The microbial dimension in insect nutritional ecology
A. E. Douglas
*Correspondence and present address: Daljit S. and Elaine Sarkaria Professor of Insect Physiology and Toxicology, Department of Entomology, 5136 Comstock Hall, Cornell University, Ithaca, NY 14853, USA. E-mail: email@example.com
*Correspondence and present address: Daljit S. and Elaine Sarkaria Professor of Insect Physiology and Toxicology, Department of Entomology, 5136 Comstock Hall, Cornell University, Ithaca, NY 14853, USA. E-mail: firstname.lastname@example.org
1Many insects derive nutritional advantage from persistent associations with microorganisms that variously synthesize essential nutrients or digest and detoxify ingested food. These persistent relationships are symbioses.
2There is strong experimental evidence that symbiotic microorganisms provide plant sap-feeding insects with essential amino acids and contribute to the digestion of cellulose in some wood-feeding insects, including lower termites. Basic nutritional information is, however, lacking for many associations, including the relative roles of microbial and intrinsic sources of cellulose degradation in many insects and B-vitamin provisioning by microorganisms in blood-feeding insects.
3Some nutritional interactions between insects and their symbiotic microorganisms vary among conspecifics and closely related species. This variation can, in principle, contribute to nutritional explanations for variation in the abundance and distribution of insects. For example, the plant utilization traits of phloem-feeding aphids and stinkbugs have been demonstrated to depend on the identity of microbial partners. Evidence that associations can evolve rapidly comes from the demonstration that the impact of the bacterium Wolbachia on natural populations of its insect host can change from deleterious to beneficial within two decades.
4Developing genomic tools, especially massively parallel sequencing and metagenomic analyses, offer the opportunity to explore the metabolic capabilities of symbiotic microorganisms and their insect hosts, from which defined hypotheses of nutritional function can be constructed. Nutritional ecology provides the appropriate framework to test these hypotheses in the relevant ecological context.
The insects are rather uniform in their nutritional requirements (Dadd 1985) but remarkably diverse in their diets (Slansky & Rodriguez 1987). There are insect predators, herbivores, generalist ‘scavengers’ and an array of specialists that utilize, for example, nectar, pollen, plant sap, fungi, fur, feathers, skin and blood. The capacity of some insects to adopt certain nutritional lifestyles cannot be attributed to traits of the insect alone but to an alliance between the insect and microorganisms with different biosynthetic or degradative capabilities from the insect. The basis for this article is that the nutritional ecology of many insects can be understood only by including microorganisms as an integral part of the ecological and physiological processes.
The microbial dimension in insect nutritional ecology is increasingly recognized, linked to the current research interest in the microbiology of healthy insects. Insects, like other animals, live in a microbial world. Repeatedly, they come into contact with microorganisms on surfaces and associated with food. Only a tiny minority of the microorganisms encountered are pathogenic. Most are of no discernible significance to the insect and are transiently associated with the insect: they, variously, fail to adhere to the insect surface, are lost as the cuticle is shed at insect moult, or pass with the food through the insect gut. Others are significant to insect nutrition in the broadest sense by contributing to the insect diet, promoting the acquisition and processing of food, and providing the insect with supplementary nutrients.
The purpose of this article is to explore the nutritional significance of living microorganisms persistently associated with insects. These insect–microbial relationships can be described as symbioses by the original definition of symbiosis as ‘any association between different species’. This article does not address the insects which depend on microorganisms as food (e.g. the fruitflies which consume yeasts associated with fruit, various aquatic detritivores that selectively feed on microbial biofilms on decaying vegetation, and collembolans which graze on fungi). However, the distinction between microbial food and microbial symbionts is not clear-cut for some insects. In particular, fungi are cultivated for food by attine ants, termites of the subfamily Macrotermitinae and ‘ambrosia beetles’ (Scolytinid and Platypodinid weevils), and these associations offer important contributions to our understanding of the insect nutritional interactions with living microorganisms.
This article addresses three topics. I first consider insects which derive specific nutrients, such as vitamins, essential amino acids and sterols, from resident microorganisms, and then review the contribution of microorganisms to the digestion and detoxification of certain foodstuffs, notably plant products. The third topic is the significance of symbiotic microorganisms as a determinant of the food utilization traits of insects. Throughout, I identify topics where information is lacking or data are inconclusive or contradictory, as pointers for further research opportunities.
