Identification and synthesis of a kairomone mediating host location by two parasitoid species of the cassava mealybug Phenacoccus herreni

Article abstract of DOI:10.1023/A:1012274703197

Two encyrtid species, Acerophagus coccois and Aenasius vexans, parasitoids of the cassava mealybug Phenacoccus herreni use a contact kairomone from the body surface of their host as a host-location stimulant. The kairomone was synthesized and identified as O-caffeoylserine based on a combination of chromatographic methods. The synthetic compound was determined to be active.

Full text of DOI:10.1023/A:1012274703197

c
Journal of Chemical Ecology, Vol. 27, No. 11, November 2001 ( 2001)
IDENTIFICATION AND SYNTHESIS OF A KAIROMONE
MEDIATING HOST LOCATION BY TWO PARASITOID
SPECIES OF THE CASSAVA MEALYBUG Phenacoccus herreni
PAUL-ANDRE CALATAYUD,1,* JACQUES AUGER,2 ERIC THIBOUT,2
SEVERINE ROUSSET,2 ANA MILENA CAICEDO,1
SABINE CALATAYUD,1 HOLGER BUSCHMANN,3 JOSETTE GUILLAUD,4
NICOLE MANDON,2 and ANTHONY C. BELLOTTI5
1IRD (formerly ORSTOM)/CIAT, Cassava Entomology, CIAT
AA 6713, Cali, Colombia
2School of Sciences
Franc¸ois Rabelais University, IRBI, UMR CNRS
6035 Parc de Grandmont, F-37200 Tours, France
3Department of Agroecology in the Tropics and Subtropics
University of Hohenheim
Fruwirthstrasse 12, D-70593 Stuttgart, Germany
4INSA/INRA, Laboratory of Applied Biology, INSA
20, Avenue A. Einstein
F-69621 Villeurbanne Cedex, France
5CIAT, Cassava Entomology, CIAT
AA 6713, Cali, Colombia
(Received January 26, 2001; accepted July 10, 2001)
Abstract—Two encyrtid species, Acerophagus coccois and Aenasius vexans,
parasitoids of the cassava mealybug Phenacoccus herreni use a contact
kairomone from the body surface of their host as a host-location stimulant. The
kairomone was synthesized and identified as O-caffeoylserine based on a com-
bination of chromatographic methods. The synthetic compound was determined
to be active.
Key Words—O-Caffeoylserine, Acerophagus coccois, Aenasius vexans,
Hymenoptera, Encyrtidae, Homoptera, Pseudococcidae.
To whom correspondence should be addressed. E-mail: P.CALATAYUD@CGIAR.ORG
2203
C
0098-0331/01/1100-2203$19.50/0 2001 Plenum Publishing Corporation

2204
CALATAYUD ET AL.
INTRODUCTION
The cassava mealybug, Phenacoccus herreni William & Cock (Homoptera:
Pseudococcidae), is an important pest of cassava in South America (Bellotti
et al., 1983; Noronha, 1990). Two encyrtid parasitoids, Acerophagus coccois
Smith and Aenasius vexans Kerrich (Hymenoptera: Encyrtidae), are being stud-
ied to control P. herreni populations at the International Center for Tropical
Agriculture (CIAT) in Cali, Colombia. A. coccois is considered a generalist be-
cause it parasitizes different mealybug species among the Pseudococcidae
family. A. vexans is a specialist, parasitizing only P. herreni (Dorn et al.,
2001).
During host selection, parasitoids use a variety of cues to assess the quality of
their hosts, such as shape, surface structure, and both external and internal semio-
chemicals (waxy secretions and honeydew) (Lewis and Martin, 1990; Vinson,
1991; Godfray, 1994). The structure and quantity of these semiochemicals influ-
ence host acceptance by parasitoids and vary according to species, age or stage of
development, size, condition, and diet of the parasitoid host (Vinson, 1991; Powell,
1992; Takabayashi and Takahashi, 1993; Ro¨se et al., 1997). One important class of
these subtances used by parasitic Hymenoptera to identify their hosts is kairomones
(Lewis and Martin, 1990). For example, in the California red scale Aonidiella
aurantii Maskell (Hemiptera: Diaspididae), a kairomone isolated from their cover
and identified as O-caffeoyltyrosine, the ester of caffeic acid and tyrosine, mediates
host recognition and induces oviposition by the parasitic wasp Aphytis
melinus DeBach (Hymenoptera: Aphelinidae) (Hare et al., 1993; Millar and Hare,
1993).
In cassava mealybug Phenacoccus manihoti Matile-Ferrero (Homoptera:
Pseudococcidae), a serious pest of cassava in Africa (Neuenschwander et al.,
1990; Le Ru¨ et al., 1991), preliminary analyses indicated the presence of a com-
pound composed of a serine and a caffeic acid moiety, probably corresponding
to O-caffeoylserine (Calatayud, unpublished data). This compound was also de-
tected in P. herreni; however, it could not be identified in the mealybug because
it was always hydrolyzed during the purification process. It was, therefore, neces-
sary to synthesize the ester and then verify whether the synthetic compound was
similar to the compound assumed to be O-caffeoylserine. Because this putative
compound is similar to the contact kairomone O-caffeoyltyrosine, important in
another parasitoid–homopteran interaction as reported by Hare et al. (1993) and
Millar and Hare (1993), we also hypothesized that this substance would act as a
contact kairomone for the cassava mealybug parasitoids. Therefore, the objective
of this study was to confirm the identification of O-caffeoylserine in P. herreni after
its synthesis and to study its influence on host-location behavior of A. coccois and
A. vexans.

