Journal of Chemical Ecology, Vol. 19, No. 12, 1993

R E P E L L E N C Y TO TWO-SPOTTED SPIDER MITE,

Tetranychus urticae KOCH, AS RE L AT E D TO

LEAF S U R F A C E CHEMISTRY OF

Lycopersicon hirsutum ACCESSIONS

ZHENHUA GUO, ~ PAUL A. WESTON, 2 and JOHN C. SNYDER*

Department of Horticulture and Landscape Architecture N318 Ag. Sci. N., University of Kentucky

Lexington, Kentucky 40546-0091

(Received June 24, 1993; accepted August 12, 1993)

Abstract--Bioassays employing Tetranychus urticae Koch were used to examine repellency of leaves and trichome secretions of one cultivar of tomato, Lycopersicon esculentum, and 11 accessions of L. hirsutum consisting of seven accessions of L. hirsutum f. typicum and four accessions of L. hirsutum f. glabratum. Leaves of f. typicum were more repellent to mites than were those of f. glabratum or L. esculentum. Removal of trichomes and their secretions by wiping leaves of f. typicum accessions with 95 % ethanol ren- dered them less repellent, indicating that trichomes or their secretions were potentially responsible for repellency. Gas and thin-layer chromatography of leaflet washes obtained by steeping leaflets in hexane indicated that the chem- ical composition of the washes differed among accessions. The main difference between botanical forms was the presence in leaflet washes from f. typicum accessions of compounds eluting late on the gas chromatograph and migrating further on reverse-phase thin-layer chromatography; compounds having these characteristics were absent in washes of f. glabraturn and L. esculentum. The compositions of secretions taken directly from type I, type IV, and type VI trichomes were similar to that of the leaflet wash obtained from the same f. typicum plant, indicating that leaflet washes contained mainly compounds of trichomal origin. Leaflet washes from f. typicum plants were generally more repellent to T. urticae than were washes from f. glabratum when tested by choice and nonchoice bioassays. Taken together, the data indicate that tri-

*To whom correspondence should be addressed. I Current address: Department of Agronomy, University of Kentucky, Lexington, Kentucky 40546- 0091.

2Current address: Community Research Service, Kentucky State University, Frankfort, Kentucky 40601-2355.

2965

0098-0331/93/1200-2965507.00/0 �9 1993 Plenum Publishing Corporation

2966 Guo ~'1" AL.

chome secretions of f. typicum are likely responsible for potent repellency to two-spotted spider mites observed in this taxon; the likely active compounds are sesquiterpene acids or other polar constituents of trichome secretions.

Key Words--Tomato, Lycopersicon esculentum, L. hirsutum f. glabratum, L. hirsutum f. typicum, deterrence, bioassay, sesquiterpene, 2-tridecanone, volatiles, host-resistance, trichomes, antixenosis, allomone.

INTRODUCTION

Lycopersicon hirsutum Humb. & Bonpl., a relative of the common tomato, L. esculentum, is a genetically diverse species (Miller and Tanksley, 1990) and is highly resistant to at least 19 species of arthropod pests of tomato (Farrar and Kennedy, 1992). In those cases where causes of resistance have been investi- gated in detail, trichomes and the chemical composition of trichome secretions have often been implicated (Farrar and Kennedy, 1992).

In the United States approximately 150 accessions of L. hirsutum are main- tained for distribution to scientists. Of these accessions, approximately one third are maintained and distributed by the USDA Plant Introduction Station, formerly at Ames, Iowa (USDA) and now at Geneva, New York. The remainder are maintained and distributed by the Tomato Genetic Stock Center (TGSC) at Davis, California. Accessions distributed by the USDA carry a prefix of PI and those distributed by the TGSC carry a prefix of LA. Although resistance to many of the major insect and mite pests of tomato occurs in L. hirsutum, nearly 80% of the accessions available in the United States have never been system- atically examined for insect resistance. In fact, a single accession of L. hirsutum f. glabratum, PI 134417, has received most of the attention related to pest resistance of Lycopersicon spp. (Farrar and Kennedy, 1992, and references therein). The presence of 2-tridecanone and 2-undecanone in type VI trichomes of PI 134417 has been associated with antibiosis to several major pests of tomato. Factors present in leaves have also been implicated in resistance.

