Reference Work Entry

The Prokaryotes

pp 439-474

Symbiotic Associations Between Termites and Prokaryotes

  • Andreas Brune


The symbiotic associations of termites with microorganisms comprise different levels of interaction, ranging from the extracorporal cultivation of fungus gardens to the most intimate associations, where bacteria reside intracellularly in dedicated bacteriocytes. However, the majority of prokaryotic symbionts of termites are located in the intestinal tract, where they are free-swimming, attached to the gut epithelium, or associated with the intestinal protozoa (Fig. 1). Although it is suggestive that the gut microbiota of termites is directly or indirectly involved in the digestion of lignocellulose or has other nutritional implications, the exact nature of the associations and possible benefits for the partners of each particular symbiosis are often far from clear. Therefore, this chapter will use the term “symbiosis” in its broader sense, as originally defined by Anton de Bary (de Bary, 1878). A definitive classification of the associations into the different categories of symbiosis, such as mutualism, parasitism, or commensalism, would require a level of understanding that is yet to be reached.
Fig. 1.

Examples of microbial symbionts in the hindgut of Reticulitermes flavipes (Isoptera: Rhinotermitidae), a wood-feeding lower termite. A) Preparation of anaerobic protozoa from the hindgut of a worker larva, showing the large hypermastigote flagellate Trichonympha agilis, filled with wood particles, and numerous smaller flagellates (mainly oxymonads, Dinenympha spp.). Differential interference contrast photomicrograph taken by U. Stingl. B) Transverse section through the peripheral hindgut, showing the diverse bacterial microbiota associated with the thin cuticle of the hindgut wall (bottom left). Transmission electron micrograph provided by J. A. Breznak. C) Preparation of the hindgut wall, showing the dense colonization of the cuticle by numerous rod-shaped and filamentous bacterial morphotypes. Scanning electron micrograph provided by J. A. Breznak. Reproduced from Brune (2003).

In view of the enormous body of literature on the intestinal microbiota of termites and its role in lignocellulose digestion, the subject cannot be covered exhaustively. This chapter will attempt to summarize the current state of knowledge on the prokaryotic communities within the intestinal tracts of termites, the major populations and their metabolic activities, and their interactions. In addition, it will focus on the gut as a microbial habitat. The chapter will touch only briefly on the intestinal flagellates, which are most important in the phylogenetically lower termites, the exosymbiotic fungi in fungus-cultivating Macrotermitinae, and the intracellular bacteria in termite tissues. For details on these subjects and for many other aspects of the termite gut symbiosis, the reader will be referred to the pertinent review articles.

Symbiotic Digestion

Termites, like other insects thriving on a lignocellulosic diet, possess a pronounced gut microbiota housed in specially adapted regions of the alimentary tract (Fig. 2). The symbionts convert a substantial portion of the dietary components to microbial fermentation products, which are then eventually resorbed by the intestinal epithelia. It is generally assumed that the intestinal symbioses provide metabolic capacities that are otherwise not available to the host. For reviews, see Breznak and Brune (1994b), Kane (1997), Brune (1998), Brune (2003), Bignell (2000), Breznak (2000), Brune and Friedrich (2000a), and Ohkuma (2003).
Fig. 2.

Structure of the intestinal tract of Reticulitermes species (A) and Cubitermes species (B). All lower termites harbor the gut microbiota in a single, strongly dilated hindgut paunch (Pa) that tapers out via the colon into the rectum (R). In most higher termites, especially the soil-feeding species, the hindgut is more elongated and has additional dilatations. Abbreviations: crop (C), midgut (M), mixed segment (ms), and proctodeal segments (P1–P5).

The symbiotic digestion of lignocellulose by termites is a complex series of events involving both the host and its gut microbiota (Fig. 3). While the events in the foregut and midgut seem to be mainly due to host activities, the digestive processes in the hindgut are largely controlled by the symbionts. Many aspects of lignocellulose digestion are common to all termites, but there are also several noteworthy differences between the phylogenetically lower and higher taxa.
Fig. 3.

Major events in the symbiotic digestion of lignocellulose by wood-feeding lower termites. The black lines show the path of the insoluble material whose lignin-rich residues are released as feces, whereas the red lines represent soluble degradation products, which are eventually resorbed by the host. The green lines indicate the cycling of nitrogenous compounds. Blue arrows mark the sites where cellulolytic enzymes are secreted. Lower-case letters refer to the different groups of bacteria, which are either endobionts (a) or epibionts (b) of the protozoa, suspended in the gut lumen (c) or attached to the gut wall (d). The scheme has been simplified for the sake of clarity; not all possible interactions are shown. Adapted from Brune (2003).

Fiber Degradation

The degradation of plant cell walls requires the synergistic action of many different enzymes and, in the case of lignified substrates, also a mechanism to break up the lignocellulose complex (Breznak and Brune, 1994b). Microorganisms, i.e., bacteria, protozoa, and fungi, are the most efficient cellulose and hemicellulose degraders in nature, and fungi and certain actinomycetes are also the only organisms that have developed a strategy for the chemical breakdown of lignin (Béguin and Aubert, 1994; Jeffries, 1994). Not surprisingly, termites and other animals have made use of these capacities by employing microbial symbionts in the digestion of lignocellulosic food (Martin, 1983).

Role of Intestinal Protozoa

The presence of protozoa in termite guts was recognized very early, although they were initially considered to be parasites (e.g., Leidy, 1881; Koidzumi, 1921). The American protozoologist L. R. Cleveland recognized that the wood-feeding lifestyle in the evolutionarily “lower” termites is based on a mutualistic association with their intestinal protozoa (Cleveland, 1925a; Cleveland, 1926). In a series of elegant experiments, he established that the ability of lower termites to live on a diet of wood or cellulose depends on the digestive capacities of their intestinal flagellates. His pioneering work, published in 1923–1928, paved the way for many later studies (see reviews by Honigberg [1970], Inoue et al. [2000], and Brune and Stingl [2005]).

About 15 years later, Hungate elucidated the biochemical basis for this symbiosis (Hungate, 1939; Hungate, 1943). He showed that the gut flagellates depolymerize and ferment lignocellulose to short-chain fatty acids, which are resorbed and oxidized by the host (reviewed by Hungate [1955] and Breznak and Brune [1994b]). The importance of protozoa for the metabolic processes in the hindgut of lower termites is most impressively evidenced by their enormous numbers, which may constitute more than one-third of the body mass in Zootermopsis species (Katzin and Kirby, 1939).

There is a large body of literature on the decomposition of wood and cellulose by termite gut flagellates (for references, see O’Brien and Slaytor [1982], Breznak and Brune [1994b], and Inoue et al. [2000]). Apparently, the different flagellate species are nutritionally specialized, and each species might fill a specific niche in lignocellulose digestion (Yoshimura et al., 1996; Inoue et al., 2000). Most of the endoxylanase and β-xylosidase activity in the lower termite Reticulitermes speratus is located in the anterior hindgut and is lost upon defaunation (removal of protozoa by ultraviolet irradiation; Inoue et al., 1997), and the effects of artificial diets on the composition of the protozoan community corroborate that different gut flagellates are involved in xylan and cellulose degradation (Inoue et al., 1997; Cook and Gold, 2000).

The protozoa possess their own cellulase genes, which fall into different glycosyl hydrolase families (Ohtoko et al., 2000; Nakashima et al., 2002a;Watanabe et al., 2002; Li et al., 2003; Inoue et al., 2005) and may even exploit host cellulases that are secreted in the anterior gut regions (Li et al., 2003). There is no evidence that the prokaryotic symbionts of the gut flagellates (see the section Interactions Between Prokaryotes and Protozoa in this Chapter) confer cellulolytic activity to their hosts.

The Role of Fungi

The role of fungi in the digestion of lignocellulose by termites is less clear. There are termites that can thrive on sound wood, but many species show a strong preference for decaying wood colonized by saprophytic fungi, which may either precondition the wood for digestion or provide metabolic products important for termite nutrition (Sands, 1969; Rouland, 2000; Cornelius et al., 2002). Schäfer et al. suggested that the yeasts and other fungi present in the guts of the lower termites Zootermopsis angusticollis and Neotermes castaneus are involved in hemicellulolytic degradation (Schäfer et al., 1996).

Higher termites of the subfamily Macrotermitinae have established a unique exosymbiosis with basidiomycetes of the genus Termitomyces, which are maintained on predigested plant litter in so-called “fungus gardens” within the nests. The symbiosis has rendered fungus-cultivating termites independent of the intestinal protozoa, which probably allowed for the obvious diversification in their diet (Sands, 1969; Rouland, 2000). The specificity of this symbiosis, whose enormous evolutionary success is impressively evidenced by the huge nests of fungus-cultivating termites populating the grasslands of Africa, is documented by several instances of coevolution between the termites and their fungal partners, indicating both horizontal and vertical transmission of the fungal symbionts (Aanen et al., 2002; Katoh, 2002; Rouland-Lefevre et al., 2002; Taprab et al., 2002).

The association with the lignin-degrading fungus enables the fungus-cultivating termites to utilize lignocellulose nearly completely, as reflected in the small volume of their final feces (Darlington, 1994). The key activities attributed to the fungal partner in this mutualistic symbiosis are extensive delignification of the substrate (Hyodo et al., 1999, 2000; Johjima et al., 2003) and the conversion of plant fiber to fungal biomass, as proposed earlier in the lignin degradation hypothesis of Grassé and Noirot (1958). Evidence for an activity within the gut of fungal cellulases ingested by the termites together with the fungus comb material (Abo-Khatwa, 1978; Martin and Martin, 1978; Martin and Martin, 1979) gave rise to the acquired enzyme hypothesis of Martin (1983). However, claims that the fungal cellulases are essential for cellulose digestion in the termite gut remain controversial (Slaytor, 1992; Bignell et al., 1994b; Crosland et al., 1996), especially in view of the recently discovered ability of termites to produce their own cellulases (see the section The Role of Host Enzymes in this Chapter).

The Role of Host Enzymes

Since phylogenetically higher termites (family: Termitidae) lost their gut flagellates in the course of termite evolution, it was initially assumed that either ingested fungal enzymes (see the section The Role of Fungi in this Chapter) or prokaryotic symbionts took over the function of the cellulolytic protozoa. However, there is still no clear evidence that bacteria play a major role in cellulose degradation in any of the termites investigated (see the section Cellulolytic Bacteria in this Chapter), which may find its explanation in the recently discovered ability of termites to produce their own cellulases.

In all insects, the digesta are exposed to a variety of digestive enzymes secreted by the salivary glands and the midgut epithelium (Terra, 1990). Cook (1943) had already demonstrated that termites are able to absorb sugars directly, and evidence is accumulating that termites secrete a full complement of enzymes necessary for the digestion of plant structural polysaccharides, including cellulose, into the midgut (e.g., Rouland et al., 1989; Slaytor, 1992; Rouland, 2000). The presence of protease and lysozyme activities has been documented for several termites, which indicates that also microbial cells can be digested (Rohrmann and Rossman, 1980; Fujita et al., 2001; Fujita and Abe, 2002a; Fujita et al., 2002b). Although experimental evidence is scarce, one can safely assume that—as in other insects—most of the easily digestible material has been mobilized and resorbed by the time the digesta reach the end of the midgut (Fig. 3).

The persisting dogma that higher animals do not possess their own cellulases has been unequivocally refuted by the demonstration of endoglucanase genes in the termite genome and their expression in the cells of the midgut epithelium and in the salivary glands (reviewed by Watanabe and Tokuda, 2001). Even in lower termites, host cellulases secreted by the salivary glands and complement (and surpass) the cellulolytic activities of the intestinal protozoa in the hindgut (Nakashima et al., 2002b; Tokuda et al. 2004). There is evidence that glycosyl hydrolase family 9 (GHF9) cellulases present in the genomes of termites are ancient and widespread in Metazoa (Lo et al., 2003b; Davison and Blaxter, 2005).

Soil-Feeding Termites

The majority of termite species are humivorous, yet little is known about the exact nature of the dietary components exploited by these ecologically important soil macroinvertebrates (Brauman et al., 2000). Besides fragments of plant tissue, fungal hyphae, and numerous microorganisms, their diet consists largely of undefined humic material intimately associated with the mineral soil matrix (Donovan et al., 2001). While the aromatic component of humus was initially assumed to be the principal substrate of soil-feeding termites (Noirot, 1992; Bignell, 1994a), feeding trials with soil-feeding Cubitermes spp. have shown that peptidic soil components—free or polymerized into humic model compounds—are preferentially digested and mineralized (Ji et al., 2000; Ji and Brune, 2001; Ji and Brune, 2005).

The anterior hindgut of soil-feeding termites is extremely alkaline (Bignell and Eggleton, 1995; Brune and Kühl, 1996), which favors the extraction of organic matter from the soil (Brune, 1998; Kappler and Brune, 1999). Microbial biomass and its structural components are assimilated more efficiently than cellulose, which supports the hypothesis that soil microorganisms and the nitrogenous components of humus are an important dietary resource for humivorous soil macroinvertebrates (Ji et al., 2000; Ji and Brune, 2001).