Microbial synthesis of essential nutrients
nitrogen ‘hunger’ of insects
Many insects live on low-nitrogen diets, and microorganisms have been suggested to promote insect utilization of these food stuffs in various different ways. Some microorganisms simply concentrate the nitrogen. For example, the fungi cultivated by Macrotermitinae produce nitrogen-rich nodules, which the termites feed on preferentially. Many microorganisms are valuable to the insect for their wider metabolic capabilities, including their capacity to utilize insect nitrogenous waste compounds (e.g. uric acid), synthesize ‘high value’ nitrogenous compounds (e.g. essential amino acids) and fix nitrogen.
Evidence for the microbial utilization of nitrogenous waste products has been obtained for termites, cockroaches and hemipterans. The microorganisms are bacteria, including Bacteroides and Citrobacter species in the hindgut of termites (Potrikus & Breznak 1981); clavicipitacean fungi associated with the haemolymph and fat body of some plant-hoppers, notably Nilaparvata lugens (Hongo & Ishikawa 1997); and the bacterium Blattabacterium cuenoti in cockroaches (Mullins & Cochran 1975). This interaction conserves nitrogen, because the microorganisms utilize nitrogenous compounds that would otherwise be excreted, so reducing their use of nitrogenous compounds valuable to the insect host. Furthermore, at least part of the nitrogen is metabolized by the microorganisms to compounds that are made available to the host. For example, Potrikus & Breznak (1981) demonstrated that nitrogen derived from uricolysis by termite bacteria is subsequently assimilated back into termite protein. In this way, the microorganisms recycle insect waste nitrogen.
Implicit in these comments is the fact that all nitrogenous compounds are not nutritionally equivalent for an insect. This is a reflection of the limited metabolic capabilities of insects, particularly their inability to synthesize nine of the amino acids which contribute to protein. The insect requirement for these essential amino acids poses particular nutritional problems for insects feeding on plant sap. In most plants, the nitrogen of both xylem sap and phloem sap is dominated by non-essential amino acids, with the essential amino acids accounting for < 20% of the total (Douglas 2006). There is now extensive evidence that the microorganisms associated with these insects provide essential amino acids. Plant sap feeding through the life cycle is restricted to insects of the order Hemiptera, and all sap-feeding taxa bear symbiotic microorganisms. The microorganisms borne by most species are restricted to a particular cell type, the mycetocyte.
Hemipterans are not the only insects to bear microorganisms with a nutritional function in mycetocytes. Mycetocyte symbioses have evolved on multiple occasions (Table 1). Many bacteria in these associations have very small genomes, generally < 1 Mb (compared to Escherichia coli at nearly 4·5 Mb); the smallest currently known is Carsonella ruddii, the symbiont in psyllids, with 145 kb genome (Nakabachi et al. 2006). The mycetocytes are located in the haemocoel, fat body and epithelium of gut caeca, and the microbial symbionts are invariably transmitted vertically from mother to offspring, usually by insertion into the developing eggs in the female ovary (Buchner 1965). The early entomologists who discovered these associations described the insect cells as mycetocytes, whether they contained bacteria or yeasts, and I follow this convention. The cells bearing bacteria (but not fungi) are referred to as ‘bacteriocytes’ by some authors.
Table 1. Survey of mycetocyte symbioses in insects
Douglas (1989) provides a full list of coleopterans with mycetocyte symbioses; the microbial symbionts in most coleopteran symbioses have not been investigated by modern methods.
Baumannia cicadellinicola (γ-proteobacteria) and Sulcia muelleri (Bacteroidetes); Clavicipitacean fungi in some plant-hoppers
Buchnera aphidicola (γ-proteobacteria) or clavicipitacean fungi
Portiera aleyrodidarum (γ-proteobacteria)
Carsonella ruddii (γ-proteobacteria)
Tremblaya principes (β-proteobacteria)
(c) Vertebrate blood feeders
Anoplura (sucking lice)
Riesia pediculicola (γ-proteobacteria) in human lice
Wigglesworthia spp. (γ-proteobacteria) in tsetse flies
The nutritional role of mycetocyte symbionts has been investigated extensively for phloem sap feeding aphids, most of which bear the γ-proteobacterium Buchnera aphidicola in mycetocytes in the haemocoel. Experimental evidence that B. aphidicola synthesizes essential amino acids and provides them to the aphid [reviewed in Douglas (2006)] have been confirmed by annotation of the complete genome sequence of several B. aphidicola isolates, revealing that these bacteria have the genetic capacity to synthesize these nutrients (Shigenobu et al. 2000; Tamas et al. 2002; van Ham et al. 2003).
