Description of an early Cretaceous termite (Isoptera: Kalotermitidae) and its associated intestinal protozoa, with comments on their co-evolution

Background The remarkable mutualistic associations between termites and protists are in large part responsible for the evolutionary success of these eusocial insects. It is unknown when this symbiosis was first established, but the present study shows that fossil termite protists existed in the Mesozoic. Results A new species of termite (Kalotermes burmensis n. sp.) in Early Cretaceous Burmese amber had part of its abdomen damaged, thus exposing trophic stages and cysts of diverse protists. Some protists were still attached to the gut intima while others were in the amber matrix adjacent to the damaged portion. Ten new fossil flagellate species in the Trichomonada, Hypermastigida and Oxymonadea are described in nine new genera assigned to 6 extant families. Systematic placement and names of the fossil flagellates are based on morphological similarities with extant genera associated with lower termites. The following new flagellate taxa are established: Foainites icelus n. gen. n. sp., Spiromastigites acanthodes n. gen. n. sp., Trichonymphites henis n. gen., n. sp., Teranymphites rhabdotis n. gen. n. sp., Oxymonas protus n. sp., Oxymonites gerus n. gen., n. sp., Microrhopalodites polynucleatis n. gen., n. sp., Sauromonites katatonis n. gen., n. sp., Dinenymphites spiris n. gen., n. sp., Pyrsonymphites cordylinis n. gen., n. sp. A new genus of fossil amoeba is also described as Endamoebites proterus n. gen., n. sp. Fourteen additional trophic and encystid protist stages are figured and briefly characterized. Conclusion This represents the earliest fossil record of mutualism between microorganisms and animals and the first descriptions of protists from a fossil termite. Discovering the same orders, families and possibly genera of protists that occur today in Early Cretaceous kalotermitids shows considerable behaviour and morphological stability of both host and protists. The possible significance of protist cysts associated with the fossil termite is discussed in regards the possibility that coprophagy, as well as proctodeal trophallaxis, was a method by which some termite protozoa were transferred intrastadially and intergenerationally at this time.


Background
Termites are one of the most successful eusocial insect groups today and certainly the most notorious as a result of their damage to human dwellings. Their success can be attributed in large part to microbial (especially protozoa and bacterial) symbionts harbored in their alimentary tract. Especially important are gut protists, which are essential for the survival of termites feeding on lignocelluloses. While termites do produce endogenous cellulases from salivary glands and gut cells, cellulolytic protists are crucial for the complete digestion of cellulose in wood-feeding termites [1][2][3]. In the lower termites, these symbionts are mostly flagellates belonging to the Oxymonadida, Trichomonada and Hypermastigida [4][5][6][7]. Flagellates associated with an Early Cretaceous lower termite of the family Kalotermitidae are described and compared with mutualistic flagellates of extant kalotermitids. This discovery shows that, while the protist species represent different genera and species, mutualistic associations between protists and termites had already been established some 100 million years ago. The present study represents the earliest fossil record of mutualism between microorganisms and animals [8].

Description of host termite
Damage to the termite included the loss of the left hind wing, the tip of the abdomen (including the cerci), the right side of the metathorax and the right side of the first three abdominal segments. All wings were flexed at their bases. However, it was possible to determine the venation in the distal three quarters of the right forewing and the complete right hind wing.
Based on the morphological characters of the termite (Sc absent from the hind wing, R simple in both wings, no reticulation between the veins posterior to R, R separate from the costal margin, presence of a number of regularly spaced oblique forward branches between the Rs and C), the fossil is placed in the family Kalotermitidae [9,10].
Its small size, low number of antennal segments, strongly sclerotized radial sector, unsclerotized M and Cu veins, complete M vein positioned halfway between Rs and Cu, wing membrane densely covered with minute, pigmented nodules, tibial spurs lacking on the shaft of the mid-tibia and the presence of ocelli and arolia, align the fossil with the genus Kalotermes Hagen 1853 [9,10]. Since the fossil differs from previously described termites in Burmese amber [10][11][12][13], it is described below as a new species. It should be noted that the placement of the fossil in the extant genus Kalotermes is tentative since some diagnostic characters (number of apical spines on the tibiae, front wing venation, etc.) were obscured.  Comments: The present species differs from Kalotermes swinhoei [11] and K. tristis [12], previously described from Burmese amber. While Emerson [14] and Krishna [9] felt that K. swinhoei and K. tristis could be synonymous, Williams [10]  The wing venation, head shape, presence and size of arolia, presence of ocelli, absence of spines on the shaft of the mid-tibia, number of tarsal and antennal segments and presence of pigmented wing nodes separates K. burmensis from the four Burmese amber specimens described by Engel et al, [13], as well as all of the Tertiary species of kalotermitids [14].

