Chapter 12 

Phylogeny

 

Historically, almost all early publications on trichomycetes included some discussion of the possible affinities of newly described species with fungi, algae, or protozoans. From time to time the taxonomic organization of known genera into families and higher taxa were suggested or revised as new information became available and taxonomic concepts developed (e.g., Léger and Duboscq, 1929a; Duboscq et al., 1948; Manier, 1950, 1955b; Manier and Lichtwardt, 1968). Less constructively, Sörgel (1953) sought to design a phylogenetic arrangement for the families of trichomycetes, but he based his scheme in part on erroneous data and ignored reproductive features of these organisms that must be considered in such an attempt. A patently unacceptable taxonomy was proposed by Locquin (1974) in which he allied the eukaryotic Amoebidiales with the prokaryotic Actinomycetes, and, curiously, the other orders of trichomycetes (and the Kickxellales) with the Endomycetes. Some of the early studies involving taxonomic judgments are of historic interest, but it would not be useful to review them here. It should be noted, however, that in their 1948 monograph Duboscq, Léger, and Tuzet introduced the name Trichomycetes to include in the broad sense two classes: the Eccrinides (Amoebidiales and Eccrinales) and the Harpellides (Harpellales and Genistellales). In their view the Eccrinides were the true Trichomycetes. The reader will note that this original interpretation has completely shifted, because molecular studies now lead to the exclusion of Amoebidiales and Eccrinales from the fungal class Trichomycetes. In this chapter the authors will summarize the available data that bear upon phylogenetic hypotheses. 

Earlier Concepts of Relationships


It is reasonable to assume that ancestral forms of trichomycetes lived independent of or in loose external association with arthropods before they evolved the specialized features demanded for successful gut habitation. It is also likely that the trichomycetes have been evolving over a considerable period of time, for reasons outlined in the introductory paragraph of Chapter 8. As a consequence of these expectations, and considering that the gut habitat has undoubtedly exerted strong selective pressures on the evolution of the trichomycetes and has thus caused them to diverge considerably from their free-living ancestral forms, there is little reason to expect to find a close morphological relationship with extant groups of fungi, which may share their ancestry but not their ecologically specialized niches.

Several lines of morphological evidence suggest there are similarities between the Harpellales and the Zygomycetes order Kickxellales Kreisel ex Benjamin [sensu Moss and Young (1978) and Benjamin (1979)] that includes only the Kickxellaceae. The Kickxellales in this restricted sense is a relatively small order, consisting of eight genera with mostly saprobic species (often in soil or dung) and a few weakly mycoparasitic ones (Benjamin, 1979). The mycelium is extensively branched and septate, with some coenocytism present such as in the multinucleate sporocladia of Linderina spp. Unispored sporangiola are produced on the septate or nonseptate sporocladia, and spherical zygospores have been found in some of the species. The possible relationship with trichomycetes and the supporting evidence was presented by Moss and Young (1978), together with a suggested phyletic scheme for the genera of Kickxellales, Harpellales, and Asellariales (Fig. 12.1).

The similarity first observed between the Harpellales and Kickxellales is the unusual perforate crosswall with a plug that is present in both the vegetative and reproductive parts of their thalli (Farr, 1965; Farr and Lichtwardt, 1967; Manier and Coste-Mathiez, 1968; Reichle and Lichtwardt, 1972; Lichtwardt, 1973a; Manier, 1973a,b; Moss, 1972, 1975, 1976; Moss and Lichtwardt, 1976, 1977; Young, 1969; Benny and Aldrich, 1975). The harpellid septum has been described in Chapter 7 (Fig. 7.7), and it seems to be the same as the kickxellid septum in basic structure and development. An exception is that micropores, in addition to the larger central pore, have been detected in one species of Kickxellales (Coemansia aciculifera Linder), although not yet in any species of Harpellales (Moss and Young, 1978).

A second line of evidence for a relationship between the two orders resulted from the serological studies of Sangar et al. (1972). They found that some anti-Smittium antibodies precipitated when reacted with antigens of two Kickxellales, namely Linderina pennispora and Dipsacomyces acuminosporus (see Chapter 9 for data), thus indicating some immunological affinity between the two orders. Antigens of four other taxa of Zygomycetes produced no reactions, or they were negligible.