The genomic approach has also been adopted to explore the nutritional capabilities of symbiotic bacteria in a xylem-feeding hemipteran, the auchenorhynchan Homalodisca vitripennis, also known as the glassy-winged sharpshooter. This insect bears two bacteria: Baumannia, a γ-proteobacterium and Sulcia muelleri, a member of the Bacteroidetes, each with restricted but complementary predicted metabolic capabilities. In broad terms, Sulcia can synthesize essential amino acids and Baumannia can synthesize various vitamins and cofactors (McCutcheon & Moran 2007). Furthermore, the putative capabilities of the two bacteria are interdependent. For example, Sulcia synthesizes homoserine, the substrate for synthesis of the essential amino acid methionine by Baumannia, and Baumannia provides the poly-isoprenoids required for menaquinone synthesis by Sulcia (Fig. 1). These data suggest that there is cross-feeding of metabolites between the two bacteria, a condition that is known as a consortium. In other words, the genomic data suggest that the insect host is in symbiosis with a consortium of two bacteria with complementary functions.
Microorganisms in insects have also been suggested to fix atmospheric nitrogen, enabling insects to subsist on low-nitrogen diets (Nardi, Mackie & Dawson 2002). Insects can gain access to atmospheric nitrogen only through associations with bacteria because the capacity to fix nitrogen is widely distributed among bacteria but apparently absent from all eukaryotes. Nitrogen-fixing bacteria have been demonstrated, for example, in a minority of termites (Benemann 1973; Potrikus & Breznak 1977; Ohkuma, Noda & Kudo 1999), the wood-feeding larvae of the stag beetle Dorcus rectus (Kuranouchi et al. 2006) and in fruitflies, particularly the ‘medfly’Ceratitis capitata which contains large populations of nitrogen-fixing enterobacteria in their guts (Behar, Yuval & Jurkevitch 2005). The nutritional significance of nitrogen fixing bacteria to insects is, however, uncertain. The product of nitrogen fixation is ammonia, a potentially toxic compound that animals can metabolize to only a small range of organic nitrogen compounds. It is likely that some insects are colonized by nitrogen-fixing bacteria that are of no nutritional significance to the insect; and that, where bacterial nitrogen-fixation is important to insects, the ammonia is assimilated into other compounds (e.g. amino acids) prior to being made available to the insect. These issues have barely been considered experimentally.
microbial production of vitamins
Over the years, microorganisms have been implicated repeatedly in the provisioning of B vitamins to insects (Dadd 1985). The primary focus of these analyses has been insects feeding on vertebrate blood through the life cycle, including the tsetse flies Glossina, other Diptera Pupiparia, the anopluran ‘sucking’ lice, and the cimicids (bed bugs). All of these groups bear mycetocyte symbioses. Insects that utilize blood solely as adults (e.g. fleas, female mosquitoes) obtain sufficient B-vitamins from other dietary sources and are not dependent on symbiotic microorganisms.
Experimentally, these nutritional interactions are difficult to demonstrate because the vitamins are micronutrients, required in very small amounts. Traditional evidence has hinged on the performance of insects on diets with different vitamin contents, and this has depended on the amenability of insects to dietary analysis and purity of diet constituents. The current literature treats B vitamin provisioning by the mycetocyte bacteria in Glossina and other insects dependent on vertebrate blood as ‘a fact’. This is unfortunate because the experimental evidence is very limited. Puchta (1956) reports that the louse Pediculus suffers high larval mortality when experimentally deprived of its symbiotic bacteria, but that this effect is reduced when the diet of blood is supplemented with nicotinic acid (vitamin B3). Nogge (1976) eliminated the bacteria Wigglesworthia from tsetse fly by injecting the flies with lysozyme, and found that the fecundity of these flies was partially restored when their diet was supplemented with ‘B vitamins’ of undisclosed identity; the validity of this result is uncertain because lysozyme causes much nonspecific damage to the insects and is discredited as a treatment for eliminating bacteria from insects.
Genomic data are consistent with the expectation that the microorganisms play a role in B vitamin provisioning. The genome of Wigglesworthia, the mycetocyte symbiont of Glossina brevipalpis, has been sequenced (Akman et al. 2002) and the annotation has revealed the presence of genes coding for the synthesis of pantothenate (vitamin B5), biotin (vitamin B7), thiamin (vitamin B1), riboflavin and FAD (vitamin B2), pyridoxine (vitamin B6), nicotinamide (vitamin B3) and folate (vitamin B9). These data should not, however, be treated as definitive evidence that the bacteria provide the insect with these vitamins. Wigglesworthia requires all these vitamins as cofactors for its own metabolism. In an environment where these cofactors are scarce, it is to be expected that the bacterium would retain the capacity for their synthesis.
