Termites and flagellates
- Termites and flagellates
- What Kills the Hindgut Flagellates of Lower Termites during the Host Molting Cycle?
- 1. Introduction
- 2. Why Is Death of Protists in Termites of Significance?
- 2.1. Interdependence
- 2.2. Misconceptions
- 3. Molting in Cryptocercus
- 4. Molting in Termites
- 5. Why Is the Partnership Dissolved at Every Molt in Termites?
- 6. Flagellate Differentiation during the Cryptocercus Molting Cycle
- 6.1. Cryptocercus: Variation in Flagellate Sexual Cycles
- 6.2. Cryptocercus: Variation in Flagellate Mechanism of Resistance
- 7. Link between Social Behavior and Symbiont Transmission
- 8. What Changed?
- 9. Titer of Ecdysone and Timing of Events during the SE Cycle
- Period of Vulnerability
- 10. The ‘Smoking Gun’
- 11. Exceptions to the ‘Rule’ in Termites
- Two Types of Encystment?
- 12. Hypothesis I. Protists as Victim
- 13. Hypothesis II. Coordination of Life Cycles
- 14. Future Prospects
- Conflicts of Interest
Termites and flagellates
Mol Ecol. 2009 Jan;18(2):332-42. doi: 10.1111/j.1365-294X.2008.04029.x.
Symbiotic flagellates play a major role in the digestion of lignocellulose in the hindgut of lower termites. Many termite gut flagellates harbour a distinct lineage of bacterial endosymbionts, so-called Endomicrobia, which belong to the candidate phylum Termite Group 1. Using an rRNA-based approach, we investigated the phylogeny of Trichonympha, the predominant flagellates in a wide range of termite species, and of their Endomicrobia symbionts. We found that Trichonympha species constitute three well-supported clusters in the Parabasalia tree. Endomicrobia were detected only in the apical lineage (Cluster I), which comprises flagellates present in the termite families Termopsidae and Rhinotermitidae, but apparently absent in the basal lineages (Clusters II and III) consisting of flagellates from other termite families and from the wood-feeding cockroach, Cryptocercus punctulatus. The endosymbionts of Cluster I form a monophyletic group distinct from many other lineages of Endomicrobia and seem to have cospeciated with their flagellate host. The distribution pattern of the symbiotic pairs among different termite species indicates that cospeciation of flagellates and endosymbionts is not simply the result of a spatial separation of the flagellate lineages in different termite species, but that Endomicrobia are inherited among Trichonympha species by vertical transmission. We suggest extending the previously proposed candidatus name ‘Endomicrobium trichonymphae’ to all Endomicrobia symbionts of Trichonympha species, and estimate that the acquisition by an ancestor of Trichonympha Cluster I must have occurred about 40-70 million years ago, long after the flagellates entered the termites.
What Kills the Hindgut Flagellates of Lower Termites during the Host Molting Cycle?
Subsocial wood feeding cockroaches in the genus Cryptocercus, the sister group of termites, retain their symbiotic gut flagellates during the host molting cycle, but in lower termites, closely related flagellates die prior to host ecdysis. Although the prevalent view is that termite flagellates die because of conditions of starvation and desiccation in the gut during the host molting cycle, the work of L.R. Cleveland in the 1930s through the 1960s provides a strong alternate hypothesis: it was the changed hormonal environment associated with the origin of eusociality and its concomitant shift in termite developmental ontogeny that instigates the death of the flagellates in termites. Although the research on termite gut microbial communities has exploded since the advent of modern molecular techniques, the role of the host hormonal environment on the life cycle of its gut flagellates has been neglected. Here Cleveland’s studies are revisited to provide a basis for re-examination of the problem, and the results framed in the context of two alternate hypotheses: the flagellate symbionts are victims of the change in host social status, or the flagellates have become incorporated into the life cycle of the eusocial termite colony. Recent work on parasitic protists suggests clear paths for exploring these hypotheses and for resolving long standing issues regarding sexual-encystment cycles in flagellates of the Cryptocercus-termite lineage using molecular methodologies, bringing the problem into the modern era.
