Journal of Chemical Ecology, Vol. 15, No. 6, 1989

ISOLATION, IDENTIFICATION, AND BIOASSAY OF CHEMICALS AFFECTING NONPREFERENCE CARROTROOT RESISTANCE TO CARROT-FLY LARVA

A . M A K I , l J. K I T A J I M A , 2 F . A B E , 3 G . S T E W A R T ,

and M.F.

RYAN

Department of Zoology, University College Dublin, Belfield, Dublin 4, Ireland (Received March 21, 1988; accepted September 20, 1988) Abstract--Roots of the carrot cultivars Vertou L.D. (resistant) and Long Chantenay (susceptible) were subjected to detailed chemical analysis to identify extracts and compounds influencing larval host-finding (preference/nonpreference) behavior and to compare concentrations of these compounds in resistant and susceptible cultivars. Vertou yielded threefold less volatile material in headspace extracts of pureed roots. Extracts of chopped root in methanol, steam, bexane, and chloroform were inactive in behavioral assays. However, ether extracts were active and their hydrocarbon and carbonyl-rich fractions contained potent attractants. The principal constituent of the carbonyl-rich fraction of each cultivar was the carotatoxin complex comprising the neurotoxin falcarinol (carotatoxin), falcarindiol, and falcarindiol monoacetate, the latter compound here reported for the first time from carrot. Falcarinol (50 and 100 tzg) was active in a behavioral assay, and all three ingredients of the complex were potent electrophysiological stimuli, eliciting maximum single unit responses to source concentrations of 10 ng. Furthermore, the complex was more abundant by about 1000/zg/root in Long Chantenay. As this comprised extra amounts of 70, 862, and 110 t~g of falcarinol, falcarindiol, and falcarindiol monoacetate, respectively, the observed differences seem both behaviorally and physiologically relevant. It is generally accepted that coevolution has transformed the role of many toxins into hostlocation cues, but this seems a relatively rare example of a neurotoxin (falcarinol) evincing, in decreased concentrations, nonpreference host resistance.

1present address: Department of Biology, A1-Mustansiryah University, Baghdad, Iraq. 2Present address: Showa College of Pharmaceutical Science, Tokyo, Japan. 3present address: Faculty of Pharmaceutical Sciences, Fukuoka University, 8-1 9-1 Jonan-Ku, Fukuoka, P14-01, Japan.

1883 0098-0331/89/0600-1883506.00/0 9 1989 Plenum Publishing Corporation

1884

MAKI ET AL,

This evolvedresponse to a toxin present in large concentrations is contrasted with that to trans-2-nonenal, which paralysesand kills the larvaand is present in only trace amounts in the root. Key Words--Carrot Daucus carom, Psila rosae, Diptera, Psilidae, carrotfly larva, root chemicals, carbonyl-richfraction, falcarinol, falcarindiol, falcarindiol monoacetate, electrophysiology,neurotoxin, carrot resistance factor. INTRODUCTION The carrot root influences three principal mechanisms of carrot resistance to carrot fly, namely, nonpreference oviposition, independent nonpreference by the invading larva, and antibiosis against the feeding larva (Guerin and Ryan, 1984; Maki and Ryan, 1989). Of these, decreased oviposition is inadequate for plant protection, as cultivars with significantly fewer eggs nevertheless sustained significantly more root damage (Guerin and Ryan, 1984; Maki and Ryan, 1989). Thus, root resistance to the larva is critical for successful carrot resistance. There is an association between increased carrot resistance and decreased release of root volatiles, as indicated both by solvent extraction and headspace entrapment. Specifically, the following five compounds significantly preferred by the larva were released in smaller concentrations by roots of resistant cultivars: bornyl acetate; biphenyl, a-ionone, /3-ionone, and 2,4-dimethylstyrene (Guerin and Ryan, 1984). This report further investigates the relationship between root chemistry and root resistance to the larva by examining larval responses to a variety of root extracts, fractions, and individual compounds using both behavioral and electrophysiological assays. METHODS AND MATERIALS