As discussed above in relation to plant sap feeding insects, vitamin synthesis has also been implicated from genomic data for the bacterium Baumannia in the xylem feeder H. vitripennis. The bacterium B. aphidicola in aphids also possesses the biosynthetic pathway for the synthesis of riboflavin (Shigenobu et al. 2000; Tamas et al. 2002; van Ham et al. 2003; Perez-Brocal et al. 2007); and its role in riboflavin synthesis is supported by dietary experiments. Consistent with early dietary studies of aphids (Dadd 1985), Nakabachi & Ishikawa (1999) found that the pea aphid Acyrthosiphon pisum could be maintained on chemically-defined diets lacking riboflavin; in fact, they performed slightly better on the riboflavin-free diet than the diet with the riboflavin supplement (Table 2). When the symbiotic bacteria were eliminated experimentally, the aphids grew and developed more slowly, but their performance was markedly improved by dietary riboflavin, which promoted weight gain and supported development to adulthood.
Table 2. Impact of dietary riboflavin on the performance of the pea aphid Acyrthosiphon pisum*
Riboflavin content of diet (µg mL−1)
Time to adulthood (days)
Maximum life span (days)
Mean body weight at 12 days
*‘Aposymbiotic aphids’ are aphids from which the symbiotic bacteria Buchnera aphidicola have been eliminated by antibiotic treatment. ‘Symbiotic aphids’ are the controls with unmanipulated bacterial complement [Data taken from Fig. 3 of Nakabachi & Ishikawa (1999)].
B-vitamin provisioning has also been proposed for various wood-feeding insects associated with mycetocyte symbionts but it has proven difficult, to date, to test these proposals experimentally.
microorganisms and the sterol nutrition of insects
Insects need sterols for various roles including a structural role in cellular membranes and as precursors for insect hormones, including 20-OH ecdysone. Unlike many other animals, however, they cannot synthesize sterols because they lack the enzymes mediating the cyclization of isoprene units to form the steroid ring structure. Most insects derive the sterols required from their diet, but there is excellent evidence that some insects obtain sterols from fungal symbionts. Bacteria generally have no capacity for sterol synthesis, and symbiotic bacteria do not contribute directly to the sterol requirements of insects.
The sterols of fungi are generally dominated by Δ5,7-sterols such as ergosterol, and the presence of these sterols in the insect sterol profile is a firm indication of fungal source. The leaf-cutting ant Acromyrmex octospinosus and the ambrosia beetle Xyleborus ferrugineus possess these distinctive fungal sterols, suggesting that they derive their sterol requirements from the fungi that they cultivate (Chu, Norris & Kok 1970; Maurer et al. 1992). Anobiid beetles Lasioderma serricorne and Stegobium paniceum also derive ergosterol and related sterols from their fungal partners (ascomycetes assigned to the genus Symbiotaphrina), which are located in gut caeca, and they metabolize the fungal sterols to the dominant insect sterols, 7-dehydrocholesterol and cholesterol (Nasir & Noda 2003). The clavicipitacaean fungi associated with the fat body of the plant-hopper N. lugens have a truncated sterol biosynthetic pathway, due to nonsense mutations in genes mediating the terminal steps of the ergosterol biosynthetic pathway (Noda & Koizumi 2003), resulting in the accumulation of ergosta-5,7,24(28)-trienol in the fungal cells (Wetzel et al. 1992). This compound is also probably transferred to the insect, where it is metabolized to 24-methylenecholesterol, cholesterol and other steroids.
Microbial processing of food
Cellulose is an abundant source of carbon, accounting for more than 50% of foliage and > 90% of woody tissues of plants. It is also a stable polymer that is relatively inaccessible to enzymatic attack. This is because it is generally organized into crystalline microfibrils and enclosed within a matrix of hemicelluloses, pectins and lignin. The physical disruption of plant material by chewing insects increases the availability of cellulose to enzymes. Despite this, cellulose degradation is a slow process involving multiple enzymes and can only be achieved in animals with a capacious gut through which the food material passes slowly.