Host Nutrition

Irrespective of their contribution to polymer degradation, the majority of prokaryotes in termite guts are probably involved in the fermentation of the soluble products released into the gut (see the section Microbial Fermentations in this Chapter), which are derived either directly from the food by the digestive enzymes (see the section The Role of Host Enzymes in this Chapter) or by the fermentative activity of the intestinal protozoa (see the section Role of Intestinal Protozoa in this Chapter). The major products of the hindgut metabolism are acetate and, to a smaller extent, other short-chain fatty acids (mainly propionate and butyrate), which accumulate in the hindgut fluid and are eventually resorbed by the hindgut epithelium (Fig. 3). Termites—like other insects—cannot use acetate as a substrate for gluconeogenesis, but as long as the digestive processes in the midgut release sufficient amounts of soluble sugars and amino acids, this is not a problem. Breznak calculated that acetate produced by the hindgut microbiota of Reticulitermes flavipes would suffice to support the respiratory activity of the host (Breznak, 2000).

Besides being difficult to degrade, lignocellulose is also an extremely nutrient-poor substrate. While non-lignified plant cells are usually rich in protein and other nitrogenous compounds, the C-to-N ratio of sound wood is up to 100-fold higher than that of the insect body (La Fage and Nutting, 1978). Moreover, a lignocellulosic diet lacks most of the essential nutrients required by animals, such as amino acids, vitamins, and sterols. Many microorganisms are capable of fixing dinitrogen, assimilating nitrate and ammonia, or synthesizing those amino acids and vitamins essential for the host. Animals, including termites, have developed means of exploiting these biosynthetic capacities, which include—in the simplest case—the digestion of the intestinal symbionts.

Wood-feeding termites, especially those feeding on sound wood, have an extreme shortage of nutrients in their diet, and the digestion of microbial biomass acquired in the course of anal trophallaxis supplies them with high-quality nutrients (Machida et al., 2001). The gut microbiota supplies essential precursors for the biosynthesis, e.g., of methyl-branched hydrocarbons (Guo et al., 1991), and might play a role in nestmate recognition (Matsuura, 2001). To date, the lack of knowledge on the individual components of the prokaryotic microbiota and their metabolic capacities and activities in situ still makes it difficult to define the essential functions and understand the complex interactions.

The Gut Microenvironment

The intestinal tract of insects is organized into three major gut regions: a short foregut, a midgut (which is the main site of digestion), and a usually short hindgut (proctodeum). The hindguts of all termites, however, have immensely increased in length and volume over the course of evolution (Fig. 2). In the more primitive, lower termites, the hindgut is still relatively simple, consisting of a dilated “hindgut paunch” that tapers out into the colon and ends in the rectal compartment (Noirot, 1995). While this organization has been retained in the fungus-cultivating termites (Termitidae: Macrotermitinae), all other lineages of higher termites show a trend towards a further elongation and additional compartmentalization of the hindgut (Noirot, 2001), which is most pronounced in the soil-feeding representatives. The gut morphologies of lower and higher termites and the significance of the adaptations for the digestive process have been reviewed exhaustively (Noirot, 1995; Noirot, 2001).

The proctodeal dilatations increase the residence time of the digesta, thereby prolonging the exposure to the activities of the intestinal microbiota. Moreover, host factors and microbial activities give rise to physicochemical gradients that create distinct microenvironmental conditions in each gut compartment. This has been shown for oxygen, hydrogen, redox potential, and intestinal pH and has to be expected also for any other metabolite when source and sink are spatially separated (Brune and Friedrich, 2000a), especially since the microbiota are not randomly distributed within the gut (see the section Spatial Organization in this Chapter).

Physicochemical Gradients

Redox Conditions and Oxygen Status

The general concept of termite guts as anoxic habitats had been based on several pieces of circumstantial evidence (outlined by Veivers et al., 1980): 1) the oxygen sensitivity of the intestinal protozoa, already recognized by Cleveland (Cleveland, 1925a; Cleveland, 1925b); 2) the demonstration of a fermentative metabolism of cellulose by these flagellates (Hungate, 1939; Hungate, 1943) and the high concentrations of microbial fermentation products in the hindgut of all termites investigated; and 3) the presence of oxygen-sensitive or strictly anaerobic processes, such as nitrogen fixation and methanogenesis (Breznak et al., 1973; Breznak, 1975). Also, the subsequent isolation of anaerobic bacteria from termite guts (see the section Isolates and Major Metabolic Activities in this Chapter) supported the general assumption that the principle of symbiotic digestion in termite guts was analogous to that in the rumen.

Slaytor and coworkers were the first to question the anoxic status of termite guts. Following the color reaction in the hindgut of redox indicator dyes fed to Nasutitermes exitiosus and Coptotermes lacteus, they initially claimed that the hindgut paunch was “aerobic,” since methylene blue remained oxidized (Eutick et al., 1976). In a later study, however, using a more refined technique, they obtained E h values between –230 mV and –270 mV in the hindgut paunch of these and seven other termite species (Veivers et al., 1980). The initial error had been caused by the color of the reduced dye within the gut being obscured by that of the oxidized dye, which had also impregnated the oxic gut epithelium.

Moreover, Slaytor and coworkers demonstrated that the vitality of Nasutitermes exitiosus and Coptotermes lacteus depended on the presence of their prokaryotic gut microbiota (Eutick et al., 1978b) and that bacteria play an important role in maintaining the low redox potential of the hindgut paunch (Veivers et al., 1982), which led to the postulation that they maintain anoxic conditions by removing oxygen from the hindgut.

Bignell (1984) pointed out that arthropods are relatively small animals with surface-to-volume ratios higher than those in practically all vertebrates and that they are likely to reach equilibrium with their environment unless efficient permeability barriers for oxygen are established or oxygen is sequestered by the animal or by intestinal microorganisms. In a series of studies employing oxygen microsensors, Brune and coworkers clarified the situation (Brune et al., 1995a; Ebert and Brune, 1997; Schmitt-Wagner and Brune, 1999) by demonstrating that the steep gradient in oxygen partial pressure between the oxic gut epithelium and the anoxic gut contents drives a continuous influx of oxygen into the hindgut (Fig. 4). In all termites investigated, oxygen penetrated 50–200 µm into the periphery of the hindgut lumen, leaving only the central portion of the dilated compartments anoxic (Ebert and Brune, 1997; Brune, 1998; Kappler and Brune, 1999; Schmitt-Wagner and Brune, 1999).
Fig. 4.

Radial gradients of oxygen (•) and hydrogen (○) in an agarose-embedded hindgut (paunch region) of Reticulitermes flavipes worker larva. A schematic cross-section through the paunch illustrates the relative sizes of the oxic and anoxic zones. Modified after Brune (1998).

The maintenance of anoxia in the termite hindgut is not a trivial issue. Since the removal of oxygen in the gut periphery is fueled by the fermentative processes in the hindgut lumen, there must be a lower size limit for arthropods with a symbiotic digestion. However, even the smallest of all termites investigated to date (Anoplotermes pacificus, Termitidae: Apicotermitinae) seems to possess a symbiotic gut microbiota, although the spectrum of fermentation products in the hindgut differs from that of other termites (Bauer et al., 2000).

Fine-scaled redox measurements with platinum microelectrodes in wood- and soil-feeding termites (Ebert and Brune, 1997; Kappler and Brune, 2002) have shown that the redox potential in the gut mirrors the oxygen gradients (Fig. 5), and also the absolute values are in good agreement with the previously published data based on feeding experiments with redox dyes (Veivers et al., 1980). The most negative redox potentials are found in the regions of high hydrogen partial pressure, although in soil-feeding termites of the genus Cubitermes, parameters other than oxygen or hydrogen partial pressure seem to control the redox status of the intestinal contents (Kappler and Brune, 2002). There is evidence for ferric iron reduction in the gut of soil-feeding and wood-feeding species, which may be a microbial process, possibly mediated by the presence of humic acids or other phenolic polymers (Kappler and Brune, 2002; Vu et al., 2004).
Fig. 5.

Profiles of physicochemical conditions along the gut axis of Reticulitermes flavipes (A) and Cubitermes orthognathus (B). Oxygen (•) and hydrogen (○) partial pressures, intestinal pH (▪), and apparent redox potential (□) (against a standard hydrogen reference electrode [SHE]) were measured with the respective microsensors. Guts were embedded in agarose-solidified Ringer’s solution. The borders between the different gut regions (see legend to Fig. 2) are indicated by the vertical lines. Data from Brune et al. (1995a), Brune and Kühl (1996), Ebert and Brune (1997), Schmitt-Wagner and Brune (1999), and Kappler and Brune (2002).

Hydrogen Partial Pressure

Despite the massive hydrogen production by the intestinal protozoa (Odelson and Breznak, 1985b), the hydrogen emission rates of termites are relatively low (Odelson and Breznak, 1983; Ebert and Brune, 1997; Sugimoto et al., 1998; Schmitt-Wagner and Brune, 1999). Originally, it had been assumed that the situation in the intestinal tract of termites was similar to that in other methanogenic habitats, where low hydrogen partial pressures result from a tight coupling between hydrogen-producing and hydrogen-consuming processes (Breznak, 1994a; Breznak and Brune, 1994b). However, hydrogen microsensor measurements revealed that the situation in termites is quite different, giving rise to steep radial gradients of hydrogen towards the gut epithelium and enormous differences in hydrogen partial pressure along the gut axis (Ebert and Brune, 1997; Schmitt-Wagner and Brune, 1999; Kappler and Brune, 2002), which gave rise to the hypothesis that the spatial organization of the hydrogen-producing and hydrogen-consuming populations controls the hydrogen partial pressure in different gut regions.

Intestinal pH

The intestinal pH in the hindgut of most phylogenetically lower termites seems to be around neutral (Eutick et al., 1976; Bignell and Anderson, 1980a; Veivers et al., 1980; Brune et al., 1995a). Kovoor (1967) was the first to report the an alkaline region in the anterior hindgut of wood-feeding termites; this observation was later extended also to soil-feeding species (Bignell and Anderson, 1980a). Since then, a large body of data has accumulated (Bignell, 1994a; Bignell and Eggleton, 1995), documenting a tendency towards strong alkalinity in the anterior hindgut of all higher termites except the Macrotermitinae (Anklin-Mühlemann et al., 1995).

While the initial measurements (performed mostly by spotting pooled, disrupted samples of individual gut regions on pH indicator paper) still lacked accuracy and resolution, studies with pH microsensors allowed alkaline regions to be precisely located (Brune et al., 1995a; Brune and Kühl, 1996). The latter measurements are not biased by homogenization and documented that guts of soil-feeding termites are even more alkaline than reported previously. The most alkaline values (pH 11–12.5) were found among soilfeeding Termitinae and represent the highest values ever encountered in biological systems (Brune and Kühl, 1996). In all species tested, the pH of the gut contents increases sharply from circumneutral in the midgut to highly alkaline between the midgut–hindgut junction and the first proctodeal dilation (P1), which coincides exactly with the location of the mixed segment. (Fig. 5), a morphologically unique gut region present in all higher termites except the Macrotermitinae (Noirot, 2001).

Gut Compartmentation and Microhabitats

Each gut compartment provides various microhabitats differing in many environmental parameters (see the section Physicochemical Gradients in this Chapter). The small size of the guts results in large surface-to-volume ratios (Brune, 1998), and the epithelial surfaces provide ample attachment sites for gut microorganisms, which are thus protected from washout (Bignell, 1984). Additional compartmentalization is created by the protozoa inhabiting the hindgut lumen of lower termites.

Midgut Epithelium, Ectoperitrophic Space

As in other insects, the midgut epithelium is not protected by a cuticle, but a peritrophic membrane separates the epithelial surface from the digesta (Terra, 1990). The ectoperitrophic space harbors a distinct bacterial microbiota, which can be intimately associated with the microvilli of the brush border (Breznak and Pankratz, 1977). In addition, the so-called “mixed segment” in many Termitidae, a region where midgut and hindgut epithelia overlap (Noirot, 2001), is a microhabitat that harbors a specific bacterial microbiota. Kovoor (1968) described a “pure culture of spore-forming fusiform bacteria” in the mixed segment of Microcerotermes edentatus, located outside of the peritrophic membrane in a posterior pocket formed by the mesenteric side. Also Potts and Hewitt (1973) observed a prominent flora of “thin long filaments with terminal spores” in the mixed segment of the harvester termite, Trinervitermes trinervoides (Nasutitermitidae), located in the ectoperitrophic space posterior to the Malphigian tubules. Other authors described dense populations of different but also relatively uniform microorganisms in the mixed segment of Nasutitermes exitiosus (Czolij et al., 1985) and of soil-feeding Termitinae (Procubitermes aburiensis and Cubitermes severus; Bignell et al., 1980b; Bignell et al., 1983). Recently, Tokuda et al. (2000) demonstrated that the bacteria populating the mixed segment of Nasutitermes takasagoensis are phylogenetically within the radiation of the Clostridiales (see the section Clostridiales in this Chapter). Electron microscopy confirmed their close association with the mesenteric epithelium, suggesting that there is some kind of interaction with the gut tissue (Tokuda et al., 2001).

Hindgut Cuticle

Electron-microscopy studies revealed intimate associations of microorganisms with the cuticle of the hindgut epithelium in all termites investigated (Breznak and Pankratz, 1977; To et al., 1980; Czolij et al., 1985; Yara et al., 1989; Fig. 1). Bacteria are associated with cup-like indentations on the epithelial surface of the hindgut in Reticulitermes flavipes (Breznak and Pankratz, 1977). Although Mastotermes darwiniensis and Nasutitermes exitiosus possess similar structures, they seem not to be associated with microorganisms (Czolij et al., 1984). In certain soil-feeding termites, cuticular spines protrude from the hindgut wall into the lumen of the P4 compartment and form additional attachment sites for the gut microbiota (Procubitermes aburiensis; Bignell et al., 1980c).