CASSAVA MEALYBUG KAIROMONE
2205
METHODS AND MATERIALS
Insect Rearing. The cassava mealybug, P. herreni, was reared at CIAT on
30 to 40-cm potted cassava plants (cv. CMC 40) in a glasshouse at 27–33 C and
12L : 12D photoperiod. Adult females, the stage most parasitized by A. coccois
and A. vexans (Seligmann, 1998), were collected from the colony for chemical
analysis and bioassay.
Both parasitoids species, A. coccois and A. vexans, were reared continuously
on mealybug-infested cassava plants (cv. CMC 40) as described above for the
cassava mealybug. Emerging adults (both males and females) were placed for
48 hr in transparent polystyrene tubes (15 1.5 cm) containing honey droplets as
a food source. The wasps were sexed, and females were selected for bioassay.
Synthesis of Kairomone (Figure 1). The method used to synthesize O-caff-
eoylserine was that described by Millar and Hare (1993), for synthesizing
O-caffeoyltyrosine. Caffeic acid (1) (10 g, 55 mmol) was added to ethyl chlo-
roformate to produce 13.17 g (73.9% yield) of 3,4-diethoxycarbonyl-caffeic acid
(2). 1H NMR (CDCl3): ∂ 7.72 (d, 1H, J = 15.9 Hz, CHCOOH), 7.24–7.46 (3H, m
aromatic), 6.42 (d, 1H, J = 15.9 Hz, olefinic H), 4.28 (2 overlapped quartets, 4H,
J = 7.1 Hz, CH2s), 1.33 (2 overlapped triplets, J = 7.1 Hz, methyls).
The acid (2) (13.17 g, 40.6 mmol) was stirred with 36.5 ml of thionyl chloride
to obtain 10.23 g (73.7% yield) of 3,4-diethoxycarbonyl-caffeoyl chloride (3).
1H NMR (CDCl3): ∂ 7.73 (d, 1H, J = 15.6 Hz, olefinic H), 7.23–7.48 (3H, m
aromatic), 6.56 (d, 1H, J = 15.6 Hz, olefinic H), 4.30 (overlapped quartets, 4H,
J = 7.1 Hz, CH2s), 1.36 (overlapped triplets, 6 H, J = 7.1 Hz, methyls). Then
1.71 g (5 mmol) of the chloride (3) was coupled with the N-t-BOC-(L)-serine (4)
(1.03 g, 5 mmol; Sigma Chemical Co., St. Louis, Missouri) to produce 2.39 g
(quantitative) of the compound (5). 1H NMR (CDCl3): ∂ 7.70 (d, 1H, J = 16.2 Hz,
olefinic H), 7.22–7.58 (3H, m aromatic), 6.40 (d, 1H, J = 15.9 Hz, olefinic H),
5.62 (br. d, 1H, NH), 4.50 (m, 1H, NH-CH), 4.26 (br. quartet, 4H, CH2s), 3.75
(m, 2H, -O-CH2-CH-), 1.27–1.39 (m, 9H, t-BOC methyls, 6H, methyls).
Afterstirringthecompound(5)(752mg)withamixtureofmethylenechloride
and trifluoroacetic acid, the resulting dicarbonate (6) was treated with methanolic
ammonium hydroxide. O-Caffeoylserine (7) was purified by TLC using a migra-
tion solvent of n-butanol–water–AcOH (4 : 1 : 1) for a total yield of 328 mg (87%).
Synthesized O-caffeoylserine was used to confirm the presence of the compound
in P. herreni and bioassayed at several concentrations to determine its biological
activity on A. coccois and A. vexans.
To validate the synthetic method, O-caffeoyltyrosine was also synthesized,
substituting the N-t-BOC-(L)-serine with N-t-BOC-(L)-tyrosine. The chloride (3)
(1.71 g, 5 mmol) was coupled with the N-t-BOC-(L)-tyrosine (1.41 g, 5 mmol;
Sigma) to produce 2.41 g (quantitative) of the compound with the following NMR