Even though resistance of f. glabratum accession PI 134417 has been extensively researched, this accession does not represent the entire spectrum of resistance available in the species. Previously we reported that among 10 acces- sions of L. hirsutum, which exists as two botanical forms (f. typicum and f. glabratum Mull.), accessions of f. glabratum were less resistant to two-spotted spider mites (Tetranychus urticae Koch) than were accessions of f. typicum (Weston et al., 1989). Resistance was manifested by more movement of mites from an infested bean leaf onto leaves of f. glabratum accessions compared with movement onto leaves of the more resistant f. typicum accessions. Differ- ences in resistance between botanic forms were not explained by differences in type IV trichome density, a primary determinant of mite resistance in f. typicum (Good and Snyder, 1988). Composition of trichome secretions, in terms of the

Lycopersicon REPELLENCY OF TWO-SPOTTED MITE 2967

major components--2-undecanone and 2-tridecanone for f. glabratum accessions and the sesquiterpene hydrocarbons zingiberene and elemene for f. typicum accessions--differedbetween forms, but quantities of these major components of secretions were not highly correlated with resistance (Weston et al., 1989), indicating that these major components were probably not responsible for the greater resistance of these f. typicum accessions. Subsequently we demonstrated, by use of the thumbtack bioassay, that leaflets of one accession of f. typicum, P1251303, were more repellent to spider mites than were leaflets of f. glabratum PI 134417 (Weston and Snyder, 1990).

Because of unanticipated similarity in composition of trichome secretions among accessions of a botanical form acquired from the USDA (Weston et al., 1989), we decided to investigate the potential for greater variation in the com- position of trichome secretions among accessions available from the TGSC. An unpublished, limited survey of some of these accessions revealed that variation in composition of trichome secretions among these accessions was greater than that observed in accessions obtained from USDA. Greater variability among accessions acquired from the TGSC was not unexpected because these acces- sions were collected throughout the native range of L. hirsutum and have since been increased and maintained in a manner that ensures minimal loss of genetic variation. Similar assurances can not be provided for the accessions acquired from USDA.

Because we had identified heterogeneity in composition of trichome secre- tions among accessions acquired from the TGSC and had also demonstrated that leaves of one accession of f. typicum, PI 251303, were more repellent to T. urticae that were leaves from one accession of f. glabratum, PI 134417, we undertook this research to test the hypothesis that leaves of f. typicum are gen- erally more repellent to spider mites than are those of f. glabratum and to determine whether or not trichome secretions were responsible for the differences in repellency between botanic forms of L. hirsutum.

M E T H O D S AND M A T E R I A L S

Plant Material. Seeds of seven f. typicum accessions (LA 1298, LA 1353, LA 1363, LA 1777, LA 1927, LA 2155, and LA 2329) and four f. glabratum accessions (LA 407, LA 1223, LA 1265, and LA 2144) were obtained from the TGSC and those of L. esculentum Ace 55 were obtained from W. Atlee Burpee Co. Seeds were sown in the greenhouse in April 1989 in styrofoam cells filled with Pro Mix BX (Premier Brands, New Rochelle, New York) and set in the field in May. In August, leaflet washes were obtained from these plants using procedures outlined below.

To obtain additional washes and to provide leaves for bioassay, the 11

2968 Guo ET At..

accessions and Ace 55 were also grown in the greenhouse in fall 1990. Seeded on August 17, seedlings were set individually in 8-in pots filled with Pro Mix BX. Greenhouse temperature was set at 25/18 ~ (day/night). No artificial light was provided. Leaflet washes were prepared as outlined below. When leaflets from these plants were selected for bioassays, leaf development was noted by designating the youngest leaf with the smallest major leaflet 2 cm wide as position A, with each successively older leaf designated B, C, etc.

Mite Colony. Mites were lab-reared and handled using the methods of Weston and Snyder (1990). When needed, mites of uniform age (staged mites) were raised by transferring about 30 adult female mites to a mite-free bean seedling (Phaseolus vulgaris Dwarf Horticultural) having half-expanded unifo- liate leaves. Transferred females were allowed to lay eggs for 5 hr and then were blown off with compressed air. Tanglefoot (The Tanglefoot Co., Grand Rapids, Michigan) applied to stems of these seedlings prevented uncontrolled mite infestation. Plants were examined once a day for three days after inocu- lation, and any adult mites discovered were removed. Mites hatched in about twe days. Female mites 6 days old were used for the ring bioassay described later.

Thumbtack Bioassay. These bioassays followed procedures described by Weston and Snyder (1990). In August 1991, four leaves from position C were detached from plants of each of the 11 accessions and Ace 55 grown in the greenhouse. To maintain turgor during bioassay, leaf petioles were inserted into small vials filled with water. Leaves were placed abaxial surface down on sty- rofoam boards and were fastened to the board with a thumbtack in the middle area of each of four leaflets. Ten mites were transferred to each thumbtack. The number of mites remaining on the thumbtack at 30 min, 4 hr, and 8 hr was recorded. After bioassay, trichomes were counted on the adaxial surface of each assayed leaflet in four l-ram 2 interveinal areas with the aid of microscope (100x). ANOVA was performed on transformed mite response data [(X + 0.5)1/2)] and densities of type IV and VI trichomes. Least significant difference (LSD) was used to compare means among accessions.

