Chapter 7

Life Cycles

HARPELLALES - The Harpellaceae and Legeriomycetaceae, commonly referred to as "harpellids," have similar asexual and sexual reproductive structures and life cycles (Fig. 7.1). In the unbranched Harpellaceae, asexual reproduction commences after the maximum but limited number of cells has been produced. Each compartment becomes a generative cell, giving rise exogenously to an appendaged trichospore. In the branched Legeriomycetaceae, only the terminal parts of branches normally become fertile, and vegetative growth and sporulation may continue as trichospores are maturing on older branches of the thallus.

Trichospores are in reality unispored sporangia. Sporangiospores normally do not extrude from the sporangium (trichospore) immediately within the same gut that produced them, but rather are released to the outside aquatic environment where they are presumed to lie dormant until ingested by a suitable host. Extrusion then occurs rapidly, the sporangiospore breaking through the trichospore wall so that it can attach to the peritrophic membrane (in Harpellaceae) or to the hindgut lining (in Legeriomycetaceae) during the relatively quick passage through the host gut. The attached spore body then grows and matures into a new thallus. The process of sporangiospore extrusion (Fig. 9.12) has been studied in vitro by Horn (1989a, b, c, 1990), and is presented in some detail in Chapter 9.

The sexual state in some genera of Harpellales does not develop until the host is about to molt, and in some instances trichospore development may become completely arrested as zygospores begin to form. In other (branched) genera, both zygospores and trichospores may be produced simultaneously on the same thallus. The zypospores may or may not be appendaged on release, depending on the type. Like trichospores, zygospores are thought to germinate after passage from the gut and on being subsequently ingested. Zygospores are believed to produce thalli similar to those that develop from trichospores, but this development has never been observed.. Details of conjugation and zygospore structure are given in a later section of this chapter.

Given that Harpellales live in larval stages with limited mobility, it was not understood how they could be so geographically widespread in aquatic insects and present in the headwaters of many streams. It is now known in several species of Harpellales that from time to time the ovaries of blackflies become filled with fungal cysts that suppress egg development. These cysts are carried by the female to new sites where the cysts are "oviposited." Upon germination, the cysts produce cystospores that can be ingested by nearby blackfly larvae. This phase of the life cycle is described in the Pathogenicity section of Chapter 8 (Fig. 8.1). At the present time dissemination of Harpellales by means of ovarian cysts is known only in a few species, but it is hypothesized that this mechanism of dispersal may be present in one form or another in all Harpellales.

ASELLARIALES - The Asellariales, which may be closely related to the Harpellales, produce arthrospore-like cells by fragmentation of the branches (Fig. 7.2). In Asellaria ligiae and A. unguiformis the arthrospores develop outgrowths that resemble trichospores in shape and size, and therefore arthrospores may be the equivalent of disarticulated generative cells. No further development of these structures has been directly observed; it is assumed they are ingested by their hosts and develop into new thalli. Conjugations among thalli have been seen, but zygospores have not been formed in those instances.

ECCRINALES - The Eccrinales, often called "eccrinids," have several patterns of life cycle (Fig. 7.3, Fig. 7.4, Fig.7.5). In two of the three families (Eccrinaceae and Parataeniellaceae), there are two principal types of endogenous spores produced. One is multinucleate and capable of germinating immediately in the gut to produce a new thallus. The other is usually uninucleate at first, is thick walled in some species, and it serves as the propagule to transmit the fungus to another host. Except in the Parataeniellaceae, where many uninucleate spores are produced in a common sporangium, all other spores are produced singly in separate, terminal series of sporangia. The Palavasciaceae produce but one spore type, which functions to spread infestation to other appropriate hosts. No sexual reproduction has been confirmed in the Eccrinales. In the genus Enteropogon, apparent fusions between cells and their nuclei have been observed (Hibbits, 1978) (Fig. 11.36), but it is not known at present if this is a normal sexual process or aberrant behavior.

AMOEBIDIALES - The ectocommensal genus Amoebidium has two distinct phases in its cycle (Fig. 7.6) (Manier and Raibaut, 1969). The common means of propagation is by the production of elongate sporangiospores in the mature thallus, which serves as a sporangium. When these mature they elongate and burst through the sporangial wall, normally remaining in this position until another host contacts the exposed tip of a spore. The spore adheres to its new host and is pulled from the sporangium and develops into a new thallus. When the host molts or is injured, however, the protoplasm in developing thalli cleaves into uninucleate membrane bound portions with no cell walls. The thalli then rupture and release a number of motile, amoeboid cells. The amoebae of Amoebidium parasiticum crawl about the substrate for about 1-6 hr (Manier and Raibaut, 1970), then round up and encyst. [In Paramoebidium spp. the motility of amoebae may last only 15-30 min before encystment (Lichtwardt, unpublished)]. The cysts enlarge and eventually, sometimes several weeks later, produce one to more than a dozen elongate cystospores, each in its own chamber. These pop out of the cyst and are presumed to infest new hosts that make contact with them.

Species of the genus Paramoebidium live in the hindgut or rectum of their hosts, and therefore have no opportunity to transmit sporangiospores to other hosts by direct contact. Only the amoeba-cyst phase is present, and this part of the cycle is in all general respects identical to the development in Amoebidium. We must assume, in contrast to Amoebidium, that either the entire cysts or released cystospores are ingested before reinfestation of another host occurs. Dang and Lichtwardt (1979) have suggested that the cyst, in addition to its possible role as a resistant structure, may function as a timing device to insure discharge of the cystospores in the hindgut after ingestion. So far as is known, amoebae of Paramoebidium spp. do not release and emerge from the gut during the intermolt period, but rather the thalli complete their maturation and amoebae escape only after molting or injury of the host such as by dissection. Mature thalli apparently can be held in the ready stage until ecdysis, because it is sometimes possible to find amoebae being released within minutes after the exuvia is shed.

The amoebae do not appear to function as gametes. Poisson (1931b) stated that he on some 10 occasions observed amoebae fuse in pairs, but this has not been verified by other investigators who have studied this organism rather intensively; it would be interesting indeed if fusion could be observed again and nuclear behavior followed. Nor do amoebae appear to feed during their period of migration. Rather, their function may be to serve as a means of dispersal and to select an appropriate substrate upon which, or within which, to encyst.

It should be noted here that Whisler (1966) reported removing three thalli of "Paramoebidium" from the hindgut of a stonefly nymph and that they gave rise to cultures of thalli identical to A. parasiticum when plated on cornmeal agar. These axenic cultures readily infested Daphnia as ectocommensals and produced both the sporangiospore and amoeba-cyst phases. He viewed this as evidence that the two genera may be synonymous, the morphological differences being related to different growing conditions. The authors have tried unsuccessfully to culture Paramoebidium spp. from larval stoneflies, mayflies, and blackflies on many occasions. It would be most interesting and significant if Whisler's observations could be repeated. At the same time it should be borne in mind that nonsporulating thalli of A. parasiticum can be essentially indistinguishable from many young thalli of Paramoebidium spp. Lichtenstein (1917a) found A. parasiticum growing in the rectum of dragonfly nymphs, as stated in Chapter 4, and Chatton (1906b, 1920) described A. recticola growing attached to the hindgut lining of two species of Daphnia. Chatton noted that rhythmic dilations and contractions of the rectum of daphnids create a pumping action that brings water into the rectum. He demonstrated that granules of carmine could be drawn into the rectum, and surmised that in this way spores of A. recticola passively invaded the gut and became attached. In both Lichtenstein's and Chatton's reports, sporangiospores were produced in some of the thalli, leaving little doubt that these were species of Amoebidium and not Paramoebidium. Perhaps aquatic insects such as stonefly nymphs in rare instances can aspirate spores of A. parasiticum via the anal route by a similar mechanism, resulting in growth in, or at least attachment to, the rectum.
 

Thallus Types and Development

The gross morphology of trichomycete thalli varies not only from simple to highly branched, but also in relative size. Width extremes range from about 2 µm in the microthalli of Astreptonema gammari to well over 100 µm in some robust thalli of Arundinula hapalogaster from marine decapods. The length of mature thalli is as short as 30-55 µm in Stachylina minuta and 1 cm or more in length in some species of Enterobryus and Arundinula. In several genera of Eccrinaceae (Alacrinella, Astreptonema, Enteromyces, Paramacrinella, Ramacrinella), a dimorphism of thalli exists. That is, there are two rather discrete size ranges into which the thalli fall. In some species of other genera (e.g., Enterobryus borariae, Arundinula hapalogaster) the range of thallial sizes may be just as great, but the measurements intergrade to a considerable extent and are consequently not strictly dimorphic. In rare instances (e.g., Enteropogon sexuale) the species appear to be polymorphic, with the different thallial types occupying specific regions of the gut (Hibbits, 1978).

The walls likewise differ in their dimensions, from thin and delicate to exceptionally thick (3 µm or more) in older thalli of some species. Some Paramoebidium spp. become sufficiently thick walled to prevent the escape of amoebae, and the amoebae end up encysting within the sporangium. Electron microscopic studies reveal that the thallus wall usually consists of at least two distinct, fibrous layers in the Harpellales, the innermost being more electron transparent (Reichle and Lichtwardt, 1972; Moss, 1976; Moss and Lichtwardt, 1976), whereas in the Eccrinales there may be three or more discernible layers (Grizel, 1971; Manier and Grizel, 1972; Moss, 1975; Dang, 1979). The authors have observed that outer layers of the wall occasionally slough off in older thalli of a number of species of Enterobryus and some other eccrinids. Also see Arundinula galatheae in Chapter 11.

The chemical composition of the wall has been studied in only a few species (Table 7.1). Sangar and Dugan (1973), using mechanically isolated walls from an axenic culture of Smittium culisetae, determined that chitin was a major component. The studies of Trotter and Whisler (1965) on an isolate of Amoebidium parasiticum revealed no glucose or glucosamine in their chromatographic analysis of acid hydrolysates of wall material. They concluded that the walls possibly contain hemicellulose, but no chitin or cellulose. The fact that A. parasiticum and thalli of Paramoebidium sp. are soluble in hot dilute (0.225%) KOH (Whisler, 1963) rules out the likelihood that either chitin or cellulose are significant contributors to wall structure in these organisms. The results of chemical analyses of cultured trichomycetes are given in greater detail in Chapter 9.