           The appendaged trichospores (sporangia) of the Harpellales remain a structurally unique fungal spore type, but Moss and Young (1978) have presented a case for a degree of similarity between the harpellid asexual apparatus of a genus such as Smittium and that of the kickxellid genus Coemansia (Fig. 12.1). Coemansia produces branches with terminal and septate sporocladia bearing a number of unilateral pseudophialides, and from each of these a single, elongate sporangiole (unispored merosporagium) forms and later detaches. In this comparison the sporangiole-sporocladiumn arrangement is considered to be homologous with the trichospore-generative cell complex of the Harpellales, with the pseudophialides being comparable to the collar region of a Smittium. Moss and Young noted that in the Harpellid Pteromaktron protrudens a subsidiary cell grows from the generative cell and gives rise to the trichospores, and the subsidiary cell might represent the equivalent of the pseudophialide in CoemansiaNothing comparable to the structurally and functionally distinct trichospore appendage has been found in the Kickxellales. However, Young (1969) discovered a "labyrinthiform organelle" in Linderina pennispora in the pseudophialide at the base of the sporangiole, and a similar structure was found in the same species by Benny and Aldrich (1975), which they called an "abscission vacuole." It may represent a rudimentary type of appendage. At least it seems to correspond in position and appearance with the early stages of harpellid appendage formation (Moss and Young, 1978), although it has no known function and does not develop into a well-defined structure like the typical harpellid appendage. Additional detailed studies will certainly be needed before these putative homologies are substantiated, but at least the present evidence is suggestive of some relationship. Both the Harpellales and Kickxellales produce zygospores, signifying that they have a fundamental sexual similarity, but beyond this they differ in several significant aspects. Zygospores of the Kickxellales, like other Zygomycetes, are spherical, and develop from a homothallic fusion of either small gametangia borne on two undifferentiated sexual branches or a branch with an intercalary cell (Benjamin, 1958). Zygospores of the Harpellales typically develop from heterothallic or homothallic conjugations of mostly intercalary cells. From one of the conjugants, or from the conjugation tube between them, a cell called the zygosporophore develops. The zygosporophore produces initially a more or less spherical apical swelling, which then enlarges into a biconical zygospore. Limited fine structural studies on Harpella melusinae and Trichozygospora chironomidarum (Moss and Lichtwardt, 1977) suggest that meiosis possibly occurs within the conjugation apparatus, resulting in zygospores having it single haploid nucleus. The biconical shape and the possibly haploid condition of the harpellid zygospore may be adaptive to rapid germination and establishment of thalli in the host gut, as discussed in Chapter 7. Unfortunately, there are no fine-structural or cytological studies of kickxellid zygospores (Benjamin, 1979); consequently, at present comparisons of these structures in the two orders are not possible beyond gross morphology.  

            The foregoing paragraphs have stressed the morphological similarities that exist between the Harpellales and Kickxellales. In Chapter 9, a study by Porter and Smiley (1979) is reported in which they compared ribosomal RNA molecular weight of four species of Smittium and three species of Kickxellales, as well as of a variety of other Zygomycetes and Amoebidium parasiticum. The rRNA weights of the Smittium isolates were higher than those of the other test organisms. Porter and Smiley concluded that the differences were biologically significant, and that on this basis Smittium is not closely related to either A. parasiticum or any of the Zygomycetes.

            Reproductive features and thallial morphology have been studied more thoroughly in the Harpellales than in the other three orders of trichomycetes (sensu lato) (Lichtwardt et al., 2001). The protist orders Amoebidiales and Eccrinales share a common ancestry (Cafaro, 2005) but differ in their reproduction, even though both produce one or more forms of sporangiospores. Amoeboid cells are completely lacking in Eccrinales. Species of Eccrinales, unlike Amoebidiales, have more varied types of sporangia and have septate thalli. Cavalier-Smith (1998) erected a new class, Enteromycetes, to include Amoebidiales and Eccrinales, based in part on their possessing Golgi dictyosomes. Asellariales reproduce by arthrospore-like propagules, and have perforate septa that resemble those of Harpellales (Fig. 7.7). Based on morphological features and host adaptation, it is probable that the two genera of Asellariales, Asellaria and Orchesellaria, do not form a monophyletic clade. These relationships will be best determined by means of DNA analyses.