Extraction and Quantification o f Root Volatiles. Headspace extracts were derived from roots of Vertou and Long Chantenay by separate entrapment of their volatiles on the absorbents Porapak Q and Tenax GC (60-80 mesh). Weighed batches of each cultivar's roots, pureed in a food blender with periodic addition of ice, were placed in foil-covered glass tubes 62 cm long x 9 cm diam. and simultaneously extracted. Air, purified over activated charcoal, diverged to pass over each puree, then through two flow meters set at 0.1 liter/ min, and finally through two foil-covered tubes each containing 5 g absorbent. After 48 hr, the purees were discarded and replaced by fresh batches in the washed tubes. Thus ca. 1 kg of each cultivar was extracted for 70-80 hr. Pora-

CARROT ROOT RESISTANCE

1885

pak was activated as previously described (Ryan and Guerin, 1982) and Tenax GC was activated by heating at 275 ~C for 5 hr in a nitrogen stream at 0.1 liter/ min. Volatiles from each source were separately extracted in a Soxhlet apparatus for 24 hr in diethyl ether and for a further 24 hr in methanol, solvents of choice for nonpolar and polar compounds, respectively. After concentration by rotary evaporation and exposure to a nitrogen stream, such extracts were quantitatively examined by gas-liquid chromatography (details in Table 1). The following extractions of chopped root were also made: steam-distillation, separation in water-hexane and water-chloroform mixtures, extraction in methanol followed by fractionation (Figure 1), and extraction in ether; only the latter yielded consistently active fractions (Figure 2). The active carbonylrich fractions from the ether extracts were further subdivided by repeated column chromatography on silica gel. The eluants were n-hexane-diethyl ether in the successive ratios 19 : 1, 4 : 1, 2 : 1, and 1 : 1. Eluant samples were monitored by TLC or GLC and isolates were identified by GC-MS (VG Analytical 7070E mass spectrometer, PYE 204 gas chromatograph; Finnigan-Mat INCOS 2400 data system) and [1H]NMR (Perkin Elmer 12 RB), supplemented by high-field [1H]NMR (Jeol GX 270). (+_)-2-Methoxy-3-sec-butylpyrazine was an individual compound of particular interest as it has been claimed to produce the distinctive odor of carrot (Cronin and Stanton, 1976). It was not commercially available and has been recorded in concentrations too small for easy extraction ( < 1 ppm) (Cronin and Stanton, 1976). Accordingly, it was synthesised as a racemic mixture (F. Abe, In preparation). TABLE 1. YIELD OF VOLATILES EXTRACTED WITH VARIOUS ADSORBENTS AND SOLVENTS FROM ROOTS OF CARROT CULTIVARS VERTOU (V, RESISTANT) AND LONG CHANTENAY ( L C , SUSCEPTIBLE) b

Volatile yield (/~g/80 g pureed root tissue/24 hr) of cultivarsa Adsorbent

Solvent

Vertou

Long Chantenay

Porapak Q

Diethyl Ether Methanol. Diethyl Ether Methanol

539 18 1251 273

1410 21 4019 503

Tenax G C

Ratio of V to LC 1 1 1 1

: : : :

2.6 1.4 3.2 1.8

a80 g represents the mean weight of a mature root for carrots sown in June 1984 and harvested in March 1985. Values represent means of two replicates differing by less than 10%. bTypical operating conditions: Hewlett-Packard 5890 GC linked to automatic computing integrator 3380 A ; column, glass 10 m long, 530 # m diameter, coated with Carbowax 20 M, which gives superior resolution of carrot volatiles (Ryan and Guerin, 1982); injection port temperature 220~ detector temperature 220~ oven temperature 100~ for 1 min followed by a programmed rise of 10~ for 10 rain to 170~

1886

MaYa ET AL.

ILEAST ES'STA 1T CULTIVAR

IMOSTRES'1STANT

LONG CHANTENAY~eKg

g)

CULTIVAR

VERTOULD

METHANOL EXTRACT ~

2BKg

~s~

n-hexane

HEHANOLIHEXANE~ExTACT me~hon~l I H2o L HET~NOLIHETHANOL ~" EXTRACT(NO I )

h

J

I

. . . . . .