2206
CALATAYUD ET AL.
FIG. 1. Scheme of the various molecules obtained during synthesis of O-caffeoylserine.
data. 1H NMR (CDCl3): ∂ 7.75 (d, 1H, J = 15.9 Hz, olefinic H), 6.95–7.49 (6H, m
aromatic), 6.53 (d, 1H, J = 15.9 Hz, olefinic H), 5.00 (br. d, 1H, NH), 4.58 (m, 1H,
NH-CH), 4.30 (br. quartet, 4H, CH2s), 3.11 (m, 2H, benzyl CH2), 1.39 (br. s, 9H,
t-BOC methyls), 1.34 (t, 6H, J = 6.5 Hz, methyls). From 536 mg of this compound,
O-caffeoyltyrosine was obtained for a total yield of 245 mg (81%). Proton NMR
spectra were obtained on an Advance DPX 200 Bruker NMR spectrometer at
200 MHz.

CASSAVA MEALYBUG KAIROMONE
2207
Isolation and Identification of O-Caffeoylserine in Cassava Mealybug. Adult
females (2.4 g) were crushed in 5 ml of 70% ethanol. After 8 hr at 10 C, the extract
was centrifuged at 21,000 g for 3 min, and the aqueous supernatant was washed
twice with chloroform to remove lipids. The lyophilisate of the supernatant was
suspended in 500 l of 70% ethanol and banded on to a reverse-phase TLC plate
(RP-18, F254s, pre-coated sheet 20 20 cm, Merck), developed with acetonitrile–
water (50 : 50), and visualized using UV at 254 nm. Two bands with Rf values
of 0.75 and 0.69 (shown to contain the putative O-caffeoylserine by autoanalysis)
were recovered by scraping the plate and extraction in 70% ethanol. After 8 hr at
10 C and centrifugation in 21,000 g for 3 min, the supernatants were lyophilized.
The lyophilisate was suspended in 70% ethanol and submitted to the same TLC
purification as above, but developed in n-butanol–2-propanol–water (60 : 20 : 20).
Three bands with Rf = 0.81, 0.73, and 0.67 (shown to contain the putative
O-caffeoylserine by autoanalysis) were recovered as above, let stand at 10 C for
8 hr, centrifuged, and lyophilized. A mixture of compounds including the putative
O-caffeoylserine remained in the extract. It was not possible to repeat this TLC
purification because the putative O-caffeoylserine disappeared, probably due to
hydrolysis.
The compound was first identified by LC-MS from a portion of the extract.
LC-MS experiments were performed on a Finnigan TSQ 700 under atmospheric
pressure chemical ionization (APCI) as described elsewhere (Renukappa et al.,
1999). In this case, a reversed-phase column (Grom-Sil 120 ODS-5, 250 4.4 mm,
5 m, Grom, Germany) was used. Separation was achieved by using 0.1% aqueous
trifluoroacetic acid in a linear gradient of 5–100% with acetonitrile over 35 min
and a constant flow of 0.1ml/min.
In the second part, known quantities of the synthetic O-caffeoylserine were
analyzed, alone or added to the remaining extract, by HPLC and autoanalysis.
Samples were analyzed by RP18-HPLC by using the Pico-Tag method from Waters
after amino acid derivatization with phenyl isothiocyanate (PITC). The compounds
were identified by their retention time and quantified by their UV absorption at
240 nm.
Samples were submitted also to ion-exchange chromatography on an au-
tomatic amino acid analyzer (Beckmann 6300). Amino acids were detected by
ninhydrin reaction, identified by their retention time and wavelength ratio, and
quantified by their absorption at 570 nm.
Determination of O-Caffeoylserine as Host-Location Kairomone. To verify
that O-caffeoylserine is present on the body surface of P. herreni, 500 adult fe-
males (representing about 500 mg of total fresh weight) without prior maceration
were soaked in 500 l of aqueous 70% ethanol for a final proportion equiva-
lent to 1 l of extract per individual. Insects without wounds were used, and the
suspension was not stirred to limit leaching of internal insect constituents. The
suspension was left at 10 C for five days. An aliquot of 100 l was collected

