1Plants have developed various mechanisms that reduce their likelihood of being consumed or damaged by herbivores. Plant defence mechanisms can be categorized into direct defence, which include antixenosis and antibiosis (also called intrinsic mechanisms), and indirect defence (extrinsic mechanisms) (Price et al., 1980). Direct defensive mechanisms include morphological features (e.g. trichomes or hairs, thorns, waxes) and chemical adaptations (production of toxins, repellents, digestibility reducers) that have a direct negative impact on herbivore attack and survival. Earlier investigations and assessments have resulted in the incorporation of this knowledge into agricultural production practices for years, as demonstrated by the use of resistant crop varieties.

2Indirect defence mechanisms are related to those traits promoting the effectiveness of the natural enemies of herbivores; they involve the provision of alternative food (nectar), shelter (domatia) or information on herbivore presence and identity (semiochemicals). The efficacy of these natural enemies is sometimes so spectacular that they are referred to as the plant’s “bodyguards” (Dicke & Sabelis, 1988). Indirect defence mechanisms have been well documented over the past 20 years and this has led to the development of biological control practices in many crops. It is now accepted that there is a strong link between intrinsic and extrinsic defense mechanisms, thereby underscoring the importance of managing plant attributes from a tritrophic perspective (Thomas & Waage, 1996; Cortesero et al., 2000).

INTRINSIC MECHANISMS

3Painter (1951) categorized the intrinsic mechanisms of plant defense into three, namely

• antixenosis (or non-preference), which arises from the plant to deter or reduce initial colonization by insects. This mechanism can be due to biophysical or biochemical factors or a combination of both;
• antibiosis, which operates after the insects have colonized and have started utilizing the plant. This mechanism of resistance generally affects the insect’s growth, development, reproduction and survival. Chemicals and/or physical factors can be involved in antibiosis;
• tolerance, which refers to the resultant effects and not the mechanisms. This is a genetic trait of adaptation of a plant to insect infestation which prevents economic yield loss or lowering of the quality of the plant’s marketable product.

4With reference to cultivated cassava, Manihot esculenta, no total resistance resulting in a complete absence of mealybugs has been identified (Le Rü & Calatayud, 1994; Bellotti et al., 1999). Varietal differences in partial resistance to P. manihoti, including antixenosis, antibiosis and tolerance, have been observed, however (Le Rü & Tertuliano, 1993; Tertuliano et al., 1993). These results suggest that there are many resistance mechanisms operating in cultivated cassava, and that cassava resistance is of the horizontal type and is polygenic, as previously proposed for resistance to P. herreni by Bellotti and Kawano (1980).

5It is well documented that secondary compounds in addition to nutritional factors can play important roles in the resistance of plants to insect pests (Fraenkel, 1969; Kogan, 1977; Pickett et al., 1992). In Sternorrhyncha, numerous studies on aphids have illustrated the involvement of alkaloids (Smith, 1966; Dreyer et al., 1985; Wink & Witte, 1991), phenolic acids and flavonoids (Todd et al., 1971; Dreyer & Jones, 1981; Mc Foy & Dabrowski, 1984; Leszczynski et al., 1985) or cyanide compounds (Schoonhoven & Derksen-Koppers, 1976; Dreyer & Jones, 1981) in plant resistance. The role of secondary compounds (metabolites) is generally highly dependent on their localization within the plant tissues. Plant compounds strictly located in the mesophyll are thought to have a greater chance of playing a deterrent role during stylet penetration, an example of antixenotic resistance. On the other hand, if located in the phloem, these compounds will influence settling or feeding depending on their behavioural or metabolic effectiveness —a case of either antixenotic or antibiotic resistance (Givovich et al., 1992). As reported in Chapter 2 on host-plant selection, the secondary compounds found in M. esculenta are cyanogenic glucosides (mostly linamarin) as well as phenolic acids and flavonoids (Conn, 1980; Calatayud et al., 1994b). No alkaloid has been detected.