Description of protists
A variety of protists were associated with K. burmensis. Some were still attached to the gut lining, while others were free in the amber matrix adjacent to the exposed gut. Those that revealed morphological characters similar to protist families and genera found in extant lower termites are described below. Since the amber matrix normally contains a variety of small, oval-spherical bodies, care was taken to select for descriptions only protozoa attached to the gut intima or with features aligning them with extant groups associated with lower termites. None of the forms presented below are considered to be insect parasites or pathogens [15]. Stages that could not be identified due to insufficient characters are documented with photographs and briefly characterized.  Except for one species described in the extant genus Oxymonas, all of the protists are placed in fossil genera, with names based on extant genera with similar morphological characters. However, it is acknowledged that since not all diagnostic characters could be determined in the fossil protists and features that were present could align the fossil with more than one extant protist lineage, the similarities so indicated between the fossils and extant genera are only tentative.
Terminology and classification generally follows that of Brugerolle and Lee [4,5] and Patterson et al. [16]. A synopsis of the described fossils with their higher level systematic positions are summarized below.            Comments: The fossil resembles the body shape of extant species of Foaina Janicki 1915, however, it could be related to one of the other genera of small devescovinids. The straight parabasal body slightly longer than the nucleus distinguishes it from extant species. All known species of Foaina, whose size range extends from 6 μm to 59 μm, occur in kalotermitids [5,9].    Etymology: From the Greek "henos" for old.
Comment: This genus is placed in the family Trichonymphidae based on its relatively short rostral flagella, absence of an axostyle, and ribbon-like parabasals associated with the nucleus. One of three cysts of this or a related species in the same family is shown in Fig. 19. Members of the genus Trichonympha occur in at least 16 termite genera worldwide, including Kalotermes [9,17,18]. Because of its wide distribution in the termite families Hodotermitidae, Rhinotermitidae and Kalotermitidae, Kirby [17] suspected that species of Trichonympha were already present in the various termite lineages when they first appeared and then underwent a period of co-evolution that is continuing today.