‘One in general does not kill the goose that laid the golden egg’
Overwhelming evidence is now available from morphological characters and genomic analyses of both hosts and symbionts that termites are a clade nested within the cockroaches, with the wood-feeding, subsocial cockroach Cryptocercus as their undisputed sister group [2,3,4,5]. A prominent feature uniting Cryptocercus with the termites is a symbiotic relationship with single-celled anaerobic eukaryotes living in the enlarged hindgut paunch of the insects. These gut protists fall into two main groups traditionally known as flagellates—the Phylum Parabasalia and the Order Oxymonadida (Phylum Preaxostyla) , with the former historically divided into the trichomonads and hypermastigotes. Nearly 450 distinct flagellate species have been described in the termite and Cryptocercus hosts that have been investigated (see [7,8] for a catalogue of those reported). The morphological characters historically used to classify species, however, are proving in many cases to be inadequate, and new species and relationships are continually being described based on molecular sequencing techniques, e.g., [9,10,11,12,13].
The flagellates of the two host insect groups have similar taxonomic affiliations, and their distributions are either limited, or they cannot be found in other hosts or in the free-living state [2,8,14]. It has been long thought [15,16] and is now strongly established that divergence of the flagellates took place in a common ancestor of Cryptocercus and termites, currently estimated as occurring
170 mya . Intergenerational transmission is vertical [18,19]: they are transferred from parents to offspring via proctodeal (anal) trophallaxis (the hatchlings feed on parental hindgut fluids) in both Cryptocercus and termites; in the latter during the incipient, subsocial stage. In both taxa, the parents are the source of all flagellates for a given social group. In termites, however, the flagellates are re-circulated among colony members via trophallaxis once the colony becomes established. As expected in vertically transmitted symbionts, flagellate communities of different host lineages indicate co-speciation, with possible horizontal shifts due to stochastic, dietary or ecological effects [20,21,22,23].
Functionally, the gut microbiota partner with the host and each other to interactively degrade plant polymers, fix atmospheric nitrogen, synthesize amino acids, and recycle nitrogenous waste products [24,25,26]. Despite the dominance of the single celled eukaryotes, the Cryptocercus/termite hindgut system is clearly a multiple symbiosis that involves flagellates, bacteria, and archaea ; all organisms involved are symbionts as well as mutualists . The vast majority of prokaryotes in the gut (up to 85%) are associated with the protists [28,29] and exhibit a high level of integration with their flagellate hosts: bacteria or spirochaetes are attached to the cell surface, or located in the cytoplasm or nucleoplasm. Further prokaryotes live free in the gut fluid or are attached to the gut wall of the insect host [27,30,31,32,33,34,35,36,37,38,39]. Clearly the Cryptocercus/lower termite gut microbiome is an interdependent, multifactorial, synergistic system whose subtle dynamics cannot be analyzed as simple binary interactions. That being said, the focus here is on the flagellates, which not only drive cellulose degradation [40,41], but also are more exacting in their requirements than prokaryotes .
Recently, significant advances have been made in the understanding of these host-symbiont relationships, prompted by the application of new, sophisticated technologies that are independent of microbial culturing. Well over 3000 primary articles and 150 reviews that address termite symbioses were published in the period 2000–2014, inclusive . Emphasis has been primarily on identification of the components of the microbiota in a phylogenetic context, and of individual biochemical functionality and interdependence. Our increased understanding of the sophistication of the microbiome, however, is not matched by an understanding of how it is integrated into the biology, behavior and life history of the insect host.
The environment in which the flagellates live is unusually homeostatic and closely comparable from host to host . The protists are housed in the gut of an individual insect, living within a family (Cryptocercus) or colony (termites), which is lodged within the buffered environment of a nest. They are supplied with a steady stream of food, in a liquid, temperature-controlled, safe haven. Their world, then, is fairly constant, with one notable exception: the molting cycle of their insect host.