Antixenosis

6Linamarin, present on cassava phylloplane, is involved in initial host recognition by P. manihoti (see Chapter 2 for details). This compound is also present in the phloem sap, and plays a phagostimulation role to the insect (Calatayud et al., 1994a; Renard, 1999; Calatayud, 2000). These authors reported also that cassava varieties with higher levels of linamarin are generally more preferred by the mealybugs. In contrast, less preferred cassava genotypes have higher levels of phenolic acids in their cell wall structures which most likely inhibit the salivary enzyme activities during the penetration process of the stylets. This explains in part the delay in reaching the phloem by P. manihoti stylets on such cassava genotypes (Calatayud et al., 1994a).

7Antixenosis is generally regulated by a balance between attractant and repellant compounds from the plants (Schoonhoven et al., 1998). In cassava, the balance between linamarin and phenolic acid levels in the leaves seems to be a good indicator in predicting mealybug preference for the plant variety.

Antibiosis

8The phenomenon of cyanogenesis which occurs in cassava was formerly assumed to provide protection to the plant against phytophagous insects (Hruska, 1988). Following mechanical damage to the tissues by insects, hydrocyanic acid (HCN) is released as a result of linamarin coming into contact with linamarase (Conn, 1980). The location of linamarase in cassava leaves is a matter of debate. Mkpong and colleagues (1990) showed that this enzyme has an apoplastic location, but later Pancoro and Hughes (1992) found the enzyme to be located principally in laticifers (latex vessels). Whatever the location of this enzyme, the substrate (linamarin) is present in other tissue locations, indicating that the substrate and enzyme are compartmentalized and come into contact after tissue damage.

9Cyanogenesis in cassava is not a mechanism that protects the plant against mealybugs, however, because it is probably not induced by insect infestation. The penetration process of the insect’s stylets induces little cellular damage (Calatayud et al., 1994a; 2001a); it is therefore unlikely that linamarase can come into contact with linamarin for cyanogenesis to occur after stylet penetration. Studies on P. manihoti report that the linamarin ingested is probably hydrolysed by the bacteria occurring freely in the midgut, which would explain the appearance of free cyanide found in the honeydew of mealybugs (Calatayud et al., 1994b). Free cyanide therefore does not come from cyanogenesis. Moreover, free cyanide has been shown to be not very toxic to this insect, due to its efficient excretion system (Calatayud et al., 1994b) and/or to an efficient detoxification mechanism by the gut bacteria. No relationship has been established between the levels of primary compounds (amino acids and sugars) of leaves and antibiosis to P. manihoti (Tertuliano & Le Rü, 1992), suggesting that other biochemical factors may be involved in antibiosis resistance of cassava.

10Rutin (a flavonoid glycoside, Figure 5.1) has been shown to be translocated by the phloem sap of cassava, and is therefore consumed by P. manihoti (Calatayud 1993; Calatayud et al., 1994b). No linear dose effect was found between rutin content and the antibiosis level in a study by Calatayud and coworkers (1994b). Poinsettia (Euphorbia pulcherrina Wild., Euphorbiaceae) and talinum (Talinum triangularae Jacq., Portulacaceae) are optional hosts of P. manihoti (Calatayud et al., 1994a), the latter being infested only if high infestations on adjacent cassava plants occur (F. Schulthess, ICIPE, personal communication). The two plants display marked differences in antibiosis when tested in the laboratory (Tertuliano et al., 1993). The rutin content in the leaves is 13-fold higher in ‘resistant’ poinsettia (rutin at 16 mg/g; r_c1 = 0.038) than in ‘susceptible’ talinum (rutin at 1.2 mg/g; r_c = 0.150) (Calatayud et al., 1994b). In addition, rutin levels are sensitive to the presence of the mealybug, regardless of the plant genotype (Calatayud et al., 1994b). This observation has been confirmed by independent experiments involving increased levels of infestation (P.-A. Calatayud, unpublished data).

