Evolution of dry food use as a preadaptation to becoming a stored food pest

It has been hypothesized that the human cultivation of grains in tropical/subtropical climates promoted the evolution of pests that rely on stored grains/legumes for some part of their life cycle (Cotton, Reference Cotton1956). However, recent studies suggest that changes in climate and in host plant assemblages associated with different climates were a prerequisite for the evolution of host utilization in insects that are now considered to be stored grain pests (Tuda et al., Reference Tuda, Ronn, Buranapanichpan, Wasano and Arnqvist2006b). Furthermore, there is evidence that the species split between stored bean pests and that the non-pest sister species occurred long before the human cultivation of beans (Alvarez et al., Reference Alvarez, McKey, Hossaert-McKey, Born, Mercier and Benrey2005).

The majority of seed beetles from the subfamily Bruchinae (Coleoptera: Chrysomelidae) utilize the seeds of the bean family (Fabaceae) as larval hosts (Johnson, Reference Johnson, Polhill and Raven1981). After hatching from eggs deposited on or near seeds by their mothers, the larvae feed and pupate inside the seeds. Partly because of this life cycle, some members of the subfamily are considered to be serious pests of leguminous plants. Utilization of dry beans for larval hosts is most likely, an ancestral trait in bruchine seed beetles (Tuda et al., Reference Tuda, Ronn, Buranapanichpan, Wasano and Arnqvist2006b). Interestingly though, some bruchines utilize only immature, soft seeds (Southgate, Reference Southgate1979; Decelle, Reference Decelle and Labeyrie1981; Alvarez et al., Reference Alvarez, McKey, Hossaert-McKey, Born, Mercier and Benrey2005; Tuda et al., Reference Tuda, Ronn, Buranapanichpan, Wasano and Arnqvist2006b; M. Tuda, unpublished). It is unlikely that the evolution of the use of dry seeds was driven by the use of cultivated legumes. Instead, this trait is likely a preadaptation for becoming a pest of stored beans (Watanabe, Reference Watanabe1985; Tuda et al., Reference Tuda, Ronn, Buranapanichpan, Wasano and Arnqvist2006b). We hypothesize that this unique feeding trait may be explained by climate (long, arid seasons) and the associated plant taxa, not their cultivation status. The long period of time in which seeds are available may also allow multivoltinism (Southgate, Reference Southgate and Labeyrie1981; Ishihara, Reference Ishihara1998; Alvarez et al., Reference Alvarez, McKey, Hossaert-McKey, Born, Mercier and Benrey2005). Our recent study is the first to test competing hypotheses about the evolution of dried food pests (Tuda et al., Reference Tuda, Ronn, Buranapanichpan, Wasano and Arnqvist2006b). Our data suggested that the length of the dry period contributed more to the evolution of a life cycle that incorporated the use of dry food than temperature per se. This is probably due to the long period of food availability associated with dry season (Tuda et al., Reference Tuda, Ronn, Buranapanichpan, Wasano and Arnqvist2006b). Although this host-utilization phenotype may be ancestral, it is not strictly phylogenetically constrained as it is observed sporadically throughout the bruchine subfamily (M. Tuda, unpublished). Recently, several closely related species of bruchine beetles that feed on either dry seeds or fresh seeds were identified (Tuda, Reference Tuda2003; Tuda & Morimoto, Reference Tuda and Morimoto2004). A comparison of these species affords an opportunity to test the factor(s) driving the evolution of host-utilization traits that are associated with stored bean pests (Tuda, Reference Tuda2007).

Many pest insects of stored products were already preadapted to exploit the small quantities of naturally stored dry or rotten plant and animal materials, including dry seeds, fruits, pollen, dead insects, and hairs or feathers in the nests and webs of birds, mammals, spiders and insects (Linsley, Reference Linsley1944). Most of these insects are scavengers or predaceous, unlike bruchines. Therefore, the climate effect on the preadaptation suggested for bruchine beetles may not be applied to these insects. Furthermore, besides these natural causes, humans have likely had a considerable influence on the evolutionary processes. For example, the storage of large amounts of grain for long periods of time, particularly during World War I and II, may have promoted the adaptation of potential pests to storage conditions.