If the relationship of these two insect taxa to their gut fauna during developmental ontogeny is compared, it is obvious that there has been an evolutionary change during the host molting cycle. In both Cryptocercus and lower termites, neonates acquire the flagellates from a parent or a sibling (the latter in established colonies of termites); the symbiosis is established at about the third instar in both taxa. The big difference lies in what subsequently occurs during the molting cycles of the developing juvenile host. After the symbiosis is established in Cryptocercus, it is retained through all subsequent ecdyses; it is never lost under natural conditions [15,44,45]. Developing nymphs of this cockroach are thereafter nutritionally autonomous and capable of a solitary lifestyle. In termites, however, the large flagellates die prior to subsequent ecdyses and must be re-acquired by feeding on the hindgut fluids of a nest mate. Unlike most vertically transmitted symbioses  then, developing termites have an aposymbiotic phase with respect to the larger gut flagellates. If cockroaches in the genus Cryptocercus are used as a model of the termite ancestral state, that shift in the host-symbiont relationship can be pinned to a specific node of the Dictyopteran phylogenetic tree. This paper is an attempt to explore the foundations of that evolutionary transition.
2. Why Is Death of Protists in Termites of Significance?
Among eusocial insects, the symbiosis with these gut flagellates is a feature unique to the Cryptocercus-termite lineage. Historically, the death of these flagellates at molt was thought to be a key factor associated with termite eusocial origins, e.g., [47,48] but more recent work habitually excludes its influence on social structure. Any hypothesis that ignores this termite specific life history characteristic, however, is unsatisfactory and serves to emphasize the current deep-seated bias in analyzing Isopteran eusocial origins in terms of hymenopteran attributes [49,50]. A re-examination of this symbiotic relationship during host molt is overdue for two compelling reasons: first, because the need to re-acquire the symbiosis after molt precludes independent living by termite individuals, and second, because current literature rarely describes the symbiotic relationship during the host molting cycle in either insect taxon accurately.
In both Cryptocercus and lower termites there are two levels of dependence. First, individuals rely on gut flagellates to metabolize and supplement their wood diet; and second, neonates in both taxa rely on their parents (in incipient colonies of termites) for vertical transfer of the symbionts. In termites alone, there is a third level of dependence: because the flagellates die at molt, developing termite individuals do not survive unless they have access to the hindgut fluids of a relative . The dependence on gut flagellates to help metabolize their wood diet, combined with the periodic death of these symbionts precludes independent living in termites. If the host-symbiont relationship in Cryptocercus is used as a model of the termite ancestral state, previously independent insects evolved interdependency. Up to the threshold when the protistans began dying during the termite molting cycle, eusociality in termites may have been reversible, but in interdependent organisms the fitness of one depends on the fitness of another, making cooperation the best strategy . The death of protists at molt was a ‘tipping point’ , in that beyond that threshold, intrinsic processes in the system drove accelerating change .
How the gut symbiosis is established in neonates of Cryptocercus and lower termites, and how the symbiotic bond between the host and its gut flagellates is either maintained, or broken and subsequently re-established during the host molting period are consistently misunderstood in current literature . Prior to the 1980s, the confusion was understandable because Cleveland et al. ( Figure 1 ) , the early authorities on these flagellates, did not understand the social structure and behavior of Cryptocercus; rather, these authors assumed two things about the natural history of the cockroach based on their observations of gut flagellates during Cryptocercus molting cycles. The first assumption was that neonates of Cryptocercus acquired their gut symbionts by feeding on the fecal pellets of recently molted older siblings, because these pellets contained resistant stages of the gut protists. The second assumption was that, because adults do not molt (and therefore their feces would not contain resistant stages of the symbionts), an adult pair of these cockroaches could not found a colony independently. These authors speculated that new colonies were formed by a ‘combined creeping migration’ of adults and older nymphs to a new log. Cleveland et al. ( p. 209) did not rule out that trophallaxis may play a role in establishing the gut microbiota in neonate Cryptocercus, but they did not observe the behavior during their studies.