J

.....................I

Lgl)~) ~N ................ EX'tRACT [No2) ~18~)

I

.....................

"I

FIG. 1. Yields, effects, and potencies of fractions derived from methanol extracts of Long Chantenay and Vertou carrot roots. Typically, fractions were bioassayed by the box method in concentrations of 0.01, 0.5, and 1.0 mg/disk with some additional observations at 5.0 rag/disk. Encircled values are yields. Effects and potencies: A, attractive, R, repellent; *P < 0.05; **P < 0.01, Corresponding values of R(%), derived from (T C/T + C) x 100 and rounded to the nearest integer (details in text), were as follows: methanol extract, Vertou, R = 17% at 0.1 mg/disk and - 2 0 % at 0.5 mg/disk; methanol-ether-extract, Vertou, R = 38% at 1.0 mg/disk; methanol-methanol extract No. 1, Long Chantenay R = 29% at 0.5 mg/disk; methanol-methanol extract No. 2 Long Chantenay, R = 28 % at 1.0 mg/disk and - 2 0 % at 5.0 mg/disk; methanol-water extract, Vertou, R = - 2 6 % at 1.0 rag/disk. -

Behavioral Assays. Root extracts were subjected to one of two behavioral assays. The first was based on the method used to observe larval responses o f the onion fly, Delia antiqua (Meigan), to sulfur-bearing chemicals (Soni and Finch, 1979). It employed a Petri dish (10 cm diam.) lined with a 9-cm, damp, black filter paper (Schleicher and Schull No. 551), marked with a pencil into quarters designated test, control, upper neutral, and lower neutral, respectively. Typically, test chemicals were introduced by adding 100 txl o f stimulus solution onto an elderberry pith disk (1.5 cm diam. x 0.3 m m thick, preextracted with solvent) (test) in concentrations o f 0.05, 0.1, or 1 mg/100/zl; another pith disk treated with 100 #1 solvent (control) was placed in the opposite sector. After evaporation o f solvent, 2 0 - 3 0 third-instar larvae, deprived of food for 24 hr, were placed in the center o f each test chamber using a soft paint brush. Five replicates (chambers) were used for each concentration. Test chambers containing larvae were kept in the dark for 15 min at 20 + 1 ~ after which the number o f larvae in each sector was counted and recorded.

CARROT ROOT RESISTANCE

1887

'EAsrRESS l TANr g~

("~OSrRESS l TA~Y~

CULTIVAR ONG CHANTENAY ~

LVERTOU L D ~g~J

CI

CULIVAR

[

ETHER

ETHER

128]

ESSENTIAL

FRACTO IN

SILICA GEL CHROMATOORAPH~ ether

zther,

~utt6

ALCOHOL FRACTION acetone

acetone

he~ane

CARBONYLFRACTO IN FIG. 2. Yield and potencies of fractions derived from ether extracts of Long Chantenay and Vertou carrot roots. Each batch of roots was extracted with ether and the solvent evaporated in vacuo at 40~ Each ether extract was loaded on to a column of silica gel (Kiesel Gel 60, Merck, 20 times the extract weight) and eluted with the solvents indicated. Fractions at concentrations of 0.5, 1.0, 2.0, and 4.0 mg/disk were bioassayed by the box method (details in Methods and Materials). All significant responses were elicited at 1.0 rag/disk; **P < 0.01; ***P < 0.001; values in parentheses are R(%), see text for details. The yields of ether-extracted oil are equivalent to 380 and 320 ppm for Long Chantenay and Vertou, respectively. On the basis of a mean root weight of 80 g, this is equivalent to an extra 4.80 mg oil from a Long Chantenay root.