Hindgut Protozoa

The protozoa in the hindgut of lower termites occupy the bulk of the hindgut volume (Katzin and Kirby, 1939) and represent an enormous surface area in the hindgut (Berchtold et al., 1999). Pierantoni (1936) was first to point out the association of gut flagellates with bacteria; since then, ectobiotic and endocytobiotic bacteria have been found on and in almost every flagellate investigated. For example, the hypermastigote flagellate Joenia annectens, a symbiont in the hindgut of Kalotermes flavicollis, is densely colonized by prokaryotic microorganisms (Hollande and Valentin, 1969). The body is covered with rod-shaped bacteria, and the nucleus and the cytoplasm contain various types of endocytobiotic bacteria (Radek et al., 1992; Patricolo et al., 2001). Also the oxymonadid flagellate Streblomastix strix, a hindgut symbiont of Zootermopsis species, is associated with several, morphologically distinct types of bacteria that are orderly arranged end-to-end on six or seven longitudinal vanes, lending S. strix a stellate appearance in transverse section (Leander and Keeling, 2004). Adhesion of bacteria to the flagellate surfaces is based on different mechanisms and facilitated by special surface structures (e.g., Radek et al., 1996; Radek and Tischendorf, 1999b; Rother et al., 1999; Patricolo et al., 2001). The literature has been recently reviewed by Radek (1999). Possible significance of these associations of prokaryotes with hindgut flagellates is discussed in a different section (see the section Interactions Between Prokaryotes and Protozoa in this Chapter).

Prokaryotic Gut Symbionts

In view of the variety of microhabitats and microenvironmental conditions in the intestinal tracts of termites (see the section The Gut Microenvironment in this Chapter), it is not astonishing to find an equally large diversity among the microorganisms colonizing the gut. The amount of diversity indicated already by the morphological and ultrastructural features of the microbiota is greatly exceeded by that encountered at the phylogenetic level.

Morphological Diversity

Already the observation of gut preparations with a phase-contrast light microscope reveals a wide variety of prokaryotic life forms. Several comprehensive studies of the bacterial gut microbiota of termites using transmission electron microscopy have provided detailed accounts of the morphological diversity of gut microorganisms for several termite species from different families. In addition to the abundant protozoan fauna in all so-called “lower termites” (Yamin, 1979), at least 20–30 different bacterial morphotypes have been distinguished among the microorganisms colonizing the intestinal tract of Reticulitermes flavipes and Coptotermes formosanus (Rhinotermitidae; Breznak and Pankratz, 1977) and Pterotermes occidentis (Kalotermitidae; To et al., 1980).

The so-called “higher termites” (Termitidae) lack intestinal flagellates, but the morphology of their prokaryotic microbiota appears to be equally diverse. The hindgut of the wood-feeding Nasutitermes exitiosus (Termitidae: Nasutitermitinae) contains almost 30 different bacterial morphotypes (Czolij et al., 1985), and also the hindgut microbiota of the fungus-growing termite Odontotermes formosanus (Termitidae: Macrotermitinae) comprises at least 20 different morphotypes (Yara et al., 1989). Numerous bacterial morphotypes, including many filamentous forms, colonize the intestinal epithelia and the ectoperitrophic space of the soil-feeding termite Procubitermes aburiensis (Termitidae: Termitinae; Bignell et al., 1980b).

Although the morphological features usually do not allow the affiliation of a bacterium to a specific taxon, members of the termite gut microbiota have conspicuous forms or other morphological features that are of (albeit limited) taxonomic value, or conspicuous forms that seem to occur in different species of termites. One example is the thin, spore-forming filaments described by Leidy (1881) as “Arthromitus” species, which occur in many invertebrates, including termites (Leidy, 1881; Margulis et al., 1990). Also, many of the spirochetal forms are so large and conspicuous that they can be morphologically distinguished (Breznak, 1984a; see also Termite Gut Spirochetes in Volume 7). On the basis of the detailed morphological features visualized by transmission electron microscopy, Margulis and coworkers (Bermudes et al., 1988; Wier et al., 2000; see also Large Symbiotic Spirochetes: Clevelandia, Cristispira, Diplocalyx, Hollandina, and Pillotina in Volume 7) proposed a number of new species for the larger spirochetes.

Phylogenetic Diversity and Community Structure

In the recent years, the microbiota in the intestinal tracts of termites has been investigated also with molecular tools (see Ohkuma [2002a] for a review). Most studies employed the 16S rRNA gene as a molecular marker. As in most other environments, phylogenetic diversity of the intestinal microbiota is enormous, and there is still little overlap between the phylotypes recovered with cultivation-independent techniques and the isolates obtained by cultivation (see the section Isolates and Major Metabolic Activities in this Chapter). Molecular fingerprinting has been used to compare the structure of gut communities and to follow temporal changes. The bias inherent in all polymerase chain reaction (PCR)-based approaches has been addressed by backing the results with independent methods, such as fluorescence in situ hybridization (FISH).

Bacterial Diversity

König and coworkers were among the first to use the 16S rRNA-based approach to identify the phylogenetic position of uncultivated spirochetes in the gut of Mastotermes darwiniensis (Berchtold et al., 1994; Berchtold and König, 1996); a parallel study of Paster et al. (1996) was aimed at characterizing the spirochetes in the gut of Nasutitermes lujae. At about the same time, Ohkuma and coworkers attempted to characterize the full diversity of archaea and bacteria in the intestinal tract of Reticulitermes speratus and Cryptotermes domesticus using a similar strategy (Ohkuma et al., 1995; Ohkuma and Kudo, 1996a; Ohkuma and Kudo, 1998).

Although the cultivation-independent approach documented the presence of many new, hitherto uncultivated phylotypes in the intestinal tracts of termites (for a review, see Kudo et al., 1998), these early studies lacked resolution since only small numbers of clones were investigated. Later studies documented that diversity coverage of the clone libraries was far from exhaustive, even if larger numbers of clones were used. The most comprehensive assessment of molecular diversity and bacterial community structure in termite guts to date involved the gut microbiota of Reticulitermes species. Hongoh et al. analyzed 14 clone libraries (96 clones each) of the bacterial 16S rRNA genes in the hindgut of the Japanese termite species Reticulitermes speratus to characterize phylogenetic diversity and to address the bias introduced by different primer combinations and PCR conditions (Hongoh et al., 2003a; Hongoh et al., 2003b). Yang et al. (2005) performed a similar analysis with more than 500 clones from the European termite species Reticulitermes santonensis, focusing on the differences between the bacterial communities in the four major intestinal habitats: the midgut, the wall of the hindgut paunch, the hindgut fluid, and the intestinal protozoa (see the section Gut Compartmentation and Microhabitats in this Chapter).

The intestinal community of the two Reticulitermes species is quite similar, comprising representatives of several bacterial phyla (Fig. 6). Both termite species harbor Gram-positive bacteria (mainly clostridia, streptococci, and Mycoplasmatales-related clones), members of the Bacteroidetes, spirochetes, and a number of Proteobacteria, albeit at slightly different ratios. A large number of clones fall into the so-called “Termite group 1” (TG-1) phylum, which were most abundant in Reticulitermes santonensis (Yang et al., 2005); spirochetal clones were less abundant in this termite but accounted for approximately half of the analyzed clones in Reticulitermes speratus (Hongoh et al., 2003b).
Fig. 6.

Relative abundance of the major bacterial phyla in clone libraries of 16S rRNA genes from the hindgut of the wood-feeding, lower termite Reticulitermes speratus and the soil-feeding, higher termite Cubitermes orthognathus. Data from Hongoh et al. (2003a) and Schmitt-Wagner et al. (2003b).

The situation in soil-feeding termites is quite different. In a study analyzing bacterial diversity in the highly compartmentalized intestinal tract of Cubitermes orthognathus (Schmitt-Wagner et al., 2003a; Schmitt-Wagner et al., 2003b), the authors combined clone analysis with FISH and a molecular fingerprinting analysis, which not only substantiated the data obtained by the individual approaches but also allowed differences in community structure between the different gut compartments to be investigated. In contrast to the situation in Reticulitermes speratus, the bacterial clone libraries contained no clones from the TG-1 phylum and only few spirochetal clones. In the anterior gut sections, most clones represented Firmicutes. In the posterior gut sections, clones belonging to the Bacteroidetes and different subgroups of the Proteobacteria gained some numerical significance. A study of the bacterial microbiota in the P1 compartment of several higher termites extended the presence of a compartment-specific microbiota and a predominance of Firmicutes in the highly alkaline gut regions also to representatives of other feeding guilds (Thongaram et al., 2005).

All PCR bias notwithstanding, these large datasets, together with the numerous clones obtained from other termite species (for references, see Ohkuma, 2003), allow a reasonably accurate picture of the dominant phylogenetic groups to be drawn. Most clones obtained in the different studies represent lineages of microorganisms that were exclusively recovered from the intestinal tract of termites. The termite specificity of these lineages was underscored by the finding that the closest relatives of the bacterial clones within each lineage were usually derived also from the most-closely related termites, supporting the concept of coevolution between gut microbiota and host (Yang et al., 2005).


Not astonishingly, in view of the numerical predominance of this morphologically diverse and conspicuous group in most wood-feeding termites (Breznak, 2002), the majority of the clones obtained with bacteria-specific primers from the hindgut of Reticulitermes speratus represent spirochetes (Hongoh et al., 2003a). Already the first molecular studies had indicated that termite gut spirochetes represent a lineage phylogenetically distinct from other Spirochaetes (Berchtold and König, 1996; Ohkuma and Kudo, 1996a; Paster et al., 1996). Better diversity coverage was achieved by Lilburn et al. (1999), who targeted the intestinal spirochetes of Reticulitermes flavipes with spirochete-specific primers. They demonstrated that the 12–15 spirochete morphotypes in Reticulitermes flavipes (Breznak and Pankratz, 1977) were paralleled by 21 different spirochete phylotypes, which formed two major clusters of treponemes, one of them containing only clones of termite origin (Lilburn et al., 1999). Treponema-related clones have also been recovered from a variety of other termite species (Lilburn et al., 1999; Ohkuma et al., 1999a). In several cases, the ectosymbiotic association of certain phylotypes with flagellate protozoa has been documented using FISH with group-specific oligonucleotide probes (Berchtold and König, 1996; Iida et al., 2000; Noda et al., 2003).


In their first survey of the bacterial diversity in Reticulitermes speratus, Ohkuma and Kudo (1996a) obtained a number of clones whose sequences were only distantly related to other bacteria and which were subsequently recognized as a novel bacterial phylum (Hugenholtz et al., 1998). Also many clones in the comprehensive libraries subsequently obtained with Reticulitermes speratus (Hongoh et al., 2003a) and Reticulitermes santonensis (Yang et al., 2005) belong to this lineage, indicating that members of the TG-1 phylum represent a hitherto uncultivated but numerically dominant group of prokaryotes in the gut of Reticulitermes species.

Using a full-cycle molecular approach, combined with transmission electron microscopy, Stingl et al. (2005) showed that the TG-1 bacteria in Reticulitermes species are endosymbionts that colonize—exclusively and in high abundance—the cytoplasm of the larger flagellate species. The symbionts were specific for their respective host flagellate and were provisionally classified in the candidate genus “Endomicrobium.” Members of the TG-1 phylum, for which the name “Endomicrobia” has been proposed, are phylogenetically quite diverse and seem to be present in and also restricted to the guts of those insects (lower termites and wood-feeding cockroaches of the genus Cryptocercus) that are in mutualistic association with such cellulose-fermenting flagellates (Stingl et al., 2005).


In clone libraries of lower termites, clones affiliated with the Clostridia are abundant and fall into apparently termite-specific lineages (Hongoh et al., 2003a; Yang et al., 2005). In the higher termite Cubitermes orthognathus, they dominated the clone library of the alkaline hindgut sections (Schmitt-Wagner et al., 2003b), which was confirmed using FISH with cluster-specific probes and supported by molecular fingerprints of the different gut compartments (Schmitt-Wagner et al., 2003a).

One of the clostridial clusters from the Cubitermes orthognathus clone libraries falls into the Clostridium propionicum group (Schmitt-Wagner et al., 2003b). Interestingly, this cluster comprises also clones from the termite Nasutitermes takasagoensis, which were localized in the mixed segment between the midgut epithelium and the peritrophic membrane using FISH (Tokuda et al., 2000). Other clones are affiliated with homoacetogenic isolates (see the section Homoacetogenic Bacteria in this Chapter).

Clone libraries and molecular fingerprints indicated that Clostridia dominate also the bacterial microbiota in the most alkaline hindgut compartment (P1) of Termes comis, Pericapritermes lathignathus, and a Microcerotermes species (all Termitinae), whereas Bacilli dominate the P1 of a Speculitermes species (Apicotermitinae) (Thongaram et al., 2005). Many of the clones derived from the P1 region form phylogenetic clusters that are unique to termites and are often related to clones obtained from the other insects with alkaline digestive tracts, which suggests that they represent lineages of alkaliphilic bacteria (Schmitt-Wanger et al., 2003; Thongaram et al., 2005).


Clones affiliated with the Bacteroidetes were recovered from the guts of numerous termite species (Ohkuma et al., 2002b; Schmitt-Wagner et al., 2003b). There is an enormous diversity of such phylotypes in Reticulitermes species (Hongoh et al., 2003a; Yang et al., 2005). Most clones are only distantly related to described taxa and often form monophyletic clusters with clones recovered from the gut of other termite species. While some of the phylotypes seem to be associated with the hindgut cuticle (Yang et al., 2005), others represent epibionts of protozoa (see the section Interactions Between Prokaryotes and Protozoa in this Chapter).