11The biological activity of rutin towards P. manihoti has been studied using artificial diets (Calatayud, 2000). Introducing commercial rutin at various concentrations into a liquid diet has shown that the presence of this compound in the diet affects mealybug growth and development. The authors of this book, however, suggest that rutin seems to play an anti-nutritive function in the insect through phagodeterrency, thus delaying development to the adult, rather than having a toxic effect (P.-A. Calatayud and B. Le Rü, unpublished data). The appearance of quercetin (the rutin aglycone) in the honeydew of P. manihoti (Calatayud et al., 1994b) indicates that deglycosylation of rutin probably occurs in the insect and suggests that this hydrolysis product (quercetin) should be more toxic to the insect’s physiology than the quercetin glycoside, i.e. rutin itself. Metabolism of phenolic compounds has also been proposed to occur in the aphid Macrosiphum rosae (L.) (Peng & Miles, 1991). It is probable that the mealybug’s intestinal flora is involved in the hydrolysis of rutin into quercetin by the action of rutin glycosidase, as reported for the bacterium Pseudomonas viridiflava (Hendson et al., 1992). Therefore, the fortuitous gut bacteria ingested by the insect play a negative role in mealybug–cassava interactions, being more involved in the plant’s defence mechanism rather than protecting their insect host. Additional research is needed to confirm this hypothesis.

12The above studies all indicate the probable participation of rutin in antibiosis resistance of cassava to P. manihoti, or at least its linkage to an induced (defensive?) reaction of cassava towards the insect. This defensive reaction is manifested by an increase in the total phenolic compounds (mostly flavonoids, including rutin) after mealybug infestation (Calatayud et al., 1992; Calatayud et al., 1994b; 1994c; Tertuliano et al., 1999). Such a defensive reaction of cassava is also induced after infection by cassava bacterial blight (Xanthomonas campestris pv. manihoti), and is more intense in cassava resistant to the bacterium than in susceptible varieties (Kpemoua et al., 1996; P.-A. Calatayud, unpublished data). An increase in rutin level in leaves appears to be a good indicator or prediction as to whether a cassava variety will be resistant or susceptible to insect infestation and/or bacterial infection.

13The involvement of flavonoids in the resistance of plants to phloem-feeders has been extensively documented for aphids (Todd et al., 1971; Dreyer & Jones, 1981; McFoy & Dabrowski, 1984). Rutin is an allelochemical that affects the growth, development, reproduction and metabolism of insect species other than P. manihoti (Beck & Reese, 1976). With reference to Schizaphis graminum (Sternorrhyncha: Aphididae), Dreyer and Jones (1981) suggest that in phloem-feeding insects, secondary compounds probably not only affect the insects themselves but also their symbionts. Similarly, the physiology of the endosymbiotic bacteria harboured by P. manihoti (see Chapter 3) should likely be disturbed by ingested and metabolised rutin. Additional data are needed to confirm this hypothesis.

Tolerance

14Cock (1979) reported decades ago that cassava is relatively tolerant to pest attack. He suggested that tolerance is due to the extensive production of leaves during the growing season that compensates for the reduction in leaf production caused by the pest during the dry season, when the leaf area is already reduced due to drought. The tolerance of cassava to P. manihoti has been measured by quantifying morphological and physiological changes in plant growth after infestation (Le Rü & Tertuliano, 1993). All the cassava varieties studied were generally tolerant to mealybug infestation, but with different degrees of tolerance. Some varieties exhibited both a relatively high tolerance and a strong antibiotic resistance, suggesting that such varieties could be selected for a breeding strategy for mealybug resistance.

Total resistance

15Although no total resistance to mealybugs has been evidenced in cultivated cassava, such resistance should be detected in wild cassava species or closely related species (Calatayud & Mùnera, 2002). In this context, 26 wild cassava species in addition to Jatropha gossypiifolia L. (Euphorbiaceae) were artificially infested with P. herreni under laboratory conditions. Only J. gossypiifolia exhibited a strong toxicity to the insect: after 48 hours of infestation, this plant showed 100% insect mortality. It was observed that mortality is not only due to the presence of glandular trichomes on the leaf surface, but also due to a toxicity most likely residing in the mesophyll cells. Moreover, this plant species appears to be highly phagodeterrent to P. herreni, as the larvae remain extremely mobile after infestation of the leaves (P.-A. Calatayud, unpublished observations). Therefore, J. gossypiifolia displays a total resistance to P. herreni by using both antixenosis and antibiosis mechanisms, resulting in a total mortality of mealybugs 48 hours after infestation. The resistance of J. gossypiifolia to P. manihoti has not been studied.