There is a widely held misconception regarding the significance of microorganisms in insect utilization of dietary cellulose and associated plant polymers. The problem is that the central role of symbiotic bacteria in digestion by herbivorous mammals is extrapolated to insects. Mammals apparently lack intrinsic cellulases, and cellulose is degraded by bacteria, especially Ruminococcus and Bacteroides species, and also entodiniomorphid ciliates and various chytrids in large herbivores (ruminants, horses etc.) on high-fibre diets. The microorganisms are borne in large, anaerobic chambers either proximal or distal to the digestive portion of the digestive tract. The chief waste products of cellulose degradation by microorganisms in these anaerobic chambers are short chain fatty acids, particularly acetate, butyrate and propionate, which diffuse through the gut wall and are used as substrates for aerobic respiration by the animal.
Insects feeding on cellulose-rich plant material were once assumed to be ‘herbivorous mammals in miniature’ (Hungate 1966). Some termites such as Reticulitermes flavipes were used as examples. Reticulitermes flavipes bears a dense community of protists (Oxymonads, Trichomonads and Hypermastigids) that degrade cellulose, yielding short chain fatty acids which are utilized as a carbon source by the insect. Nevertheless, R. flavipes is not a miniature mammal because it also has intrinsic cellulases. Furthermore, it is now apparent that other insects exploit cellulose-rich diets in multiple, different ways.
Many insects have intrinsic cellulases (Davison & Blaxter 2005). The distribution of cellulase genes among insects has yet to be explored in detail either from a phylogenetic perspective or in relation to insect feeding habits. Even so, all termite species tested produce intrinsic cellulases (Slaytor 2000), as do thysanurans (silverfish and firebrats) (Treves & Martin 1994) and the cockroach Panesthia (Scrivener, Slaytor & Rose 1989). Reticulitermes flavipes and other ‘lower termites’ (collectively accounting for nearly 25% of all termite species) utilize both intrinsic cellulases and cellulolytic protists to digest their cellulose-rich diet, usually of native wood (Nakashima et al. 2002). The termite and microbial genes contributing to this function have been identified and analysed (e.g. Zhou et al. 2007). Some cockroaches, notably Cryptocercus species and Parasberia boleriana, bear related protists and presumably also rely on a combination of microbial and intrinsic cellulases for the degradation of cellulose in their diets (Varma et al. 1994; Pellens et al. 2002). Most termite species, however, lack protists and it has been uncertain whether cellulose digestion in these termites (which are known as ‘higher’ termites) is exclusively intrinsic or mediated by a combination of intrinsic and bacterial-mediated processes. There is, however, persuasive evidence for cellulolysis by hindgut bacteria in two higher termites Nasutitermes takasagoensis and N. walkeri lacking the cellulolytic protists (Tokuda & Watanabe 2007); and these data are supported further by a metagenomic analysis, which has identified cellulases and xylanase genes from members of the microbiota in the paunch of Nasitutermes sp. (Warnecke et al. 2007). Bacterial-mediated cellulose degradation has also been implicated in other insects, including the rhinoceros beetle Oryctes nasicornis (Bayon & Mathelin 1980) and the bostrychid beetle Prostephanus truncates (Vazques-Arista et al. 1997).
To summarize, the balance of current evidence favours a combination of intrinsic and microbial sources of cellulolysis in many insects feeding on wood and other cellulose-rich diets. An important unresolved issue is the relative importance of the two sources in different insects. We might expect the importance of microbial sources to depend critically on the enzymatic properties of the different cellulases and diet composition. The lignin content of the ingested food is particularly important. Lignin is a complex phenylpropanoid polymer tightly associated with cellulose and other plant polysaccharides. Access of cellulases to the cellulose substrate is aided by the physical disruption of the lignocellulose by insect mastication and also by chemical disruption. The latter is an oxidative process requiring molecular oxygen. This is not compatible with microbial fermentation of plant polysaccharides and, more generally, the reactive oxygen species produced at high oxygen titres linked with lignin degradation can be inimical to microbial populations. We might, therefore, expect insect species which depend on the chemical degradation of lignin to favour intrinsic cellulases and to have a relatively insignificant microbiota or to support the spatial separation of lignin and cellulose degradation into oxic and anoxic gut regions, respectively. Structuring of the physicochemical conditions in the gut lumen can occur over very small scales. For example, although the paunch of the lower termite R. flavipes is just 1 µL in volume, it includes a fully-oxygenated periphery adjacent to the gut wall and fully anoxic central portion accounting for approximately 50% of the paunch volume (Ebert & Brune 1997). The steep gradient of oxygen tension from periphery to centre is promoted by the spatial structuring of the microbiota, with oxygen consumption by facultative anaerobes (e.g. Enterobacter, lactic-acid bacteria) in the oxic region contributing to the maintenance of the anoxic core. The magnitude of lignin degradation by oxygen-tolerant bacteria in the paunch of lower termites, however, remains to be resolved fully (Ohkuma 2003).