2208
CALATAYUD ET AL.
each day and dried for autoanalysis of amino acids. This experiment was run in
triplicate.
To confirm the function of O-caffeoylserine as a contact kairomone, 100
mealybugs were soaked in 100 l of 70% aqueous ethanol at 10 C for three days,
pipetted, and used for bioassay. The same experiment was done with pure hexane
and bioassayed as a negative control. Because the ethanolic suspension also con-
tained free amino acids, a solution duplicating the composition and concentration
of these compounds was also bioassayed. The natural host (i.e., adult female of
P. herreni) was bioassayed as the positive control. Synthetic O-caffeoylserine was
bioassayed at several concentrations to confirm its function of as a host-location
kairomone for mealybug parasitoids.
Bioassays. Wasps initially investigate mealybugs by palpating with their an-
tennae as they walk from one edge of the mealybug body to the other. They then
turn 90 and repeat this behavior one or more times. This behavior, previously
observed in several Aphelinidae species, has been termed “drumming and turning”
by Luck et al. (1982). After drumming and turning, A. vexans and A. coccois insert
their ovipositors inside the host body in most cases, indicating that this behavioral
event characterizes their host location. Therefore, the number of drummings and
turns was recorded for each wasp observed in the following experiments. The
bioassay was conducted on mealybugs or on treated cotton balls, which mimic the
mealybug body. For cotton balls, 1 l of test solution was pipetted onto cotton
balls approx. 2.0 mm diameter. One insect or treated cotton ball was placed inside a
single glass vial bioassay arena (5 mm diam. 10 mm high), and for treated cotton
balls the solvent was allowed to evaporate. One gravid female was added, and the
vial was sealed with a cotton plug. Wasp behavior was observed for 15 min with
a 6 dissecting microscope. Bioassays were conducted in the same greenhouse
where the parasitoids were reared, between 1000 and 1500 hr at 25–35 C. The
natural host (adult females of P. herreni) and several solutions were bioassayed as
described above.
Five replicate trials of five wasps were bioassayed, with new insects or fresh
cotton balls prepared for each trial. Wasps were used only once. The propor-
tion of total drummings and turns on each mealybug or cotton ball was calcu-
√
lated and analyzed by ANOVA after applying (X + 1) transformation. Fisher’s
(protected least significant difference PLSD) test was used following ANOVA to
compare means. Statistical tests were performed with Statview software (Abacus
Concept).
RESULTS
Identification of O-Caffeoylserine in Cassava Mealybug. LC-MS analysis of
P. herreniextractsrevealedacompoundwithamolecularweightof268.3(M+H)+,

CASSAVA MEALYBUG KAIROMONE
2209
close to that calculated from the formula of O-caffeoylserine, corresponding to
C12H13NO6, which is 267.2.
To confirm the identity of O-caffeoylserine in P. herreni extracts, the com-
pound was synthesized according to the method of Millar and Hare (1993), which
was used to synthesize another ester, O-caffeoyltyrosine. The proton NMRs
(CD3OD) of the two synthetic esters were readily interpreted.
For the O-caffeoylserine, we obtained: 7.43 (d, 1H, J = 15.9 Hz, olefinic H),
6.78–7.06 (3H, m aromatic), 6.29 (d, 1H, J = 15.8 Hz, olefinic H), 5.09 (br s,
1H), 4.10 (m, 1H), 3.79 (q, J = 11.1 Hz, 2H). Except for protons lost in exchange
in CD3OD, these NMR data corresponded to those obtained by Lin et al. (1999),
confirming the synthesis of this ester.
For the O-caffeoyltyrosine, we obtained: 7.77 (d, 1H, J = 15.9 Hz, olefinic H
to carbonyl), 7.41 (d, 2H, J = 8.54 Hz, aromatic H, tyrosine), 7.04–7.19 (4H, m
from aromatic H of tyrosine and caffeic acid), 6.84 (d, 1H, J = 8.14 Hz, meta H,
caffeic acid), 6.50 (d, 1H, J = 15.83 Hz, olefinic H to carbonyl), 3.82 (dd, 1H,
J = 9.1, 4.24 Hz, CHNH2), 3.35 (dd, 1H, J = 14.12, 4.10 Hz, benzylic H), 3.04
(dd, 1H, J = 14.5, 9.1 Hz, benzylic H). These NMR data corresponded to those
obtained by Millar and Hare (1993), validating the synthesis method.
The retention time of synthetic O-caffeoylserine on a reverse-phase HPLC
column matched that of the peak at 18.9 min from the P. herreni extract. This peak
increased in area when the P. herreni extract was augmented with 0.4 nmol of
synthetic O-caffeoylserine (Figure 2).
The retention time of synthetic O-caffeoylserine on an ion-exchange col-
umn also matched that of the compound isolated from P. herreni (chromatograms
not shown). This result, together with RP-HPLC data, confirms the identifiy O-
caffeoylserine in P. herreni extracts.
Determination of O-Caffeoylserine as Host-Location Kairomone. When
mealybugs were soaked in 70% ethanol without prior maceration and submitted
to autoanalysis, the O-caffeoylserine amounted to 0.026 nmol/mg (approx. 0.03
nmol/mg) or 5.2% (in moles) (Table 1). Compared to the other nitrogenous com-
pounds analyzed, this was not major. Alanine and glutamic acid were much better
represented among the compounds analyzed, at 28.4 and 17.1% (in moles), re-
spectively. Furthermore, this concentration of O-caffeoylserine was the maximum
possible when mealybugs were soaked in 70% aqueous ethanol. In fact, this con-
centration varied over time (Figure 3) from a minimum of 0.011 0.002 nmol/mg
(mean SE, N = 3) on day 1 to a maximum of 0.032 0.002 nmol/mg on day 3
and declining to 0.027 0.004 nmol/mg on day 5, probably due to hydrolysis.
The fresh weight of an adult female is about 1 mg on CMC 40. According to
the extraction proportion, 1 l of extract was obtained per individual or per 1 mg
of fresh weight. Therefore, 0.03 nmol/mg is equivalent to 0.03 nmol/ l. Synthetic
O-caffeoylserine was bioassayed based on this concentration (at 0.03 nmol/ l and
10 times higher or 2, 4, or 8 times lower).