Other Groups

Clone libraries of Reticulitermes species contained clones related to the Mycoplasmatales in a distinct and apparently termite-specific lineage (Hongoh et al., 2003a; Yang et al., 2005) that were abundant in the protozoan fraction of Reticulitermes santonensis (Yang et al., 2005) and comprised also a clone obtained from a symbiont of the termite gut flagellate Koruga bonita from Mastotermes darwiniensis by single-cell PCR (Fröhlich and König, 1999a). Among the Lactobacillales, most clones were affiliated with the genus Streptococcus and were mainly from the midgut clone library. Most clones affiliated with the Proteobacteria formed distinct, termite-specific lineages in the α-subgroup (only distantly related to other lineages of the Rickettsiales) or in the β-subgroup (most closely related to Dechlorimonas agitatus or to fermenting bacteria of the genus Propionivibrio; Brune et al., 2002). Clones belonging to the δ-subgroup were rare but virtually identical to the sequences of Desulfovibrio desulfuricans and of a sulfate-reducing isolate from Reticulitermes santonensis (Kuhnigk et al., 1996).

Only a single clone among the >100 clones retrieved from the hindgut of the soil-feeding termite Cubitermes orthognathus was affiliated with the Planctomycetales (Schmitt-Wagner et al., 2003b). However, a large fraction of the cells in the posterior hindgut of the closely related Cubitermes ugandensis hybridized with a mixture of FISH probes targeting this phylum. This severe underestimation of this phylum in the clone libraries is probably caused by the inadequacy of the commonly used Bacteria-specific PCR primers to amplify the 16S rRNA genes of planctomycetes (Derakshani et al., 2001) and underlines the importance of backing the results of PCR-based analyses with an independent method. Although FISH analysis indicates that more than one-third of the bacteria in the second hindgut compartment (P3 segment) of Cubitermes ugandensis may be planctomycetes (Schmitt-Wagner et al., 2003b), their metabolic function remains obscure.

Archaeal Diversity

Molecular phylogenetic profiling of the microbial communities by dot-blot hybridization with domain-specific probes has indicated that archaea represent 0.1–2.6% of small subunit (SSU) rRNA extracted from the guts of 24 nutritionally and taxonomically diverse termite species (Brauman et al., 2001). Interestingly, the relative abundance of archaea seems to be related to the host diet. The percentage of archaeal 16S rRNA among prokaryotic 16S rRNA in the gut of soil-feeding termite species (1.4–3.1%) was significantly higher than in wood-feeding and litter-feeding termite species (0.1–1.7%). This is in agreement with the methane emission rates, which are generally higher among soil-feeding termite species (Brauman et al., 1992), and it has been speculated that the majority of the archaea in termite guts are methanogens (Brauman et al., 2001).

Methanogenic Archaea

Partial sequences of the genes encoding for 16S rRNA and for subunit A of the methyl coenzyme M reductase (mcrA) of methanogens indicated that the methanogens in Reticulitermes speratus belong to the order Methanobacteriales (Ohkuma et al., 1995) and are closely related but not identical to the Methanobrevibacter species isolated from Reticulitermes flavipes (Leadbetter and Breznak, 1996; Leadbetter et al., 1998; also, see the subsection Methanogenic Archaea in section “Hydrogen Metabolism”). Later studies concentrated on the 16S rRNA genes and confirmed the presence of Methanobacteriales in Reticulitermes speratus (Shinzato et al., 1999), Cryptotermes domesticus (Ohkuma and Kudo, 1998; Shinzato et al., 2001), Hodotermopsis sjöstedti (Ohkuma et al., 1999c), Neotermes koshunensis, Reticulitermes kanmonensis, Coptotermes formosanus, and Mastotermes darwiniensis (Shinzato et al., 2001). All sequences cluster within the radiation of the genus Methanobrevibacter, but the sequences from termites differ from those of known methanogens, forming unique lineages in the phylogenetic trees. A single clone related to Methanomicrobiales was recovered from Reticulitermes speratus (Shinzato et al., 1999).

Dot-blot hybridization indicated that Methanobacteriales constitute one-third to more than one-half of the archaea in the guts of almost all termite species studied (Brauman et al., 2001). By contrast, Methanosarcinales seem to be present only in the guts of about half of the termite species, apparently forming the dominant group of methanogens in 4 of the 24 species studied and accounting for the total archaeal signal in the fungus-growing species Macrotermes subhyalinus (Brauman et al., 2001). Additionally, the clones retrieved from the guts of the phylogenetically higher termites Nasutitermes takasagoensis, Odontotermes formosanus and Pericapritermes nitobei clustered mostly among the Methanomicrobiales and Methanosarcinales (Ohkuma et al., 1999c). In a detailed study of archaeal diversity in the gut of the soil-feeding higher termite, Cubitermes orthognathus, most archaeal clones were affiliated with Methanobacteriales, Methanomicrobiales and Methanosarcinales, and a few clones had their closest relatives among the Methanococcales (Friedrich et al., 2001). Similar results were obtained in a study with Cubitermes fungifaber, which also corroborated that there is little overlap between the communities of methanoarchaea present in the gut and in the food soil (Donovan et al., 2004). In contrast, the similarities between the methanoarchaeal communities of congeneric termites are substantial, and many clones obtained from the intestinal tract of termites cluster with clones retrieved from other insects. However, a purely vertical transmission of the methanogenic gut microbiota is not supported (Donovan et al., 2004).

Methanobrevibacter spp. in Reticulitermes are associated with the hindgut wall (Leadbetter and Breznak, 1996; Leadbetter et al., 1998) and, in the gut of Reticulitermes speratus and Hodotermopsis sjoestedti, attached to the flagellated protist species Dinenympha and Microjoenia (Tokura et al., 2000): there are indications that the lineages attached to the flagellates are phylogenetically different from those associated with the gut epithelium. Fröhlich and König (1999a) retrieved single cells of endosymbiotic methanogens from the anaerobic flagellate Pentatrichomonoides scroa occurring in the hindgut of Mastotermes darwiniensis that were affiliated with the genus Methanobrevibacter.

Methanogens are among the few groups of organisms for which one can infer metabolic information from the 16S rRNA gene sequence. FISH with an archaea-specific probe revealed that archaea are largely restricted to the gut sections P3 and P4 in Cubitermes ugandensis, which is in agreement with the distribution of F420-fluorescent cells and methane-emission rates along the gut axis of Cubitermes species (Schmitt-Wagner and Brune, 1999). Cells hybridizing with the archaea-specific probe presented 1.6% and 3.8% of the DAPI-stained cells in the P3 and P4 section, respectively (Schmitt-Wagner et al., 2003b), but since many of the abundant and morphologically diverse F420-fluorescent microorganisms in these gut sections were filamentous forms and appeared to be fragmented or destroyed during homogenization, they were likely underestimated by the FISH analysis.

Non-methanogenic Archaea

In a dot-blot analyses of many termite species, the total combined value of the subgroup-specific probes was much lower than that of the Archaea domain probe, indicating that termite guts may contain (possibly non-methanogenic) archaeal populations whose 16S-like rRNAs do not hybridize with probes for methanoarchaeal subgroups employed in this study (Brauman et al., 2001).

Shinzato et al. (1999) provided the first evidence for the presence of Thermoplasmales in the intestinal tract of Reticulitermes speratus. Friedrich et al. (2001) obtained numerous clones of Thermoplasmales and a few clones related to the Thermococcales from the gut of the soil-feeding termite Cubitermes orthognathus and also documented for the first time the presence of crenarchaeota in an intestinal tract. Donovan et al. (2004) retrieved several clones related to the haloalkaliphile genus Natronococcus from Cubitermes fungifaber, which is quite intriguing in view of the extreme alkalinity of the anterior hindgut of Cubitermes species (Brune and Kühl, 1996).

Spatial Organization

As mentioned above, termite guts are axially and radially structured, providing numerous microhabitats with different physicochemical microenvironments (see the section The Gut Microenvironment in this Chapter). Not astonishingly, therefore, the distribution of gut microbiota is not random but seems to be spatially organized. Detailed descriptions of the spatial arrangement of the intestinal prokaryotes in situ (Breznak and Pankratz, 1977; To et al., 1978; To et al., 1980; Czolij et al., 1985; Yara et al., 1989), together with numerous other observations of certain morphotypes in particular regions of the gut (see the section Gut Compartmentation and Microhabitats in this Chapter), indicate that many microhabitats harbor characteristic microbial populations.

Until recently, most of such evidence was based purely on morphological data, and the line of evidence is far from complete. With the advent of molecular tools, however, it became possible to address not only the diversity of the termite gut microbiota but also the spatial distribution of individual phylotypes or phylogenetic groups. A study employing whole-cell hybridization in homogenates of different gut regions and in situ hybridization of gut cryosections with group-specific oligonucleotide probes provided the first results documenting differences in microbial community structure between different regions of the hindgut of Mastotermes darwiniensis at the level of phylogenetically defined microbial groups (Berchtold et al., 1999). Other studies used FISH to document the association of gut protozoa with certain phylotypes of hitherto uncultivated spirochetes (Berchtold and König, 1996; Iida et al., 2000; Noda et al., 2003), Bacteroidetes (Wenzel et al., 2003; Stingl et al., 2004), members of the TG-1 phylum (Stingl et al., 2005), or methanogenic archaea (Tokura et al., 2000).

Also, PCR-based approaches allow one to resolve differences in the community structure of different gut regions or gut compartments (Friedrich et al., 2001; Schmitt-Wagner et al., 2003b; Yang et al., 2005). Costs and effort involved in sequencing and phylogenetic analysis limit investigations based on clone analysis, but molecular fingerprinting allows expansion of the investigation of diversity and—observing the necessary cautions inherent to all PCR-based techniques—community structure to include a larger number of samples. Terminal-restriction-fragment-length polymorphism (T-RFLP) analysis in the higher, soil-feeding termite Cubitermes orthognathus documented that the different archaeal and bacterial populations are not randomly distributed along the gut and that the prokaryotic communities in the individual gut segments differ considerably with respect to diversity and abundance (Friedrich et al., 2001; Schmitt-Wagner et al., 2003a). By contrast, the bacterial community structure in homologous compartments in three different species of Cubitermes was quite similar, indicating the existence of gut-segment-specific communities (Schmitt-Wagner et al., 2003a).

Isolates and Major Metabolic Activities

The major metabolic activities of the gut microbiota have been outlined (Breznak, 2000; Slaytor, 2000; Tholen and Brune, 2000; Fig. 7), but there are still considerable gaps in our knowledge that underline the need for a refined concept (Tholen and Brune, 2000). Numerous attempts have been made to characterize the prokaryotic gut microbiota of termites by isolating microorganisms that either were numerically abundant or possessed a metabolic potential considered important in the metabolism of the hindgut. Many of these efforts have yielded results of uncertain significance, either because the methods were not fully described or because no quantitation of bacteria was made relative to the total number of microorganisms present. As there is little overlap between the existing isolates and the 16S rRNA genes obtained in the molecular studies (see the section Phylogenetic Diversity and Community Structure in this Chapter), it is apparent that many of the microorganisms responsible for the major metabolic activities remain to be cultivated.
Fig. 7.

Schematic presentation of the metabolic processes involved in the fermentative degradation of polysaccharides in the hindgut of Reticulitermes flavipes. The dashed lines indicate metabolic fluxes which seem to be strongly influenced by the continuous influx of oxygen into the gut periphery. Line thickness indicates the relative importance of the process. The major metabolic groups are gut flagellates (1), primary (2) and secondary (3) fermenting bacteria, homoacetogens (4), and methanogens (5); it remains to be clarified whether the flagellates are also a major source of lactate (?). Short-chain fatty acids are oxidized by the host (6). Adapted from Brune (2003).

Numerically Predominant Isolates

Most cultivable heterotrophic bacteria in the hindgut of Reticulitermes flavipes are Streptococcus and Enterococcus species, followed by Bacteroides species and representatives of the Enterobacteriaceae, mostly Citrobacter species and Enterobacter cloacae (Schultz and Breznak, 1978; Tholen et al., 1997). Coccoid lactic acid bacteria also dominated among the isolates obtained from the hindguts of the lower termites Mastotermes darwiniensis and Cryptotermes primus (Eutick et al., 1978a) and the higher termites Nasutitermes arborum, Thoracotermes macrothorax and Anoplotermes pacificus (Bauer et al., 2000). While Enterobacter species were found to dominate among the isolates from the rhinotermitid species Heterotermes ferox, Coptotermes acinaciformes, Coptotermes lacteus and Schedorhinotermes intermedius (Eutick et al., 1978a), most isolates from termitid species Nasutitermes exitiosus, Nasutitermes graveolus and Nasutitermes walkeri were staphylococci. Isolates from these genera have been recovered also in earlier studies (e.g., Mannesmann and Piechowski, 1989).

Although some of these studies were at best semiquantitative and many employed only aerobic techniques, the pattern of bacterial species cultivated from each host species is remarkably constant. Most importantly, the majority of the isolates obtained from termite guts are either aerobes or aerotolerant anaerobes. The absence of obligate anaerobes among the isolates in those studies that did not attempt to apply a methodology appropriate for the successful cultivation of such bacteria is not astonishing. However, even in the studies that explicitly used the Hungate technique and employed reduced media for cultivation, the fraction of obligate anaerobes was always smaller than that of aerotolerant anaerobes and aerobes (Schultz and Breznak, 1978; Tholen et al., 1997). It is not clear whether this phenomenon is caused by the oxygen status of the termite gut (see the section Redox Conditions and Oxygen Status in this Chapter) or the strong cultivation bias against certain groups of bacteria (see the section Cultivation Bias in this Chapter).