Histochemical tests are less reliable and precise indicators of wall composition, but are the most practical means of testing species that are not culturable. Whisler (1963) was able to get a positive van Wisselingh test for chitin with two uncultured species of Harpellales, Legeriomyces ramosus and Pteromaktron protrudens, but this same test was negative for Enterobryus halophilus, A. parasiticum, and Paramoebidium sp. These tests, plus others listed in Table 7.1, lead to the tentative conclusion that the Harpellales contain chitin in their walls, the Eccrinales may have cellulose in, theirs, but the walls of Amoebidiales contain neither of these subtances. It is probable that the Asellariales, believed to be closely related to the Harpellales, will be found to have walls composed of chitin whenever a suitable chemical analysis can be run. It must be stressed that wall composition, important as it is in contributing to phylogenetic schemes (Bartnicki-Garcia, 1868, 1970), is not a settled matter in the trichomycetes.

In the Harpellales and Asellariales the thalli become septate as they develop. Each cell appears to be normally uninucleate in most species, although coenocytism is the condition in the Asellariales and some Harpellales in portions of thalli that do not become regularly septate. In contrast, the coenocytic thalli of the Eccrinales do not form crosswalls except to delimit sporangia. Farr made a significant discovery during the course of his studies on Smittiumculisetae (Farr, 1965; Farr and Lichtwardt, 1967), namely that the septa are perforate and differ morphologically from the well known perforate septa of higher fungi such as ascomycetes and basidiomycetes. Smittium septa were found to be swollen around a central pore that later became occluded by an electron-opaque plug. Farr found that this type of septum also occurred at the base of trichospores. This harpellid septum (Fig. 7.7), also known to form during zygospore delimitation, has now been seen in many genera of Harpellales and in two genera of Asellariales (Manier and Coste-Mathiez, 1968; Reichle and Lichtwardt, 1972; Manier, 1973a, 1973b; Moss, 1972, 1975, 1976; Moss and Lichtwardt, 1976, 1977; Sato et al, 1989).

In median longitudinal sections of thalli the septum is seen to have three distinct layers: two outer ones that are continuous with the inner wall of the thallus, with a more electron-transparent inner layer sandwiched between. The inner layer bifurcates at the periphery of the pore to give a flared appearance in transverse sections of the septum. The mature structure superficially resembles the bordered pit of a tracheid. In younger septa there is complete continuity of protoplasm through the 0.8- to 0.9-µm-diameter pore, and endoplasmic reticulum is commonly seen linking the two adjacent cells. The septal pore later becomes occluded by a biumbonate plug, without a bounding membrane, that effectively prevents passage of cellular organelles when fully formed. Nevertheless, intercellular continuity is maintained by the plasmalemma, which is continuous from cell to cell around the periphery of the plugged pore (Moss, 1975).

This unusual type of perforate septum was also discovered by Young (1969) in the hyphae of Linderina pennispora (Kickxellales; Zygomycota), and the similarity is believed to be phylogenetically significant, as is discussed in Chapter 12. Less similar, but nevertheless having some resemblance, is the crosswall of species of Dimargaritales (Zygomycotina) with its flared, lenticular pore region enclosing a biumbonate plug attended on each side by a large globose body (see Brain et al., 1982).

Of equal interest in trichomycete morphology is the type of perforate septum constructed by species of Eccrinales during sporangium delimitation. This has been described by Manier (1979a) in Palavascia sphaeromae, but studied in detail only in Astreptonema gammari (Moss, 1972, 1975). It would be very desirable to look at septum formation and structure in other species of Eccrinales to determine if there is a common pattern. Moss described the crosswall in A. gammari as a single electron-transparent layer continuous with the inner layer of the three-layered lateral wall, with a central pore of 0.3-0.4 µm diameter around which the septum is dilated. The plasma membrane is continuous from one cell to the other through the pore, and cytoplasm appears to stream into the developing sporangium during this initial stage. One or more dictyosomes appear in the sporangial cell near the pore and produce secretory vesicles that migrate toward the pore and form a "pore-cap" on each side of the septal pore. This is followed by fusion of the vesicles with the plasmalemma and deposition of wall material in and around the pore so as to completely occlude the opening and isolate the sporangial plasmalemma from the rest of the thallus on completion of the process. This plugging of the pore by dictyosome-derived vesicles in the Eccrinales appears to be quite different from the formation of the septal plug in the Harpellales and Asellariales. Dictyosomes have not been seen in those two orders, but they have been reported in the Amoebidiales (Whisler and Fuller, 1968).

The cellular contents and organelles of trichomycete thalli are generally similar to those of other fungi. Lipid droplets are evident and sometimes abundant in electron micrographs. Patrick et al. (1973) found that lipid materials extracted from thalli of Smittium culisetae account for 9.9% of the total weight, and one can assume that lipids are a major food storage substance. Glycogen granules have also been detected and are sometimes abundant (Moss, 1972; Moss and Lichtwardt, 1977). Reichle and Lichtwardt (1972) reported finding membrane-bound microbodies containing electron-opaque crystals in Harpella melusinae. In the same species they also found a semicircular structure associated with the nuclear membrane and paralleled by 8-9 tubules or circular particles. The origin and function of these and other unusual cellular inclusions are yet to be determined. Virus-like particles have been found in two species of Paramoebidium, and are described and illustrated in the last section of this chapter. Also described in the sections that follow are special cellular features pertaining to the reproductive bodies of trichomycetes.

Very few critical cytological studies have been made of the nuclei of trichomycetes. Raabe (1911a, 191lb, 1912) described and illustrated the structure and division of Amoebidium parasiticum nuclei in great detail, but an interpretation of his studies, in light of our modern knowledge of karyology, is difficult. He described them as lacking a nuclear membrane, and having about 6-8 chromosomes that divide amitotically with the involvement of centrioles.

At interphase, and in some stages that appear to be dividing, the nuclei of many species of trichomycetes are clearly seen and stain well (Fig. 7.8). At other times the nuclei do not take nuclear stains or are obscured by other components of the cytoplasm. Mithramycin and acridine orange can reveal the nuclei clearly with fluorescence in such cases. Even with phase-contrast microscopy it is often difficult to see or study nuclei in many species at certain stages of development. Electron micrographs clearly show that the nuclei are bound by a typical nuclear envelope, but the behavior of the membrane during mitosis has not been studied in detail in any species. Manier (1979a) stated that in Palavascia sphaeromae and other (unidentified) trichomycetes the nuclear membrane persists during nuclear division and that there are no centrioles or spindles involved, but she provided no electron micrographs. The authors have seen what appear to be intranuclear divisions on several occasions with the light microscope. Moss (1974) was able to describe some Feulgen-stained mitotic figures in Stachylina grandispora and to obtain a count of about seven chromosomes in the somatic nuclei.

It is assumed that the nuclei of trichomycetes are haploid like those of most other fungi. Isozyme patterns in Harpellales confirm that they are haploid (Grigg, 1994; Grigg and Lichtwardt, 1996). Duboscq et al. (1948) described the development of resistant spores in Eccrinidus flexilis as involving a complicated sequence of steps that would appear to consist of meiotic divisions followed by karyogamy. If this interpretation is correct, the somatic complement would be diploid, but Manier (personal communication) has studied this same species and has not been able to substantiate the kind of nuclear behavior described by Duboscq et al. Moss and Lichtwardt (1977) found four nuclei in the sexual apparatus of Harpella melusinae after thallial anastomosis and presumed karyogamy that suggested a meiotic division prior to zygospore formation (Fig. 7.9). If true, this would result in the direct development of haploid thalli from zygospores. Additional studies of sexual processes in these fungi are needed before ploidy can be determined.

Thalli of trichomycetes are conveniently divisible into two forms: branched (Legeriomycetaceae and Asellariales) and unbranched (Harpellaceae, Eccrinales, and Amoebidiales). The branched forms become septate as they develop, though some of the cells may remain coenocytic, whereas the unbranched thalli produce septa only to delimit reproductive cells. One species of Eccrinales, Ramacrinella raibauti, produces multiple true branches at the base of the thallus, but the thallus remains coenocytic and develops no septa except during terminal sporangium formation on one or more of the branches. Species of another eccrinid genus, Palavascia, produce what appear to be branches from sporangia; however, these are clearly sporangiospores that "germinate" in situ (Manier, 1979a), and not true branches. In a few eccrinids the holdfast may not remain strictly basal due to some peculiar pattern of growth and development, as described in the next section, but these thalli also are not truly branched despite such an appearance in some cases.
 

Holdfasts

A feature common to all trichomycete thalli is their firm adhesion to the substrate by means of a holdfast of some kind (Fig. 7.10). The typical holdfast consists of a rigid structure at the base of the thallus, easily discernible with the light microscope. However, considerable variation in morphology and development exists among the species. As defined here, the holdfast is considered to be either the rigid, acellular structure secreted at the base of an essentially unmodified cell, or a morphologically distinct basal cell (the "holdfast cell") together with whatever substance is secreted to provide adhesion to the cuticle. The holdfast has proved to be sufficiently distinctive in some species or genera as to be of taxonomic value. It is perhaps to be expected that there be some variation in this essential apparatus, considering the morphological heterogeneity found within the fungal class and the diversity of arthropod guts to which the different species must adhere. Nevertheless, as will be discussed later in this section, several recent studies of their fine structure have revealed unexpected differences in holdfasts, even among closely related taxa in some instances.

In the unbranched trichomycetes (Harpellaceae, Eccrinales, and Amoebidiales), the holdfasts, are discrete structures that may range from a short cylindrical form, sometimes flared into a broad disk at the base, to longer forms that may be essentially cylindrical, campanulate, or may have other, sometimes less regular shapes. In some Eccrinales the holdfasts are especially large and, although usually much shorter, may measure 50 µm or more in length. Large holdfasts are often longitudinally striated. Intraspecific variation in holdfast size and shape can be found in some Eccrinales, primarily in those species that have a variable thallus morphology.