The second method employed a lidded rectangular Perspex chamber (15 x 4 x 4 cm deep). The floor was lined with damp filter paper (Whatman No. 40) marked with pencil at the midline to give test and control halves. As before, a pith disk treated with the test chemical (usually 0.001, 0.01, 0.05, 0.1, and 0.5 mg/disk) was placed at the far end o f the test half, a corresponding disk treated with solvent placed in the other end o f the control half, and 2 0 - 3 0 unfed larvae were employed per replicate. Five lidded chambers were kept in the dark for 115 min, after which time the number o f larvae in each half was recorded. Both methods gave essentially similar results when tested with the same compounds. All data were subjected, first, to a X2 test for homogeneity and homogenous data were then further assessed by the t test. It was convenient to score larval responses in particular experiments according to the formula: R ( % ) = ( T - C~

1888

MAKI ET AL.

T + C) x 100, where T is the number of larvae on the test side of the chamber, and C is the number of larvae on the control side of the chamber. Electrophysiological Assays. The experimental set-up was housed in an earthed Faraday cage into which ran a stimulus delivery system (Figure 3). Air flow was directed through three solenoids controlled by a timer switch opening one valve for 1 sec every 5 min; it could also be operated manually. The valve directed air over the stimulus chemical applied as 20/zl to a folded piece of filter paper (12 cm) contained in a glass cartridge (54 mm long and 10 mm diameter) with ground glass end. The other two lines terminated in a control (solvent) cartridge and an empty cartridge, respectively, the latter providing a continuous flow of purified air to remove stimulus molecules from the vicinity of the preparation. The principal electronic control unit was the WPI Micro-probe system model M-707 equipped with a probe, which was gold-plated, epoxy-sealed, and contained the first stage of amplification ( x 1). Plugged into this was a microelectrode holder having a molded Ag-AgC1 half cell into which the recording microelectrode was inserted through a rubber gasket. This microprobe system had the following features: current passing and low noise, drift-free performance, bandwidth limiting filter, notch filter, selectable bridge balance range, and a probe test. The system was connected to both a Phillips PM 8010 double pen recorder and to a Telequipment DM 64 storage oscilloscope. Recording and reference microelectrodes had tip diameters of 0.5 and 5.0 /zm, respec-

/ACTIVATEO'-~.~ [~D ~.[J

;,ssEo .EA,80.

..o /

L

I

,00it M-707 ] /ou~lpuClROPPOBE SYSTIEMjTS~

I[ I ~

I'-

AIR VALVES J'~

l /

I n

'IIIoo3,,

~_

I I

STIMULUS\ // /

--

[/

I

/ CONTINUOUS

I!I|

.... II

~l

""..INDIFFERENT I~ I[

/ I

~

It

j] 8prn FIG. 3. A block diagram of the electrophysiological set-up. Inset: diagrammatic representation of the larval group B sensilla indicating the ampullaceous sensillum (a), from which recordings were made.

1889

CARROT ROOT RESISTANCE

tively, and were loaded with Ringer solution identical in the following constituents to that of receptor lymph from Antheraea polyphemus (K.E. Kaissling, personal communication: glucose (22.50 mM/liter), KC1 (171.90), KH2PO4 (9.17), K2HPO4 (10.83), MgC12 (3.00), CaC12 (1.00), NaC1 (25.00), and HC1 (1.50); pH 6.5 and 475 mosm. Third-instar larvae, unfed for 4 hr at room temperature, were immobilized on a disk-shaped rubber stage (3 cm diam.) by knots of human hair. The larva bears on its cephalic lobes 24 sensilla arranged in three groups (A, B, and C), of which group B comprises six evident sensilla on each side of the ventral midline: one ampullaceous (3 /zm diam.) and three basiconic on a cuticular projection; dorsolateral to this are two additional sensilla, one styloconic and one basiconic (Ryan and Behan, 1973) (see inset to Figure 3). This group of sensilla seems analogous to the terminal organ of the house fly larva and the anterior organ of the larvae of the onion and seed corn flies (Yamada et al., 1981; Honda and Ishikawa, 1987a). The latter organ has been associated, through electrophysiological recordings, with a gustatory function, but, because the capillary tip touched the whole organ, it was not possible to distinguish the functions of individual sensilla (Honda and Ishikawa, 1987b). In the Psila larva, the dome of the ampullaceous sensillum of group B is perforated by pores above branched dendrites (Behan, unpublished results), which is consistent with an olfactory function. The recording electrode just penetrated the cuticle of this ampullaceous sensillum (3 ~m diameter) so the records obtained may be viewed as single unit responses, which we measured as DC changes; the primary electrical event in the response of a chemosensory neuron is the DC generator or receptor potential. The reference electrode was inserted into the hemolymph adjacent to the cephalic lobes. As electrophysiological responses diminished throughout an experiment, the response amplitudes to the test compounds were expressed as a percentage of the responses to a standard: 0.1 #g limonene for carrot root fractions and 0.5 tzg trans-2-hexen-l-ol for the single compounds. Consistent physical conditions are necessary in order to obtain reproducible electroantennograms (Adler, 1971). Thus, the cartridge was always positioned at the same distance (2 cm) from the preparation, the size of the filter paper within the cartridge was always identical, all cartridges had identical dimensions, and the flow rate of air was always the same (1 liter/min).