Adaptation to arid climates

The tolerance for changes in temperature and humidity varies among the developmental stages of insects, such as Drosophila (Krebs & Loeschcke, Reference Krebs and Loeschcke1995; Gilchrist et al., Reference Gilchrist, Huey and Partridge1997), the kelp fly (Klok & Chown, Reference Klok and Chown2001) and Tribolium, a pest of stored, dried foods (Oosthuizen, Reference Oosthuizen1935; Mahroof et al., Reference Mahroof, Subramanyam, Throne and Menon2003). As such, holometabolous insects and amphibians experience different habitat climates at each life-history stage. The egg and adult stages tend to be exposed to the ambient environment and may be tolerant of severe climate extremes (Howe & Currie, Reference Howe and Currie1964; Bale & Hayward, Reference Bale and Hayward2010). A number of behavioral and physiological adaptations to aridity have been noted in these life stages. For example, some bruchines lay their eggs in layers or cover them with fecal material. This prevents the eggs from desiccation, as well as from egg parasitoids (Teran, Reference Teran1962; Prevett, Reference Prevett1966; Janzen, Reference Janzen1971; Mitchell, Reference Mitchell1977; Delobel et al., Reference Delobel, Couturier, Kahn and Nilsson1995; Tuda, Reference Tuda2007). The eggs are also tolerant to a wide range of both humidity and temperature (Howe & Currie, Reference Howe and Currie1964). Adults and larvae have a greater proportion of water in the bound form than the pupae (Kagoshima, Ishibashi, Tuda, & Iwaya-Inoue, unpublished). Most bruchine beetles enter diapause (reproductive quiescence) as adults (Utida, Reference Utida1954; Sano, Reference Sano1967; Ouedraogo et al., Reference Ouedraogo, Monge and Huignard1991; Amevoin et al., Reference Amevoin, Glitho, Monge and Huignard2005; Appleby & Credland, Reference Appleby and Credland2007; review by Bell, Reference Bell1994), larvae (Donahaye et al., Reference Donahaye, Navarro and Calderon1966; Shimada & Ishihara, Reference Shimada and Ishihara1991) or both (Kurota, Reference Kurota2004). This phenomenon has also been reported in other beetle species that feed on dry, stored foods (review by Bell, Reference Bell1994). The absence of diapause during the pupal stage is common among beetles, but not in other orders (Leather et al., Reference Leather, Walters and Bale1993). This suggests that the physiological tolerance of beetles to extreme temperatures differs among the developmental stages (Mahroof et al., Reference Mahroof, Subramanyam, Throne and Menon2003). This difference may, in part, be a by-product of morphological reorganization, with little or no evolutionary significance (Bowler & Terblanche, Reference Bowler and Terblanche2008). However, there is evidence that the expression of at least one of the genes encoding heat-shock proteins is down-regulated during the pupal (and late larval) stages in the stored pest beetle, Tribolium castaneum (Mahroof et al., Reference Mahroof, Zhu, Neven, Subramanyam and Bai2005).