Leaflet Washes. Fully expanded, healthy leaflets were collected from plants of each accession and steeped in hexane (3 ml/leaflet) for 60 min, with agitation every 15 min. After decantation, the leaflets were briefly steeped in hexane (I rnl/leaflet) twice. The washes were combined, filtered through Whatman No. 1 filter paper, and then dried with anhydrous Na2SO4. After evaporating solvent with a rotary evaporator, the final concentration was brought to 1 ml of hexane per leaflet. This concentration is referred to as 1 x . Washes were stored in glass vials at - 8 0 ~

Removal of Trichome Secretions from Leaflet Surfaces. Absorbent cotton balls were dipped in one of the following: 95% ethanol, 50% ethanol, pure methanol, 50% methanol, pure acetonitrile, 50% acetonitrile, pure acetone, or

Lycopersicon REPELLENCY OF TWO-SPOTTED MITE 2969

50% acetone. Leaflet surfaces of LA 1363 were swabbed three times with each solution and then blotted with a Kimwipe (Kimberly-Clark Co.). Opposing leaflets of a pair, one wiped and the other intact, were then extracted as described above. Aliquots (2 #1) were analyzed by gas chromatography (detail follows) to determine whether or not hexane-soluble volatiles were removed from leaflets. Percentage removed by each solvent was calculated.

Repellency of Intact and Wiped Leaves. For this experiment, conducted during summer 1991, five f. typicum accessions were chosen because they had greatest resistance at the 4-hr sample in the previous thumbtack bioassay. Leaves from position C were removed from the plants, and the adaxial surface of one leaflet of a pair was wiped with cotton balls wetted with 95 % ethanol. The wiped surface and the adaxial surface of the unwiped opposing leaflet were then bioassayed using the thumbtack bioassay. Ten pairs of leaflets from each acces- sion served as replications. Numbers of mites on the thumbtacks at 10 min were recorded, and data were analyzed with the paired t test.

Spring-Board Bioassay. This bioassay, designed to assess repellency of leaflet washes, consisted of paper strips 5 mm wide and 5 cm long cut from Whatman No. 1 filter paper. In the middle area of the strips, two faint lines parallel to each other and 12 mm apart were drawn across both surfaces of the strip with a pencil. One end of each strip was clamped above a mirror, which allowed us to monitor movement of a mite on the lower surfaces of the strips. On two such strips, 2/~1 of 1 • wash was distributed with a micropipet along the pencil line nearer to the clamp on one strip, and another 2/zl was distributed on the line further away from the clamp on the other strip. Care was taken to ensure even distribution and minimal diffusion. Two microliters of hexane was distributed on the other line on each strip as control. After solvent evaporation, a mite was gently transferred with a brash to the area between the two lines on each strip. Movement of the mite was carefully observed, and the line over which it exited was recorded. The exiting mite was removed and another mite was immediately transferred to the area between the two lines. On each strip, 15-20 mites were assayed. Two such strips having the sample loaded on different positions (line nearer to or further away from the clamp) constituted an assay, so a total of 30-40 mites were used for assaying a sample. Exit ratios on the two strips were compared using X 2. In all cases, performance was similar between the two strips, so the data were combined and the significance of the deviation from an expected 1 : 1 exit ratio was determined by Xz-

Ring Bioassay. This procedure, also designed to assess the repeUency of trichome secretions, consisted of paper tings (1.3 cm ID, 1.7 cm OD) cut from Whatman No. 1 filter paper. The area of such a ring is about 1 cm 2. Six microliters of 1 x wash from field-grown plants was distributed evenly on each of three rings, and 6/~1 of hexane was distributed on another ring as a control. After the hexane evaporated, the four tings were placed in the bottom of a glass

2970 Guo E'r AL.

Petri dish lined with filter paper. One staged mite was transferred with a brush to the center of each ring. When a mite escaped over the ring or was inactive for more than 1 rain, it was replaced by another. During the bioassay (minimum duration 30 rain), mite activity was recorded on videotape. Afterwards the video- tape was played and the retention time of each mite in each ring was determined. The data for the three tings loaded with leaflet wash were pooled, and a t test was used to determine if the average retention time of mites placed in rings loaded with t • wash was statistically different from that of the hexane control.

GC Analyses. The volatile profiles of the 1 x washes were determined by gas chromatography (GC). A Hewlett Packard 5890A gas chromatograph equipped with a flame ionization detector (FID) and a 30-m x 0.25-mm RTX-5 column (Restek, Bellefonte, Pennsylvania) was used. The carrier gas was He flowing at 10.0 ml/min and the head pressure was 90 kPa. The following temperatures were used: injector, 240~ detector 300~ oven program, 100~ for 4 min, then increasing at 10~ to 250~ One-microliter samples were injected, and the split ratio was 1 : 8.