A curious multiple holdfast system, which characteristically forms in a few species of Eccrinales along with some simple holdfasts, consists of a tuft of individual thalli whose holdfasts are fused into a complex branching structure with a common base of attachment to the host cuticle. In Enteromyces callianassae (from the stomach of decapods), this apparently begins as a simple holdfast. Then, as sporulation proceeds, the unbranched thallus curves downward and sporangiospores are released near the basal part of the thallus and become attached to the original holdfast (Fig. 11.35), each new developing thallus contributing more holdfast material to the complex as it grows (Hibbits, 1978). Leidyomyces attenuatus (in the hindgut of passalid beetles) produces rosettes of individual thalli that arise from a common holdfast system. It has not been observed how the thalli become joined in such a fashion in this species, although it is probably through some kind of deposition of spores as in E. callianassae. Manier and Théodoridès (1965) illustrated an unusual multiple holdfast system in an eccrinid from passalid beetles collected in Laos that they tentatively assigned to L. attenuatus (but probably is not that species). Their fungus produced some very long (up to 70 µm) and narrow holdfasts with a loosely branched appearance when mature. In the several instances cited above, the development of multiple holdfasts is a consistent feature of some of the thalli of those species, and not an abnormal situation such as can be seen on rare occasions in other species in which the holdfasts of two thalli may become fused when they form in very close proximity.

In almost all unbranched trichomycetes the holdfast is located at the base of the thallus in line with its axis. Some exceptions to this have been found in the Eccrinales that were caused by uneven growth of the thallus from the germinating spore or due to other unusual ontogenetic factors. The most extreme example was described by Whisler (1963) in Enterobryus bifurcatus from a millipede. In this species the multinucleate secondary infestation sporangiospores have the usual basal rudimentary holdfast upon initial attachment to the cuticle, but the developing thallus bifurcates at an early stage of growth, resulting in a mature nonseptate thallus with a "lateral" holdfast and two divergent arms of approximately equal length (Fig. 11.235). Other examples of holdfasts that are not strictly basal are described in Chapter 11 (e.g., Enterobryus tuzetae).

Among the branched trichomycetes (Legeriomycetaceae and Asellariales), one finds several different basic forms of holdfast. The majority of Legeriomycetaceae produce a rigid secreted holdfast not unlike the typical holdfast of the unbranched trichomycetes. Sometimes the holdfast is inconspicuous, or it may be difficult to locate if the mature thallus has no main axis and is prolifically branched near the base (e.g., some species of Smittium).Where the thallus contains a prominent main axis (e.g., Genistellospora homothallica), the holdfast usually is also prominent.

A different type of holdfast system has evolved in species of Pennella and Stipella. Here there is a secretion of mucilaginous substance from the basal cell that serves to anchor the thallus, and the basal cell lies parallel to the substrate so as to present a greater adhesive surface. In some species (e.g., Pennella hovassi)the basal cell of larger thalli is typically branched and thereby provides an even greater surface area in contact with the host cuticle. Pennella angustispora (and possibly also other less studied species) secretes the adhesive substance through the wall onto the outer surfaces of the entire basal cell, not just the lower part of that cell. The cell wall itself is not in direct contact with the cuticle (Mayfield and Lichtwardt, 1980). The mucilage in this type of holdfast system is essentially transparent and is therefore best seen with phase-contrast microscopy. Often bacterial cells and some debris are embedded in or stuck to the mucilage. Smittium pennelli, unlike any other described species of this large genus, also produces a Pennella-likeholdfast with a mucilaginous secretion.

A third type of holdfast in the branched trichomycetes is found in species of Asellariales. It consists of an enlarged or otherwise distinctively modified basal cell, usually without a prominent secreted structure. In some species a small amount of mucilage or a short basal zone of rigid holdfast material can be detected. (Asellaria armadillidii is an exception, for a noticeable amount of mucilage is secreted such that the basal part of the thallus somewhat resembles the type found in Pennella and Stipella.)In most Asellariales the shape of the basal cell is so distinctive that it serves as the principal means by which the various species are distinguished (Manier, 1979). These basal cells of Asellaria and Orchesellaria spp. are described and compared in Chapter 11.

Modifications can be recognized in the three basic holdfast types in the branched trichomycetes just described. An interesting variation, for instance, is found in Glotzia ephemeridarum (Lichtwardt, 1972; Moss, 1979), where the primary holdfast consists of a small, rigid, peglike or disklike secretion. Then, as the thallus become larger, series of auxiliary holdfasts develop along the basal cell to strengthen further the attachment to the cuticle (Fig. 7.10). Large thalli of G. ephemeridarum may have a tapered basal cell with a row of auxiliary holdfasts somewhat resembling the suckers on the tentacle of an octopus.

Light and electron microscopic studies to date, with a few notable exceptions cited in the next paragraph, indicate that holdfasts are entirely superficial on the host cuticle. The holdfasts are in intimate contact with the gut lining and follow the contours, setae, depressions, or other relief features of the lining. Sometimes a slight distortion of the cuticle by the holdfast is evident in whole mounts or sectioned material [e.g., Genistellospora homothallica (Mayfield and Lichtwardt, 1980); Palavascia sphaeromae (Manier, 1979a)], or, where there is an enlarged and modified basal cell (Asellariales), that cell may grow in such a way that it appears to pinch the cuticle layer. The lack of penetration by the holdfast through the cuticular layers of the gut adds weight to the interpretation that holdfast structures function as anchorage devices and not for absorption of nutrients from the tissues of the host.

However, several exceptions to the strictly superficial attachment of the holdfast are known. Ultrastructural studies of the holdfast of a dozen trichomycete species have revealed that in two Harpellaceae, Stachylina grandispora and Harpella leptosa (Moss, 1972, 1979; Moss and Lichtwardt, 1980), the secreted holdfast substance may penetrate into the inner layers of the peritrophic membrane immediately adjacent to the holdfast [although another species of Harpella, H. melusinae, does not appear to do this (Reichle and Lichtwardt, 1972)]. Such penetration should enhance adhesion to the membrane, and it would not be surprising to find this situation in other trichomycetes yet to be studied ultrastructurally. More interesting are three examples of trichomycetes that actually penetrate through the chitinous lining: (1) Several Stachylina species have a base that perforates the peritrophic membrane (Fig. 7.11) and produces a slightly bulbous swelling underneath the membrane (Lichtwardt, 1984a), apparently for anchorage and as a substitution for the rigid, secreted type of holdfast that is characteristic of other species of that genus. (2) Peculiar structures in Enterobryus borariae were found located beneath the cuticle of its millipede host by Lichtwardt (1958) that consisted of pyriform, cystlike bodies (Fig. 7.11) that germinated to produce sporulating thalli of that species within the gut lumen; the origin of these bodies was not determined. (3) A third instance of cuticle penetration has been observed by Lichtwardt (unpublished) in several specimens of Ligia sp. collected in a freshwater stream in Hawaii and containing Asellaria ligiae: in heavily infested hosts a number of thalli were seen to have no bulbous basal cell characteristic of this species, but rather had penetrated the hindgut lining and had produced a branching rhizoidal growth underneath the lining (Fig. 7.11). These thalli appeared to be branches of the fungus that had broken off from other thalli in the gut and then had produced this adventitious growth from the proximal end to anchor themselves to the lining.

In none of the cases cited above have the structures in question appeared to penetrate into the underlying cellular tissues of the host. However, Sweeney (1981a) has described penetration of epithelial cells of the midgut of mosquito larvae by Smittium morbosum. This apparently unusual occurrence, and other pathological effects by species of Smittium, are presented in Chapter 8.

The vast majority of trichomycetes attach to specific and predictable substrates. Some instances of unusual attachment are known (Table 7.2). One of the most interesting is found in Simuliomyces microsporus which grows in the hindgut of blackfly larvae. While its thalli usually attach to the gut lining, not uncommonly it is seen in various stages of development, including sporulation, on thalli of Paramoebidium sp. (Fig. 7.12) or Genistellospora homothallica that share the same gut, or even on thalli of its own species. In the latter instance, it may require careful observation to see that some of the "branches" arising from thalli are in fact different thalli of the same fungus attached by a rather inconspicuous holdfast. What evidently was S. microsporus on Paramoebidium spp. was seen and published by several investigators prior to 1972 when S. microsporus was described, but it appears to have been misidentified as species of other genera: as Stipella vigilans (Manier, 1955a) and as Smittium sp. (Ingold, 1967; Moss, 1970).

Enterobryus borariae is another species that usually attaches to the hindgut cuticle of its millipede host but has been found in all stages of development attached to other thalli of its own kind (Lichtwardt, 1958). Frequently it attaches as well to the cuticle of oxyuroid nematodes that are common parasites of millipede hindguts (Fig. 7.12). The attachment of Eccrinales to nematode surfaces was first mentioned by Leidy in 1849a, and described and illustrated in 1853, in Enterobryus elegans. This phenomenon was restudied in the same species by Lichtwardt (1954), who also found a modified spore type that attached to nematodes in abundance just prior to molting of the millipede. Lichtwardt suggested that this attachment to nematodes might be a mechanism to maintain some of the thalli in the gut during ecdysis, because the nematodes, unlike the eccrinids attached to the gut lining, are not expelled with the molt.

The authors have seen several other, undetermined species of Enterobryus attached to gut nematodes, and other biologists have also observed this phenomenon in Enterobryus spp. (Udekem, 1859; Thomas, 1930; Tuzet and Manier, 1952; Dollfus, 1952). Manier (1969b, p. 581) mentioned that Eccrinidus flexilis does likewise. Most Enterobryus species rarely, if ever, attach to nematodes, however. The frequency of attachment of those that do varies from one millipede specimen to another, but in general it depends on the density of the fungal population. In heavily infested guts, it is not uncommon to find the majority of nematodes carrying some eccrinid thalli. The millipede hosts of Enterobryus elegans can have several species of gut nematodes, all of which may carry the fungus, but the percentage of infestation can vary among the species (Wright, 1979). None of the trichomycetes that attach to nematodes, to other trichomycetes, or to their own species penetrates or harms the other organism in any observable way.

The initial attachment of a trichomycete to its substrate is determined by the spore. Except in the Eccrinaceae and Parataeniellaceae, where spore types exist that germinate in the gut immediately on release, spores of trichomycetes pass from the gut and are ingested by a suitable host before they attach and germinate. Alien trichomycete spores, which one can assume are ingested by certain arthropods whose microhabitats overlap those of others (see Chapter 4), apparently pass through the gut without attachment. In fact, recognition of a suitable substrate to which the spore can attach may represent the first level by which host specificity is determined. Once in a suitable host, attachment occurs predictably either on the lining of the foregut, midgut, or hindgut, often in particular regions of those parts of the gut. [Sporangiospores of the ectocommensal Amoebidium parasiticum may have a simple adhesive property, which could account for the wide host range and lack of selectivity in that species, for they even attach to various inanimate objects under special conditions (Table 7.2).] Although a few species of trichomycetes can attach to unusual living substrates, there appear to be no instances known where spores attach to any of the many kinds of solid debris that pass through the gut, except by entanglement of trichospore appendages. These observations lead to the conclusion that there is considerable selectivity exercised by the spore. One possibility, not yet demonstrated, is that the spore produces a lectin that binds to specific carbohydrates in the appropriate living substrate. Such a system could operate even if the major component of the substrate were chitin, cellulose, collagen (as in nematode cuticles), etc., provided suitable saccharides were also present.