A classic physiological study by Robinson (Reference Robinson1928) revealed that stored grain pest beetles retain disproportionately high levels of water in their body relative to the dried grains they feed on (fig. 1). The same is true for bruchine beetles. For example, larvae of Callosobruchus beetles feed on dry beans, whereas those of Bruchus rufimanus feed only on immature beans. Despite the difference in water content of the food source, both species have a similar proportion of body water (fig. 1). Furthermore, in these and other stored-product pests, the majority of body water is conserved in the bound form to minimize its loss (fig. 2) (Robinson, Reference Robinson1928). The cells of all living tissues contain millions of minute colloidal particles, which attract and bind water as a film on their surfaces. The water, thus bound, is held by the colloidal bodies with a powerful force (Robinson, Reference Robinson1928). By conserving water in this less mobile form, the insects gain resistance against both dry and cold environments (Robinson, Reference Robinson1927; Block, Reference Block1996). Similarly, heat-resistant individuals of C. chinensis are characterized by longer adult life duration in very dry air (Kiyoku, Reference Kiyoku1962). This ‘side effect’ of adaptation to arid environments is seen in Callosobruchus beetles. In some species, the use of dry beans appears to have promoted a habitat shift (or expansion) from subtropical to temperate regions (Tuda et al., Reference Tuda, Ronn, Buranapanichpan, Wasano and Arnqvist2006b). Furthermore, the mechanisms proposed for insect mortality at high temperatures are the same as those for low temperatures (Tauber & Tauber, Reference Tauber and Tauber1986; Fields, Reference Fields1992). These include changes in membrane lipid property (Gibbs, Reference Gibbs1998; Patel et al., Reference Patel, Nelson and Gibbs2001), rate imbalances (Hochachka & Somero, Reference Hochachka and Somero1984; Fields, Reference Fields1992), perturbation of ionic activities (Cloudsley-Thompson, Reference Cloudsley-Thompson1962; Mullins, Reference Mullins, Kerkut and Gilbert1985) and desiccation (Cloudsley-Thompson, Reference Cloudsley-Thompson1962; Fields, Reference Fields1992).

Fig. 1. Relationship between the water content of insects and their food (only Coleoptera are shown, modified from Robinson Reference Robinson1928, with additional data). The examples on the upper side of the broken line indicate that the water content of the insects is greater than that of the food. Granary weevil, rice weevil (Curculionidae) and Callosobruchus beetles (Chrysomelidae: Bruchinae) are stored seed/grain pests, whereas the Colorado potato beetle (Chrysomelidae), locust borer (Cerambycidae), white grub (Scarabaeidae) (Robinson, Reference Robinson1928) and broad bean beetle, Bruchus rufimanus (Chrysomelidae: Bruchinae) (Kagoshima & Tuda, unpublished data) are pests of fresh vegetables and trees, at one or more life stages. ●, Callosobruchus chinensis (Ishii, Reference Ishii1952); ○, C. chinensis; ∆, C. maculatus; □, C. analis; ▪, C. phaseoli; +, Zabrotes subfasciatus (Utida, Reference Utida1966). The data were collected from the adults of all species except for the white grub (larvae) (Robinson, Reference Robinson1928).

Fig. 2. Relationship between total water content and the percentage of water bound by colloids (only Coleoptera are shown, modified from Robinson, Reference Robinson1928). The species that live on dry food and have a small percentage of water in their own tissues retain a large percentage of water in the bound form. The data were collected from the same developmental stages as in fig. 1 for each species.

Behavioral adaptation to aridity is observed in C. maculatus. Females mate more frequently under water deficiency, presumably because they can obtain a water source from male ejaculates (Edvardsson, Reference Edvardsson2007). They also prefer drier beans for oviposition (Hudaib et al., Reference Hudaib, Hayes, Brown and Eady2010).