Leaflet wash from the f. typicum accession PI 251303 and purchased 2-methyl ketones (Alltech Associates, Deerfield, Illinois) served as standards. Three sesquiterpene hydrocarbons, zingiberene, curcumene, and "r-elemene (Weston et al., 1989), and two sesquiterpenoid acids have been identified in the leaflet washes of PI 251303 (unpublished data).

To obtain volatile profiles for individual types of trichomes, about 10 type I, 20 type IV, or 6 type VI trichome tips from 80-day-old plants grown in the greenhouse were collected with the aid of a 10-~1 syringe and a stereomicroscope (100 x). Trichome tips were picked up with the syringe needle, which was then rinsed in a 25-/zl micropipet containing 3/xl of hexane. The volatile profile of the 3 ~1 was determined by GC-FID on a 1.2-~m, 10-m x 0.54-mm RSL-150 column (Alltech Associates). Helium was the carrier gas flowing at 3 ml/min. The following temperatures were used: injector, 240~ detector, 300~ oven, 100~ for 1 min, then increasing at 6~ to 170~ and then at 10~ to 250~ In order to compare volatile profiles from different types of trichomes on the same accession, the chromatograms for the leaflet wash and trichome secretions from an accession were visually compared and the percent- age abundances of the three most prevalent compounds detected in greatest abundance were calculated, based on integrator area units.

Thin-Layer Chromatography (TLC). Reverse-phase TLC was conducted on C18 plates (Whatman KC18F). Samples were prepared by evaporating hexane from the 1 x leaflet washes, weighing the residue, and then adding sufficient hexane to the residue to provide a concentration of 2 mg/ml. Forty micrograms of each sample was spotted on the TLC plate. After developing the plate with acetonitrile-water (80:20), visualization was accomplished by staining with anisaldehyde/H2SO4 (Krebs et al., 1963), which stains sesquiterpene hydrocar-

Lycopers icon REPELLENCY OF TWO-SPOTTED MITE 2971

bons dark blue, 2-methyl ketones blue-green, and glucose and sucrose acyl esters and sesquiterpene acids brown. The Rf and color of each detected spot was recorded. Relative abundance of each detected spot was visually estimated using a three-point scale based on density of color of the spot and its area. Known sesquiterpene hydrocarbons and their acids, isolated from L. hirsutum f. typicum PI 251303, and purchased 2-undecanone and 2-tridecanone (Pfaltz & Bauer) served as standards. When compounds were present with Rf values and colors similar to those of known compounds, the Rf regions were scraped from replicate plates and then etuted with hexane. The composition of the Rf region was determined by coelution on the gas chromatograph with authentic compounds.

RESULTS

Thumbtack Bioassay. In general, mites remained on the thumbtacks longer on f. typicum accessions than on f. glabratum accessions or on the L. esculentum entry Ace 55 (Table 1). This indicated that leaflets of f. typicum were more

TABLE 1. DENSITIES OF TYPE IV AND VI TRICHOMES ON LEAFLETS AND REPELLENCY

OF LEAFLETS FROM 1 1 Lycopers icon h irsu tum ACCESSIONS AND ONE CULTIVAR OF

L. esculentum AS MEASUI~ED BY THUMBTACK BIOASSAY a

Average number of mites remaining on thumbtacks at

indicated time C Trichome densities

(N/mm 2)

Access ion Taxon b 30 rain 4 hr 8 hr Type IV Type VI

LA 1298 typ 9.5 A 8.5 A 8.5 AB 16 C 4 EF LA 1353 typ 9.0 A 7.5 A 2.4 B 102 A 7 B

LA 1363 typ 8.5 A 6.0 B 5.5 B 6 DEF 6 C LA 1777 typ 8.5 A 8.5 A 7.5 A 30 B 9 B LA 1927 typ 8.0 A 3.5 C 3.0 C 10 CDE 10 B LA2155 typ 5.0 B 0.0 D 0.0 D 2 FG 5 D

LA 2329 typ 9.5 A 7.0 A 6.0 AB 8 DEF 8 B LA407 gla 4.0 B 0.0 D 0.0 D 11 CD 16 A LA 1223 gla 0.5 C 0.0 D 0.0 D 1 FG 4 F LA 1265 gla 1.0 C 0 .0 D 0 .0 D 4 EFG 2 F

LA2144 gla 0.5 D 0 .0 D 0.0 D I G 5 D 'Ace 55' esc 0 .0 D 0.0 D 0.0 D 0 G 4 E

aTen mites were initially placed on each tack. b typ = L. hirsutum f. typicum, gla = L. hirsutum f. glabratum, esc = L. esculentum. "Means in a co lumn sharing the same letter are not significantly different at 5%.