The process of substrate recognition and initial attachment of the spore could well be a developmental aspect distinct from the subsequent formation of the holdfast proper. The several electron microscopic studies of holdfasts to date have been based mostly on relatively mature holdfasts; as a consequence, developmental features are not well understood. Nevertheless, these studies have revealed some aspects of their probable development as well as a rather surprising difference among the few species studied so far (Table 7.3). There have been no published analyses of the chemical nature of the secreted substance that forms the holdfast. Duboscq et al. (1948) and Manier (1950) stated that, on the basis of its staining properties (particularly with cotton blue), the holdfast in Eccrinidus flexilis consists of callose. However, the stains they used are not sufficiently specific, nor is callose a likely candidate in this instance because of its solubility properties. Mayfield and Lichtwardt (1980) reported that the silver-methenamine stain gave no indication of polysaccharides being present in the multiple holdfast of Leidyomyces attenuatus, but a strong positive test for polysaccharides was demonstrated in the "ring complex" at the base of each thallus through which the holdfast material is extruded.

Holdfast formation commences soon after a spore has become attached, and in some species may continue to develop for some time along with thallial growth and maturation. In others, holdfasts are of limited size, and development appears to cease as soon as firm attachment is insured. Multinucleate secondary infestation sporangiospores of Eccrinales are capable of immediate attachment on release from the sporangium. In some species the end of the cylindrical sporangiospores that will produce the holdfast can be distinguished while still in the sporangium by virtue of a slightly swollen base, and in a few species the spore may even produce a pad of substance resembling a small holdfast before release.

Whisler and Fuller (1968) studied the attachment of cultured sporangiospores of the ectocommensal Amoebidium parasiticum to both Daphnia and cotton fibers (Table 7.3). (The attachment to cotton fibers occurred when thalli were first transferred to a dilute salt medium but not when kept in their tryptone broth growth medium.) The fusiform spores were found to have numerous conical pits that perforated the wall and were primarily concentrated at each pole, but there were some pits widely dispersed along the lateral wall as well. Within the spores they found many membrane-bound vesicles aggregated mostly near the two poles and containing homogeneous material that was electron microscopically similar in appearance to the amorphous substance of the mature holdfast. Unreleased spores showed no holdfast, but extrusion of the cementing substance apparently occurred rapidly after initial attachment. The concentration of pits at both ends of the spore may be adaptive; either end of the spore can break through the sporangial wall at maturity, with the unexposed end of the spore lying within the sporangium until (under natural conditions) the spore attaches to a passing arthropod on contact and is pulled from the sporangium. Species of the closely related endocommensal genus Paramoebidium produce no sporangiospores. However, like Amoebidium species, they produce fusiform cystospores that develop from encysted amoebae. Dang and Lichtwardt (1979) reported that at least one end of these cystospores in P. curvum contained high concentrations of cylindrical pits and that vesicles were aggregated around the plasmalemma in the same region.

Discrete wall pits have also been found at the base of the thallus in the eccrinid Eccrinidus flexilis (Manier and Grizel, 1972), with the pits restricted in this case to a small basal zone on the thallus. An even more specialized area of holdfast secretion has been found in Leidyomyces attenuatus (Mayfield and Lichtwardt, 1980) that consists of a "ring complex" at the base of each thallus, with the holdfast material apparently being extruded through a network of spaces in the middle of the complex. In Enterobryus elegans there are no discrete pores or spaces, but there are thinner areas along the inner wall at the base of the thallus through which the holdfast substance is presumed to pass.

Distinct pits at the base of thalli have not yet been found in the few species of Harpellales studied, but in trichospores of Legeriomyces ramosus (Manier, 1973a) and Genistellospora homothallica (Moss and Lichtwardt, 1976) there are some peculiar apical channels on the inner spore wall that may well be associated with the process of spore attachment and holdfast formation. These are described in the next section. In the eccrinid Palavascia sphaeromae the oval sporangiospores, while still in the sporangium, possess an area of pits in the wall at one end (Manier, 1979a). On release of the spore and attachment to the host cuticle, the holdfast material passes through these pits. The pits can be seen in the more mature thalli in the outer wall layer just above the holdfast; the inner wall layers do not form along this zone of holdfast secretion.

Membrane-bound vesicles present within cells near the region of holdfast formation suggest that these may function as carriers of the holdfast substance or some precursor. The origin of such vesicles is not known for certain, though in Palavascia sphaeromae Manier (1979a) implied they were derived from Golgi. Mayfield and Lichtwardt (1980) reported that the membranes of some of the vesicles in Enterobryus elegans fused with the plasmalemma and were seen releasing their contents just above the zone where the holdfast material is secreted through the wall, and Manier (1979a) described a similar process in P. sphaeromae. The holdfast of E. elegans continues to form and enlarge for some time during thallus maturation, in contrast to most of the other species of trichomycetes that have been studied. It is probably necessary to observe earlier stages in most species if it is to be determined whether vesicular transport is a feature common to all types of holdfast formation. It should be mentioned here that vesicles are also involved in cell wall formation, consequently it is conceivable that some (or all) of the vesicles seen in the general zone of holdfast secretion might in fact be active instead in deposition of wall material, at least in the younger stages of thallus development. This is a problem that should be addressed in future studies. However, the observed fusion of some vesicles with the plasmalemma in the immediate area above the holdfast at the base of relatively mature thalli of Enterobryus elegans and L. attenuatus (Mayfield and Lichtwardt, 1980), in cases where the cell wall seems to be already well developed, gives credence to the concept of involvement of vesicles in holdfast formation.

It is evident from observations on some species of Eccrinales with large holdfasts that the holdfast substance must be at least initially in a fluid or plastic state. For instance, thalli of Arundinula washingtoniensis have been seen where the holdfast substance has wrapped around setae on the stomach cuticle of its decapod host (Hibbits, 1978). In other Eccrinales the base of the holdfast is spread into a disk much broader than the zone at the base of the thallus through which the holdfast substance is extruded. Some thalli of E. elegans have a large and roughly conical shape with successive tiers of holdfast material heaped one on the other (Dang, 1979), looking as though the lower layers had spread out as the holdfast grew in length.

The holdfast of E. elegans is especially perplexing in terms of its development because of the numerous parallel vertical channels (containing a loose fibrous substance) that form within the holdfasts (Fig. 7.13). The upper holdfast region, just below the cell, consists of a relatively dense and homogeneous fibrous material. As one sections the holdfast serially downward, the channels, numbering around 200, become discernible and progressively more numerous. There were about 1800 in the broad base of one large thallus that was studied (Mayfield and Lichtwardt, 1980). Each channel arises from a pitlike depression in the gut cuticle. The parallel channels closer to the periphery of the conical holdfast open as pores on the outer edges, and only the more central ones extend up through the holdfast. Because of their arrangement, it appears reasonable to assume that the channels develop due to some form of condensation (polymerization?) of the holdfast substance well after it has been extruded. Wright (1979) studied this species and made the interesting observation that the holdfast channels are found only in thalli attached to millipede cuticle; those attached to nematodes (Rhigonema infecta) had no distinct channels nor any pores opening to the outside of the holdfast. Taking into account that the arthropod hindgut serves to dehydrate intestinal contents by absorption of water, Wright suggested that if the pitlike depressions on the millipede cuticle represent areas of greater permeability, then fluid flow through these points might prevent consolidation of holdfast material above them and result in channel formation in the holdfast; there is no comparable fluid flow through the cuticle of nematodes and consequently there are no channels within holdfasts attached to nematodes. This interesting hypothesis does not account for the channels in the central core of the holdfast that do not open to the outside of the holdfast.

The many variations in holdfast structure that have been detected through electron microscopic studies are not necessarily adaptations induced by substrate differences in the hosts of the respective species they presently occupy. As an example, one has but to compare the substantially different holdfasts of Pennella angustispora with Genistellospora homothallica growing side by side in blackfly hindguts. Understanding the commonalities as well as the differences that characterize trichomycete holdfasts will have to await further investigations on their development and chemistry.
 

Asexual Reproduction

Asexually produced spores are the major means of propagation in trichomycetes. Sexual reproduction is apparently lacking in most taxa, or where it does occur it is generally less frequent than asexual reproduction. In Graminella spp. a means of vegetative propagation appears to be significant in producing more thalli endogenously in the gut (Léger and Gauthier, 1937; Lichtwardt and Moss, 1981), but trichospores are still formed in relatively large numbers and may function as the major propagule in transmitting the fungus from one individual to another.

Asexual spores develop entirely within the gut lumen in all but a few instances. The exceptions include representatives of three orders. In the Harpellales, Pteromaktron protrudens thalli protrude from the anus of mayfly nymphs and produce a brushlike cluster of spores in the external aquatic environment (Whisler, 1963). The larger sporulating thalli of Zygopolaris ephemeridarum and Z. borealis often project from the anus of their mayfly hosts (Moss et al., 1975; Lichtwardt and Williams, 1984), and Moss (1970) reported that thalli of Stipella vigilans occasionally extended beyond the anus in blackfly larvae. In the Eccrinales there have also been reports of spores projecting from the gut. Thalli of Palavascia sphaeromae apparently do not sporulate until the tip reaches the anus of its marine isopod hosts. The tip then coils and produces a series of sporangiospores, which can often be seen projecting outside of the gut (Lichtwardt, 1961b). Wolf and Wolf (1947) recorded that tufts of thalli of an eccrinid, possibly Enterobryus sp., have been seen protruding from the anal opening of the mud crab Panopeus herbstii, but they did not state if any of these projecting thalli were sporulating.