Cultivation Bias

A comparison of the viable counts of heterotrophic bacteria to the direct microscopic counts of the microorganisms in the hindgut of Reticulitermes flavipes indicates that about 90% of the microbial cells have escaped cultivation (Schultz and Breznak, 1978; Tholen et al., 1997). Viable counts obtained in a similar study attempting to characterize the major gut bacteria of nine species of termites indicate an even larger cultivation bias (Eutick et al., 1978a).

When the isolates obtained in these studies are compared to the results of the cultivation-independent characterization (see the section Phylogenetic Diversity and Community Structure in this Chapter), there are enormous discrepancies between the frequencies of the phylogenetic groups dominating the clone libraries (Fig. 6) and the species recovered by cultivation. Nevertheless, many isolates are unique to the termite gut habitat, and their characterization has provided valuable information on metabolic properties and other physiological features relevant for the colonization of this particular habitat (Table 1). To increase cultivation efficiency, it will be necessary to develop new cultivation strategies that take into account the special environmental conditions within the gut, in particular the steep physicochemical gradients (see the section The Gut Microenvironment in this Chapter) and the metabolic interactions among the microbiota (see the section Microbe-Microbe Interactions in this Chapter).

Table 1.

Described species of prokaryotes unique to the intestinal tract of termites.

Group species

Termite speciesa

Unusual feature



Acetonema longum

Pterotermes occidentis (K)


Kane and Breznak, 1991

Bacillus oleronius

Reticulitermes santonensis (R)

Degrades aromatic compounds

Kuhnigk et al., 1995

Clostridium mayombei

Cubitermes speciosus (T)


Kane et al., 1991

Clostridium termitidis

Nasutitermes lujae (T)


Hethener et al., 1992

Isoptericola variabilis (formerly Cellulosimicrobium variabile)

Mastotermes darwiniensis (M)


Bakalidou et al., 2002 Stackebrandt et al., 2004

Sporobacter termitidis

Nasutitermes lujae (T)


Grech-Mora et al., 1996

Sporomusa aerivorans

Thoracotermes macrothorax (T)


Boga and Brune, 2003

Sporomusa termitida

Nasutitermes nigriceps (T)


Breznak et al., 1988

Sporotomaculum hydroxybenzoicum

Cubitermes speciosus (T)

Degrades aromatic compounds

Brauman et al., 1998


Desulfovibrio intestinalis

Mastotermes darwiniensis (M)


Fröhlich et al., 1999

Desulfovibrio termitidis

Heterotermes indicola (R)


Trinkerl et al., 1990


Sebaldella termitidis (formerly Bacteroides termitidis)

Reticulitermes lucifugus (R)


Collins and Shah, 1986


Candidatus Vestibaculum illigatumb

Neotermes cubanus (K)

Epibiont of Staurojoenina sp.

Stingl et al., 2004


Treponema azotonutricium

Zootermopsis angusticollis (Z)


Graber et al., 2004

Treponema primitia

Zootermopsis angusticollis (Z)


Graber et al., 2004


Candidatus Endomicrobium trichonymphaeb

Reticulitermes santonensis (R)

Endobiont of Trichonympha agilis

Stingl et al., 2005

Candidatus Endomicrobium pyrsonymphaeb

Reticulitermes santonensis (R)

Endobiont of Pyrsonympha vertens

Stingl et al., 2005


Methanobrevibacter curvatus

Reticulitermes flavipes (R)


Leadbetter and Breznak, 1996

Methanobrevibacter cuticularis

Reticulitermes flavipes (R)


Leadbetter and Breznak, 1996

Methanobrevibacter filiformis

Reticulitermes flavipes (R)


Leadbetter et al., 1998

aTermites belong to the families Kalotermitidae (K), Mastotermitidae (M), Rhinotermitidae (R), Termitidae (T), and Termopsidae (Z).

bCandidatus taxon: not cultivated, but well-characterized with respect to morphology, ultrastructure, phylogeny, and specific location.

Like many arthopods, termites harbor filamentous bacterial morphotypes with refractile inclusions resembling endospores. These bacteria are usually attached to the hindgut wall (Fig. 8) and were first described in 1849 as “Arthromitus” by Leidy (1849, 1881). The filaments have not been cultivated, but Margulis et al. (1998) have proposed that they represent a different life stage of aerobic, rod-shaped bacteria closely related to Bacillus cereus, a species group that occurs ubiquitously in soil. However, their conclusions were based merely on the isolation of such bacteria from the boiled intestines of ten species of soil arthropods containing "Arthromitus"-like filaments and on earlier reports on the isolation of B. cereus from arthropod guts. A phylogenetic identity of the filaments with the isolates has not been confirmed with molecular methods. In this context, it should be noted that other authors had previously identified the segmented filamentous bacteria in the gut of mice, rats, and chickens as a new lineage of Clostridiales, based on a 16S rRNA gene sequence analysis (Snel et al., 1994). The same group has proposed “Candidatus Arthromitus” as a provisional generic name for the segmented filamentous bacteria falling into this lineage (Snel et al., 1995)—unfortunately without verifying whether they are indeed related to the morphologically similar filaments in arthropods originally described by Leidy.
Fig. 8.

Phase contrast photomicrograph of the hindgut wall of Reticulitermes santonensis, colonized by Pyrsonympha flagellates and “Arthromitus”-like filaments (arrows). Reprinted from Yang et al. (2005).

Lignocellulose Degradation

There are numerous reports on the presence in termite guts of enzyme activities acting on different cellulose and hemicellulose preparations. The enzyme activities in foregut and midgut are most likely of host origin or are ingested fungal cellulases that remain active within the gut (see the section Fiber Degradation in this Chapter), whereas most activities in the hindgut are probably due to the microbiota. In the lower termites, it is also important to differentiate between protozoan and bacterial origin.

Cellulolytic Bacteria

Cellulolytic prokaryotes have been isolated from the guts of lower and higher termites on numerous occasions (for references, see Breznak [1975], Breznak and Brune [1994b], and Wenzel et al. [2002]). However, many of these efforts either were unsuccessful or have yielded positive results of uncertain significance, either because the methods were not fully described or because the population size had not been established (Schultz and Breznak, 1978). Therefore, the contribution of bacteria to cellulose degradation in the termite gut has always been a matter of debate (see the section Fiber Degradation in this Chapter).

Interestingly, successful attempts to isolate cellulolytic bacteria usually employed oxic cultivation conditions; only a few anaerobic strains have been reported. Hungate isolated an anaerobic actinomycete, “Micromonospora propionici,” from Amitermes minimus (Hungate, 1946). Clostridium termitidis was isolated from the gut of the wood-feeding termite Nasutitermes lujae (Hethener et al., 1992). All attempts to isolate anaerobic cellulolytic bacteria from the gut of Reticulitermes flavipes were negative (Schultz and Breznak, 1978). No cellulose-degrading bacteria were present among the numerically predominant isolates recovered from nine species of termites representing all major families, using both aerobic and anaerobic techniques (Eutick et al., 1978a). No anaerobic cellulolytic bacteria were isolated from the soil-feeding termite Cubitermes speciosus using serial dilutions of gut homogenates and the Hungate technique (Brauman et al., 1990b).

By contrast, numerous studies have yielded aerobic bacteria, albeit of often dubious numerical significance. Frequent isolates, found in numerous studies, were assigned to Serratia marcescens and Bacillus cereus (e.g., Thayer, 1976). Strains of both species, isolated from Reticulitermes hesperus, formed soluble, and in the case of Bacillus cereus also cell-bound, carboxymethylcellulases (Thayer, 1978). Serratia marcescens seems to be a common but minor inhabitant of the intestinal tract of insects that on occasion can become pathogenic (see the section Pathogens in this Chapter).

On the basis of the few studies where the number or density of cellulolytic bacteria in the gut had been determined (e.g., Paul et al., 1993; Wenzel et al., 2002), it appears that bacteria are not very relevant for cellulose digestion. The most abundant among 23 groups of cellulolytic isolates from Zootermopsis angusticollis were closely affiliated with Bacillus megaterium, Bacillus cereus, or Paenibacillus polymyxa (Wenzel et al., 2002). Many of the cellulolytic isolates are species known to occur also in soil and other habitats and may represent transient permanent populations of microorganisms ingested with the food, while others are apparently specific for the intestinal tracts of insects (Table 1).

Isoptericola variabilis (formerly Cellulosimicrobium variabile; Stackebrandt et al., 2004) and other closely related bacteria were isolated from the hindgut of Mastotermes darwiniensis and various other termites (Bakalidou et al., 2002) and form significant populations also in the gut of other insects (Cazemier et al., 2003).

Cellulolytic activity has been detected also among numerous actinomycetes isolated from the gut of several soil-feeding Termitidae (Pasti and Belli, 1985); some of the isolates were also lignolytic (Pasti et al., 1990). Although an unusual association of soil-feeding termites (Termitidae, Termitinae) with actinomycete-like bacteria has been documented (Bignell et al., 1979; Bignell et al., 1980b) and facultatively aerobic actinomycetes have been isolated from the gut, parent soil, and mound materials of the termites Procubitermes aburiensis and Cubitermes severus (Bignell et al., 1991), their significance in the degradation of lignocellulose remains to be established.

Hemicellulolytic Bacteria

Compared to the information on cellulose degradation, our understanding of the degradation of the hemicellulose component of lignocellulose in termite guts is rather meager. Xylanase activity has been observed in midgut and hindgut of several termites, including wood-feeding, soil-feeding, and fungus-cultivating species (Breznak and Brune, 1994b; Rouland, 2000). In the fungus-cultivating Macrotermes bellicosus, there is evidence that the source of xylanase activity may be a symbiotic fungus (Matoub and Rouland, 1995), but analogous to the host cellulases, a host origin of the midgut xylanases remains to be scrutinized.

In lower termites, the gut flagellates also seem to play a major role in the degradation of xylan (see the section Role of Intestinal Protozoa in this Chapter). However, also bacteria and yeasts might be involved in the metabolism of hemicelluloses. Aerobic and facultatively anaerobic hemicellulose-degrading bacteria and yeasts were isolated from the guts of several wood-feeding termites, with xylan-degrading bacteria (106–107 per ml) dominating in Mastotermes darwiniensis and xylan-degrading yeasts (107–5 × 108 cells per ml) in Zootermopsis angusticollis and Neotermes castaneus (Schäfer et al., 1996). Gram-positive isolates belonged to the genera Bacillus and Paenibacillus or to the Actinobacteria, while Gram-negative strains were affiliated with the genera Pseudomonas, Acinetobacter, and Ochrobactrum or with the enterobacteria.

Several strains of alkaliphilic bacteria, which were isolated from the extremely alkaline P1 compartment the soil-feeding Sinocapritermes mushae and Amitermes longignathus and represent a novel lineage of Paenibacillus, express alkalitolerant xylanase activity with a high PH optimum (Ohkuma et al., 2003; Thongaram et al., 2003). An unusual xylanase, distantly related to xylanases of bacteria and fungi colonizing the bovine rumen (GHF11), has been discovered in a genomic library to microbial DNA extracted from the intestinal tract of an unspecified higher termite belonging to the subfamily Nasutiermitinae (Brennan et al., 2004).

Lignin-Degrading Bacteria

The degradation of lignin involves the “enzymatic combustion” of the highly recalcitrant polyphenolic resin by peroxidases. Despite a few claims to the contrary, there is currently no conclusive evidence that bacteria lacking extracellular peroxidative activity solubilize or degrade polymeric core lignin significantly, and lignin degradation by the gut microbiota has been questioned (Breznak and Brune, 1994b; Kuhnigk et al., 1994). A 13C-nuclear magnetic resonance (NMR) analysis of the feces of the wood-feeding termites Microcerotermes parvus and Coptotermes formosanus indicated a preferential loss of polysaccharide during gut passage, whereas lignin accumulated and was modified only in its O-aromatic-C and O-methyl-C components were modified (Hopkins et al., 1998; Hyodo et al., 1999).

However, extensive delignification occurs in the fungal gardens of fungus-cultivating termites (Hyodo et al., 2000) and has to be expected also in any wood colonized by lignolytic fungi. Many lignin-derived aromatic compounds (representing the major subunits and lignin-carbohydrate linkages found in lignins and their depolymerization products) can be degraded by aerobic bacteria, and numerous bacteria degrading lignin monomers or certain dimeric lignin model compounds have been isolated from various wood-feeding termites (Kuhnigk et al., 1994; Kuhnigk and König, 1997; Harazono et al., 2003). They comprised a wide range of Gram-positive bacteria and Proteobacteria, with strict aerobes dominating in Nasutitermes nigriceps and enterobacteria in the lower termites. An aerobic isolate able to degrade hydroxybenzoic acids was described as a new species, Bacillus oleronius (Kuhnigk et al., 1995).

All attempts to enumerate bacteria degrading aromatic compounds indicated that termite guts contain significant numbers of aerobic bacteria capable of degrading aromatic rings, whereas anaerobic degradation of the aromatic nucleus appears to be of little significance (Brune et al., 1995b; Kuhnigk and König, 1997). Under anoxic conditions, only ring and side-chain modification seem to be of importance (Kuhnigk et al., 1994; Brune et al., 1995b). Some isolates from termite guts involved in anaerobic modification or degradation of aromatic compounds were novel taxa (Table 1).