2972 Guo ET AL.

repellent than those of f. glabratum or Ace 55. Of all f. typicum accessions, leaflets of LA 2155 were least repellent and also had the lowest type IV trichome density. Type VI trichome densities ranged from 2 to 16/mm 2, and type IV trichome densities ranged from 0 to 102/mm 2.

Deterrence of Wiped and Intact Leaflets o f f typicum. Wiping leaflet sur- faces of LA 1363 with cotton balls wetted with different solvents removed most of the hexane-soluble, volatile compounds. Acetonitrile, ethanol, methanol, and acetone, whether concentrated or diluted 50 % with water, each removed more than 96 % of leaflet volatiles (data not shown). Uniformly, mites left thumbtacks more quickly on leaflets that had been wiped with 95 % ethanol than on intact leaflets (Table 2). This indicated that intact leaflets were more repellent than those having trichomes and secretions removed by wiping.

Chemical Profiles of Leaflet Washes and Trichomes. Chemical profiles, determined by GC and TLC, differed between forms and among accessions of a form. However, some components were similar among accessions. Washes from all f. typicum accessions except LA 1353 and from the f. glabratum accession LA 1223 possessed volatiles having retention times on GC-FID (data not shown) and Rf values from TLC identical to known sesquiterpene hydro- carbons (Table 3). Washes from f. typicum accession LA 1353 and from f. glabratum accessions LA 407 and LA 1223 contained compounds having reten- tion times and Rf values identical to 2-undecanone and/or 2-tridecanone. Few compounds were detected in leaflet washes from f. glabratum accessions LA 2144 and 1265, or from Ace 55. In addition to the above differences, f. typicum accessions, with the exception of LA 1353, were distinguished from f. glabra- rum accessions by the presence of compounds in their leaflet washes having long retention times on the GC (data not shown) indicative of higher boiling

TABLE 2. AVERAGE NUMBER OF MITES REMAINING ON THUMBTACKS ON 10 PAIRS OF WIPED AND UNWlPED LEAFLETS FROM EACH OF 5 Lycopersicon hirsutum

f. typicum ACCESSIONS a

Average number of mites on tack at 10 min

Accession Intact leaflet Wiped leaflet t value b

LA 1298 6.3 0.4 2.8* LA 1353 9.5 1.5 5.4** LA 1363 5.3 1.3 3.5** LA 1777 6.9 1.3 2.3* LA 2329 5.1 0.8 2.2*

a Ten mites were initially placed on each tack. b,. ,• significant at 5% and 1%, respectively.

TA

BL

E 3

. R

f VA

LU

ES

AN

D R

EL

AT

IVE

AB

UN

DA

NC

E a

OF

CO

MP

OU

ND

S P

RE

SE

NT

IN

LE

AF

LE

T W

AS

HE

S O

F 1

1 L

ycop

ersi

on h

irsu

tum

A

CC

ES

SIO

NS

AN

D O

NE

CU

LT

IVA

R O

F Z

. es

cule

ntum

SE

PA

RA

TE

D B

Y C

18 R

EV

ER

SE

-PH

AS

E T

HIN

-LA

YE

R C

HR

OM

AT

OG

RA

PH

Y

Tax

on

b

typ

g la

LA

L

A

LA

L

A

LA

L

A

LA

L

A

LA

L

A

LA

Rf

Ten

tati

ve I

D

1298

13

53

1363

17

77

1927

2

15

5

23

29

40

7 2

14

4

1265

12

23

esc

Ace

55

0.08

Se

squi

terp

ene

hydr

ocar

bons

c +

+ +

0.13

2-

trid

ecan

one

0.18

0.22

+

0.24

2-

unde

cano

ne

0.30

0.

32

Sesq

uite

rpen

e ac

id J

0.33

0.35

Se

squi

terp

ene

acid

a +

+

0.39

0.45

0.51

0.63

+

+ +

+

++

+

+ +

+ +

+

+ +

+

+ +

+ +

+ +

+ +

++

+

++

+

+

+ +

+ +

+ +

+

++

+

++

+

++

++

+

+

+ +

++

+ ++

+

~T

he n

um

ber

of

+ sy

mb

ols

ind

icat

es r

elat

ive

abu

nd

ance

fro

m l

ow

(+

) to

hig

h (

+ +

+).

b

typ

=

L.

hirs

utum

f.

typi

cum

, gl

a =

L

. h

irsu

tum

f.

gla

bra

tum

, es

c =

L

. es

cule

ntum

. C

Det

enni

ned

by

Wes

ton

et

al.