2210
CALATAYUD ET AL.
FIG. 2. HPLCchromatogramsof(A)thesynthetic O-caffeoylserinealone(at2nmol/10 l);
(B) an extract of P. herreni showing the unknown peak at 18.9 min; and (C) the same extract
analyzed in (B) augmented with 0.4 nmol of synthetic O-caffeoylserine.

CASSAVA MEALYBUG KAIROMONE
2211
TABLE 1. CONTENT AND COMPOSITION OF FREE
NITROGENOUS COMPOUNDS IN MEALYBUGS
Free nitrogenous
compounds
Concentration
(pmol/1 mg)
% (mol)
Asp
Thr
Ser
Asn
Glu
Gln
Pro
Gly
Ala
Val
Met
Ile
Leu
Tyr
Phe
30
14
20
8
5
2
5
2
5.9 0.05
2.9 0.86
3.9 0.25
1.5 0.05
17.1 1.80
2.6 0.40
4.7 0.10
6.7 0.11
28.4 2.41
3.8 0.30
1.8 0.10
1.4 0.09
1.6 0.03
2.7 0.20
0.3 0.03
5.2 0.36
4.4 0.30
0.3 0.07
0.7 0.02
2.1 0.44
2.0 0.45
86 25
13
24
34
143 14
19
9
4
5
7
2
1
1
1
2
1
7
6
1
2
4
4
7
8
13
2
26
22
2
4
11
11
O-Caffeoylserine
Try
Orn
Lys
His
Arg
aValues are (mean SE, N = 3).
FIG. 3. Concentration of O-caffeoylserine in extracts from mealybugs (mean SE, N = 3)
soaked without prior maceration in 70% ethanol as a function of time at 10 C.

2212
CALATAYUD ET AL.
For both parasitoid species, the extracts from mealybugs soaked in 70%
ethanol were significantly more stimulating for host location than those in hexane
(Figure 4), but no oviposition probing was observed. Moreover, these ethanolic
extracts showed similar host-location stimulations when P. herreni was bioas-
sayed. No stimulation was noted with ethanol alone, suggesting that the mediation
of host location was in fact due to compound(s) present in ethanolic extracts of
FIG. 4. Percent drums and turns observed for A. coccois (A) and A. vexans (B) on P. herreni
and on cotton balls treated with solvent alone (hexane or 70% aqueous ethanol), extracts
from mealybugs suspended without prior maceration in hexane or in 70% ethanol, or a
mixture of free amino acids using the same concentration and composition as in Table 1.
Bars indicate the mean ( SE; five trials with five wasps each). Means with the same letter are
not significantly different at 5% level (Fisher’s PLSD test following the ANOVA, P < 0.05).
No statistical comparison was possible for means = 0%.