The digestibility of lignocellulose is improved if the phenylpropanoic acid residues esterified to the hemicellulose chains and the diferulic ester linkages between the hemicellulose chains in plant cell walls are hydrolyzed, since this increases the solubility of the macromolecules and reduces steric hindrance of hydrolytic enzymes (Jeffries, 1994). A strain of Clostridium xylanolyticum (producing hydroxycinnamoyl esterases) that hydrolyzes and then hydrogenates ferulic and p-coumaric acid residues has been isolated from the gut of the grass-feeding termite Tumulitermes pastinator (McSweeney et al., 1999). Several termites have been shown to possess also activities cleaving non-phenolic β-O-4 type-lignin model compounds representing the major linkage within the lignin macromolecule (Hirai et al., 2000).

Lignin-solubilizing actinomycetes have been isolated from the gut of soil-feeding termites (Pasti et al., 1990), and all isolates were Streptomyces strains. Moreover, in lower termites, the only isolate decolorizing Remazol Brilliant Blue and Azure B was a Streptomyces strain (Kuhnigk and König, 1997). Since isolation and phylogenetic characterization of actinomycetes from termite guts indicated that the actinomycete flora of termites depends largely on their geographical origin (Watanabe et al., 2003), it is not clear whether such isolates simply represent spores ingested with the food.

Oxygen Reduction

The bacteria and protozoa colonizing the gut periphery, especially those directly associated with the gut epithelium (Fig. 8), are exposed to the continuous influx of oxygen (Brune et al., 2000b). Oxygen microprofiles indicate that the anoxic status of the termite hindguts must be attributed to the oxygen consumption of the microbiota located in the gut periphery (see the section Redox Conditions and Oxygen Status in this Chapter), which can represent a substantial fraction of the respiratory activity of the host (Brune et al., 1995a). This conclusion is supported by the predominance of facultatively or even obligately aerobic isolates in all cultivation-based studies.

However, there are several indications that not all oxygen-consuming activities in termite guts are of a respiratory nature. In the extremely alkaline gut regions, the high rates of oxygen consumption might be partly attributable also to chemical reactions, such as the autoxidation of phenolic residues in lignin or humic substances (see Kappler and Brune, 1999). On the other hand, radiotracer analysis of the in situ metabolism in the hindgut of Reticulitermes flavipes demonstrated that the high oxygen fluxes also significantly influence fermentation processes in the hindgut (Tholen and Brune, 2000). This is supported by the oxygen reduction potential found in obligate anaerobes such as lactic acid bacteria (see the section Lactic Acid Bacteria in this Chapter), homoacetogenic bacteria (see the section Homoacetogenic Bacteria in this Chapter), and sulfate-reducing bacteria (see the section Sulfate-reducing Bacteria in this Chapter).

Microbial Fermentations

In lower termites, the bulk of the fermentative activity should be caused by the intestinal protozoa, which immediately phagocytize the wood particles entering the hindgut. The identity of the substrates of the prokaryotic microbiota is far from clear. Since most of the prokaryotic symbionts probably do not participate in fiber degradation, presumably they ferment either soluble products entering the hindgut from the midgut or intermediates released by the protozoa (Fig. 7). The composition of short-chain fatty acids and other fermentation products in the hindgut fluid of termites indicates that a variety of microbial fermentations occur simultaneously (Schultz and Breznak, 1979; Odelson and Breznak, 1983; Anklin-Mühlemann et al., 1995; Tholen and Brune, 2000). However, the exact nature and amount of the monomeric substrates entering the hindgut are not known, and the spectrum of fermentation products formed by the protozoa might be much larger than initially assumed (Tholen and Brune, 2000).

The apparent absence of pyruvate dehydrogenase activity in termite tissues (O’Brien and Breznak, 1984) and several other observations prompted the suggestion that the termite releases pyruvate into the hindgut to be converted to acetate and reabsorbed by the host (Slaytor et al., 1997; Slaytor, 2000). However, recent findings documented that termite mitochondria possess a pyruvate dehydrogenase complex, although the difference in activity suggests differences in pyruvate metabolism between lower and higher termites (Itakura et al., 1999; Itakura et al., 2003).


Coccoid lactic acid bacteria predominate among the carbohydrate-utilizing bacteria isolated from the hindguts of termites of the families Mastotermitidae, Kalotermitidae, and Rhinotermitidae (Eutick et al., 1978a; Schultz and Breznak, 1978; Tholen et al., 1997; Bauer et al., 2000) and have been isolated also from several representatives of the Termitidae (Eutick et al., 1978a; Bauer et al., 2000). The most abundant isolates obtained from the hindguts of Reticulitermes flavipes and Thoracotermes macrothorax belong to the genera Enterococcus and Lactococcus (Bauer et al., 2000). However, the low frequency of such clones in cultivation-independent studies indicates that lactic acid bacteria might be present only in moderate numbers (Hongoh et al., 2003a; Schmitt-Wagner et al., 2003b; Yang et al., 2005).

Physiological characterization of the isolates has revealed high potential rates of O2 reduction in the presence of fermentable substrates (Tholen et al., 1997; Bauer et al., 2000), which might represent an adaptation to variable oxygen tensions and could explain why lactococci and enterococci are regularly encountered in the intestinal tracts of termites and other insects and possibly restricted to specific compartments of the intestinal tract (Yang et al., 2005).

Clostridiales Compared to their abundance in bacterial clone libraries derived from termite guts, clostridial isolates are underrepresented in cultivation-based studies. Several isolates have been described as new species and seem to be unique to termite guts (Table 1), but their numerical significance is either low or has not been established. Bacteria distantly related to Clostridium oroticum, detected in an alkaline enrichment culture derived from gut homogenates of Pericapritermes latignathus, may prove to be first exception, since they were represented also in a clone library obtained from the P1 compartment (Thongaram et al., 2005).


Besides the coccoid lactic acid bacteria, a relatively large proportion of the isolates from the hindgut of Reticulitermes flavipes have been identified as Bacteroides species (Schultz and Breznak, 1978). The strains fall into two groups: aeroduric anaerobes fermenting lactate to propionate and acetate (Schultz and Breznak, 1979) and anaerobic strains forming butyrate and isobutyrate as characteristic products from complex media (Schultz and Breznak, 1978). The only described Bacteroides species isolated from termite guts, Bacteroides termitidis (Sebald) from Reticulitermes lucifugus, was phylogenetically misplaced (Paster et al., 1985) and recently reclassified among the Fusobacteria in the genus Sebaldella (Collins and Shah, 1986). A recent isolate, obtained from the hindgut of Reticulitermes speratus by dilution plating of gut suspensions, has a 16S rRNA gene sequence identical to that of a group of clones recovered exclusively from termite guts, representing a lineage of bacteria within the radiation of the Bacteroides subgroup but only distantly related to the genus Bacteroides (Ohkuma et al., 2002b).

The uricolytic activity of Bacteroides isolates from termite guts (Potrikus and Breznak, 1980) suggests a potential role in nitrogen cycling (see the section Nitrogen Recycling in this Chapter). Many of the clones from termite guts clustering among the Bacteroidetes fall into a lineage comprising epibionts of protozoa (see the section Interactions Between Prokaryotes and Protozoa in this Chapter).

Hydrogen Metabolism

The enormous accumulation of hydrogen at the gut center and the steep radial hydrogen gradients in the gut periphery of several termites (Ebert and Brune, 1997; Schmitt-Wagner and Brune, 1999) indicate that molecular hydrogen is a key intermediate in the microbial food chain. Hydrogen-dependent CO2 reduction by methanogens and homoacetogens is probably the most important hydrogen sink in termite hindguts. To understand the metabolism in termite guts, it is important to identify the hydrogen-producing and hydrogen-consuming populations and their functional interactions (see the section Microbe-Microbe Interactions in this Chapter).

Hydrogen-Producing Microorganisms

On the basis of the few pure culture studies available, it is generally assumed that the polysaccharides of the wood particles taken up by the gut flagellates are oxidized to acetate and CO2, and the reducing equivalents are released as H2 (see Breznak and Brune [1994b] and Brune and Stingl [2005]; Fig. 7). This is in agreement with the enormous accumulation of hydrogen in the hindgut of Reticulitermes flavipes (Ebert and Brune, 1997; Fig. 4). In view of the wide variety of clostridial clones retrieved from Reticulitermes species (Hongoh et al., 2003a; Yang et al., 2005), bacterial fermentations may also contribute to the intestinal H2 production. Such assumptions are substantiated by the considerable amounts of H2 produced in certain gut regions of soil-feeding Cubitermes species (Schmitt-Wagner and Brune, 1999), a group of termites that lack gut flagellates but contain a similar assemblage of clostridial clones (Schmitt-Wagner et al., 2003b). The cellulolytic bacterium Clostridium termitidis isolated from Nasutitermes lujae (Hethener et al., 1992) is affiliated with a number of clones from Cubitermes orthognathus guts (Schmitt-Wagner et al., 2003b), but nothing is known about the physiology of the bacteria represented by these 16S rRNA genes.

Homoacetogenic Bacteria

Reductive acetogenesis from H2 and CO2 in termite gut homogenates was first demonstrated by Breznak and Switzer (1986). In the following years, the presence of homoacetogenic activities was established for a large number of termite species from all major feeding guilds, including representatives of wood-feeding, fungus-cultivating, and soil-feeding termites (Breznak and Kane, 1990; Brauman et al., 1992; Williams et al., 1994). Although reductive acetogenesis in gut homogenates of soil-feeding termites was always outcompeted as a hydrogen sink by methanogenesis (Brauman et al., 1992; Breznak, 1994a), microinjection of H14CO3 into intact hindguts of soil-feeding Cubitermes spp. has identified a high potential for reductive acetogenesis (Tholen and Brune, 1999), which indicates that the contribution of reductive acetogenesis to the overall electron flow in the guts of soil-feeding termites might be larger than expected. Possibly, part of the explanation lies in the ability of homoacetogens to grow mixotrophically on H2 and other substrates (Breznak and Switzer, 1991; Breznak, 1994a).

Although reductive acetogenesis has been demonstrated to occur in a large number of termite species from all major feeding guilds (Brauman et al., 1992), only seven species of homoacetogenic bacteria from termites have been described to date (Table 1). Sporomusa termitida (isolated from Nasutitermes nigriceps; Breznak et al., 1988), Sporomusa aerivorans (isolated from Thoracotermes macrothorax; Boga et al., 2003b), and Acetonema longum (isolated from Pterotermes occidentis; Kane and Breznak, 1991a) are three spore-forming representatives in the Sporomusa group of Gram-positive bacteria that are characterized by a Gram-negative cell wall. All isolates are capable of H2-dependent reduction of CO2 to acetate and possess a large potential for hydrogen-dependent oxygen reduction (Boga and Brune, 2003a).

Clostridium mayombei has been isolated from Cubitermes speciosus (Kane et al., 1991b) and belongs to the Clostridium lituseburense group. A very unusual homoacetogen (Sporobacter termitidis, isolated from Nasutitermes lujae) clusters in the Clostridium leptum group together with numerous clones of 16S rRNA genes retrieved from the anterior hindgut of Cubitermes orthognathus (Schmitt-Wagner et al., 2003b). It grows exclusively by the disproportionation of methyl groups derived from methoxylated aromatic compounds but not on H2 + CO2 or other typical substrates of homoacetogens, and it methylates sulfide and cysteine if these compounds are present in the medium (Grech-Mora et al., 1996).

Recently, numerous spirochetal strains have been isolated from the gut of Zootermopsis angusticollis (Leadbetter et al., 1999). They belong to the Treponema branch of Spirochaetes that contains mainly sequences obtained from termite guts (Lilburn et al., 1999), and many are capable of H2-dependent acetogenesis from CO2 (Leadbetter et al., 1999). One of the homoacetogenic isolates, which has been described as a new species, Treponema primitia (Graber et al., 2004b), utilizes a wide range of substrates and is capable of mixotrophic growth, i.e., the simultaneous utilization of H2 and organic substrates (Graber and Breznak, 2004a). These represent important clues for the unknown metabolic function of the enormous populations of spirochetes colonizing the hindgut lumen of termites and the surfaces of many intestinal protozoa (Breznak, 2002).

Generally, there seems to be a strong cultivation bias against homoacetogens from termite guts, and their numerical abundance and contribution to reductive acetogenesis in situ are not clear. Recently, sequences clustering with the formyl tetrahydrofolate synthase (FTHFS) homologues of termite-gut spirochetes were found to dominate the diversity of genes in the hindgut of Zootermopsis angusticollis coding for FTHFS (Salmassi and Leadbetter, 2003). FTHFS is a key enzyme in reductive acetogenesis, which makes the FTHFS gene a good functional marker for homoacetogenic bacteria (Leaphart and Lovell, 2001; Leaphart et al., 2003). Quite likely therefore homoacetogenic spirochetes are responsible for the large potential activities of H2-dependent acetogenesis encountered in most wood-feeding termites (Breznak, 1994a; Breznak, 2002).

All investigated homoacetogens isolated from termite guts catalyzed hydrogen-dependent oxygen reduction (Boga and Brune, 2003a); the isolate Sporomusa aerivorans has the highest known oxygen-reducing capacity among obligately anaerobic bacteria other than sulfate-reducing bacteria of the genus Desulfovibrio (see the section Sulfate-Reducing Bacteria in this Chapter), which possibly indicates an adaptation to the oxygen status of the gut environment. Although the location of homoacetogens relative to the oxygen gradient remains to be established, it is rather unlikely that the homoacetogens in Reticulitermes flavipes contribute substantially to the removal of oxygen diffusing into the gut: in contrast to methanogenesis, the in situ rates of reductive acetogenesis in this termite species are not affected by oxygen (Tholen and Brune, 2000).