Species of Eccrinales that are capable of producing unusually long thalli probably will be found from time to time protruding from the hindgut, as reported by Hibbits (1978) in Arundinula hapalogaster. Terrestrial beetles of the Passalidae in the neotropics sometimes contain Passalomyces compressus, which Thaxter (1920) described as "growing wholly exposed on the anal plates of a large species of Passalus. " The authors have made several collections of passalids with this fungus, but have found that the majority of thalli are relatively sheltered within the gut and enclosed within the posterior edge of the elytra most of the time. They are probably exposed only during extension of the anal plates. Nevertheless, their periodic exposure to the air may account for the rather thick walls that develop in most of the thalli and sporangia.

The order Amoebidiales includes the ectocommensal Amoebidium parasiticum whose thalli as well as all reproductive stages are, of course, continuously in the external aquatic environment. Other Amoebidiales inhabit the rectum of their hosts, but Paramoebidium curvum thalli are sometimes found growing on the retractile anal gills of blackfly larvae, rather than being in their more normal position just inside the anus, and consequently are fully exposed to the outside when the gills are extended. It is possible that a number of trichomycetes in the rectum of aquatic arthropods develop and sporulate with at least a sporadic exposure to the external water, brought about by muscular activity, which may draw water into the rectum.

The remainder of this section is devoted to describing the development, structure, and function of asexual spores in each of the orders.

HARPELLALES - The sole asexual spore of the Harpellales is the trichospore. This structure can be defined as an exogenous, deciduous sporangium containing a single uninucleate sporangiospore and normally having one to several basally attached filamentous appendages (Moss and Lichtwardt, 1976). The sporangial nature of the trichospore is particularly apparent when it is it ingested by a suitable host and extrudes its sporangiospore within the gut, which then attaches to the peritrophic membrane or hindgut cuticle and begins to develop into a thallus. This process, and the stimuli that induce extrusion in Smittium culisetae and S. culicis, was studied in vitro by Horn (1989a, b, c, 1990), and is described in some detail in Chapter 9.

The term trichospore was first introduced in this sense by Manier and Lichtwardt (1968) [although its first use in the trichomycete literature was by Chatton (1925), who used it once to refer to the elongate spores of Amoebidium]. Other terms that have been applied to the trichospore are conidium (Léger and Duboscq, 1929a), azygospore (Poisson, 1936), mastigocyst (Manier, 1955b), trichosporangium (Moss, 1972), merosporangium (Moss and Young, 1978), and sporangiole or sporangiospore (Moore-Landecker, 1982). Thalli of almost all unbranched Harpellaceae are completely converted to reproductive cells on reaching maturity, whereas in the branched Legeriomycetaceae only part of the thallial biomass normally becomes reproductive. In Orphella coronata, the trichospores were described as being produced singly at the tips of branches (Léger and Gauthier, 1931), but in all other species of Legeriomycetaceae they develop in basipetal series from generative cells located at the distal region of branches. An unusual situation can be seen in some strains of Smittium culisetae wherein the production of trichospores is not restricted to a short series of generative cells at the tips of branches, as is normal. Rather, many of the thallial cells that normally would remain vegetative become fertile, and in extreme cases virtually the entire thallus becomes reproductive through production of spurious generative cells (Fig. 7.14). This development may be preceded by septation in some of the longer cells of the thallus. In some species of Smittium where cells remain longer than normal they can give rise to trichospores markedly larger than the norm (Fig. 7.14).

Appendages have not been seen, or are not produced, in 7 of the 33 genera of Harpellales: Carouxella, Bojamyces, Caudomyces, Ejectosporus, Gauthieromyces, Orphella, and Zygopolaris. In all other genera the number of appendages on trichospores varies, according to the genus, from one to six or seven. Two types of trichospores based on current electron microscopic studies of appendage formation (Moss and Lichtwardt, 1976) are illustrated in Fig. 7.15. Type A consists of one or several appendages that develop in a corresponding number of appendage sacs, which are invaginations of the plasmalemma. located in the central region of the distal end of the generative cell. In Type B, the appendages form adjacent to the generative cell wall. The common feature of all appendages that have been sufficiently studied is that they form outside of the plasmalemma, rather than within the generative cell protoplast. This process involves synthesis and transport of appendage material, or some precursor, from the cell by means of membrane-bound vesicles that fuse with the plasmalemma at specific sites and deposit appendage material to the outer side of the membrane where the appendages are constructed. Trichospores thus appear to be unique fungal propagules in terms of development and fine structure of the appendages. The chemical composition of appendages has not been determined. Descriptions and comparisons of appendage formation in several genera follow.

Stachylina grandispora (Harpellaceae) has been studied by Moss (1972, 1974, 1976). Each generative cell contains a nucleus that divides mitotically. One nucleus migrates into the developing trichospore, and the other remains in the generative cell, after which a perforate septum forms to separate the two cells. Appendage vesicles appear in the cytoplasm of the generative cell, and seem to originate in the perinuclear space and bleb off from the outer nuclear membrane. The vesicles have an electron opaque core and a less dense peripheral layer. They migrate to the septal region next to the plugged pore and fuse with the plasmalemma, the process eventually causing that membrane to become invaginated. As additional appendage material is deposited near the septum, a larger invagination develops to become the appendage sac containing a single appendage folded several times upon itself. Within the appendage sac is a matrix surrounding the ribbon-like appendage of electron-opaque amorphous material. The initially formed portion of the appendage at the base of the spore, however, is rod shaped. While the appendage is taking shape, the generative cell becomes progressively more vacuolated, and eventually all the cellular organelles appear to degenerate. At maturity the trichospore breaks loose, retaining a small collar of wall material below the septum, and at this moment the appendage unfurls and quickly leaves the generative cell as though forced from the inside. The appendage of the released spore provides no motility. It is about 0.3-0.4 µm wide near the base of the spore and 200-500 µm long, although the length is often difficult to determine because it tapers toward the terminal end and cannot always be resolved optically with the light microscope.

The actual process of appendage formation has not been studied in the genus Smittium, but the mature appendage in both S. mucronatum (Manier and Coste-Mathiez, 1968) and S. culicis (Moss and Lichtwardt, 1976) forms within a large appendage sac that fills the outward extension of the generative cell beneath the trichospore that later becomes the spore collar. Transverse sections of appendages in both of these species appears in electron micrographs as a series of concentric electron-opaque rings. This unusual structure was not found by Preisner (1973) in either S. culisetae or S. simulii, however, and there may be variation in appendage substructure within the genus.

A single appendage is formed on trichospores of Smittium spp., but in Trichozygosporachironomidarum there are about 5-7 appendages, each produced in a separate appendage sac located primarily in the collar region but extending into the main body of the generative cell (Moss and Lichtwardt, 1976). Trichozygospora appendages consist of an amorphous electron-opaque material surrounded by a lighter matrix that fills the appendage sac, and, as in Smittium spp., the appendages are twisted about within the sac.

A study of Genistellospora homothallica trichospores by Moss and Lichtwardt (1976) revealed yet another modification of appendage formation (Fig. 7.16). Appendages begin to form as soon as trichospore growth commences. As the developing spore begins to bulge out of the generative cell, membrane-bound vesicles are seen in the apical and subapical regions of the generative cell and appear to be involved both in deposition of wall material in the enlarging spore as well as in formation of appendages at its base. These vesicles originate from areas of convoluted endoplasmic reticulum. Approximately six appendages develop; they initiate adjacent to the generative cell wall in invaginations of the plasmalemma, and continue their development straight downward. The appendage sacs enlarge, and in transverse section a matrix is seen filling the appendage sac and completely surrounding each of the semicircular to lunate appendages.

The most complex appendage substructure occurs in Harpella melusinae (Reichle and Lichtwardt, 1972). In longitudinal section the appendage has an alternating light- and dark-banded structure (Fig. 7.17). The lighter bands are approximately equal in width, but the dark bands are alternately wider and narrower. The periodicity of the banding is about 136 nm. Light microscopic observations (Lichtwardt, 1967) indicate that there may be a 10-fold longitudinal expansion of the appendages after release from the generative cell. Four appendages normally form at the base of each H. melusinae spore, and these are seen more or less tightly spiraled against the inner generative cell wall and outside the plasmalemma, (Fig. 7.17). Active fusion of vesicles with the plasmalemma can be seen at certain stages along the length of the spiral, but the origin of the forming vesicles has not been determined. Nor is it known how these structurally complex appendages are constructed after vesicles fuse with the membrane.

In some genera (e.g., Legeriomyces, Genistelloides) appendages begin to form well before the trichospore starts to grow out from the generative cell. Thus, the time at which appendages begin to develop seems to depend on the particular genus or species.

Terminal trichospores in Genistellospora homothallica arise by holoblastic outgrowth from the generative cell; that is, the wall of the developing trichospore is an extension of the two wall layers of the generative cell (Moss and Lichtwardt, 1976). Subterminal trichospores develop indirectly from their generative cells: First, an outgrowth develops laterally from the terminal end of the generative cell by production of new wall layers laid down internally, in the same manner in which branches develop in vegetative hyphae of this species. Then, after a short (1-4 µm) extension of this branch, the trichospore develops by holoblastic outgrowth, as in the terminal trichospores. A septum develops at the base of the trichospore, and this is followed by the manufacture within the original trichospore wall of a new wall around the protoplast consisting of two layers of wall material similar to those of vegetative hyphae and generative cells. In young spores the new inner wall is appressed to the outer wall, but the two wall layers are not fused, and sometimes can be seen physically separated. Thus, the trichospore can be interpreted as a monosporous sporangium consisting of a sporangial wall with an inner, separate sporangiospore wall surrounding the spore proper.

Two unusual internal features of trichospores characterize several genera of Harpellales whose mature trichospores have been studied by electron microscopy (Manier, 1973a; Preisner, 1973; Moss and Lichtwardt, 1976). The apex of the inner wall contains an annulus consisting of a thickened structure with several canals that traverse both inner wall layers (Fig. 7.18). Immediately behind this perforated thickening and extending almost to the nucleus are several (1-10) elongate or spherical membrane-bound "apical spore bodies." These undoubtedly correspond to the refringent bodies seen with the light microscope in trichospores of many genera of Harpellales. In Genistellospora homothallica, numerous membrane-bound vesicles 30-69 nm in diameter have been found located between and near the apical spore bodies and the annulus. The membranes of these vesicles are sometimes continuous with the membrane of the apical spore bodies or with the plasmalemma lining the annulus. The implication is that vesicles carry material from the apical spore bodies and pass it through the plasmalemma to the canals of the annulus. In Legeriomyces ramosus, considerable secreted material appears to accumulate between the annulus and the outer trichospore wall (Manier, 1973a). Manier, and also Moss and Lichtwardt (1976), have suggested that this apparatus is involved in release of the sporangiospore from the sporangial wall, or attachment of the spore to the gut lining of the host, or both. Extruded sporangiospores of G. homothallica and L. ramosus have been seen only rarely; therefore electron microscopic studies of these spores during extrusion have not been feasible. Only future fine-structural studies will indicate whether apical spore bodies and annuli are found in all trichospores.