           

Molecular Systematics

 

Just as molecular systematics has revolutionized our views of what phyla constitute the true fungi, so it has transformed our understanding of relationships of trichomycetes within the group and with other organisms. Data that follow will provide evidence that (1) Amoebidiales and Eccrinales are protists and not fungi; (2) Harpellales are true fungi related to the Kickxellales (Zygomycetes); (3) several of the larger genera of trichomycetes are not monophyletic and must be revised; and (4) the families within Harpellales and Eccrinales that were based on reproductive criteria do not represent monophyletic clades, and consequently are not justified as currently delimited.

            Molecular studies have now confirmed that Amoebidiales are not fungi, but belong to a group of protozoans, the Mesomycetozoea (Mendoza et al., 2001),  that were collectively known by the acronym DRIPs, a clade of mostly pathogenic protists believed to lie in the tree of life where animals and fungi may have diverged (Benny and O’Donnell, 2000; Ustinova et al., 2000; Cafaro, 2005). More significantly, species of the larger order Eccrinales also are Mesomycetozoea and closely related to Amoebidiales (Cafaro, 2005) (Fig. 12.6). Though not fungi and consequently not belonging to the fungal Class Trichomycetes, Amoebidiales and Eccrinales will continue to be studied by those interested in gut symbionts of arthropods as members of the ecological group trichomycetes (lower case t), as in this monograph.

            Most species of Harpellales remain intractable to culture attempts (as is true of all species of the other three orders, except for Amoebidium parasiticum). As a consequence, the DNA of such species must be obtained from thalli taken from the gut, and discriminated from the host DNA and that of  other organismal gut inhabitants and transients. White (2002, 2005) used mostly DNA extracted from unculturable species, and from some cultured species, to construct phylogenies of Harpellales (Fig. 12.7). He confirmed other studies indicating a relationship with the Kickxellales and other Zygomycetes (Keeling et al., 2000; O’Donnell at al., 1998; Gottlieb and Lichtwardt, 2001). White also demonstrated that the two largest genera of Harpellales, Smittium and Stachylina (no species of the latter genus is culturable) are polyphyletic and consequently need revision. Gottlieb and Lichtwardt (2001), using selected isolates of Smittium and a few other genera from the University of Kansas culture collection, also found that Smittium was polyphyletic, breaking down into five lineages (Fig. 12.8). Likewise, the largest genus of Eccrinales, Enterobryus, does not appear to be monophyletic (Cafaro, 2005) (Fig. 12.6). These data indicate that the placement of species in some of the larger genera of trichomycetes cannot be done with confidence based solely on morphology. This said, it appears that species of most of the smaller genera of trichomycetes have distinctive characters that suggest the genera are reliable.

            White (2005) also found that classifying Harpellales into two families, Harpellaceae and Legeriomycetaceae, is not supported by DNA evidence. Likewise, Cafaro’s (2005) trees did not support separation of Eccrinales into two of the three families established in this monograph, Eccrinales and Palavasciaceae. (Representatives of the third eccrinid family, Parataeniellaceae, were not included in his analyses.) Cafaro was not able to get definitive DNA sequences of a representative of the fourth trichomycete order, Asellariales, but there is some evidence that Asellariales and Harpellales share their ancestry (James and White, unpublished).

            It is quite clear that studies on trichomycete evolution will require the use of a wider range of genes and the inclusion of a better representation of taxa before more robust phylogenetic hypotheses are possible. Particular problems remain. For instance, the placement of Baltomyces styrax in the Asellariales is provisional, and placement of Orchesellaria spp. (in Collembola) and Asellaria spp. (in Isopoda) in the same order and family is probably not justified, given their different modes of reproduction and host adaptations. Within the Harpellales and Eccrinales the evolutionary relationships of most of the genera remain undetermined using modern DNA analytical techniques, because there are problems in finding infested host specimens in their respective geographic regions of the world and the difficult process of extracting, amplifying, and sequencing DNA of these microscopic gut organisms when axenic cultures are not available. 
 