Methanogenic Archaea

Methane formation by termites, first documented by Breznak (Breznak, 1975; Odelson and Breznak, 1983), has received considerable attention owing to its implication for the global methane budget (Sanderson, 1996). Methane formation is restricted to the archaea; therefore, the methane production by all termites investigated indicates that methanogenic archaea are present in virtually all termites (Brauman et al., 1992). Furthermore, the gas bubbles around termites entrapped in amber (Fig. 9), which have been attributed to continuing methanogenesis in the fresh resin, are palaeontologic evidence for the presence of methanogenic archaea in termite guts.
Fig. 9.

Termites entrapped in a block of copal (young amber) from the Andean uplift region of Boyaca Province, Colombia (Pleistocene). The photograph shows gas bubbles around the termites, presumably methane that escaped from the body after the termites had been engulfed by the resin.

The autofluorescence of coenzyme F420 allows easy visualization of methanogens by epifluorescence microscopy. They are either located free in the hindgut lumen, attached to the hindgut cuticle, or associated with other microorganisms (Breznak and Brune, 1994b). In Zootermopsis, methanogens are almost exclusively associated with the gut flagellates (Lee et al., 1987). On the basis of their F420-like autofluorescence, filamentous bacteria colonizing the cuticular spines in the fourth proctodeal segment of soil-feeding termites (Bignell et al., 1980b) also appear to be methanogens (Schmitt-Wagner et al., 2003b).

Despite their relative abundance and phylogenetic diversity in termite guts (see the section Archaeal Diversity in this Chapter), to date only a few methanogens from Reticulitermes flavipes have been isolated in pure culture. The isolates represent new species in the genus Methanobrevibacter (Table 1), grow best with H2 and CO2, and form large populations that are attached to the luminal side of the gut epithelium or adhere to other prokaryotes colonizing the hindgut wall (Leadbetter and Breznak, 1996; Leadbetter et al., 1998).

Already the location in the microoxic gut periphery and the presence of catalase activity in Methanobrevibacter cuticularis indicated a considerable oxygen tolerance (Leadbetter and Breznak, 1996). This was substantiated by the finding that – despite a general inability among Methanobacteriales to synthesize heme – the wetwood isolate Methanobrevibacter arboriphilus expresses a catalase when the medium is supplemented with hemin (Shima et al., 2001). The documentation of a F420H2 oxidase in M. arboriphilus (Seedorf et al., 2004) finally provided also a biochemical basis for the high rates of H2-dependent O2 reduction exhibited by Methanobrevibacter species colonizing the periphery of the termite hindgut (A. Tholen and A. Brune, unpublished results) and their apparent ability to cope with the continuous influx of oxygen (Leadbetter and Breznak, 1996; Ebert and Brune 1997).

Sulfate-Reducing Bacteria

Sulfate-reducing bacteria have been isolated from the intestinal tracts of many different termite species (Brauman et al., 1990a; Brauman et al., 1990b; Trinkerl et al., 1990; Kuhnigk et al., 1996; Fröhlich et al., 1999b). All isolates are members of the genus Desulfovibrio and seem to form substantial populations in the gut of certain termites. The relevance of sulfate reduction under in situ conditions is not clear, but the sulfate-reducing bacteria in termite guts might partake also in the removal of oxygen or play a role in syntrophic fermentations (Kuhnigk et al., 1996). Like other Desulfovibrio species isolated from sediments (for references, see Cypionka, 2000), all strains isolated from termite guts show extremely high rates of O2 reduction in the presence of H2 (Kuhnigk et al., 1996; Fröhlich et al., 1999b).

Nitrogen Transformations

The diet of the termites ranges from sound wood to lignocellulosic plant materials in various stages of humification, including soil and animal dung. Owing to the high C-to-N ratio of sound wood, many xylophagous termites are strongly limited by nitrogen (Collins, 1983). They conserve nitrogen by recycling—a strategy that has been termed “carbon elimination” (Higashi et al., 1992). Another way to improve the C/N balance of the colony is through nitrogen fixation.

This section will summarize the present knowledge on these processes and the microorganisms involved. For details, the reader is referred to the comprehensive review of Breznak (2000).

Nitrogen Recycling

The best way to deal with a rare resource is conservation and recycling. Termites, like other insects, secrete uric acid and urea, the waste products of nucleic acid and protein metabolism, via the Malpighian tubules into the intestinal tract (Terra, 1990; Fig. 3). Potrikus and Breznak (1981) demonstrated that termites lack uricase and that uric acid is recycled by anaerobic bacteria in the hindgut of Reticulitermes flavipes, including Streptococcus and Citrobacter species and Bacteroides (now Sebaldella) termitidis (Potrikus and Breznak, 1980).

Despite the low nitrogen content of the diet, nitrogen recycling creates high ammonia concentrations in the hindgut of wood-feeding termites, which allows the maintenance of an active gut microbiota and thus ensures high rates of carbon mineralization. The concentration of ammonia in the paunch of Nasutitermes walkeri is in the range of 3 mM (Slaytor and Chappell, 1994). The efficient assimilation of ammonium into microbial biomass (and possibly also resorption of ammonium in the posterior hindgut) avoids the loss of nitrogen via the feces (Breznak, 2000). The transformation to high-quality microbial protein also leads to an upgrading of any low-quality nitrogen in the diet.

The nitrogen cycle is closed by the digestion of microbial cells. Since termites cannot access the microbes in the hindgut directly, worker larvae solicit hindgut contents from their nestmates. This behavior, termed “proctodeal trophallaxis,” is unique to this group of social insects and increases in frequency with the level of nitrogen limitation (Machida et al., 2001). Digestion of the hindgut contents and resorption of the nitrogenous products probably take place in the foregut and midgut (Fig. 3), as indicated by lysozyme and protease activities in these gut regions (Fujita et al., 2001; Fujita and Abe, 2002; Fujita, 2004). The efficiency of nitrogen conservation within the colony is increased further by the consumption of exuviae and dead nestmates.

Nitrogen Fixation

Though nitrogen cycling is efficient in termites, termite colony growth is limited by the net nitrogen taken in with food (Breznak, 2000). Therefore, many termites show a preference for lignocellulosic substrates that are colonized by fungi and therefore have a decreased C-to-N ratio (Amburgey et al., 1980; Cornelius et al., 2002). Termites living on sound wood, however, rely on the exclusive capacity of their prokaryotic gut microbiota to fix atmospheric nitrogen.

After the first convincing demonstrations of N2 fixation in termites using the acetylene reduction assay (Benemann, 1973; Breznak et al., 1973) and the incorporation of 15N into biomass by termites incubated under an 15N2-enriched atmosphere (Bentley, 1984), a large body of data on nitrogen fixation in termite guts has accumulated. A few aspects will be mentioned; for more details, the reader is again referred to the excellent review of this subject by Breznak (2000).

Nitrogen fixation is widespread among termites, although the rates differ considerably. Nitrogenase activity in certain Nasutitermes spp. would be sufficient to double the nitrogen content of a colony within a few years (Breznak, 2000), and stable isotope analysis has revealed that 30–60% of the nitrogen in Neotermes koshunensis workers is derived via this pathway (Tayasu et al., 1994).

However, the nitrogen content of the diet increases with humification of the organic matter, and peptides, amino sugars, and microbial biomass are potential sources of nutrition for soil-feeding Cubitermes spp. (Ji et al., 2000; Ji and Brune, 2001). The natural abundance of the 15N isotope indicates that dietary nitrogen and not atmospheric nitrogen fixation is an important nitrogen source in humivorous termite species (Tayasu et al., 1997; Tayasu, 1998). The enormous ammonium concentrations in feces and nests (constructed from feces) of soil-feeding termites (Ndiaye et al., 2004; R. Ji and A. Brune, unpublished results) suggest that termites in this feeding guild are not nitrogen-limited.

There are also many indications that the nitrogen fixation rates cannot easily be extrapolated to the colony level or ecosystem level. Apart from methodological aspects, there are interspecific differences, intraspecific variations, seasonal patterns, and effects of laboratory maintenance (Curtis and Waller, 1995; Curtis and Waller, 1998) to consider. Additionally, the reduced oxygen partial pressure in subterraneous nests and galleries has to be taken into account (Curtis and Waller, 1996). The problems are discussed in detail by Breznak (2000).

Dinitrogen-fixing bacteria isolated from termite guts include enterobacteria (French et al., 1976; Potrikus and Breznak, 1977; Eutick et al., 1978a), sulfate-reducing bacteria (Kuhnigk et al., 1996), and spirochetes (Leadbetter et al., 1999; Graber et al., 2004b). Molecular approaches, using the nifH gene as a functional marker, have revealed a much wider spectrum of potentially N2-fixing microorganisms, including clostridia, Proteobacteria, and methanogenic archaea, comprising several sequence clusters unique to termites (Ohkuma et al., 1996b; Ohkuma et al., 1999b). However, a gene-expression study at the community level has revealed that only a few nifH homologs, probably representing alternative nitrogenases (anf genes), were actually transcribed under in situ conditions in the gut of Neotermes koshunensis (Noda et al., 1999). The preferential expression of anf genes might be related to an insufficient molybdenum (Mo) supply with the lignocellulosic diet (Ohkuma, 2002a). The number of unrelated nifH homologs in a single spirochete species and the variety of apparently spirochete-related nifH homologs in hindgut clone libraries indicate that spirochetes are potentially important nitrogen-fixing microorganisms in termite guts (Lilburn et al., 2001).

Symbiotic Interactions

The enormous complexity of this subject emerges from the large number of species present within the gut, the different organizational levels of the potential partners, their metabolic capacities and topological orientation, and the numerous possibilities for metabolic interactions especially within anaerobic communities. In addition, symbiotic interactions can be extremely specific with respect to the partners involved or may simply consist of a cross-feeding of metabolites or nutrients between populations that are not even in direct contact with each other.

Microbe–Microbe Interactions

Trophic cascades and cross-feeding of metabolites can form a network of interactions between the individual populations of any microbial community. In termite guts, the resulting uneven distribution of sources and sinks of microbial metabolites within the system gives rise to metabolic gradients (see the section Physicochemical Gradients in this Chapter), which are indicators of the location of microbial activities in situ. Additionally, the spatial organization of different populations, including intimate associations between prokaryotes and protozoa, underlines the importance of studying the nature of the respective interactions.

Interactions Among Prokaryotes

At least three metabolically different groups of microorganisms are involved in the metabolism of hydrogen in the gut of lower termites: the protozoa, which produce H2 as a product of their fermentative metabolism, and methanogens and homoacetogens, which reduce CO2 to methane or acetate, respectively (see the section Hydrogen Metabolism in this Chapter). It is generally assumed that hydrogenotrophic methanogens outcompete homoacetogens owing to their higher affinity for the common substrate (Cord-Ruwisch et al., 1988). Nevertheless, both metabolic groups occur simultaneously in the hindgut of termites. In Reticulitermes flavipes, where methanogens are mostly restricted to the gut periphery (Leadbetter and Breznak, 1996), in situ rates of reductive acetogenesis surpass those of methanogenesis considerably (Tholen and Brune, 2000), whereas in Zootermopsis angusticollis, where methanogens are located mostly inside the gut protozoa (Lee et al., 1987), methanogenesis appears to be the major hydrogenotrophic process (Brauman et al., 1992).

The methanogens in termites seem to be hydrogen-limited in situ, as indicated by the stimulation of methane emission by externally supplied H2 (Odelson and Breznak, 1983; Messer and Lee, 1989; Ebert and Brune, 1997; Schmitt-Wagner and Brune, 1999). Also, the per cell rates of methanogenesis determined in vitro with Methanobrevibacter cuticularis, multiplied by the viable counts of methanogens in Reticulitermes flavipes, are much higher than the methane emission rate of living termites (Leadbetter and Breznak, 1996). The relative rates of methane and hydrogen emission by different termites vary considerably (Sugimoto et al., 1998), and evidence is accumulating that the spatial organization of the hydrogen-producing and hydrogen-consuming microorganisms determines the competition and coexistence of different populations (Ebert and Brune, 1997; Tholen and Brune, 2000).

Although reductive acetogenesis is outcompeted as a hydrogen sink by methanogenesis in gut homogenates of soil-feeding termites (Brauman et al., 1992), intact guts of soil-feeding Cubitermes spp. show a large potential for reductive acetogenesis (Tholen and Brune, 1999). Taking into account the possibility of an intercompartmental transfer of H2 (Schmitt-Wagner and Brune, 1999), as has been demonstrated between the midgut and hindgut compartments of cockroaches (Lemke et al., 2001), the contribution of reductive acetogenesis to the overall electron flow in the guts of soil-feeding termites might be larger than expected.

A cross-feeding of hydrogen could also be the basis for the interaction between the prokaryotic epibionts on filamentous bacteria in the gut of Reticulitermes flavipes (Breznak and Pankratz, 1977), some of which seem to be methanogenic archaea (Leadbetter and Breznak, 1996). Schultz and Breznak (1979) had demonstrated a cross-feeding of lactate between Lactococcus lactis and a propionigenic Bacteroides species, to consolidate the predominance of these species among the cultivable bacteria in the hindgut fluid of Reticulitermes flavipes with the absence of lactate accumulation. Microinjection of 14C-labeled lactate confirmed that lactate is an important intermediate under in vivo conditions but raised several new questions regarding the source of lactate and the role of oxygen in shifting the product spectrum from propionate to acetate (Tholen and Brune, 2000).

An example for cross-feeding of nutrients has been provided by Graber and Breznak (2005), who made a convincing case that Treponema primitia, a homoacetogenic spirochete in the gut of Zootermopsis angusticollis that requires folate for growth, benefits from the excretion of 5-formyl-tetrahydrofolate by other community members (Lactococcus lactis and Serratia grimesii).