The appendages at the base of released trichospores appear to have a passive but possibly important function in transmission. They have been seen to entangle with various external objects as well as host fecal material (Fig. 7.19), and it is believed that this helps to prevent spores from floating or washing away, in this way keeping many of them on the substrate of their hosts and increasing their chances of being ingested. Most species of Harpellales grow in insect larvae that inhabit streams with fast-flowing water, and a mechanism to restrain spores from floating downstream may have been critical in the evolution of these fungi. In the case of still waters, where mosquitoes and bloodworms may be found for instance, the entanglement of spores in fecal matter may allow them to settle more rapidly or remain at the bottom where such hosts usually feed.

Species of two monotypic genera apparently have no appendages associated with their trichospores. Zygopolaris ephemeridarum (Legeriomycetaceae) trichospores remain attached more tenaciously to their generative cells than do most Harpellales. Occasional loose trichospores have been seen with a small blob of material at the base of the spore that might represent incompletely formed appendage material (Moss et al., 1975), and Moss and Lichtwardt (1976) reported finding in their electron microscopic study of this species a rudimentary type of appendage structure at the base of the spore still attached to its generative cell. It thus appears that in Z. ephemeridarum the ability to form discrete appendages may have been lost, or at least they are not completely developed in many instances. Tufts of trichospores on the brushlike end of the thallus of this species often project from the mayfly nymph anus, leading Moss et al. (1975) to suggest that one method of transmission may be by direct ingestion of trichospores by another nymph, without the spores being first released. Carouxella scalaris from ceratopogonid larvae is a curious member of the Harpellaceae in that it has no appendaged trichospore and an unusual type of dispersal. The spores do not detach from their generative cells; rather, the generative cells break apart from each other at maturity and the two structures are disseminated as a unit, referred to as a "diaspore" by Manier et al. (1961). According to these authors, the generative cell on reestablishment in a host produces the holdfast, after which the trichospore, still attached, develops into a new thallus. Disarticulation and detachment of generative cells has not been seen in Carouxella coemeteriensis, nor have trichospores been seen to detach from their generative cells.

Sporangiospores normally extrude from trichospores only after being ingested. In Ejectosporus extrusion of the sporangiospore occurs on slides prepared from thalli removed from the gut while trichospores are still attached to their generative cells. Occasional spores of several other genera have been seen to germinate in water on microscope slides. In some cases this may be due to changes in osmotic conditions or other artificial factors. Extrusion typically involves an extension of the inner wall resulting in the spore body breaking through a tear in the sporangial wall. This process is described in more detail in Chapter 9 in connection with studies on cultured species of Smittium.

ASELLARIALES - The branched thalli of this order reproduce asexually by cells that disarticulate like arthrospores. Immature thalli are septate, the septa being the perforate harpellid type discussed earlier in this chapter (Manier, 1973b; Moss, 1975; Moss and Young, 1978). At this stage the cells typically are relatively long in most species, and contain one to several nuclei. On maturation, and possibly in response to onset of molting in the host (Poisson, 1932a; Manier, 1958, 1964c), nuclei may begin to divide in longer cells, and secondary crosswalls develop to delimit uninucleate segments of more or less uniform length. It is not known whether the secondary septa are perforate or consist of nonperforate walls. This process is essentially basipetal, and in some species (e.g., Asellaria ligiae, Orchesellaria lattesi) the complete thallus may disarticulate, leaving only the specialized holdfast cell intact. As a consequence, the molt of the arthropod may contain large numbers of arthrospores, or, if these have already disseminated, only the basal cells remain still firmly attached to the cuticle.

On first consideration this type of asexual reproduction would appear to be quite distinct from that found in other orders of trichomycetes, were it not for the observations (Manier, 1963a, 1969b; Lichtwardt, 1973a) that the arthrospores of Asellaria ligiae, when kept in water, "germinate" to produce a lateral cell whose size and shape resembles that of a trichospore (see Fig. 11.26). Lichtwardt (1973a) called attention to the remarkable similarity of this "germinated" arthrospore to the trichospore-generative cell unit of Carouxella scalaris (Harpellaceae) (Fig. 11.3) referred to previously. Unfortunately, material has not been available for an electron microscopic study to ascertain whether in A. ligiae (or in C. scalaris for that matter) the trichospore-like bodies have the same double wall and internal structures that characterize the typical trichospore of the Harpellales.

Arthrospores of other Asellariales have not yet been seen to develop in this manner. Poisson (1932a) kept arthrospores of Asellaria cauleryi in water, but saw no development similar to that found later in A. ligiae. Rather, he stated that only on ingestion by its isopod host did the arthrospore of A. cauleryi germinate, with one end of the arthrospore giving rise to the specialized holdfast cell while the other grew into the highly branched thallus. Manier (1958, 1964c, 1979b) has described a similar direct germination of ingested arthrospores in several species of Orchesellaria.

ECCRINALES - Species of this order reproduce asexually by producing one or more kinds of sporangiospores. These typically develop singly in sporangia that form basipetally. Two basic types can be recognized. One is the primary infestation sporangiospore (Fig. 7.20). It is uninucleate in most species and bi- or quadrinucleate in others, often thick walled (and presumably "resistant"), and is believed to germinate only after passage from the gut and ingestion by another suitable host. It is a propagule for transmitting infestation from one host to another. The other type is the secondary infestation sporangiospore (Fig. 7.20). This type is multinucleate and capable of immediate germination on release from the sporangium, thereby serving to increase the population of thalli in the already infested gut without requiring the ingestion of additional spores. The two basic spore types can be recognized in most instances by their structure. As used in this treatise, any eccrinid spore that germinates immediately within the gut where it was produced, regardless of its morphology, is considered to be a secondary infestation spore. Reproduction in the Eccrinales is complicated by the fact that several morphological variants of both types exist; furthermore, there are cells produced in some species whose function is not known and which may in fact be nonfunctional. Several of these variants will be described later in this section.

First, let us consider the typical secondary infestation spore which is capable of immediate germination, for in the majority of eccrinid species these are the most commonly encountered reproductive unit during the intermolt period. Prior to their formation in the coenocytic thallus the terminal nuclei become aligned more or less equidistantly, and a septum forms to delimit the most terminal nucleus. This is followed by the production of successive septa in such a way that a series of initially uninucleate cells is formed. While this process goes on, the more terminal nuclei begin to divide mitotically and in synchrony within each cell so that the cells or sporangia usually end up with 4 or 8 nuclei aligned roughly in a row. The number is relatively constant in any given species, but may be as few as 2 (e.g., Parataeniella spp.)or as many as 16 (e.g., Enteromyces callianassae). Within each sporangium a new wall is laid down around the protoplast to form a single sporangiospore.

The mature spore usually escapes from the sporangium through a hole that develops in the lateral wall of the sporangium at either the distal or proximal end. Presumably the hole is produced by enzymatic dissolution or weakening of the wall material by the spore. The distal or proximal position of the exit hole on the lateral wall is generally constant for a species, but in some thalli it can be reversed. Invariably, however, it is the basal end of the spore-the end that will attach to the cuticle-that emerges first. This can be identified in most species by a slight swelling at that end of the spore or, more rarely, by a rudimentary structure resembling a small holdfast. In some cases spore emergence may entail a tearing of the sporangial wall around the exit hole. After release of the spore the sporangial wall disintegrates, possibly by bacterial decomposition (Fig. 7.21). (Empty sporangia, in time, almost consistently have bacteria attached to their walls, sometimes in large numbers. It is interesting that bacteria of similar appearance often are seen attached to the walls of immature sporangia and nonreproductive parts of the thallus, but they do not appear to inhibit development of the eccrinid.) Occasionally it is possible to find all of the sequential stages, from sporangial delimitation to spore release, at one time in individual thalli of some species.

The released multinucleate spore attaches immediately to the host cuticle and begins to germinate. In a few eccrinid species, release of several of the spores can occur within a short space of time with the result that a row of attached spores or developing thalli from one mother thallus can be seen on the host cuticle (e.g., Arundinula spp.). Germination commences by extension of the wall at the base of the spore immediately above the holdfast so that the spore maintains a terminal, usually reflexed, position on the elongating thallus, changing little in shape or size as the thallus grows. The fate of the spore mother-cell follows one of several patterns: (1) it remains in its apical position only during the initial stages of thallus elongation; a wall then forms to cut off the spore case, and it disintegrates and disappears (e.g., Enterobryus spp., Fig. 7.21); (2) it is persistent through sporulation of the thallus, eventually deteriorating as the sporangiospores mature (e.g., Palavascia spp.); or (3) it is persistent and eventually functions as the terminal sporangium (e.g., Enteromyces callianassae). These patterns are more or less uniform in any given genus, and often are useful taxonomic characters. What is presently called the spore mother-cell has been referred to in the literature under various other names: gland, spore case, mother spore, appendage, etc.

The primary infestation spores, which serve as dissemination propagules, are in most instances uninucleate on release and are thick walled in many, but not all, species. Their development within thalli is similar to that just described in the secondary spore type, except that mitotic divisions do not occur in the young sporangium. Frequently the spores assume an oval or ellipsoidal shape on reaching maturity. These spores form only in response to molting in some hosts, as evidenced by the fact that mature ones are found only during ecdysis. The shed molt may contain large numbers of primary infestation spores in such cases.

In some genera (Astreptonema, Eccrinoides) the developing primary spores do not, in fact, remain uninucleate, but may be quadrinucleate at the time of release. Or, as in Eccrinidus flexilis, there is a further development after release to produce two chambers within the spore, each containing a quadrinucleate protoplast. In Enterobryus, as this genus is currently conceived, the primary infestation spores remain uninucleate, but are not thick walled and do not seem to form in response to ecdysis. These examples illustrate that several forms of primary spore exist.