Acclimation and adaptation to high temperatures

Mellanby (Reference Mellanby1954) may have been the first author to demonstrate thermal acclimation at high temperatures in insects. Mealworm, Tenebrio molitor, larvae that were acclimated to temperatures of 37 and 30°C died at 44 and 42°C, respectively, supporting the ‘beneficial acclimation’ hypothesis (Leroi et al., Reference Leroi, Bennett and Lenski1994). Using C. chinensis, Kiyoku (Reference Kiyoku1960) observed acclimation and adaptation to high temperature (35°C) in support of this hypothesis. However, the ‘beneficial acclimation’ hypothesis has been challenged in the latest decade. For example, thermal adaptation (or acclimatization) to a habitat was examined by exposing the bruchine beetle Stator limbatus to a range of temperatures (Stillwell & Fox, Reference Stillwell and Fox2005). Exposure to (and acclimation to) high temperatures during the immature stages of several species of bruchines leads to a general reduction in fecundity, irrespective of the tested oviposition temperature (Ishikura, Reference Ishikura1939; El-Sawaf, Reference El-Sawaf1956; Stillwell & Fox, Reference Stillwell and Fox2005). This reduction in fecundity seemingly supports ‘colder is better’ hypothesis (Huey et al., Reference Huey, Berrigan, Gilchrist and Herron1999), which is induced by a reduction in body size as is typical in ectothermic organisms (Atkinson, Reference Atkinson1994). Overall, either ‘colder is better’ hypothesis or ‘optimal developmental temperature’ hypothesis (Cohet & David, Reference Cohet and David1978; Huey et al., Reference Huey, Berrigan, Gilchrist and Herron1999 or ‘deleterious acclimation’ hypothesis (Loeschcke & Hoffmann, Reference Loeschcke and Hoffmann2002; Wilson & Franklin, Reference Wilson and Franklin2002) seems to be supported both in S. limbatus (Stillwell & Fox, Reference Stillwell and Fox2005) and Zabrotes subfasciatus (Howe & Currie, Reference Howe and Currie1964). Acclimation of other insects that feed on stored products depends on the range of the highest temperatures, and this is ascribed to thermal effect on biochemical process (Fields, Reference Fields1992). Survival at very high, sublethal temperatures (40–50°C) may be improved by acclimation at high temperatures. However, survival at extremely high, lethal temperatures (55–80°C) is not affected by acclimation. Within this temperature range (56–60°C), denaturation and inactivation of pyruvate kinase, a key enzyme in glycolysis, is induced and is likely responsible for mortality (Hochachka & Somero, Reference Hochachka and Somero1984).

Effect of climate and the host plants on polyphenism, population density and competitive/predatory interactions

Climate does not act alone on phytophagous insects. The host plants also alter the fitness of insect herbivores in a predictable manner. By selecting different host plants, C. maculatus is able to survive over a wider range of temperatures and humidity than is optimal in a limited range of climate (Howe & Currie, Reference Howe and Currie1964). Stillwell et al. (Reference Stillwell, Wallin and Fox2007b) noted an increase in fitness at high temperatures for traits associated with body size (i.e. growth rate, fecundity and body size itself) but not for larval survival or development time. Temperature has a significant effect on the success of small males during intraspecific competition for mates (Moya-Larano et al., Reference Moya-Larano, El-Sayyid and Fox2007). Both temperature and the host plant type affect the size of the nuptial gift, as seminal fluids (ejaculates), offered by the males (Fox et al., Reference Fox, Stillwell, Wallin and Hitchcock2006). Furthermore, it appears that there is an interaction effect between these two variables. Variation in the body size of S. limbatus throughout its range was attributed to differences in host plant seed size, moisture (humidity) and seasonality (variation in humidity, precipitation and temperature), but not to mean temperature (Stillwell et al., Reference Stillwell, Morse and Fox2007a). Moisture also partially explains the geographic variation in sexual dimorphism (Stillwell et al., Reference Stillwell, Morse and Fox2007a). Conversely, in C. maculatus, experiments at a constant temperature suggest that higher temperatures induce greater sexual dimorphism in body size (Stillwell & Fox, Reference Stillwell and Fox2007). This is probably caused by sexual conflicts related to fitness optimization strategies, which may increase population fitness (Rankin & Arnqvist, Reference Rankin and Arnqvist2008; Arnqvist & Tuda, Reference Arnqvist and Tuda2010). Early-emerging (smaller) males are more likely to find mates, particularly if the females are ready to mate soon after emergence (Rönn et al., Reference Rönn, Katvala and Arnqvist2008) and are reluctant, or refuse, to mate more than once (Edvardsson & Tregenza, Reference Edvardsson and Tregenza2005; Sakurai & Kasuya, Reference Sakurai and Kasuya2008). However, larger females tend to take longer to develop and are more fecund and, thus, fitter (Yanagi & Miyatake, Reference Yanagi and Miyatake2002; Vamosi, Reference Vamosi2005; Vamosi & Lesack, Reference Vamosi and Lesack2007), especially when resource replacement cycle is long (Sibly et al., Reference Sibly, Smith and Moller1991).