CASSAVA MEALYBUG KAIROMONE
2213
P. herreni and well solubilized in 70% aqueous ethanol. A mixture of free amino
acids, similar in concentration and composition to the one obtained in Table 1,
showed no stimulating effects as compared to the extracts of P. herreni in ethanol,
indicating the presence of other compound(s) in the extracts, different from free
amino acids, that have a stimulatory function for host location by A. coccois and
A. vexans.
Among the concentrations in O-caffeoylserine tested, 0.03 nmol/ l was
highly attractive for A. coccois (Figure 5A). Nevertheless, no difference was found
between 0.03 and 0.015 nmol/ l. After removing the carbonate-protecting groups
FIG. 5. Percent of drums and turns observed for A. coccois (A) and A. vexans (B) on cotton
balls treated with synthetic O-caffeoylserine at several concentrations or with mixture of
serine and caffeic acid at 0.03 nmol/ l each compound. Bars indicate the mean ( SE; five
trials with five wasps each). Means with the same letter are not significantly different at
5% level (Fisher’s PLSD test following the ANOVA, P < 0.05, except for A. vexans (B),
P > 0.05 for ANOVA). No statistical comparison was possible for means = 0%.

2214
CALATAYUD ET AL.
(ultimate phase in Figure 1), using Millar and Hare’s (1993) synthesis method,
there is inevitably a partial ester cleavage, resulting in the presence of free serine
and caffeic acid in the synthetic O-caffeoylserine solution. Therefore, a mixture
of these compounds was also bioassayed at 0.03 nmol/ l, i.e., at the same con-
centration of O-caffeoylserine where the wasps responded better (Figure 5). It
appeared that the A. coccois female wasps were less stimulated by the hydroly-
sis products of the O-caffeoylserine (caffeic acid and serine) confirming that the
ester, O-caffeoylserine, is the biologically active compound. Moreover, for this
parasitoid species, the ester also induced oviposition probing because 24% (i.e.,
6 of the 25) of the wasps bioassayed at 0.03 nmol/ l of O-caffeoylserine probed
the cotton ball with their ovipositor. Oviposition probing was not observed with
the other concentrations of O-caffeoylserine tested.
Synthetic O-caffeoylserine at 0.03 nmol/ l elicited more drumming and turn-
ing of A. vexans, but differences were not observed among the concentrations tested
(Figure 5B, P = 0.0547 for ANOVA), indicating no dosage-dependent response
in this parasitoid species. Nevertheless, O-caffeoylserine did induce a stimulating
effect in this parasitoid species at 0.03 nmol/ l because the wasps exhibited sig-
nificantly higher drummings and turns than when the solvent alone or the mixture
of serine and caffeic acid were bioassayed.
DISCUSSION
The identification of O-caffeoylserine in P. herreni was confirmed. To our
knowledge, this represents the first report of O-caffeoylserine isolated from bio-
logical material, although this ester was recently synthesized by Lin et al. (1999).
Without prior maceration, O-caffeoylserine amounted to approx. 0.03 nmol/mg
of mealybug fresh weight, representing 5.2% (in moles) of the nitrogenous com-
pounds analyzed (Table 1). Although it can be assumed that the procedure used
(i.e., soaking the mealybugs in 70% ethanol for three days at 10 C) is sufficient
to leach out a number of internal insect constituents, some arguments support
the fact that the method tends to dissolve surface rather than internal constituents
and that O-caffeoylserine should, in fact, amount to 0.03 nmol/mg on the body
surface of P. herreni females in the adult developmental stage. First, when mealy-
bugs were crushed, O-caffeoylserine amounted to about 9.5 nmol/mg of mealy-
bug fresh weight, i.e., approx. 916 times higher, being much more concentrated
than that obtained without maceration. Furthermore, when whole insects were
soaked in ethanol, the concentration of O-caffeoylserine reached a maximum at
three days and was three times higher than on day 1 (Figure 3). If the proce-
dure were efficient in leaching out internal insect constituents, the concentration
would increase much more than three times based on the internal concentration
of O-caffeoylserine mentioned above. Figure 5 shows that both parasitoid species