Lemuel Roscoe Cleveland (1892–1969). Smithsonian Institution Archives, Image #SIA2008-0153. The Smithsonian has granted permission for use.
Nutting  was the first to question the described mode of symbiont transfer in Cryptocercus. A subsequent study was done on adults collected in the field as mated pairs , suggesting that adults may be able to found colonies independently. In 1983, field work by Seelinger and Seelinger  demonstrated that molting older siblings were not present in the nest when Cryptocercus egg cases hatch; the typical family consists of a pair of adults with a single cohort of offspring. This social structure, together with the observation of parent to offspring vertical transmission of gut symbionts via proctodeal trophallaxis, was subsequently supported by Nalepa  and Park et al. . The subsocial, semelparous life history of Cryptocercus, including parent to offspring vertical transmission of gut symbionts, is now well established, and it is generally accepted that extended parental care in an ancestral cockroach modeled after Cryptocercus was the key basal condition in the pre-eusocial termite ancestor . Misconceptions regarding the host-symbiont relationship in this lineage nevertheless continue to cloud interpretations of its evolutionary trajectory, e.g., [58,59,60].
3. Molting in Cryptocercus
The gut microbiota of insects have a unique challenge not faced by those in vertebrate taxa. Because the insect hindgut lining is ectoderm, it is shed along with the rest of the exoskeleton during the molting period throughout host development. Consequently, insect gut microbiota may require measures to cope with the conditions of the digestive tract during this time. As Cryptocercus enters its molting period, rising titers of the molting hormone ecdysone trigger the sexual/encystment (SE) cycles (addressed below) of the flagellates in the gut. During the SE cycle, the protistans become more resistant by producing cysts, pseudocysts, or by employing other mechanisms. The lining of the cockroach hindgut is then shed whole, but retained within the host body. The resistant flagellates are then held within the old gut lining (intima) until formation of the new lining is completed. The old intima ruptures, and the protists escape into and repopulate the gut lumen. The old hindgut lining subsequently breaks into small pieces, and is voided in fecal pellets along with some resistant protists [15,61]. Although these protists may pass with host feces shortly before or after molting (in some hosts half or more, in others just a few), they play no role in intergenerational transmission under natural conditions.
Prior to the work on social structure of Cryptocercus in the 1980s, it was thought that SE cycles in flagellates of Cryptocercus were rooted in a fecal-oral mechanism of transmission between generations. The cysts were considered mandatory to withstand the stress of being subjected to the environment outside the insect host, i.e., in fecal pellets lying on the floor of the nest. We now know that flagellates are transferred to neonates via ingestion of the trophozoite stage in the hindgut fluids of parents. After their hindgut microbiota have become established at about the third instar, trophallactic behavior in young nymphs of Cryptocercus ceases. The fact that some cysts may be found in the fecal pellets of post-molt juveniles of Cryptocercus, however, suggests that contemporaneous members of the family may ingest some of these; trophozoites that subsequently emerge from these cysts are probably re-integrated into the existing gut microbiota of the coprophage. Feces are a part of the diet in most cockroaches, including Cryptocercus [18,62]. However, because all flagellates in a given family originate with the parents, no new flagellate taxa would enter the social group via coprophagy.