The optimal temperature for rapid development in insects tends to be higher than the optimal temperature for survival and reproduction (Howe & Currie, Reference Howe and Currie1964; Braby & Jones, Reference Braby and Jones1994). The optimal humidity is typically high for development, survival and reproduction. However, very high humidity levels enhance fungus development and increase mortality in immature beetles. The effects of temperature/humidity on populations and/or species should, therefore, be compared on lifetime fitness (Deutsch et al., Reference Deutsch, Tewksbury, Huey, Sheldon, Ghalambor, Haak and Martin2008; Yanagi & Tuda, Reference Yanagi and Tuda2010).

There are two adult morphs, the flight form (or active form in reproductive diapause) and non-flight form, in some bruchine seed beetles. High or increased temperature, long thermoperiods, extreme day-lengths, high seed water content and high larval density induce the flight form of C. maculatus (Utida, Reference Utida1954, Reference Utida1965, Reference Utida1969; Sano, Reference Sano1967; Sano-Fujii, Reference Sano-Fujii1984; Ouedraogo et al., Reference Ouedraogo, Monge and Huignard1991). The flight form is partially heritable (Caswell, Reference Caswell1960; Sano-Fujii, Reference Sano-Fujii1986). These environmental factors also have a similar effect on induction of the active morph in other species congeneric to C. maculatus (temperature effect in C. subinnotatus (Appleby & Credland, Reference Appleby and Credland2007); density effect in C. subinnotatus and C. rhodesianus (Amevoin et al., Reference Amevoin, Glitho, Monge and Huignard2005; Appleby & Credland, Reference Appleby and Credland2007); and seed water content in C. chinensis (Nahdy et al., Reference Nahdy, Silim and Ellis1999)).

Most phytophagous arthropods engage in predator-prey or parasitoid-host interactions in nature. Predictions regarding how climate change will alter these interactions are crucial for conservation of biological diversity and for biological control practices. An understanding of the effects of climate change on these interactions in wild populations relies on direct observation and analysis of field populations with the associated climate data (Tuda & Shima, Reference Tuda and Shima2002; Tuda et al., Reference Tuda, Matsumoto, Itioka, Ishida, Takanashi, Ashihara, Kohyama and Takagi2006a). Alternatively, demographic experiments that use model species provide a useful tool for understanding the general rules governing the effects of climate on whole ecosystems.

Studies have shown that just a 2°C increase in the temperature reduces the population size of seed beetles via a decrease in density-dependent rates of egg hatching and larval survival (Tuda, Reference Tuda1993; Tuda & Shimada, Reference Tuda and Shimada1993). Furthermore, such an increase in temperature also affects the coexistence of the seed beetle and its parasitoid populations (Tuda & Shimada, Reference Tuda and Shimada1995, 2005). However, the effect is highly species specific. Both the host and the parasitoid will coexist if the functional response of the parasitoid stabilizes the host-parasitoid population dynamics following an increase in temperature (M. Tuda, unpublished data).

Phytophagous insects also participate in competitive interactions, although competition may occur less frequently than predation (Jermy & Szentesi, Reference Jermy and Szentesi2003). Changes in climate (e.g. a 2°C rise and humidity change) alter the interspecific competitive success of two Callosobruchus beetle species, which may be ascribed by different environmental effects on development time and fecundity between the two species (Fujii, Reference Fujii1967). Furthermore, Davis et al. (Reference Davis, Lawton, Shorrocks and Jenkinson1998) showed that allowing dispersal along temperature clines altered the outcome of competition and species distribution in laboratory-reared Drosophila metapopulations.