Family Teranymphidae Koidzumi, 1917
Teranymphites Poinar n. gen. (Figs. 10A,B) Description. Medium sized oval cell, with short rostrum attached to intestinal cells of host; rostral flagella short; post-rostral area with parallel flagellar rings without connections and giving the cell a segmented appearance; nucleus located under rostral area; putative axostyles (possibly 2) narrow, fiber-like.
Comment: The presence of an anterior rostellum with associated fibers and the uninucleate condition are diagnostic characters of the genus Oxymonas. However members of Oxymonas usually have 4 flagella arising from the shoulder area, which are not evident in the fossil [4,19]. Species of Oxymonas occur in at least 10 genera of kalotermitids and range from 5-165 μm in length [5,9]. The holdfast is used to attach the cell to the chitinous lining of the termite gut and the size of the rostellum has been used to estimate the density of populations in termite guts, with a lengthy rostellum indicating a crowded condition [19].
Oxymonites Poinar n. gen. (Figs.12A,B) Description. Uni-nucleated flagellate with body comprising a single karyomastigont; body wider than long; axostyle short, not extending more than half body length; nucleus large, located near middle of body; rostellum with several fiber bundles. Etymology: From the Greek "geros" for old age.
Comments: The presence of an anterior rostellum with fibers arising from the base of the attachment point and the single nucleus is why this fossil was placed in the Oxymonadidae. While oxymonidids typically have an anteriorly placed nucleus, during certain phases, the nucleus may be located in the posterior portion of the body [19]. The shape and size of the rostellum vary greatly in extant oxymonadids [19]. The family is widely distributed in kalotermitids, with some 27 species described worldwide [9]. Possible cysts of this fossil and/or Oxymonas protus are shown in Figs. 19 and 20. Each cyst contains a nucleus and the one in Fig. 19 shows an axostyle in the lower portion of the body, which is characteristic of some oxymonadid cysts [20].   [4]. It is difficult to say if the rounded rostral region is characteristic of the mature trophozoite or if the cell was entering a resting stage. Superficially, this fossil has the appearance of an eugregarine sporont, with the rostral area corresponding to the protomerite and the basal portion the deutomerite. However, there is no scar on the top of the hemispherical rostral area that would indicate the attachment of an epimerite, no evidence of a septum separating the two parts and no evidence of a large nucleus in the portion that would correspond to the deutomerite. The putative flagella and small nuclei associated with the rostral area are not characteristic of eugregarine sporonts, certainly not those from extant termites [21,22].
Description. Axostyle prominent, thickened anteriorly, extending length of body; 4 flagella emerging from rostral area; nucleus located in lower half of body; large dark body at base of rostral area (macronucleus?); network of apparent fiber bundles throughout body; minute dark area at posterior end may represent tip of axostyle. Etymology. From the Greek "katatonos" meaning broader than high.
Comments: This species is roughly the same size and shape as extant species of Sauromonas in kalotermitids [4].
Type species: Dinenymphites spiris n. sp.  Etymology: From the Greek "speira" for twisted in reference to the body structure.

Dinenymphites spiris
Comments: The fossil resembles extant species of the genus Dinenympha Leidy, which range from 24 μm to 64 μm in length. Members of this genus are now considered motile forms of Pyrsonympha Leidy [4].
Pyrsonymphites Poinar, n. gen. (Figs. 16A,B) Description. Large club-shaped, slightly spirally twisted flagellate with two nuclei positioned slightly below midbody; with several flagella (some adhering to cell body) arising from anterior end of cell; anterior tip with thickened tubular area; axostyle not detected.

Pyrsonymphites cordylinis
Description. Spherical nucleus in process of dividing; nucleolus not apparent; body amoeboid-like, nearly spherical, with broad pseudopodia and short protuberances. Etymology: From the Greek "proteros" for earlier.
Comments: The size and shape of the body and nucleus resemble those of extant species of Endamoeba Leidy from termites and roaches [23] which are the only known hosts [16]. The nucleus appears to be dividing and some putative chromatin threads connect the two adjacent nuclear zones. A possible cyst of Endamoebites (Fig. 21) is 53 μm in diameter, possesses a thick membrane, contains 13 nuclei in the focal plane shown and resembles cysts of extant Endamoeba [16].

Figure 21
Spherical cyst with two or more nuclei. Bar = 14 μm.

Additional protists
Additional unknown protist stages associated with K. burmensis are illustrated in Figs. 18, 19, 20, 22, 23, 24, 25, 26, 27, 28, 29, 30, Fig. 13b]. Since these cysts were found in the amber matrix adjoining the termite, it is likely that they were already formed when the termite was entombed, especially since Cleveland et al. [6] showed that they required a period of 4 days to form in Cryptocercus. It is highly unlikely that any of the cysts were formed after the termite entered the resin, since terpinols and other chemicals would have killed the precystic, trophic stages instantly.
Unknown stages, some of which could be ciliates, are shown in Figs Ciliates have been reported from extant termites [24] but have not been well documented.
A pair of dauer juvenile rhabditid nematodes were also associated with K. burmensis. These will be described in a separate study.

