (Schaeffer et al, 960). A single feeding test with intact plants was con. ducled with E. fabric·ii. The test consisted of 8 repllcatiolls of 4 genotypes of sweetclover offered to a POpulation of 2 beetlesjplant in a screened cage, 86X5liX38 em. A separate feeding cage was used for each replication, III each cage. each genotype was represt'llIed by 5 plants growing in soil ill a I-pint p!astk-cmlled milk carton. Plants averaged ca. 15 em high having been unifonnly cut back 2 weeks prior to the feeding test. The extent o[ feeding was meas~red by visual estimation of the percentage consumption of each leaf. All other tests utllized cut swcetclover shoots which Were trimmed to a uniform number of leaves and placed in vials of water or various solutions to be tested. Shoots in which a high content of C'is-o-HCA glucoside was desired were exposed to sunlight prior ---...... • We lhank R. B. Selander, DeJ)arlment of Entomolc,gy, Uo;. ver.tll· "f JIIInoll, UrbatlB, fot' Idetlli[ylol the Ipedel of E/.;ca IIIa.. Table I.-Fel.'ding by blister beetles on ,"lelilotus leaves of 4 genOlypes diftering in content of o-hydroxydnnamic acid and p-glucosidase actil·ity. Sweet% of leaves consumed" clover -------_.genoE·1um?lSylt)'pe"" E. [abricii E. pestliera E.vi/tala Vll7lICa --_.._-CllrolJ/J 9i a 100a 100 a 88 a ruc.ullB !)ij a 100a 100 a 81 a C"C"blJ 311 b 52b !II> 21ib CuCuJJB 12 c 23 c Oc 19b .. I't:mt> were grown in growth chambers and exposed to sunlighl prior to initiation of the feeding tests, h Values shown are mea". of ~ rCl,licadons of c:UI shoots held in '....ter except for the a.,hgl'al' blister beetles. in whkh 8 tepllcatton» of inial'" plants were used. Within each spcdel, means not foIIoWl~1 by the same leiter arc signific.aJJlly di lIerent al the 5% level of probabllirr. to the feeding tests. Feeding chambers consisted of I-gal ice-cream cartons containing a 5-(111 layer of silica sand and covered with a plastic screen. Vials containing the shoots were pushed into the sand to prevent tipping, and shoots of the various treatments were closely intertwined. One replication consisted of a single shoot of each treatment, A population of 2-1J beetles/shoot was sufficient in most cases to provide an adequate level of 'feeding. Feeding damage was recorded as described for the intact plants. RESULTS AND DL~(;USSloN.-BeetIesof each of the 4 species fed preferentially on leaves of the cuc:u genotype, which contain a very low level of both the transand cis-o-HCA glucosides (Table I). Thus. it a.ppears that in CUC1l leaves, one or both of the isomers of o-HCA glucoside have a strong deterring effect on feeding by blister beetles. Furthermore, the apparent preference pf the beetles for CuC",I]/} leaves over leaves of the CuCvBB genotype suggests that free coumarin also serves as a feeding deterrent. Both genotypes contain high levels of the o-HCA glucosides, but only CuCuBD plants contain the p-glucosidase necessary to hydrolyze cis-o-HCA glucoside, thereby per. mitring the formation of free coumarin during beetle feeding. In tests designed to determine whether blister beetles contain a fJ-glucos.idase which hydrolyzes d.r-o·HCA glucoside, assays were positive only when the beetles had fed previously on plants of the BB genotype. Prior feeding on cucubb plants for 2 days resulted in uniformly negative assays of various insect parts. Thus. the observed effects of /I-glucosidase activity in the feeding tests were due entirely to plant-derived enzyme. Table 2 shows the combined effects of tramand c:is-o-HCA glucoside content and p-glucosidase presence on blister beetle feeding. ShOOlS of the CuCubb and CuellBB genotypes, grown in a growth chamber, were exposed to sunfight for 2 hr lO effect conversion of most of the traru-o-HCA glucoside to tile cis isomer (Haskins et al. 1964), while similar shoots were held in vials of water 10 the dark. Assays for content of cisand tTllns·o-HCA glucosides (Haskins and Gorz 1970) in similarly treated shoots revealed that 90-94·% of the total o-HCA glucoside content of the leaves was present as the trans isomer in shoots held in the dark, whereas only 18-20% was present as the trans isomer in leaves exposed to sunlight. In each of the !l tests, there was a highly significant