Genetics of thermal adaptation

Development time and metabolic rate of C. maculatus are affected by temperature and by the interaction between mitochondrial haplotype and nuclear genotype (or mitonuclear interaction) (Dowling et al., Reference Dowling, Abiega and Arnqvist2007; Arnqvist et al., Reference Arnqvist, Dowling, Eady, Gay, Tregenza, Tuda and Hosken2010). Such mitonuclear interactions (Ballard & Rand, Reference Ballard and Rand2005) may be ubiquitous in insect thermal response. There is evidence that mtDNA transmission rates of Drosophila is temperature dependent (Nagata & Matsuura, Reference Nagata and Matsuura1991) and that nuclear genome is involved in the temperature dependency (Doi et al., Reference Doi, Suzuki and Matsuura1999). Although these studies using a model beetle shed light on the genetic background of the thermal response in beetles, our understanding of the genetic mechanisms controlling thermal adaptation is poor. Thus, this area of research offers considerable promise in the future.

Future directions

This review focused on climate effects on the evolution, developmental stages, variation within and among taxonomic groups, species interactions and coexistence in holometabolous insects, particularly beetles. Genetic effects on physiological plasticity and adaptation tend to be underestimated in previous studies, except for some that have used model organisms such as Drosophila (Hoffmann et al., Reference Hoffmann, Sorensen and Loeschcke2003) and Arabidopsis (Siomos, Reference Siomos2009). For well-studied laboratory model insects, however, extra caution must be paid to eliminate bottleneck events and artificial selection due to laboratory conditions and rearing. Populations that were reared in a large batch for a small number of generations (<<20) (for timescale of selection, see Tuda (Reference Tuda1998) and Tuda & Iwasa (Reference Tuda and Iwasa1998)) or several smaller populations with a moderate gene flow may be used.

Recent advances in molecular methods will improve our understanding of adaptation and acclimation to climate. On one hand, studies on molecular mechanisms of adaptation/acclimation to climate changes in beetles might benefit from DNA sequence data of a stored pest beetle, T. castaneum (Tribolium Genome Sequencing Consortium, 2008). For example, an evolutionary process may be uncovered for the close correlation between enzyme genotypes and the expression of a heat shock protein in a leaf beetle, Chrysomela aeneicollis (Dahlhoff & Rank, Reference Dahlhoff and Rank2000). In addition, approaches using phylogeny-based inference may improve the ability to distinguish between ecological plasticity and phylogenetic signals in climate-dependent traits. Comparative methods, such as have been used to describe the distribution of the variation in thermal tolerance at different levels of taxonomic hierarchy (Chown, Reference Chown2001), may employ the molecular-phylogeny-based hypothesis testing.

Quantification of the effects of global climate changes will be needed to predict future changes in phenology (Doi et al., Reference Doi, Takahashi and Katano2010; Musolin et al., Reference Musolin, Tougou and Fujisaki2010), dispersal (Berg et al., Reference Berg, Kiers, Driessen, van der Heijden, Kooi, Kuenen, Liefting, Verhoef and Ellers2010; Srygley et al., Reference Srygley, Dudley, Oliveira, Aizprua, Pelaez and Riveros2010), their synchronization with interacting species (Yukawa & Akimoto, Reference Yukawa and Akimoto2006; Hance et al., Reference Hance, van Baaren, Vernon and Boivin2007; van Asch & Visser, Reference van Asch and Visser2007; Klapwijk et al., Reference Klapwijk, Groebler, Ward, Wheeler and Lewis2010) and their population-dynamic consequences (Gilg et al., Reference Gilg, Sittler and Hanski2009) at local and regional scales (Kiritani, Reference Kiritani2006; Tabachnick, Reference Tabachnick2010). A number of climate models predict that several regions of the world will experience more or more severe droughts as a result of rises in atmospheric CO₂ and global temperature (Solomon et al., Reference Solomon, Plattner, Knutti and Friedlingstein2009; Xie et al., Reference Xie, Deser, Vecchi, Ma, Teng and Wittenberg2010). In these regions, particular attention should be paid to monitor any increase in the density and diversity of herbivorous insects, including stored crop pests.