CASSAVA MEALYBUG KAIROMONE
2215
responded better to 0.03 nmol/ l of O-caffeoylserine. This concentration, equiv-
alent to 0.03 nmol/mg of mealybug fresh weight, is similar to that found when
mealybugs were soaked in 70% ethanol for three days at 10 C. Given the fact
that higher biological activity was obtained at this concentration and that it is the
maximum that can be extracted by soaking unmacerated insects in ethanol, we
suggest that O-caffeoylserine amounts 0.03 nmol/mg on the body surface of adult
P. herreni females.
The presence of free amino acids in the sample supports the idea that some in-
ternal constituents can be leached out as well (Table 1). However, this is more likely
due to honeydew secretions containing free amino acids that could stick to the body
surface of neighboring mealybugs, given their proximity under mass-rearing con-
ditions. O-Caffeoylserine was not found in the honeydew secretion of P. herreni.
Using several concentrations of synthetic O-caffeoylserine, it was confirmed
that this compound influences host-location behavior of two parasitoid species
of the cassava mealybug: A. coccois and A. vexans. Perhaps because of its gen-
eralist nature, this was more clearly shown with A. coccois due to the fact that
the dosage-dependent response was well observed, with higher biological activ-
ity at 0.03 nmol/ l (Figure 5A). Although no such response was observed for
A. vexans, O-caffeoylserine also was stimulatory at 0.03 nmol/ l (Figure 5B).
In conclusion, for both parasitoid species, 0.03 nmol/ l is the concentration of
O-caffeoylserine that induces more drummings and turns of wasps (approx. 40%
of relative response among the concentrations bioassayed, Figure 5). This con-
centration, equivalent to 0.03 nmol/mg, is similar, to what can be extracted from
the body surface of mealybugs, induces significant drummings and turns (approx.
40% of relative response as compared to the natural host, Figure 4), and indi-
cates that synthetic O-caffeoylserine is as active as the chemical isolated from the
body surface. All these results are consistent with O-caffeoylserine being a con-
tact kairomone that mediates host-location behavior by the two cassava mealybug
parasitoids.
Except for A. coccois at 0.03 nmol/ l of O-caffeoylserine, oviposition prob-
ing was not observed with any of the extracts bioassayed, regardless of parasitoid
species. It may be that the female wasp behavior for oviposition probing is influ-
enced by other factors such as color and texture of the mealybug body. In fact,
in preliminary experiments we observed that for both parasitoid species, texture
influenced wasp behavior for host location more than color (Calatayud, unpub-
lished data). Some wasps inserted their ovipositor inside the untreated cotton plug,
whereas oviposition probing was not observed with the treated cotton ball inside
the glass vial. This was probably due to the fact that the cotton plug was more
compact than the cotton ball, suggesting that a good combination between the
concentration of O-caffeoylserine and texture could enhance oviposition probings.
Unfortunately, increasing the pressed texture of the cotton ball was not possible
given its size.

2216
CALATAYUD ET AL.
O-Caffeoylserine is not provided directly by the host plant; it is either syn-
thesized by the insect itself or with the help of symbiotic bacteria. In fact, it was
detected in adult mealybugs reared on holidic diets at 14.9 nmol/mg fresh weight,
a higher level than in plant-reared mealybugs (about 9.5 nmol/mg), but not for ex-
tracts of host plant (Manihot esculenta) leaves. Nevertheless, enzymes commonly
involved in the synthesis of phenolic compounds in plants such as phenylalanine or
tyrosine ammonia-lyase were not evidenced in P. herreni (Calatayud, unpublished
data).
The physiological role of O-caffeoylserine in the cassava mealybug is not
clear. The caffeic acid moiety, as suggested by Hare et al. (1993) for O-caffeo-
yltyrosine, could be a precursor of the quinone compound involved in cuticle
sclerotization and tanning. The serine moiety could be a precursor of the bris-
tle (serine polymer) in the ovisacs. The fact that this phenolic compound is
present in higher concentration in eggs (at about 26.4 nmol/mg) than in adult
females suggests that it may also have a protective function (e.g., antibacterial
or antiviral activity) in the insect. All these hypotheses need to be demonstrated
explicitly.
Acknowledgments—We thank Juan Christian Hamon for his technical help and Trudy
Brekelbaum for her English corrections. This research was supported by a special project at CIAT,
co-financed by IRD, and a donation to CIAT from the French Ministry of Foreign Affairs.
REFERENCES
BELLOTTI, A. C., REYES, J. A., VARELA, A. M., and CASTILLO, J. 1983. El piojo harinoso (Phenacoccus
sp.) de la yuca; una de las plagas agr´ıcolas ma´s importantes del mundo. 27 pp. Centro Internacional
de Agricultura Tropical, Cali, Colombia.
DORN, B., MATTIACCI, L., BELLOTTI, A. C., and DORN, S. 2001. Host specificity and comparative
foraging behaviour of Aenasius vexans and Acerophagus coccois, two endoparasitoids of the
cassava mealybug. Entomol. Exp. Appl. In press.
GODFRAY, H. C. J. 1994. Parasitoids: Behavioral and Evolutionary Ecology. Princeton University Press,
Princeton, New Jersey.
HARE, J. D., MILLAR, J. G., and LUCK, R. F. 1993. A caffeic acid ester mediates host recognition by a
parasitic wasp. Naturwissenschaften 80:92–94.
¨
LE RU, B., IZIQUEL, Y., BIASSANGAMA, B., and KIYINDOU, A. 1991. Variations d’abondance et facteurs
de re´gulation de la cochenille farineuse du manioc Phenacoccus manihoti (Hom., Pseudococcidae)
cinq ans apre`s l’introduction d’Epidinocarsis lopezi (Hym., Encyrtidae) parasitoide ne´otropical
au Congo en 1982. Entomophaga 36:499–511.
LEWIS, W. J. and MARTIN, W. R., JR. 1990. Semiochemicals for use with parasitoids: Status and future.
J. Chem. Ecol. 16:3067–3089.
LIN, Z., NEAMATI, N., ZHAO, H., KIRYU, Y., TURPIN, J. A., ABERHAM, C., STREBEL, K., KOHN, K.,
WITVROUW, M., PANNECOUQUE, C., DEBYSER, Z., DE CLERCQ, E., RICE, W. G., POMMIER, Y.,
and BURKE, T. R., JR. 1999. Chicoric acid analogues as HIV-1 integrase inhibitors. J. Med. Chem.
42:1401–1414.