The above outline of the two basic spore types (primary and secondary) is based on the largest of the three eccrinid families, the Eccrinaceae. In the Palavasciaceae, only primary (but multinucleate) infestation spores are formed. In the Parataeniellaceae both types occur; the secondary (binucleate) type is produced individually in sporangia, whereas the primary (uninucleate) type develops in larger numbers within a thallus that functions as one multispored sporangium. In the Parataeniellaceae there appear to be no septa separating the primary spores even in early stages of cleavage, but it should be pointed out that this unusual multispored type of eccrinid sporangium has not been studied at the electron microscope level of magnification. In several Eccrinaceae the primary infestation spores clearly develop individually within sporangia; however, at maturity the crosswalls separating the spores may disappear, leaving the spores loose and scattered in the thallus tip until they are released from the apical end or the remaining thallus wall material disintegrates around them.

In most species of Eccrinales much of the thallus remains nonreproductive. In a few (e.g., Enteromyces callianassae) only a short basal portion of the thallus may remain unconverted to sporangia at the end of the reproductive process. Complete conversion of the coenocytic thallus to spores is found in the Parataeniellaceae (primary spore type only), as mentioned above, but apparently no complete conversion is found in the Palavasciaceae and only rarely in the Eccrinaceae. [In Arundinula washingtoniensis, Hibbits (1978) reported finding the entire protoplast cleaved into uninucleate spores, and she saw a similar cleavage (but no spores) in A. hapalogaster on one occasion.] In almost all sporulating thalli of Eccrinaceae, one finds either the primary or the secondary spore type, but occasionally thalli that have begun to form one type may shift to the other (e.g., Enterobryus borariae, Parataeniella spp.), resulting in a thallus bearing both types simultaneously, as in Fig. 7.12.

An interesting structural feature has been found in the thick walled, oval, primary infestation spores of several genera of Eccrinaceae, namely that on release some have appendages at one or both ends. These are all associated with molting marine or freshwater Crustacea. One may assume that the appendages serve to entangle the spores in aquatic substrate materials, thus increasing their chances of successful transmission. At least this may be true in those eccrinid spores where the appendages are long and filiform, as in appendages of harpellid trichospores. Eccrinid primary spores with appendages were first reported in Arundinula capitata from hermit crabs by Duboscq et al. (1948), who found them in only 3 of 150 molting crabs. They were described and illustrated as being uni- or binucleate with one short appendage emanating from each pole of the oval spore. Duboscq et al. thought the spores developed from a fusion of uninucleate cells functioning as gametes within the thallus. These spores in A. capitata have not subsequently been found for restudy. Hibbits (1978) saw similar spores with one long appendage at each pole in Arundinula washingtoniensis; however, she did not consider them to have arisen by a sexual process, nor has sexuality been suggested in the formation of appendaged spores in other species.

Primary spores with a single appendage at each pole are now known in Astreptonema gammari (Fig. 11.30) and A. typica (Manier, 1964b), and Hibbits (1978) described an unnamed species of Astreptonema from Exosphaeroma amplicauda that had a single appendage at one pole. (She also found a single appendage on a spore of Alacrinella sanjuanensis, but stated that it might have been an artifact of fixation.) In Taeniella carcini, Moss (1979) discovered that there are two appendages at each pole of the oval primary spores, and Hibbits (1978) found two attached to only one pole in T. grandis. Species of Taeniella often have gelatinous masses at the ends of the spore in the sporangium, and it is probable that in species of this genus, and in others as well, additional appendaged primary spores will be found on further study. Appendages unfurl only after release from the sporangium. They are often difficult to see without phase-contrast microscopy or other special techniques, and probably have been overlooked in some cases.

The significant question is whether appendaged spores of Eccrinales and Harpellales have evolved separately through convergent evolution, or whether they are in fact structurally and ontogenetically similar and phylogenetically related (see Chapter 12). At the present time little is known about the development of eccrinid appendages. Moss (1972, 1979) found little similarity between these structures in the two orders based on his investigation of the eccrinid Astreptonema gammari. He showed that in this species the appendages are extensions of an outer, mucilaginous sporangiospore wall formed within the sporangium by deposition of Golgi derived material. Further electron microscopic studies of this kind of spore in the Eccrinales would be most valuable in order to compare their development with the better known trichospore of the Harpellales.

A different form of spore appendage was briefly reported by Lichtwardt (1973a) in Palavascia sphaeromae, and is described here in more detail (Fig. 7.22). This widespread species of Eccrinales occurs in the hindgut of marine isopods. A collection of Sphaeroma serratum from the Bassin de Thau near Sète, southern France in July of 1968 was well infested with the eccrinid. When thalli bearing mature primary infestation spores were placed on a microscope slide in 50% seawater and gentle pressure was applied to the coverslip to release spores from their sporangia, the spores were seen in phase contrast to bear two adjacent short, hyaline, hornlike gelatinous appendages, each with a small terminal knob, arising from an inconspicuous flat pad on the outer surface of the spore. Most commonly they were attached to the side of the oval spore, but they also occurred in pairs on other parts of the spore. These sometimes adhered to the glass slide. Within an hour or less, many of the released spores elongated and burst forth from a thin, elastic, outer wall that surrounded the spore, and the outer wall immediately collapsed in a pleated fashion. At the base of each of these ejected spores, and attached to the collapsed outer wall, was seen a hyaline body of similar appearance but larger than the outer appendages, possibly consisting of a hygroscopic substance that had swollen to aid in ejecting the spore. The hyaline body and the spore did not remain attached to each other. This process was observed many times on slides prepared from different isopods, and undoubtedly was not an artifact.

Whether spores of all P. sphaeromae populations regularly behave in this fashion in vitro has not been determined. It is possible that the two hornlike appendages serve to cause the spore to adhere to the substratum until ingested, and that the sudden shedding of the outer wall in natural circumstances occurs only after ingestion and just preceding attachment of the spore to the host cuticle. Manier (1979a) did a study of the fine structure of P. sphaeromae collected from the same general site and detected two distinct walls in the sporangiospores within the sporangium, as expected, but she did not see fine-structural evidence in her sections for the appendages on the outer wall of the unreleased spores.

The terminology one finds in the literature in reference to the two basic spore types in the Eccrinales is extensive and confusing. At one time or another they have been called conidia, endoconidia, oidia, etc., as well as sporangiospores, which of course they are. Most often the prefixes "macro" and "micro" have been used to differentiate between the secondary infestation spores and the primary ones, respectively. However, a size difference is not consistent in all species. The thicker walled primary spores have also been called "resistant spores" or "durable spores" by the French workers, and while these designations are descriptive for particular genera of eccrinids, they are not applicable in a general way to all primary infestation spores. In past publications we have distinguished between the two types by calling them multinucleate and uninucleate sporangiospores, but there are known exceptions to the uninucleate condition in the latter type because of either species differences or the stage of development seen after spore formation. In her 1969 monograph, Manier (1969b) called the secondary infestation sporangiospores "immediate-development spores" and the primary infestation sporangiospores "dissemination spores," thus placing their terminology on a functional basis, which seems preferable in view of the structural variation among the taxa. Despite some evidence to the contrary (Hibbits, 1978), there are no convincing data at present to indicate that the function of each of the two basic spore types is not distinct.

Another complicating factor in studies of reproduction in the Eccrinales, particularly in the Eccrinaceae, is the presence in thalli of some species of cells whose function is not known and which, in fact, may be nonfunctional. It is not uncommon in some Eccrinaceae to find series of cells at the tips of thalli that resemble sporangia, but do not appear to produce spores [Hibbits (1978) called these "partitioned" thalli]. In Enterobryus spp., such cells may be larger than normal sporangia and are sometimes slightly swollen, with nuclei that are not arranged in a row as is typical in true sporangia; or they remain uninucleate but are morphologically different from the true uninucleate primary infestation spores. It is surprising that Leidy (1853), who was the first to study eccrinids and a competent microscopist, never found spores in Enterobryus elegans and other species he described. He only saw what he called "secondary cells" in six instances out of several thousand thalli from over 150 millipedes collected from early spring until late fall. In Eccrina longa (nom. nud.) he found a few such cells detached singly or in groups of two or three, and assumed that reproduction in the eccrinids was by segmentation. Since Leidy never identified nuclei in these fungi, it is not known for certain whether they were a modified type of sporangium or immature ones. The disarticulation he reported has not been confirmed in subsequent studies of some of his same species (Lichtwardt, 1954b, 1957a,b), but one cannot discount this type of occurrence. It has been said to happen in other eccrinids, such as Eccrinoides (Eccrinopsis) helleriae (nom. nud.) (Duboscq et al., 1948) and Trichellopsis schizophylli (nom. dub.) (Maessen, 1955).

An unusual cell type that has been seen occasionally in a number of species of Enterobryus (Lichtwardt, 1954b, 1958, and unpublished) consists of apical cells tightly packed with dozens of nuclei. Thalli that produce them are quite distinct, even before such cells have formed, for they contain more numerous nuclei than are normally found, and these are scattered randomly throughout the thallus. In time, some of the nuclei become densely aggregated at the tip of the thallus, and a septum then forms to isolate them. This process continues until up to about half a dozen such cells are formed. What may be nuclear divisions in some of these terminal cells have been observed on rare occasions, and, judging from observations of the static condition but involving sequentially produced events in cells connected to both ends of those with dividing nuclei, the divisions appear to result in nuclei with about one-half their previous volume. The function of these cells, if any, is not known. Thalli bearing this cell type can be found interspersed with other thalli reproducing normally, and consequently it does not seem that their development is necessarily affected by some unusual environmental factor in the gut.

Another spore type whose function is unknown is frequently found in both species of Palavascia. The reproductive unit in the genus, as described previously in this section, is a multinucleate primary infestation sporangiospore. However, some of the many sporangia germinate in situ, putting out one or more long, narrow thalli ("microthalli") that resemble a small branch. Manier (1979a) has clearly shown in P. sphaeromae that the walls of these narrow thalli are extensions of the spore wall contained within the sporangial wall and are not outgrowths of the outer, sporangial wall itself. After breaking through the outer wall, the narrow thalli elongate considerably and eventually divide into a series of small uninucleate cells. Spores that germinate in situ in this manner do not normally emerge from the sporangium, and the uninucleate cells they produce appear to have no function.

Other modified spores of Eccrinales are described in Chapter 11 in the descriptions of individual species.