(198

9).

aDet

erm

ined

by

Co

ates

et

al.

(198

8).

to

--...I

2974 Guo ET AL.

points, compared to the sesquiterpene hydrocarbons or the methyl ketones, and larger Rf values on reverse-phase TLC, indicating their more polar character (Table 3).

The volatile profiles of type I, type IV, and type VI trichomes for an accession were similar to each other and to that of the leaflet wash from the same plant. Visual inspection of chromatograms indicated that chemical profiles

of trichome tips closely matched those of the leaflet washes from the same plant (data not shown). Relative abundance of major components was similar among secretions obtained directly from the three types of trichomes and leaflet wash from the same accession (Table 4), suggesting that trichome secretions were similar among trichome types and supporting the finding of Weston et al. (1989) that volatiles present in leaflet washes contain compounds mainly associated

with trichomes.

Spring-Board Bioassay. Leaflet washes from f. typicum accessions except LA 1353 were strongly repellent; all washes from f. glabratum accessions and

that from Ace 55 were not repellent (Table 5). The amount of residue per leaflet differed among accessions, but no consistent trend between amount of residue per leaflet and repellency was detected.

TABLE 4. RETENTION TIME AS DETERMINED BY GAS CHROMATOGRAPHY AND RELATIVE ABUNDANCE OF THREE COMPOUNDS PRESENT IN GREATEST ABUNDANCE IN LEAFLET

WASHES AND SECRETIONS TAKEN DIRECTLY FROM 3 DIFFERENT TYPES OF TRICHOMES ON 5 ACCESSIONS OF L. hirsutum f. typicum

Origin of compound

Trichome source

Accession Retention time I IV VI Leaflet wash

Percent abundance

LA 1298 8.00 66 60 64 62 9.03 13 9 9 11

14.29 17 21 12 15 LA 1363 9.18 3 6 2 4

12.51 89 65 73 75 13,23 3 5 5 5

LA 1777 9.03 15 13 12 14 13.38 30 33 33 32 13.86 31 25 25 28

LA 1927 9.05 11 6 15 12 12.49 81 65 66 78 13.25 4 3 3 5

Lycopersicon REPELLENCY OF TWO-SPOTTED MITE 2975

TABLE 5. REPELLENCY OF t X L~AFLET WASHES FROM 11 Lycopersicon hirsutum ACCESSIONS AND ONE Z. esculentum CULTIVAR AS MEASURED BY

SPRING-BOARD BIOASSAY a

Accession Taxon b Residue per leaflet (mg) Exit ratio X 2 value

LA 1298 typ 0.4 7 : 23 4.4 **C LA 1353 typ 0.3 11 : 19 1.1 "s LA 1363 typ 0.2 3 : 27 9.6** LA 1777 typ 1.3 2 : 28 11.3"* LA 1927 typ 0.2 4:26 8.1"* LA 2155 typ 0.4 5:25 6.7** LA 2329 typ 1.2 1:29 13.1"* LA 407 gla 0.2 11 : 19 1.1 "S LA 1223 gla 0.2 10 : 20 1.7 "s LA 1265 gla 0.2 12 : 18 < 1 "s LA 2144 gla 0.3 14 : 16 < 1 "S

'Ace 55' esc 0.2 14 : 16 < 1 "S

~ exit ratio is the ratio of the number of mites exiting over the leaflet wash and the number exiting over the hexane control. The expected exit ratio used for calculating X 2 was 1 : 1.

btyp = L. hirsutum f. typicum, gla = L. hirsutum f. glabratum, esc = L. esculentum. cn~.** Not significant, and significant at 1%, respectively~

Ring Bioassay. Results f rom the ring bioassay (Table 6) general ly co inc ided

with those f rom the spring-board bioassay. A m o n g the f. typicum accessions,

the wash f rom the except ional f. typicum access ion L A 1353 was not repellent;

washes f rom the other six f. typicum accessions were repellent . Washes f rom f.

glabratum accessions except L A 2144 were not repel lent , nor was that f rom

Ace 55. The average t ime spent by mites inside t ings treated with the wash

f rom L A 2144 was very s imilar to those for o ther f. glabratum accessions and

was much shorter than those for f. typicum accessions; the significant difference

in t ime spent inside control and treated t ings for this access ion arose f rom the

atypical ly short tenure t ime of mi tes wi thin the control ring.

DISCUSSION

Results f rom the thumbtack bioassay indicated that, in general , leaflets o f

f. typicum accessions were more repel lent to T. urticae than were leaflets o f f.

glabratum accessions or those o f L. esculentum. L A 2155 was an except ion to

this general izat ion, having leaves that were less repel lent than other f. typicum accessions and no more repel lent than the most repel lent f. glabratum accession,

L A 407. These results conf i rm the observat ions o f Wes ton et al. (1989) that f.

typicum accessions are general ly more repel lent to T. urticae than are f. gla-

2 9 7 6 G v o ET AL.