Morphologically Similar but Unrelated Organisms


Attached to the guts or external cuticle of many arthropods are various kinds of microorganisms, some of which have a superficial resemblance to trichomycetes. These, as well as certain free-living forms, have from time to time been compared with trichomycetes or sometimes have been confused with them taxonomically. A list of a few such organisms and a discussion of their characteristics follows below.
 

Arthromitus

The hindguts of many arthropods, especially the millipedes and certain beetles, may have numerous narrow unbranched filaments attached in clusters by means of a small common holdfast. They are prokaryotes. The filaments divide into cells, and the more terminal series of cells at one stage may each contain an oval endospore. The genus Arthromitus was established for such microorganisms by Leidy (1849a) in the same paper in which he described the first trichomycete genus, Enterobryus. The type species is Arthromitus cristatus. Leidy later (1850b) described a second species, A. nitidus, but in his illustrated major paper of 1853 on gut organisms he made the second species a synonym of the first.

There is considerable morphological variation in these filaments (Fig. 12-2). They are most often attached directly to the cuticle of the hindgut, but may as well attach to nematodes and even to eccrinid thalli. Their taxonomy is uncertain and confused. The 8th edition (1974) of Bergey's Manual of Determinative Bacteriology places Arthromitus, along with some other septate filamentous genera of endospore-producing bacteria, in the category of "uncertain taxonomic position." However, previous editions of Bergey's Manual placed several species of Arthromitus in a separate family of the order Caryophanales Peshkoff. Manier (1961a) studied Arthromitus in the guts of species belonging to several millipede families, and described phases of its development and cytology. We have frequently encountered one or more species of the genus, and has been unable to culture them axenically. The filaments one observes among hosts may be so variable that it is possible that different taxa are involved. The most distinctive feature of the nonsporulating phase is the minute but obvious holdfast that connects the cluster of unbranched filaments at their base. In some cases the filaments are wound together in tight helical strands very much like a piece of rope, but they tend to uncoil when slides are prepared. The individual thalli are 0.6-2.3 µm wide and up to 1 mm or more long (Manier, 1961a), although usually considerably shorter. They are more robust than most of the other, often numerous, filamentous, and sometimes branched bacterial thalli that proliferate in the guts of certain arthropods. More recently, Margulies et al. (1998) managed to cultivate Arthromitus from different arthropods, and were able to identify this organism as a particular stage of the bacterium Bacillus cereus through 16S ribosomal sequences.

Maessen (1955) apparently confused Arthromitus with eccrinid thalli. She established the genus Microeccrina and eight new species taken from millipede guts and one isopod. We also believe, like Manier (1961a), that these are almost certainly filaments of Arthromitus species. The writers also believe that Maessen's new genus and species, Microtrichella hydrophilorum, found in a species of hydrophilid beetle, may be an Arthromitus.
 

Harpochytrium

Harpochytrium Lagerheim is a genus of zoosporic fungi normally attached epiphytically to living freshwater algae but occasionally to other substrates as well. The thalli have some resemblance to Amoebidium, and could possibly be confused with that genus (Fig. 12.3). Reproduction in Harpochytrium, however, does not resemble that of Amoebidium. Jane (1946) stated that Harpochytrium spp. were fungi, and in his revision of the genus Jane recognized six species, some with several different forms. The type species is H. hyalothecae Lagerheim. Emerson and Whisler (1968) cultured H. hedinii Wille axenically and established a new order of posteriorly uniflagellate lower fungi, the Harpochytriales, to contain this genus and another trichomycete look-alike, Oedogoniomyces. Based on molecular studies, Harpochytrium has now been established as a Chytridiomycota (James et al., 2000), whereas the superficially similar genus Amoebidium belongs to the Mesomucetozoea (Cafaro, 2005).
 