RESULTS

Extraction and Behavioral Assays of Root Volatiles. In each of four comparisons of total volatiles released into the headspace, resistant Vertou yielded less than susceptible Long Chantenay (Table 1). Tenax was threefold more

1890

MAKI ET AL.

efficient than Porapak as adsorbent, and washes with diethyl ether yielded 8.8fold more volatiles than subsequent washes with methanol, so the ether wash of the Tenax extraction is the most appropriate one for comparison. In this, total volatiles released by Vertou were less than one third the quantity released by Long Chantenay (Table 1). Behavioral assays were made with a range of extracts including steam distillates, hexane-water, and chloroform-water extracts derived from Clause's Original Sytan (resistant), Long Chantenay (susceptible), and Danvers Half Long 126 (intermediate in resistance) (0.05-0.1 rag/disk). The aqueous layers from these extracts of the cultivars were all inactive. Only the steam distillate of Sytan and Danvers (5 tA/disk), the hexane extract of Long Chantenay (0.1 rag/disk), and the chloroform extract of Sytan (0.05 rag/disk) were significantly preferred to their corresponding controls. Batches of Vertou and Long Chantenay extracted and fractionated in methanol (Figure 1) elicited too few significant responses to merit further detailed investigation. In contrast, ether extracts were quite active, especially at 1 mg/disk. Although the essential oil extract of Vertou was inactive, that from Long Chantenay elicited a highly significant positive response (details of subfractions, yields, and responses in Figure 2). Neither the alcohol nor more polar fractions, collectively representing 26 % and 60 % of Vertou and Long Chantenay extracts, respectively, elicited any significant responses. In contract, the fractions rich in hydrocarbon and carbonyl compounds elicited significantly positive responses (Figure 2). As a-ionone, /3-ionone, and bomyl acetate are attractants (Ryan and Guerin, 1982) and contain a carbonyl group, we focused on the carbonyl-rich fractions that were further fractionated by repeated silica gel chromatography monitored by thin-layer and gas-liquid chromatography (Figure 4)" Elution by hexane-diethyl ether 19:1 followed by 4:1 afforded falcarinol (carotatoxin); further elution with the same solvents in the ratios 2 : 1 and 1 : 1 afforded falcarindiol monoacetate and falcarindiol (Figure 4). These compounds are represented by CH 3 ( C H 2 ) 6 C H = C H - C H - C - = C - C = C - C H - C H = C H [ I R OR 1

2

where for falcarinol (carotatoxin) R = R 1 = H; for falcarindiol monoacetate R = OH, R 1 = Ac; and for falcarindiol R = OH, R ~ = H (Figure 4). Of the three compounds, falcarindiol monoacetate is newly recorded from carrot. For completeness, the foliage of the two cultivars was subjected to an identical extraction, which also afforded falcarindiol monoacetate and falcarindiol but not falcarinol (Figure 4). Quantitatively, these three compounds collectively

CARROT ROOT RESISTANCE

1891

...~.~-.: (">