TABLE 6 . REPELLENCY OF | X LEAFLET WASHES OF 11 Lycopers icon ACCESSIONS AND

ONE L. escu len tum CULTIVAR AS MEASURED BY RING BIOASSAY a

Leaflet wash Control (hexane)

Mean Mean retention retention

time + SE t ime _+ SE

Accession Taxon b N (see) N (sec) t value

L A 1298 typ 9 257 + 250 10 55 • 39 2 .4 **c

L A 1353 typ 30 92 + 78 11 73 + 79 < 1 n~

L A 1363 typ 21 124 ___ 120 26 32 • 41 3.4**

L A 1777 typ 8 664 + 585 19 53 + 38 3.0**

L A 1927 typ 16 257 + 242 10 28 _ 19 3.8**

L A 2155 typ 10 312 + 218 14 29 ___ 41 4 . 1 " *

L A 2329 typ 23 217 + 189 10 28 + 19 3.8** L A 407 gla 22 100 + 83 9 124 + 143 < 1 nS

L A 1223 gla 28 110 + 76 10 122 • 105 < 1 nS

L A 1265 gla 32 61 + 50 20 51 + 33 < 1 n~

L A 2144 gla 26 84 + 77 12 40 + 34 2 .5*

'Ace 55 ' esc 15 38 + 27 7 23 _+ 12 1.6 "~

~The number of mites assayed = N. b typ = L. hirsutum f. typicum, gla = L. hirsutum f. glabratum, esc = L. esculentum. ..... ** Not significant, significant at 5 % and 1%, respectively.

bramm accessions, extending the conclusion to a more diverse sample of germ- plasm.

A sequence of observations supports the conclusion that the quality and/or quantity of trichome secretions were at least partly responsible for the greater repellency of f. typicum leaflets. Wiped leaflets of f. typicum were less repellent than intact leaflets, implicating trichomes or trichome secretions as possible causes of repellency. The compositions of secretions collected directly from trichomes on an accession were similar to each other and similar to the com- position of the leaflet wash from that accession. Leaflet washes from f. typicum, except that from LA 1353, were more repellent to T. urticae than were those from f. glabratum accessions or L. esculentum.

Because composition of leaflet washes differed among accessions and between botanical forms, it is likely that more than one compound was respon- sible for repellency. The main difference in the composition of washes between forms was the presence of compounds in washes from f. typicum accessions having higher boiling points, as indicated by their longer retention time on the GC column, and greater polarity, as indicated by their greater mobility on reverse-phase TLC. Compounds having these characteristics were largely absent,

Lycopersicon REPELLENCY OF TWO-SPOTTED MITE 2 9 7 7

as were compounds having characteristics of sesquiterpene hydrocarbons, in washes obtained from f. glabratum accessions. It seems likely that these more polar compounds account for the greater repellency of f. typicum accessions. Diversity in composition of washes from f. typicum, as judged by GC and TLC, was mainly associated with diversity among these more polar compounds. Diversity in composition of washes among f. glabratum accessions was mainly associated with presence or absence of 2-undecanone and 2-tridecanone. In fact, abundance of methyl ketones and volatiles in general was low in two f. glabra- tum accessions (LA 2144 and LA 1265).

The presence of compounds in washes from f. typicum accessions having greater polarity than the 2-methyl ketones was anticipated because carboxylic acids of sesquiterpene hydrocarbons are known to occur in trichome secretions of LA 1777 (Coates et al., 1988) and PI 251303 (unpublished data). The f. typicum accession LA 1353 was an exception to the generalization that leaflet washes from f. typicum leaves contained these more polar compounds. Washes from this accession contained 2-undecanone and probably a sesquiterpene hydro- carbon, indicating that the accession may have characteristics of both botanic forms. In fact, the flower structure on this accession is intermediate, lending credence to this hypothesis. Furthermore, like washes from all f. glabratum accessions tested, washes from this accession were not repellent, supporting the earlier suggestion that the presence of the polar compounds may be a cause of repellency.

Taken together, the data support the hypothesis that f. typicum accessions are more repellent to T. urticae by virtue of the quality of their trichome secre- tions, especially in regard to the presence of polar constituents in the secretions. The lack of a clear trend between quantity of secretions and repellency suggests that quantitative differences in trichome secretions among the accessions tested play little role in the degree of repellency. The relationships between quantity, composition, and repellency are under investigation and necessitate the devel- opment of a quantitative bioassay for repellency.