The amount of plant material that can be degraded in insect guts is constrained by gut volume. This limitation is circumvented by insects that cultivate cellulolytic microorganisms in their nests. For example, Macrotermes species maintain the cellulolytic fungus Termitomyces, which degrades the cellulose and associated plant polysaccarides (Hyodo et al. 2003); the insects feed on both the fungal tissues and plant material degraded by the fungus. Martin & Martin (1977) have obtained evidence that the Termitomyces also contributes to cellulolysis in the gut of Macrotermes natalensis. Specifically, the fungus ingested by the insect has a high content of cellulases that are, first, resistant to proteases in the insect gut and, second, contribute to the hydrolysis of cellulose-rich plant material ingested by the insect. Martin (1991) has also suggested, with some supportive evidence, a role for ingested enzymes in cellulose degradation by siricid woodwasps and cerambycid beetles.
Virtually all research on cellulose degradation in insects has concerned species feeding on wood and other diets of very high cellulose content. What about foliage and grass feeders? Unlike mammalian browsers and grazers, these insects do not generally digest ingested cellulose. Instead, they eat very large amounts of plant material, digest and assimilate the easily utilized compounds (sugars, starch, protein etc.) and void the remainder, including the cellulose. For these insects, cellulose is a dietary ‘filler’. Two factors can be invoked to explain the difference between insects and mammals feeding on foliage and grasses. First, insects have a lower demand for carbon than mammals, which as persistent endotherms, need large amounts of carbon to fuel their high metabolic rates. Despite failing to use ingested cellulose, insects on these diets are nitrogen-limited, and not carbon-limited (McNeil & Southwood 1978). Second, the low activity of cellulases, whether intrinsic or microbial, means that cellulolytic animals bear a large volume and weight of slowly degrading plant material. This represents a greater cost for small, active insects, especially those capable of flight, than the larger mammals. The outcome of the cost : benefit equation generally favours cellulase-independence in insects and cellulase-dependence in mammals feeding on plant tissues without extensive secondary thickening. This distinction is not, however, a fixed rule. The giant panda is a cellulase-independent mammal, assimilating < 10% of ingested cellulose (Dierenfeld et al. 1982), while the locust is reported to have some cellulolytic capability (Cazemeier et al. 1997). Future research may identify other foliage/grass-feeding insects with some microbial or intrinsic capacity to degrade cellulose.
Most foods contain compounds (or elements) at concentrations which are potentially deleterious. The hazards are particularly great for herbivores because many plant tissues contain toxic secondary compounds which function as defence against consumers. Insects, like other animals, can protect themselves from such compounds by various routes, including elimination, detoxification and sequestration. These capabilities are generally intrinsic traits of the insect, independent of any microbial associates.
A few insects, however, use symbiotic microorganisms in detoxification. Research on this topic has focused principally on detoxification prior to digestion. Among herbivorous mammals, this is achieved by the pre-gastric symbiosis (i.e. microbiota housed in a chamber proximal to the enzymatic portion of the gut), for example in ruminants which are reputedly more tolerant of some plant toxins than mammalian herbivores with a post-gastric symbiosis (e.g. horses). This route to microbial detoxification is not available to insects, among which pre-gastric microbial symbioses are virtually unknown. The attine ants, however, obtain microbial detoxification by a pre-digestive symbiosis through a different route: cultivation of their microbial partner externally, in the nest. The microbial partners are fungi of genus Leucoagaricus (North, Jackson & Howse 1997). The ‘higher’ attines, such as Atta and Acromyrmex species, harvest leaf fragments and other plant material. They chew and bite it, and then deposit it on the fungus in their nests. This physical disruption promotes access by the fungal hyphae. The fungus subsists on the plant material, detoxifying plant secondary compounds by hydrolases and other enzymes. The fungal mycelium is the only food consumed by ant larvae and also contributes to the diet of adults. In this way, the ants can be considered to exploit the superior detoxifying capability of the fungus to gain access to plant food. Importantly, these ants have a very wide plant range. For example, the classic analysis of Cherrett (1968) revealed that Atta cephalotes use nearly half of the plant species in a Guyanese rainforest. This is an important factor contributing to the dominant position of attine ants as herbivores in the tropical/subtropical habitats of South America.