4. Molting in Termites
The most common error regarding the disappearance of the protistans in molting termites is that the hindgut lining is excreted whole with the protists still in it; prevalent fallacies regarding how the symbiosis is re-established after molt is that it is restored by coprophagy or by feeding on an intact, shed hindgut lining. With a few possible exceptions (discussed below), however, the flagellates are dead and gone from the hindgut prior to the molt of their termite host [44,63,64,65,66,67]. The symbionts are not ‘shed’, ‘discarded’ or ‘cast’ with the hindgut lining. The gut is devoid of flagellates seven days before molt in Kalotermes flavicollis  and six days before molt in Coptotermes formosanus . The physical process of molt in termites is the same as in Cryptocercus. The lining of the hindgut is shed but retained within the body for two or three days; it later breaks up and is voided. Consequently, the gut protists cannot be re-acquired by a post-molt termite via feeding on its own shed exuvium, that of a nestmate, or by feeding on feces. The hindgut lining is not attached to the remainder of the shed exuvia , and termite feces never contain active flagellates, cysts, or other resistant stages [15,43,47,61,64,71]. The sole mechanism of refaunation in a newly molted termite is via interaction with a conspecific: repeated proctodeal trophallaxis with intermolt donor nestmates . If Cryptocercus is used as a model of the termite ancestor, the host-symbiont relationship during molt changed from retention by the individual host via the behavioral and physiological responses of the flagellates (SE cycles), to retention by the colony via host behavioral interaction with other colony members .
One explanation offered for the death of termite flagellates during host molt was that because termites exhibit colony-wide proctodeal trophallaxis, the SE cycle in their flagellates was superfluous and expendable, as the flagellates were never exposed to the outside environment ; this explanation is still popular [72,73]. It is the premise of this paper that the directionality of cause-and-effect of this common assumption should be reversed. Instead of colony-wide trophallaxis leading to the death of flagellates at host molt, the early stages of eusociality led to dysfunctional SE cycles in the flagellates and required a behavioral solution on the part of the host: colony-wide trophallaxis [18,51].
5. Why Is the Partnership Dissolved at Every Molt in Termites?
It is commonly assumed that abiotic factors in the gut environment are responsible for killing termite protistans during molt, and that encystment by the flagellates in Cryptocercus protect them during this period. The flagellates are known to be extremely sensitive to changing conditions of temperature, food, and water ; indeed, subjecting Cryptocercus and termites to heat, starvation or oxygen is a standard technique for rendering them aposymbiotic for experimental purposes, e.g., [75,76]. The assumption that such conditions kill termite flagellates during molt is therefore understandable, since these symbionts pass through a gauntlet of harsh conditions each time the insect undergoes a molting cycle. The host ceases to feed, and the gut fluid increases in viscosity to a heavy paste [61,77]. Large bubbles also appear in the hindgut ( Plate 7), suggesting potential vulnerability to oxygen poisoning. In free living and parasitic protists, stimuli such as these are often the first triggers of a cell differentiation process that leads to encystment, thus enabling resistance to a deteriorating environment [78,79,80]. Starvation, desiccation or oxygen, however, play no role in initiating the SE cycles of the flagellates in Cryptocercus [44,81]. The cycle is triggered by rising titers of host ecdysone long before there is a change in the feeding habits of the host or a change in viscosity of the gut fluid [81,82]. Furthermore, SE cycles can be induced with ecdysone in the flagellates of intermolt Cryptocercus nymphs as well as in adults, both of which have normal, wood-and fluid-filled guts . The relevant question, then, is why does a rising titer of the molting hormone ecdysone trigger a differentiation process in the flagellates of Cryptocercus, but lead to disintegration of the flagellates in termites? To address this question, the host-symbiont relationship at molt displayed by Cryptocercus will be examined in detail.
6. Flagellate Differentiation during the Cryptocercus Molting Cycle
Protists in Cryptocercus exhibit two morphologically distinct stages in the life cycle. Intermolt Cryptocercus typically harbor the trophozoite stage, which reproduces by mitosis. Flagellates change from asexual to sexual methods of reproduction (i.e., initiate an SE cycle) at the time their host enters its molting period. The molting period of Cryptocercus begins with the first appearance of ecdysone in the insect and ends with the disappearance of this hormone. Ecdysone can begin rising as much as 50 days prior to ecdysis, and persists until three days after [81,83,84]. Ecdysis is defined as the behavioral act of shedding the exoskeleton. Note that for simplicity, I am here calling the differentiation events in flagellates during host molt a sexual-encystment (SE) cycle, regardless of the genetic consequences of the cycle and the mechanism of resistance. There is extraordinary variation in the sexual cycle of the flagellates during the Cryptocercus molting cycle; likewise, the morphological indications of flagellate resistance vary, both in degree and in the timing of appearance.