16Jatropha gossypiifolia was initially confused with a wild species of cassava, and in fact it is genetically close to the Manihot genus (M. Fregene, CIAT, personal communication). The plant reproduces well by seeds, possibly making it useful in a breeding programme to obtain hybrids (with cassava varieties) totally resistant to the cassava mealybug, P. herreni.

Influence of antixenosis and antibiosis on natural enemies

17Natural enemies of phytophagous insects function and develop in a multitrophic context (Price et al., 1980; van Emden, 1991). Souissi and Le Rü (1997) studied the relationship between the cassava plant and the parasitoid Apoanagyrus (Epidinocarsis) lopezi De Santis (Hymenoptera: Encyrtidae) (Figure 5.2a) used in the biological control of P. manihoti in Africa. They showed that cassava with a high level of antibiosis resistance has a deleterious effect on A. lopezi survival and development. These authors later observed that the size of this parasitoid species negatively correlates with plant resistance (Souissi & Le Rü, 1998). Apoanagyrus lopezi appears to perform better if cassava cultivars are selected for their strong antixenosis (non-preference) but low antibiosis characteristics.

18Other studies have revealed that most life-history parameters of the generalist ladybird predator Exochomus flaviventris Mader (Coleoptera: Coccinellidae) (Figure 5.2b), the most active predator of P. manihoti in Central Africa (Fabres & Kiyindou, 1985; Iziquel & Le Rü, 1989), are greatly influenced by the host plant, showing that mealybugs feeding on higher-antibiosis cassava are less suitable prey for the coccinellid (Le Rü & Mitsipa, 2000). However, the voracity of E. flaviventris in this study was high on the cassava variety with the highest antibiosis resistance. This suggests that such a cassava variety may synergistically interact with the coccinellids to provide a significant level of mealybug control. The conclusions drawn are that antibiosis in cassava significantly affects the life-history parameters of the coccinellids without modifying their fitness, and thus provides better cassava mealybug control (Le Rü & Mitsipa, 2002).

Influence of abiotic factors on direct plant defences

19While drought stress positively affects the biological performance of the mealybugs via the superior nutrition provided (see Chapter 4), water limitation negatively influences the cassava’s resistance to the insect. During the long dry season when the P. manihoti population increases and spreads, the linamarin content in the leaves increases, producing cassava plants that are phagostimulant to the mealybug. Simultaneaously, the defensive response of cassava to P. manihoti infestation is less important during the dry season, explaining in part the changes in P. manihoti populations during this period (Calatayud et al., 1994c) and suggesting the occurrence of a temporary alteration in partial plant resistance against the insect.

20Infestation by P. manihoti is rarely observed in forest regions and in soils rich in organic elements (Matile-Ferrero, 1978; Nwanze et al., 1979; Neuenschwander et al., 1989; 1990). The impact of improved soil fertility in diminishing cassava infestation by P. manihoti has been reported also by Schulthess and colleagues (1997), and this finding has been supported by the evidence of Tertuliano and colleagues (1999) that mulch and manure are the best fertilizers in enhancing cassava resistance to P. manihoti, as shown by a higher defensive response (i.e. the increase of rutin level in leaves after infestation).