CASSAVA MEALYBUG KAIROMONE
2217
LUCK, R. F., PODOLER, H., and KFIR, R. 1982. Host selection and egg allocation behaviour by A. meli-
nus and A. lingnanensis: Comparison of two facultatively gregarious parasitoids. Ecol. Entomol.
7:397–408.
MILLAR, J. G. and HARE, J. D. 1993. Identification and synthesis of a kairomone inducing oviposition
by parasitoid Aphytis melinus from California red scale covers. J. Chem. Ecol. 19:1721–1736.
NEUENSCHWADER, P., HAMMOND, W.N.O., AJUONU, O., GADO, A., ECHENDU, N., BOKONON-GANTA,
A. H., ALLOMASSO, R., and OKON, I. 1990. Biological control of the cassava mealybug, Phenacoc-
cus manihoti (Hom., Pseudococcidae) by Epidinocarsis lopezi (Hym., Encyrtidae) in West Africa,
as influenced by climate and soil. Agric. Ecosyst. Environ. 32:39–55.
NORONHA, A. C. 1990. Insectos e acaros que atacam a cultura da mandioca. Empresa Brasileira de
Pesquisa Agropecuaria (EMBRAPA), Centro Nacional de Pesquisa de Mandioca e Fruticultura
(CNPMF), Cruz das Almas-BA, Brasil. 27 pp.
POWELL, W. 1992. Synomones and kairomones from plants and honeydew in susceptible and resistant
varieties: Their importance to natural enemies. in Proceedings of the XIX International Congress
of Entomology, 1992, Beijing, China.
RENUKAPPA, T., ROOS, G., KLAIBER, I., VOGLER, B., and KRAUS, W. 1999. Application of high-
performance liquid chromatography coupled to nuclear magnetic resonance spectrometry, mass
spectrometry and bioassay for the determination of active saponins from Bacopa monniera Wettst.
J. Chromatogr. A 847:109–116.
SELIGMANN, C. D. 1998. Influencia de la nutricio´n del piojo harinoso Phenacoccus herreni de la yuca
Manihot esculenta en el desarrollo de sus parasitoides Acerophagus coccois, Aenasius vexans,
Apoanagyrus diversicornis. MS thesis. University of Los Andes, Bogota´, Colombia, 109 pp.
¨
ROSE, U. S. R., ALBORN, H. T., MAKRANCZY, G., LEWIS, W. J., and TUMLINSON, J. H. 1997. Host
recognition by the specialist endoparasitoid Microplitis croceipes: Role of host and plant related
volatiles. J. Insect Behav. 10:313–329.
TAKABAYASHI, J. and TAKAHASHI, S. 1993. Role of the scale wax of Ceroplastes ceriferus Ander-
son (Hemiptera: Coccidae) as a defence against the parasitic wasp Anicetus ceroplastis Ishii
(Hymenoptera: Encyrtidae). J. Insect Behav. 6:107–115.
VINSON, S.B. 1991. Chemical signals used by parasitoids. Redia 74:15–42.