Repellency is probably not determined solely by the composition or amount of trichome secretion. There is likely an interaction between trichome density and repellency, especially if trichome secretions are mainly responsible for chemical repellency. For example, leaflets of LA 2155 were not highly repellent, but its secretions were. This accession had very few type IV trichomes, implying that some threshold of type IV density may exist for leaflet repellency. Fur- thermore, leaflets of LA 1353 had very high type IV trichome densities and were highly repellent, but leaflet washes of this accession were not repellent, suggesting that very high densities of type IV trichomes, even in the absence of repellent trichome secretions, may repel or deter mites.

A threshold response of mites to type IV trichome density has been reported previously (Carter and Snyder, 1985, 1986), and Good and Snyder (1988) dem-

2978 G~o ET AL.

onstrated that type IV density is a primary determinant of the interaction of L. hirsutum leaves and mites placed directly on the leaves. Furthermore, it is not surprising that there is an interaction between density and repellency if trichomes are the primary repository of chemical repellents.

The diversity of trichome secretions present in this small number of acces- sions of L. hirsutum is much greater than that reported by Weston et al. (1989)

for a similar number of accessions. The reasons for the greater diversity remain unexplained but may relate to site of original collection, to how the accessions have been maintained since collection, or simply to which accessions were chosen for examination. These results confirm that composition of trichome secretions varies among accessions, a feature that needs additional investigation especially as it relates to modification of plant-insect interactions.

In summary, our data indicate that repellency is a mechanism of resistance in this species, which has been reported as immune to insect attack (Rick, 1982). The evidence also suggests that repellency is at least partially associated with

trichomes and their secretions and may ultimately be related to the quality of trichome secretions as well as trichome density. Our data indicate that causes of repellency may not be similar among accessions. We must point out that

repellency is likely one layer of resistance in this highly defended species, but may be a very useful mechanism for introgression into crops such as tomato, in which small amounts of insect damage can lead to great economic loss.

Acknowledgments--The authors want to thank Mr. Richard Thacker for preparing the leaflet washes and offering other technical assistance. Part of this work was funded by USDA competitive agreement No. 87-CRCR-1-2348 to P.A.W. and J.C.S. Contribution No. 93-10-99 of the Kentucky Agriculture Experiment Station, published with the approval of the Director.

REFERENCES

CARTER, C.D., and SNYDER, J.C. 1985. Mite response in relation to trichomes of Lycopersicon esculentum • L. hirsutum F 2 hybrids. Euphytica 34:177-185.

CARTER, C.D., and SNYDER, J.C. 1986. Mite responses and trichome characters in a full-sib F z family of Lycopersicon esculentum • L. hirsutum. J. Amer. Soc. Hortic. Sci. 111:130-133.

COATES, R.M, DENISSEN, J.F., JUVIK, J.A., and BABKA, B.A. 1988. Identification of c~-santan- lenoic and endo-~-bergamotenoic acids as moth oviposition stimulants from wild tomato leaves. J. Org. Chem. 53:2188-2192.

FARRAR, R.R., JR., and KENNEDY, G.G. 1992. Insect and mite resistance in tomato, pp. 122-142, in G. Kalloo (ed.). Genetic Improvement of Tomato. Monographs on Theoretical and Applied Genetics 14. Springer-Verlag, New York.

GOOD, D.E., and SNYDER, J.C. 1988. Seasonal variation of leaves and mite resistance of Lycoper- sicon interspecific hybrids. HortScience 23:891-894.

KREBS, K.G., HEUSER, D., and WIMMER, H. 1963. Z. Spray reagents, pp. 854-909, in E. Stab1 (ed.). Thin-Layer Chromatography. A Laboratory Handbook, 2nd ed. Translated by M.R.F. Ashworth, Springer, New York, 1969.

Lycopersicon REPELLENCY OF TWO-SPOTTED MITE 2979

MILLER, J.C., and TANKSLEY, S.D. 1990. RFLP Analysis of phylogenetic relationships and genetic variation in the genus Lycopersicon. Theor. Appl. Genet. 80:437-448.

RICK, C.M. 1982. The potential of exotic germplasm for tomato improvement, pp. 1-28, in I.K. Vasil, W.R. Scowcroft, and K.J. Frey (eds.). Plant Improvement and Somatic Cell Genetics. Academic Press, New York.

WESTOrq, P.A., and SNYDER, J.C. 1990. Thumbtack bioassay: A quick method for measuring plant resistance to two-spotted spider mites (Acari: Tetranychidae). J. Econ. Entomol. 83:500-504.

WESTON, P.A., JOHNSON, D.A., BURTON, H.T., and SNYDER, J.C. 1989. Trichome secretion com- position, trichome densities, and spider mite resistance of ten accessions of Lycopersicon hirsutum. J. Am. Soc. Hortic. Sci. 114:492-498.