EXTRINSIC MECHANISMS

21Plant and host-derived volatiles play a fundamental role both in herbivore habitat location and herbivore location by carnivores. The informational value of stimuli used by carnivores in the location of herbivores depends on two inversely correlated characteristics: detectability (i.e. degree of stimuli reception) and reliability in indicating herbivore presence and suitability. Volatile chemicals released in relatively large amounts from the food plant of the host provide long-range olfactory information to the carnivores; these compounds are easily detectable and mainly used in host habitat location. Volatile chemicals originating from the host itself and from its by-products and released in small amounts, appear to operate at short range and are involved in host location once the carnivore reaches the microhabitat of the herbivore (Vet & Dicke, 1992).

22Some authors consider the release of chemical compounds that stimulate the attraction of the herbivore’s natural enemies as an indirect mechanism of plant defence (Price et al., 1980; Dicke et al., 1990; Takabayashi et al., 1995). Others propose this process to be a kind of antixenosis resistance of the plant against herbivores (Storer & van Emden, 1995; Petterson et al., 1996). This indirect defence mechanism or antixenosis resistance can be constitutive (Price et al., 1980) or can be induced by herbivory (Dicke et al., 1990; Takabayashi et al., 1994). It is well documented that many natural enemies discriminate between volatiles produced by uninfested and infested plants (Turlings et al., 1990; Reed et al., 1995), by different plant species (Elzen et al., 1986; Dicke et al., 1990; Turlings et al., 1990; Gouinguene et al., 2001), and by different cultivars of the same plant species infested by the same herbivore (Elzen et al., 1986; Dicke et al., 1990; Takabayashi et al., 1991, 1994; Gouinguene et al., 2001). These complex behavioural responses indicate that olfactory information may be specific for different herbivore–plant interactions, suggesting that the quality and quantity of induced infochemicals may vary with the plant species or variety as well as with the herbivore. According to Takabayashi and colleagues (1991), the variation with the plant appears to be more important than with the herbivore.

Extrinsic mechanisms involving parasitoids

23The responses of female parasitoids to volatile chemicals emitted by cassava infested by mealybugs have been investigated with the cassava – P. manihoti – Apoanagyrus lopezi model (Nadel & van Alphen, 1987; van Baaren & Nénon, 1996; Souissi & Le Rü, 1997), and with the cassava – P. herreni – Apoanagyrus diversicornis Howard / Aenasius vexans Kerrich / Acerophagus coccois Smith (Hymenoptera: Encyrtidae) model; the last three are the endoparasitoid species used to control P. herreni in South America (Bertschy et al., 1997; 2001) (Figure 5.3). All these authors have clearly demonstrated that mealybug-infested cassava plants chemically mediate attraction of the parasitoid species.

24Souissi and co-workers (1998) have shown that cassava plants damaged by P. manihoti are the main source of volatiles that attract females of A. lopezi to the microhabitat of its host. There is evidence that unparasitized mealybugs are more attractive than parasitized mealybugs, thus preventing the risk of superparasitism. Infested plants with unparasitized mealybugs attract more female parasitoids than plants infested with parasitized mealybugs. The data of these authors suggests that the source of attractants for female parasitoids at long distance originates from the interactions between the plant and feeding by the mealybug. These attractants are detected at longer distances than those produced by undamaged plants or kairomones . chemicals that convey information between two species and for which the receiver alone benefits from the exchange of information – left by unparasitized hosts, thus guiding the parasitoids into areas that harbour suitable hosts.

25Bertschy and colleagues (1997) showed that A. coccois, a more generalist feeder than A. diversicornis and A. vexans, does not distinguish between healthy or infested leaves, suggesting that A. coccois may respond to general plant odour, whereas specialized parasitoids (A. diversicornis and A. vexans) respond more specifically to odours associated with mealybug-infested plants. After landing on an infested plant, the mealybugderived chemicals can play a role in host location. Calatayud and co-workers (2001b) have shown that A. coccois and A. vexans use a contact kairomone identified as O-caffeoylserine from the body surface of P. herreni as a host-location stimulant.