6.1. Cryptocercus: Variation in Flagellate Sexual Cycles
Cleveland  thought that the flagellates in Cryptocercus exhibited a variety of sexual forms as extensive as all other protistan groups known at his time. In general, his accounts of the SE cycles in the flagellates indicate little opportunity for genetic recombination. There are reports of remarkable abbreviations of, and variations on, standard sexual processes that are difficult to interpret (summarized in [85,86,87]). In Pyrsonympha, for example, meiosis-like reductional divisions can be considered either degenerated meiosis or as precursors of conventional meiosis . Haploid lineages undergo postzygotic meiosis, some with one, others with two divisions . Diploid flagellates show pre-gametic meiosis requiring two divisions in some species and just one in others. Autogamy occurs in a few species. One-divisional meiosis probably exists only in flagellates from Cryptocercus (reviewed by Raikov ). Even within a given flagellate species, SE cycles do not always follow a uniform pattern. Leptospironympha eupora, for example, may encyst either before or after gametogenesis; the usual pattern being encystment following fertilization .
Although meiosis is highly conserved in eukaryotes, deviations from the norm are ubiquitous and the evolutionary significance of these deviations is largely unknown. In many protists meiosis is not tightly associated with reproduction [88,92,93], and may have a functional basis in DNA restoration rather than in recombination, i.e., elimination of deleterious mutations, direct repair of DNA breaks, or removal of oxidative DNA damage . Other apparent advantages include recombination fidelity, ploidy reduction and epigenetic resetting (refs. in ).
6.2. Cryptocercus: Variation in Flagellate Mechanism of Resistance
The flagellates in Cryptocercus exhibit the entire spectrum of the morphological indications of resistance during host molt, ranging from those that produce a thick, double-walled cyst to those that undergo a differentiation cycle during host molt but display no physical evidence of a resistant outer wall. Individuals of more than one species can skip an SE cycle altogether but are nonetheless retained during host molt.
True cysts are defined as having a thick hyaline wall formed from the outer layer of the cytoplasm, with an empty space between the inner margin of the cyst wall and the living cytoplasm . Examples of flagellates that clearly form cysts include Trichonympha and Macrospironympha; Cleveland  considered the latter as the best flagellate for studying encystment. The definition of a pseudocyst is more nebulous. A general definition is that it is a compact, resistant, resting form, where the cell rounds up without forming a cyst wall . Cleveland, however, struggled with defining this mechanism of resistance. In Barbulanympha, for example, he described the organism as losing water, taking on a hyaline appearance, discarding its flagella, becoming circular and immobile, and becoming capable of withstanding conditions outside the host body. He considered it a ‘clear cut case of protozoa becoming resistant without formation of an external, protective cyst membrane or wall’ but refrained from naming the condition . He later, however, indicated that Barbulanympha forms a pseudocyst . Rhynchonympha was described as undergoing ‘partial or pseudoencystation’, but it was also noted as ‘sometimes’ forming a cyst [96,99]. Variation within a genus was noted: Oxymonas nana forms a cyst, but O. dorsalis forms a pseudocyst . Leptospironympha typically forms a cyst, but there is no encystment or pseudoencystment in L. wachula [91,96]. As in L. wachula, Eucomonympha, Notila and Saccinobaculus apparently undergo sexual cycles with neither encystment nor pseudoencystment [96,100,101].