Symbiotic microorganisms as a determinant of food utilization traits of insects
From an ecological perspective, the nutritional interactions between insects and microorganisms that are of greatest interest are those which are variable. Such variation can contribute to nutritional explanations for variation in the abundance and distribution of insects. Different individuals of one species or related species may be infected or symbiont-free, or bear microorganisms with different nutritional capabilities. As a result, the microbiological status of the insect might influence the type or amount of food consumed and the fitness and life-history traits of the insect.
To date, variation in the nutritional ecology of insect–microbial interactions has been studied principally in phloem-feeding hemipterans. In particular, aphids bear bacteria collectively known as secondary symbionts that can influence plant utilization. The effects are, however, variable. The bacterium Regiella insecticola has been reported to promote, and have no effect on, the utilization of Trifolium repens by the pea aphid, presumably varying with aphid and bacterial genotype (Leonardo 2004; Tsuchida, Koga & Fukatsu 2004); Serratia symbiotica and a rickettsial secondary symbiont depress aphid fecundity on various plants (Fig. 2a), albeit to a variable extent (Chen, Montllor & Purcell 2000); and both R. insecticola and a further bacterium Hamiltonella defensa depress the growth rate of Aphis fabae on the red dead nettle Lamium purpureum (Fig. 2b) (Chandler, Wilkinson & Douglas 2008). Further research is required to establish the mechanistic basis of these variable and probably complex interactions between bacteria, insect and plant.
A second hemipteran system where symbiotic microorganisms influence plant utilization traits is the plataspid stinkbugs of the genus Megacopta. These insects feed on phloem sap and depend on bacteria known as Ishikawaella in the distal gut. The bacteria are deposited in a modified faecal pellet next to each egg and, as the larva emerges, it feeds on the pellet and is thereby infected with Ishikawaella of maternal origin. Hosokawa et al. (2007) exchanged the pellets between two insect species, M. punctatissima and M. cribraria. The larvae fed on the heterologous pellets and formed apparently normal symbioses that supported growth and reproduction. In one set of experiments, the insects were reared on soybean, a crop plant that is naturally infested by M. punctatissima. The insects performed well through the first generation and the females deposited eggs. However, many of the eggs produced by insects bearing bacteria from M. cribraria subsequently failed to hatch (Table 3). The embryos developed normally but the neonate insects were frail and could not escape from the egg shell. Although the reason for this effect is not yet established, these data demonstrate that the symbiotic bacterium Ishikawaella, and not the insect, determines the capacity of Megacopta species to utilize soyabean plants. As with the studies of bacterial determinants of aphid plant utilization, the mechanism underlying this effect is not known.
Table 3. Impact of symbiont type on the fitness of plataspid stinkbugs Megacopta reared on soybean*
There are indications that the microbial relationships in other insects may vary between closely related species, with possible implications for the wider nutritional ecology of the insect. In particular, the mode of cellulose digestion apparently varies among conspecific termite species that cultivate Termitomyces fungi. As considered above, Martin & Martin (1977) obtained evidence that the gut cellulose complement of one species, M. natalensis, includes enzymes derived from ingested Termitomyces. Termitomyces-derived cellulases and xylanases have also been demonstrated in several other Macrotermes species [see Table 1 of Licht, Boomsa & Aanen (2007)], but intrinsic cellulose activity can account for the total cellulolytic capability of M. subhyalinus and M. michaelseni (Bignell et al. 1994). Many factors, including the stability of the fungal enzymes, physicochemical conditions in the termite gut and the details of the diet, may contribute to this variation (Rouland-Lefèvre 2000).
Although the data are fragmentary, they do suggest that microbial factors can contribute to intra- and interspecific variation in the nutritional ecology of insects. The implications differ between associations that are horizontally and vertically transmitted.