Extrinsic mechanisms involving predators

26The importance of chemical signals emitted by damaged plants to the ability of natural enemies to detect hosts or prey has been extensively studied with reference to insect parasitoids and predatory mites (Vet & Dicke, 1992, Dicke & Vet, 1999; Cortesero et al., 2000). However, it has been poorly documented in insect predators such as the coccinellids, although there is some evidence of their use in the case of the latter (Obata, 1986; Zhu et al., 1999). The response of experienced gravid females of the generalist ladybird E. flaviventris to odours from the cassava–mealybug (P. manihoti) complex has been investigated by Le Rü and Makaya Makosso (2001). The odour of cassava plants alone was not attractive, but in dual-choice tests, the mealybug-infested plants were preferred to mealybugs alone and mealybug-damaged plants, and were the major sources of volatiles that attracted the female coccinellids to the microhabitat of their prey.

27The authors emphasize that the emission of volatiles does not appear to be limited to the infested parts of the plant, but occurs systematically throughout the plant. The release of such induced compounds might be elicited by a factor present in the mealybug’s saliva or by the feeding process of the insect. It has been revealed also that E. flaviventris females use herbivore-induced plant volatiles during foraging, and detect via olfaction the presence of conspecific gravid females and parasitized prey, thus assessing patch suitability from a distance. The presence of conspecific coccinellid larvae or adult males did not modify the attractiveness of the mealybuginfested plants, but the E. flaviventris females showed more preference for mealybuginfested plants with or without conspecific males than for mealybug-infested plants with conspecific predator females (alone or with conspecific males). Moreover, the female predators preferred the plant infested with unparasitized mealybugs over plants infested with parasitized mealybugs. These results confirm that ladybirds, which are very active predators with a full set of sensory organs, are capable of detecting small patches of mealybugs even when the pest is scarce in the habitat.

28The involvement of plants in herbivore finding by the generalist coccinellid shows remarkable similarities with those of the specialized parasitoids. There are great differences in the foraging behaviour of the parasitoid and predator, however: the specialized parasitoid A. lopezi shows an innate response to the systemically released cassava volatiles induced by mealybug infestation, while the generalist predator needs first to learn about them (J. P. Makaya Makosso, unpublished observation). This behaviour provides the parasitoid with a great advantage over the predator, as it can locate infested cassava immediately after emergence, whereas the female coccinellid needs at least 8 days before it can recognize cassava infestation.

Influence of abiotic factors on indirect plant defences

29Indirect plant defences are thought to be modulated by environmental factors, however the mechanism of this is poorly understood (Cortesero et al., 2000). Recently Koumba (2000) investigated the role of water stress and soil nitrogen on cassava’s ability to attract both parasitoid (A. lopezi) and coccinellid (E. flaviventris). For both natural enemies, infested cassava under water limitation was preferred over infested cassava with a good supply of water, and both P. manihoti natural enemies preferred the odour of infested cassava supplied with nitrogen over that of infested cassava alone. Previous studies on tritrophic interactions have reported an increase of natural enemy densities on plants supplied with nitrogen (Hagen et al., 1971; Fox et al., 1990). These results suggest that both water stress and a good nitrogen supply increase the amount of synomone—chemicals that convey information between two species and for which both the receiver and the signaler benefit from the exchange of information—released by the infested plant and confirm that abiotic factors have a strong influence on the plant’s ability to attract natural enemies. However, water stress appeared to enhance the encapsulation—immune response of mealybugs to parasitoids eggs or larvae—and to effect the success of parasitism and, depending on the parasitoid species, the parasitoid development as well. This has been observed for P. herreni and its associated parasitoids (Calatayud et al., 2002c).

30These studies also demonstrate that plants with good nitrogen supply have a beneficial impact on natural enemies. For example, soil fertility exhibits strong tritrophiclevel effects via changes in the size of cassava mealybugs and the sex ratio of the parasitoid A. lopezi (Schulthess et al., 1997). This corroborates the results of van Dijken and colleagues (1989) on sex allocation according to size of the mealybug. A strategy put forward by Schulthess and colleagues (1987; 1997) to enhance biological control consists of selecting for cassava with high leaf nitrogen in order to increase the size of cassava mealybugs and thereby influence a female-biased sex ratio in A. lopezi.