Oedogoniomyces

Shortly after Kobayashi and Okubo (1954) discovered and described the new genus and species Oedogoniomyces lymnaeae (Fig. 12.4), which they found attached to living freshwater snails in Japan, Emerson and Whisler (1959) reported the discovery and axenic isolation of several strains of Oedogoniomyces in Costa Rica and later (1968) in California as well. Their isolates came from freshwater ponds and soil, and were found growing naturally on substrates in the water (banana, twigs) or were obtained by baiting with apples and hemp seeds. Emerson and Whisler included this genus, along with Harpochytrium, in the new order of posteriorly uniflagellate fungi, the Harpochytriales, established in 1968.

The unbranched (or rarely branched) thalli of Oedogoniomyces with their basal holdfasts and chains of zoosporangia formed in basipetal succession have a remarkable gross similarity to an eccrinid thallus such as Enterobryus sp. Emerson and Whisler called attention to this similarity (1959, 1968), but on the basis of the production of zoospores and the chitinous nature of the thallus walls, they did not conclude there was a natural affinity between the two groups. In addition to the matter of motility of spores in Oedogoniomyces, it should be noted that while both types of coenocytic thalli form sporangia in basipetal succession, those of the eccrinids normally delimit uninucleate sporangia within which develop single uni- or multinucleate aplanospores, while the Oedogoniomyces sporangia presumably are multinucleate from the start (the karyological status at this stage has not been described), and each sporangium produces many uninucleate zoospores.
 

Mononema

This genus was established by Balbiani (1889) for a curious fungus-like growth attached to the stomodeum (foregut) of centipedes (Chilopoda). Mature specimens of Mononema moniliforme Balbiani are unbranched filaments of up to 1 cm or more in length consisting of swollen pairs of cells, each pair connected to the others by a narrow zone or separated from them by an isthmus (Fig. 12-5). The basal cell is slightly tapered and is attached to the cuticle by the narrower end. The pairs of cells appear to arise by intercalary division, and according to Manier and Ormières (1980), the cells are uninucleate and there is no coenocytic stage of development. The pairs of cells apparently disarticulate in the manner of arthrospores and function as propagules. This species has been found in Cryptops hortensis L. by Balbiani and by Manier and Ormières, and in C. anomolans lusitanicus Verg by Léger and Duboscq (1903). Patterson (1999) classifies Mononema as a parasitic protist.

A number of millipedes (Diplopoda) harbor similar organisms in their oesophagus (foregut), which Manier and collaborators have placed in the genus Mononema, but further studies are needed to determine whether they are not, in fact, sufficiently different to warrant a separate generic designation. Two species from millipedes have been named: M. perapolydesmi Manier 1964 ex Manier, Gasc & Bouix 1972b from Perapolydesmus progressus (Bröl.) (a small cave-dwelling millipede) and from Polydesmus complanatus L.; and M. demangei Manier, Gasc & Bouix 1972b from Orthomorpha coarctata (Saus.) and Cordyloporus ornatus (Pet.). Mature filaments are unbranched and tapered toward the base, and consist of swollen binucleate cells that appear to develop at least partially by intercalary divisions. Filaments of M. demangei may attain a length of 7 mm (Manier et al., 1972b). The tapered basal cell of both species has a distinctive holdfast and is multinucleate. They propagate, like M. moniliforme, by separation and dispersal of cells. Chains of attached cells may also break loose.

Mononema species from millipedes have some resemblance to eccrinid thalli, although their development by intercalary divisions and their arthrospore-like disarticulating cells would seem to make them unrelated. Lichtwardt (1954, 1960) found detached chains of cells and even some entire detached filaments of Mononema sp. among thalli of Enterobryus apheloriae in the hindgut of Apheloria iowa and E. oxidi in Oxidus gracilis, and he mistakenly thought they were uncommon forms of development of those respective species. Subsequently, Lichtwardt (unpublished) has found Mononema sp. in the foregut of a number of neotropical millipedes. These same hosts are often infested at the same time with Enterobryus spp.

Several other fungus-like organisms have been found in the foregut of centipedes, but they have less of a resemblance to the trichomycetes. These include Omphalocystis plateaui Balbiani 1889, Rhabdomyces lobjoyi Balbiani 1889, and Oesophagomyces lithobii Manier and Ormières 1980.