For horizontally transmitted microorganisms, the nutritional traits of an insect may be influenced by the identity of the colonizing microorganisms, and this is likely to be determined only partly, or not at all, by the insect genotype. The importance of microbial determinants of nutrient utilization is increasingly recognized in mammalian, including human, physiology. In particular, parallel research on humans and mice has demonstrated that the composition of the gut microbiota both influences and is influenced by obesity. Genetically obese mice use food more efficiently than lean, wild-type mice. The gut microbiota is a contributory factor, since germ-free mice inoculated with the microbiota from obese mice accumulate more body fat than mice inoculated with microorganisms from lean mice (Turnbaugh et al. 2006). The gut microbiota of obese and lean people also differs. When obese people dieted for a year, the microbial composition of their stools became progressively more like that of lean people, and the people with the greatest weight reduction displayed the greatest change in bacterial profile (Ley et al. 2006). Does the composition of horizontally transmitted gut microbiota of insects play a role in the nutrition of insects? This is an important topic for future research. We should, however, be cautious in translating directly from mammals to insects. Some important differences have been suggested. In particular, McFall-Ngai (2007) has proposed a fundamental difference in animal–gut microbiota interactions between animals, including mammals, with an adaptive immune system and animals, such as insects, exclusively dependent on an innate immune system. McFall-Ngai argues that the adaptive immune system can ‘manage’ diverse and complex microbial communities, supporting multi-way metabolic interactions among different microbial taxa and the host. Animals without the adaptive immune system are predicted to support a microbiota of relatively low diversity. Nevertheless, it is disputed whether the taxonomic diversity of microorganisms varies according to immune system function, and uncertain how this might impact on the nutritional ecology of the animal. Research on the nutritional ecology of insects with different dietary habits should contribute to resolving the validity of these ideas.
In vertically transmitted symbioses, the microorganisms and their interactions with the insect host can evolve rapidly, as illustrated by the relationship between Drosophila simulans and the bacterium Wolbachia. In D. simulans populations in North America, Wolbachia causes cytoplasmic incompatibility, that is, crosses between uninfected females and infected males are infertile thereby increasing the prevalence of infected females. Cytoplasmic incompatibility has driven the Wolbachia infection through D. simulans populations in California between 1985 and 1994, despite a negative impact on host fitness; when the Wolbachia was eliminated with antibiotics, egg production by the flies increased by 10–20%. However, once the prevalence of Wolbachia in the population is high, its fitness depends on the fecundity of the female host, meaning that the Wolbachia is under selection pressure to reduce the deleterious effects of the Wolbachia infection. Weeks et al. (2007) have obtained persuasive evidence for rapid evolutionary response to this selection pressure. The fecundity of Wolbachia-infected D. simulans collected from the field in 2002 was either unaffected or reduced when Wolbachia was experimentally removed (Fig. 3), and supplementary experiments demonstrated that this change could be attributed to genetic changes in the Wolbachia and not the insect.
It is not known whether Wolbachia in this system is beneficial through nutritional or other interactions. Wolbachia can affect insect phenotype in multiple ways (Iturbe-Ormaetxe & O’Neill 2007), including a reported interaction with insulin signalling and glucose metabolism (Clark et al. 2005). Nevertheless, this analysis reveals the high rate at which the impact of microorganisms on their insect hosts can change.
The nutritional importance of microorganisms is profound for many insects. The likelihood is that some foodstuffs, including plant sap, vertebrate blood and sound wood, would be unavailable to insects through the life cycle without the nutritional input of microorganisms. Even so, any global assessment of the significance of microorganisms to the nutritional ecology of insects is constrained by a paucity of information on many interactions, particularly the significance of microorganisms in vitamin provisioning to blood-feeding insects, and the relative importance of microbial and intrinsic processes in insect utilization of cellulose and other plant cell wall components. These are long-standing issues that are important to resolve.
The current research priorities for the nutritional ecology of insect–microbial relationships are not, however, framed exclusively by ‘old problems’. The field also has ‘new solutions’ offered by ongoing advances in genomics. The massively parallel sequencing technologies and bioinformatics tools for metagenomic analysis represent tremendous opportunities, as is illustrated by the elucidation of the complementary metabolic capabilities of the two bacterial symbionts in the sharpshooter H. vitripennis (McCutcheon & Moran 2007; see Fig. 1) and the great diversity of metabolic capabilities represented in the total gene pool of the paunch microbiota in the termite Nasutitermes sp. (Warnecke et al. 2007). We should bear in mind that this genomic information refers to potential capabilities. Such analyses generate hypotheses of function that the nutritional ecologist is well-positioned to test. The priorities are to establish nutritional function and to explore the contribution of microorganisms to variation in nutritional function among insects in an ecologically-relevant context. These analyses should not neglect that the nutritional function is integrated with other physiological systems, especially the immune system. The contribution of microorganisms to the nutritional ecology of insects can only be understood in the context of the totality of the insect physiology and insect interactions with the wider environment.
Author thanks Professor Simon McQueen Mason for valuable discussions. This article was written while in receipt of a BBSRC Research Fellowship.