Discussion
The present study represents the first descriptions of protists from a fossil termite and the earliest fossil record of mutualism involving microorganisms and animals [8].
The occurrence of an Early Cretaceous kalotermitid termite with a variety of flagellates that includes members of the same orders, families and possibly genera that occur today in kalotermitids shows considerable behavior and morphological stability of both host and protists.
Successful establishment of protozoa in xylophagous insects necessated particular attributes. The protozoa had   to withstand the chemical and physical conditions inside the insect's alimentary tract, utilize the gut contents as a food source, cause no damage to the host, and to be carried through successive insect stages and generations. These protists apparently underwent a long period of coevolution with their hosts, since some lineages found today in the intestines of xylophagous cockroaches and lower termites are thought to have been established in Carboniferous Blattida [25].
Today, the composition of intestinal protozoa tends to be correlated with the phylogenetic position of their termite hosts and it is possible to classify families, genera and even species of termites based on their flagellates [7]. This also applies to K. burmensis, since it contains representatives of the three most abundant and widespread groups of protozoa in kalotermitids today (Trichomonada, Hypermastigida and Oxymonadea). If the systematic assignments are correct, some of the protists in K. burmensis appear to be restricted to other families of lower termites today. However in the Early Cretaceous, partitioning of host habitats was probably less fixed and the choice of host by a protozoan was    predicated largely on availability and chance encounter. Some of the protozoa described here may have been thwarted by a change in the habits of the host and since they could not adapt, disappeared from the colonies. Since angiosperms were becoming more diverse by the mid-Cretaceous, some protists in K. burmensis may have succumbed when the host diet shifted from gymnosperm to angiosperm wood. While it is assumed that most of the protozoa described here had a mutualistic relationship with K. burmensis, as members of the Trichomonada, Hypermastigida and Oxymonadida do with extant lower termites, some, as for instance Endamoebites proterus, may have been simply commensals and provided no benefit to K. burmensis.
One striking difference in the behavior of the protist symbionts in extant termites and K. burmensis is the presence of encysted stages in the latter. In extant termites, mature protist cysts are rarely formed [26,27]     and never in alates since the flagellates are transferred to nest mates and hatchlings by proctodeal trophallaxis (intrastadial and intergenerational transfer). However it has been hypothesized that the distant ancestor of termites passed flagellate cysts to nest mates and hatchlings by coprophagy [6,7,[27][28][29][30].
Several possible scenarios could explain the presence of cysts in K. burmensis: A), the colony was subsocial and ingesting cysts in fecal pellets (coprophagy) was the main method for the intrastadial and intergenerational transfer of protists; B), the colony was subsocial or eusocial and both coprophagy (ingesting cysts in fecal pellets) and proctodeal trophallaxis served to transfer the flagellates to nest mates and hatchlings; C), the colony was eusocial and protozoa were transferred by proctodeal trophallaxis. The cysts were an ancestral carry-over and represent an evolutionary dead-end; D), some of the cysts could have been ingested while feeding and had no trophic association with K. burmensis.

Specimens
The amber with the fossil termite containing the protists is roughly semi-circular in outline, measuring 13 mm along the longest edge, 10 mm in width and 1 mm in thickness.
Observations, drawings and photographs were made with a Nikon SMZ-10 R stereoscopic microscope and Nikon Optiphot compound microscope. Since all photographs were taken through the thickness of the amber matrix, it was not possible to get as close as desired to individual protists without polishing away adjacent ones as well as portions of the termite host. Therefore all photos were taken at 20×. With such a small image, it was only possible to obtain a single clear photo of the specimen since further fine adjustment produced blurry images.
Parasites & Vectors 2009, 2:12 http://www.parasitesandvectors.com/content/2/1/12 Adobe Photoshop was used to enlarge the photos and to obtain several modified images by using different settings of contrast, light intensity and resolution. The drawings were made from a combination of the modified images and that is why they contain more detail than the corresponding photographs, which represent the best single image obtained under the various settings. Thus Photoshop manipulation was used to replace optical sections, that were not possible to make with such small objects and at such a low magnification.

Competing interests
The author declares that he has no competing interests.