Finally, Cleveland described instances of flagellates that do not undergo an SE cycle but are nonetheless retained by Cryptocercus during molt. Although in most flagellates all individuals undergo an SE cycle when their host molts , a small percentage of those in the genera Trichonympha, Leptospironympha and possibly others do not. These individuals are actively motile, free of wood particles, and location specific: they are always found in the very anterior part of the hindgut [15,61,91]. The number of these individuals that do not undergo an SE cycle is always sufficient to repopulate the gut after molt. Consequently, Cleveland et al. ( p. 262) concluded that the function of flagellate encystment was to produce resistant organisms capable of withstanding conditions external to the host, so that they may be transmitted in feces to neonates. Since we now know that these flagellates are vertically transmitted between generations via proctodeal trophallaxis, another possibility suggests itself: cysts in flagellates of Cryptocercus are currently of no functional consequence, but are vestiges from a distant gregarious ancestor when externally excreted cysts functioned in horizontal transmission . Fewer than half of the flagellates studied by Cleveland exhibit the thick outer wall associated with a true cyst ( Table 2), and the morphological variation in resistance exhibited by the flagellates in this host may represent stages in dispensing with the costly encystment process. Encystment in the infectious vertebrate parasite Giardia lamblia, for example, involves pulsed production, processing and secretion of cyst wall material, and neogenesis of encystment specific vesicles [80,102], indicating substantial metabolic and material investment.
7. Link between Social Behavior and Symbiont Transmission
The host-symbiont patterns currently seen in Cryptocercus and extant lower termites suggests a logical evolutionary sequence ( Figure 2 ). In a distant gregarious cockroach ancestor, the host-symbiont relationship was not obligate, and the flagellates likely had two distinct differentiated states: the vegetative state inside of the insect host, and the encysted state in fecal pellets in the outside environment. Encystment was a crucial part of the life cycle, as it was the mechanism of horizontal transmission via the fecal-oral route (coprophagy) . As in well-studied parasitic protists from diverse lineages, meiosis and resistance to the outside environment were linked to dispersal from the host ; it allowed the flagellates to withstand external conditions if new hosts were not immediately available. Cysts within fecal pellets served as an ‘infection bank’, much like the gregarines that parasitize extant gregarious cockroaches [104,105]. Aggregation behavior of the distant cockroach ancestor served as a mechanism that concentrated both the cysts and the hosts in a limited space.
Sexual-encystment cycles of the flagellates in molting individuals of: (A) a hypothetical gregarious cockroach ancestor, (B) the subsocial cockroach Cryptocercus, and (C) eusocial lower termites. The symbiont-host relationship logically evolved in sequence, along with shifts in host social behavior, from a distant gregarious cockroach ancestor, to a subsocial Cryptocercus-like ancestor, to eusocial termites . In both Cryptocercus and termites, the SE cycle has degenerated from the strong encystment pattern likely shown by an ancestor, but in both taxa, protistan numbers and functions are reliably restored after molt, albeit in different ways (retention of protists in individuals of Cryptocercus, anal trophallaxis from nestmates in termites). Neither Cryptocercus nor termite protists currently have a functional relationship with the external environment. Inspired by Weedall and Hall ( Figure 1).
With Cryptocercus as model, proctodeal trophallaxis evolved from this pre-existing intraspecific coprophagous behavior when termite ancestors became semelparous and subsocial, because the physiology of encystment in the flagellates precludes their transfer via cysts in adult feces: adults don’t molt, and older developmental stages are not present in the social group. This was a key point in their evolutionary history, as it assured vertical intergenerational transmission of all pro- and eukaryotic members of the gut microbiome, led to growing interdependence of the host and symbionts, and established the behavioral basis of trophallactic exchanges [18,51]. Parent to offspring proctodeal trophallaxis in both Cryptocercus and termites during colony initiation is currently a mechanism of both inoculating neonates with the symbionts and of supplying them with nourishment until their microbiome becomes established. Table 1 summarizes the similarities and differences in the host-symbiont relationship in Cryptocercus and lower termites.
Comparison of the host-flagellate relationship in a hypothetical cockroach ancestor, the wood-feeding cockroach Cryptocercus, and the lower termites. Characteristics of Cryptocercus are assumed to be the state immediately ancestral to lower termites. Key differences between Cryptocercus and lower termites are in red.