Interactions Between Prokaryotes and Protozoa

Most of the flagellate protozoa in the guts of lower termites are associated with prokaryotic microorganisms, either as epibionts on their cell surface or as endobionts in the cytoplasm and in the nucleus. The abundance and variety of such associations are documented by numerous electron microscopy studies (for references, see Honigberg [1970], Radek [1999a], Inoue et al. [2000], and Dolan [2001]), but virtually nothing is known about the physiological basis of these associations, especially the metabolic interactions (Brune and Stingl, 2005).

Cook (1932) was the first to observe the emission of a gas other than CO2 by Termopsis nevadensis (syn. Zootermopsis nevadensis) and speculated that intestinal protozoa could produce hydrogen, methane, or possibly a mixture of both. Many decades later, termites were shown to emit both hydrogen and methane (Breznak, 1975; Odelson and Breznak, 1983), and axenic cultures of Trichomitopsis termopsidis (a gut flagellate from Zootermopsis) were shown to produce only hydrogen after methanogenic symbionts were eliminated by treating the cultures with bromoethanesulfonate, an inhibitor of methanogenesis (Odelson and Breznak, 1985a; Odelson and Breznak, 1985b).

Methanogenic symbionts of protozoa can be easily visualized by epifluorescence microscopy (Fig. 10). Lee et al. (1987) reported that methanogenic bacteria in the hindgut of Zootermopsis angusticollis are associated only with the small flagellated protozoa Trichomitopsis termopsidis, Tricercomitus termopsidis and Hexamastix termopsidis, whereas they were not observed in the large, hypermastigid protozoa. On the basis of the results of various treatments that selectively eliminated or affected certain protozoa from the gut of this termite, Messer and Lee (1989) concluded that the large protozoa of the genus Trichonympha were the most important hydrogen source in the hindgut, whereas the methanogenic symbionts of Trichomitopsis termopsidis produced most of the methane.
Fig. 10.

Phase contrast (A) and epifluorescence (B) photomicrographs of a small trichomonad in Schedorhinotermes lamanianus and its attached epibionts, showing the typical F420-autofluorescence of methanogenic archaea. Photomicrographs taken by M. Pester.

The large gut flagellates are often colonized by epibiotic bacteria. The presence of special attachment sites on the cell envelope of the flagellates (e.g., Tamm, 1980; Radek et al., 1992; Radek et al., 1996; Dolan and Margulis, 1997; Stingl et al., 2004) indicates a tight association. In some cases, the epibionts seem to be responsible for the locomotion of the host.

The polymastigote flagellate Mixotricha paradoxa, which occurs exclusively in the gut of Mastotermes darwiniensis, uses its four flagella only for steering. The cell is propelled by the many thousands of spirochetes that cover most of the body surface. The spirochetes are attached to projecting brackets of the cell surface in a manner that allows the helical movement of the individual cells to travel in metachronal waves along the cell surface of the host, resulting in locomotion (Cleveland and Grimstone, 1964). The epibiotic spirochetes were recently identified as members of the Treponema cluster by 16S rRNA gene sequence analysis (Wenzel et al., 2003). In addition, a second, rod-shaped epibiont is intimately associated with the cell surface by specific attachment sites (Fig. 11); it is affiliated with other uncultivated Bacteroidales.
Fig. 11.

Transmission electron micrograph (A) of the epibiotic bacteria on the cell surface of Mixotricha paradoxa, a large flagellate occurring exclusively in the gut of Mastotermes darwiniensis that lives in a motility symbiosis with prokaryotes. The schematic presentation (B) illustrates the regular arrangement of the spirochetal (S) and rod-shaped (B) epibionts and the special attachment brackets (br) at the cell surface. Reproduced from Cleveland and Grimstone (1964).

The devescovinid flagellate Caduceia versatilis (d’Ambrosio et al., 1999) carries two different, rod-shaped and filamentous, epibionts. In this case, the host is propelled by the self-synchronizing movement of the flagella of several thousand rod-shaped bacteria, which are deeply embedded into its the cell surface (Tamm, 1982). The epibionts of Staurojoenina flagellates (Fig. 12), recently assigned to the candidate taxon “Vestibaculum illigatum” (Stingl et al., 2004), have a similar morphology, although in this flagellate, motility is not due to the bacteria but to their own flagella. Nevertheless, “Vestibaculum illigatum” falls into the same cluster of Bacteroidales as the rod-shaped epibiont of Mixotricha paradoxa (Stingl et al., 2004), a lineage that also contains numerous clones obtained from other termites (Ohkuma et al., 2002b; Yang et al., 2005).
Fig. 12.

Scanning (A, B) and transmission (C) electron micrographs of Staurojoenina sp. from Neotermes cubanus. A) Overview of a cell, showing two of the four flagellar tufts (f), numerous bacterial rods (b), and occasional spirochetes (s) attached to the surface. B) Close-up of the cell surface, showing single spirochetes between the ectobiotic rods. Arrows point to early stages of cell division. C) Cross-section of the epibiotic rods. In addition to an inner membrane (im) and outer membrane (om), the cell is surrounded by a diffuse layer (dl). Electron-dense material supports the plasma membrane of the flagellate below the attachment sites (arrowhead). The arrow points to a phagocytized rod-shaped bacterium with remnants of attachment complex. Reprinted from Stingl et al. (2004).

The larger flagellate species in the gut of Reticulitermes species harbor numerous prokaryotic endobionts within their cytoplasm (Bloodgood et al., 1974; Bloodgood and Fitzharris, 1976). The endobionts of Trichonympha and Pyrsonympha species in Reticulitermes santonensis (Fig. 13) belong to the candidate phylum “Endomicrobia”, which seem to occur exclusively in termite gut flagellates (Stingl et al., 2005). To date, there is no indication of their possible function. An involvement in hydrogen metabolism is unlikely since the oxymonadid Pyrsonympha species are phylogenetically distant from the hypermastigid flagellates (Moriya et al., 2003; Stingl and Brune, 2003) and do not seem to possess hydrogenosomes (see Brune and Stingl, 2005).
Fig. 13.

Transmission-electron micrographs of ultrathin sections of Trichonympha agilis (A) and Pyrsonympha vertens (B) showing the ultrastructure of the endosymbiotic bacteria in the candidate genus “Endomicrobium,” which are very abundant in the cytoplasm of these flagellate species. Cells are surrounded by two membranes; the outermost membrane forms tube-like elongations at the tapered cell poles (arrow in A). g, glycogen. Bars = 0.2 µm. Reprinted from Stingl et al. (2005).

Microbe-Host Interactions

Although the host provides a favorable habitat for the intestinal microbiota and thrives on their metabolic products, there is not always a necessity or clear indication for any specific interactions. However, the intimate associations of bacteria with microvilli of the midgut epithelium or with the epithelial cups in the hindgut cuticle (Breznak and Pankratz, 1977) indicate a closer integration of certain bacteria with the host tissues. Another example suggesting interactions with the gut tissue are the clostridia-related bacteria located in the ectoperitrophic space between the midgut wall and the peritrophic membrane in the mixed segment, closely associated with the mesenteric epithelium (Tokuda et al., 2000; Tokuda et al., 2001).


A few reports indicate that the gut microbiota of termites is not always beneficial. Serratia marcescens, which has been isolated from the guts of Reticulitermes hesperus (Thayer, 1976) and Mastotermes darwiniensis (Kuhnigk et al., 1994), causes septicemia in Coptotermes formosanus (Osbrink et al., 2001). Serratia marcescens seems to form part of the normal microflora of insects since it can be isolated from both healthy and diseased specimens. Usually nonpathogenic when present in the digestive tract in small numbers, it multiplies rapidly once it enters the hemocoel and causes death in 1–3 days (for references, see Lysenko [1985] and Sikorowski and Lawrence [1998]). Also, the fungi associated with Reticulitermes flavipes include both cellulolytic species and potential pathogens (Zoberi and Grace, 1990). It is not clear whether entomopathogenic fungi form a part of the natural gut community or are only found among diseased insects (Rath, 2000).

Intracellular Symbionts

Termites, like many other insects, are also associated with intracellular prokaryotes that are vertically transmitted via the ovaries (Breznak, 2004). Flavobacteria of the genus Blattabacterium reside in specialized cells (bacteriocytes) of cockroaches and the most primitive termite, Mastotermes darwiniensis (Bandi et al., 1995). The close phylogenetic relationship between endosymbionts from Mastotermes darwiniensis and members of the wood-feeding cockroach genus Cryptocercus supports the hypothesis that termites evolved from subsocial, wood-dwelling cockroaches (Lo et al., 2003). All other termites examined carried endocytoplasmic bacteria that are affiliated with the Wolbachia group and are located in nonspecialized fat body cells (Bandi et al., 1997). Endonuclear bacteria have been observed in the trophocytes of Reticulitermes lucifugus and Kalotermes flavicollis (Grandi et al., 1997).

Mutualists or Commensals?

In the hindgut symbiosis, the host creates a rather constant environment for its symbionts, provides a continuous supply of substrates, and by the transfer of the microorganisms to other nestmates, ensures their propagation within the ecosystem. In the symbiosis between lower termites and fiber-digesting flagellates, both partners are indispensable and the mutual advantage is obvious. However, in the case of most prokaryotic gut symbionts, the symbionts appear to be of little advantage to the host.

If the benefit of the association is unidirectional, a symbiont would be classified as a commensal, and the host might even benefit from its elimination (Brune, 2003). Unfortunately, it is not easy to eliminate specific prokaryotes from the intestinal microbial community of termites selectively, and it is difficult to distinguish between the direct and indirect consequences of their elimination. Removal of the spirochetes from the gut of Nasutitermes exitiosus by feeding metronidazole or exposing the termites to pure oxygen kills the termites almost as rapidly as the complete removal of all bacteria by antibiotics (Eutick et al., 1978b).

In view of the constant and massive inoculation with microorganisms in their diet, it is clear that termites cannot keep their gut sterile. The molecular diversity studies have provided ample evidence for a specific microbiota, composed of lineages that occur exclusively in termites (Yang et al., 2005). The absence of a “normal gut microflora” would allow the uncontrolled proliferation of ingested foreign microorganisms and increase the danger of colonization by potential pathogens. Serratia marcescens failed to colonize the gut of normal Coptotermes lacteus, but transient colonization occurred when the protozoa and spirochetes were killed by exposure to pure oxygen (Veivers et al., 1982).

Under oxygen-limited conditions, the energy loss caused by the activity of the gut microbiota is relatively small (Table 2). The hypothetical, nonsymbiotic oxidation of cellulose to CO2 would allow the termite to exploit 100% of the free energy of the reaction (Eqn. 1). A homoacetogenic degradation of one glucose equivalent to three acetate molecules by the gut microbiota (Eqn. 2 + Eqn. 3) releases only 10.8% of the free energy contained in the substrate; the rest is still available for the host (Eqn. 4), which also benefits from the nutritionally valuable microbial biomass. However, the situation is quite different if the H2 formed in the microbial fermentations is converted to methane (Eqn. 2 + Eqn. 5). Although the amount of free energy released in methanogenesis is not much larger than that released in reductive acetogenesis, the host receives only two-thirds of the acetate available under homoacetogenic conditions (Eqn. 6); 28.4% of the free energy in the substrate remains in methane, which cannot be exploited by the host (Eqn. 7) and is lost to the environment.

Table 2.

Free energy of important reactions involved in symbiotic digestion.a


ΔG°′ (kJ/mol)b

Relative changec (%)

(1) Glucose + 6 O2 → 6 CO2 + 6 H2O



(2) Glucose + 2 H2O → 2 Acetate + 2 H+ + 2 CO2 + 4 H2



(3) 4 H2 + 2 CO2 → Acetate + H+ + 2 H2O



(4) 3 Acetate + 3 H+ + 6 O2 → 6 CO2 + 6 H2O



(5) 4 H2 + CO2 → CH4 + 2 H2O



(6) 2 Acetate + 2 H+ + 4 O2 → 4 CO2 + 4 H2O



(7) CH4 + 2 O2 → CO2 + 2 H2O



aFor explanations, see text.

bGibbs free energy under standard conditions at pH 7 is calculated after Thauer et al. (1977).

cFree energy change relative to the aerobic oxidation of glucose (Reaction 1).

Such energetic considerations suggest an obvious advantage if a termite uses reductive acetogenesis as the hydrogen-consuming process, and why especially the more highly evolved termites (Termitidae) show a tendency towards increasing methane emission rates is still a mystery (Brauman et al., 1992; Bignell et al., 1997). Nevertheless, considering the added value of metabolic properties such as nitrogen fixation, ammonia assimilation, and provision with vitamins, the advantages for the host may be well worth the investment. The exploitation of the nutritive resources contained in the microbial biomass requires the digestion of the gut symbionts, which is realized by proctodeal trophallaxis (Machida et al., 2001), a behavioral trait that was probably fundamental to both the establishment of the gut microbial community and the evolution of sociality in termites (Nalepa et al., 2001).


Termite guts are minute but efficient bioreactors for the conversion of lignocellulose to short-chain fatty acids and microbial biomass. However, termite guts are not simply anoxic fermentors, but axially and radially structured environments with physicochemically distinct microhabitats, and we are just beginning to understand the complex interactions within the intestinal microbial communities. Microbial diversity in the termite gut is enormous, and the existing isolates represent only a negligible fraction of the untapped diversity of prokaryotes in the guts of the more than 2600 described species of termites (Kambhampati and Eggleton, 2000).

Copyright information

© Springer-Verlag 2006
Show all