AMOEBIDIALES - In Amoebidium parasiticum the entire protoplast cleaves to produce uninucleate sporangiospores or, on a molting or injured host, uninucleate amoeboid cells. Whisler and Fuller (1968) found no significant difference in development of the two cell types, except that sporangiospores become enveloped in a rigid wall whereas amoebae are merely membrane bound. There is a question remaining as to whether the thallus functions as a single sporangium, or whether spores form in separate, walled chambers within the thallus. The authors on many occasions have studied A. parasiticum under the phase-contrast microscope and frequently, but not always, has seen what appear to be walls or partitions within emptied thalli. In her 1970 thesis, Coste-Mathiez stated that the empty walls of thalli often show an internal partitioning as though the sporangiospores are isolated individually or in groups within chambers, and two of her electron micrographs of a sporulating thallus (her Plate II, Figs. 11 and 12) clearly show a separate wall within the thallial wall surrounding the spore walls. On the other hand, electron micrographs by Whisler and Fuller (1968) of cultured A. parasiticum, and by Dang (1979) of the same species occurring naturally on mosquito larvae, show no internal partitioning of the thallus around the spores. It is possible that the formation of chambers may or may not occur, depending on factors such as the environment during development, thallial size, and possibly strain differences.

It seems clear, however, that amoeba production in both A. parasiticum and Paramoebidium spp. does not involve internal wall formation, so that the entire thallus functions as a single reproductive structure. This phase of the life cycle is similar but not identical in both genera. The amoebae of the two genera are essentially indistinguishable. (On a few occasions the authors have seen thalli of Paramoebidium curvum and other undetermined species of Paramoebidium release unusually large amoeboid cells that appear to be multinucleate, along with normal amoebae, apparently due to abnormal cleavage of the protoplast.) Manier and Raibaut (1970) observed that the cystospores of A. parasiticum emerge from the cyst wall as a compact cluster within a membranous, partitioned sac from which they later are released. Lichtwardt (1973a) found no such membranous sac surrounding cystospores of P. curvum; rather, the cystospores are ejected forcibly from individual chambers within the cyst wall. These differences in cyst development, if indeed they exist, may be attributable to the different functions of the cysts: Amoebidium cystospores are released in the external environment where they can eventually attach promiscuously to the external cuticle of various arthropods, whereas it is believed the cysts of Paramoebidium are ingested and may release their cystospores suddenly and only when they reach the proper location in the arthropod hindgut (Dang and Lichtwardt, 1979).
 

Sexual Reproduction

Of the four trichomycete orders, only the Harpellales produce structures (zygospores) that appear to arise from a true sexual process. There is no cytological or genetic proof of sexuality currently available. Nevertheless, the sequence of zygospore development, beginning with thallial conjugation, and the thickened walls and storage materials contained within the spore, leave little doubt that they are comparable to zygospores of other fungi, although structurally and ontogenetically different. No zygospores have been found yet in the putative closely related order Asellariales, but observed conjugations identical to those seen in the Harpellales suggest that in time zygospores will be discovered in that order too (Lichtwardt, 1973a). Poisson (1931b) reported observing fusion of amoeboid cells in Amoebidium parasiticum (Amoebidiales), but this has not been verified by subsequent investigators, as mentioned in the first section of this chapter. In the Eccrinales a peculiar form of thallial conjugation and fusion has been seen in several instances in crustaceans that is suggestive of an attempt at sexuality, and this will be described following an account of zygospore development in the Harpellales.

HARPELLALES - Zygospores are known in 24 of the 33 genera of Harpellales, but not in all species of two of these genera. Zygospores may be unknown in some species due to insufficient investigations, or the host specimens may have been collected at the wrong season or stage of maturity. In other more studied species zygospores appear in fact to be rare. For instance, Stachylina grandispora has been seen conjugating but never producing zygospores (Lichtwardt, 1972; Moss, 1972). Conjugations are frequent at times in some populations of blackfly larvae containing Harpella melusinae, but zygospores have been found in only some seven larvae out of thousands dissected from different climatic regions (Lichtwardt, 1967; Moss and Lichtwardt, 1977; and unpublished). It may be that in H. melusinae there is frequently some impotence factor present involving an inhibition or loss of genetic control over the sequence of events necessary for completion of the sexual process. Other possibilities, not mutually exclusive, are that larvae may sometimes develop too rapidly for zygospores to form prior to molting, or that environmental factors within the gut lumen are not favorable for complete sexual development. Fortunately, zygospores are not difficult to find in many other harpellids.

It is clear that in some species of Harpellales zygospores are associated only with the prepupal molting process. Mature aquatic insect larvae about to molt can often be identified externally by the appearance of pupal structures beneath the larval skin; internally, the gut linings become loosened as the epithelial cells lay down new cuticle material prior to ecdysis. Since the last larval stage is normally the longest, it is possible that the development of zygospores just before pupation is due primarily to temporal factors. However, it seems evident in some harpellid species (e.g., Legeriomyces ramosus)that there is a rather abrupt morphogenetic shift from asexual to sexual reproduction, with the result that the exuviae may contain for the most part masses of zygospores and no remaining trichospores. In other species (e.g., Genistelloides hibernus), when most trichospores are mature but while some are still maturing, clumps of thallial branches begin to conjugate and produce zygospores. In a few species (e.g., Pennella simulii, Trichozygospora chironomidarum, Genistellospora homothallica), trichospore and zygospore development can be concurrent. Nevertheless, sexual reproduction in the majority of Harpellales does seem to be influenced by the molting processes of the host or at least is correlated with ecdysis.

Currently there is no understanding of the stimuli that effect the morphogenetic change in these gut fungi. Is it due to the direct influence of molting hormones or other chemicals activated during the molting process, or do these fungi sense less directly the physiological changes occurring in the host? The only practical means of resolving this intriguing biological problem is through the use of axenically cultured fungus material, but unfortunately at present zygospores have not been induced to form in any of the cultured harpellid species.

The onset of sexuality in some species of both families of Harpellales is marked by uneven swellings in some of the cells of the thalli. This is especially noticeable in the unbranched thalli of Harpella melusinae, where this initial stage of sexuality can be readily apparent at low magnification in well-infested peritrophic membranes, even before actual conjugations are detected at higher magnification. This suggests that there may be some kind of hormonal interplay that makes certain regions of the thalli receptive to adhesion and fusion. It should be noted that these gut fungi, unlike other sexually reproducing fungi on more stable substrates, are frequently in motion due to peristalsis and the movement of host-ingested particles through the gut. Also, when the gut is dissected and cleansed of debris, the positions of the thalli are considerably disturbed. Consequently, it is difficult to determine whether the swellings on adjacent simple thalli (or branches in Legeriomycetaceae) can develop when they are still separated but in close proximity, or only when in actual contact. Both partners on contact produce a short protuberance contributing to the formation of the conjugation tube, and the walls between them quickly dissolve completely and allow the cytoplasm to fuse. The nuclei from each of the uninucleate mating cells have been seen juxtapositioned in the conjugation area, and single, presumably diploid, nuclei have been seen (Lichtwardt, 1967), but actual karyogamy has not been observed.

Multiple conjugations between two branched or two unbranched thalli may produce a scalariform pattern, but conjugations also may be promiscuous and involve many branches or thalli. Lichtwardt (1967) reported as many as seven unbranched thalli of H. melusinae interconjugated. Analysis of such complexes in H. melusinae did not give any indication of a simple mating system, such as the plus-minus type characteristically found in heterothallic Zygomycetes. Some branched thalli of harpellids (e.g., Pennella spp.)appear to mate only with other thalli, and therefore may be truly heterothallic, but it is often too difficult to trace the intermingled branches to determine accurately their thallial source (Williams and Lichtwardt, 1971; Lichtwardt, 1972). Stipella vigilans is reported to be either "homothallic" or "heterothallic" (Manier, 1963b). On the other hand, Genistellospora homothallica differs in that it forms its zygospores without any conjugations.

Unlike the spherical zygospores of other Zygomycota, those of the Harpellales are biconical at maturity (although spherical very early in their formation) and each develops as an outgrowth from a supporting cell, the zygosporophore. Moss et al. (1975) recognized four basic types (Fig. 7.23):

Type I. The axis of the zygospore lies perpendicular to the zygosporophore, and its attachment is median (midway between the poles). This type is found in Allantomyces, Harpella, Genistelloides, Simuliomyces, Spartiella, Stachylina, and Stipella.

Type II. The zygospore position is oblique to the zygosporophore, and its attachment is submedian. Found in Austrosmittium, Capniomyces, Furculomyces, Glotzia, Graminella, Harpellomyces, Legerioides, Legeriomyces, Legeriosimilis, Smittium, and Trichozygospora. In these genera, released zygospores have a collar and, in some genera, one or many appendages.

Type III. The zygospore lies parallel to the main axis of the zygosporophore, and attachment is median. Found in Genistellospora and Pennella.

Type IV. Attachment is at one pole, so that the zygospore and zygosporophore are coaxial. Found in Carouxella, Lancisporomyces, Plecopteromyces, and Zygopolaris.

The genera whose zygospores presently are unknown are Bojamyces, Caudomyces, Coleopteromyces, Ejectosporus,Gauthieromyces, Graminelloides, Orphella, Pteromaktron, and Stachylinoides.

The location of the developed zygospore and its subtending zygosporophore in relation to the zone of conjugation is fairly predictable. In Capniomyces, Glotzia, and Genistelloides these usually arise directly from the conjugation tube. In other genera these structure arise from one of the conjugated cells, and usually close to the zone of conjugation except in Pennella, Simuliomyces, and Trichozygospora. Where the conjugations are scalariform, the zygospores normally all arise from the cells of one of the thalli or branches. In a few genera (Pennella, Stipella, Trichozygospora, sometimes Zygopolaris and others), the fertile cells from which the zygospores arise may become noticeably swollen, whereas the "donor" cells retain their shape. Conjugations may occur between terminal cells and intercalary cells, or (more commonly) two intercalary cells. Williams and Lichtwardt (1971) noted that in Pennella simulii donor branches could conjugate with more than one receptor branch, and sometimes a donor branch terminated in the production of an apical trichospore.

Zygospore development in H. melusinae (Type I) begins with the budding out of the zygosporophore from one of the conjugated cells (Lichtwardt,1967). Its two-layered wall originates from new wall layers formed internal to those of the cell from which it grows, in a manner similar to lateral branch formation in the Legeriomycetaceae (Moss and Lichtwardt, 1976, 1977). After the zygosporophore achieves its normal size, it bulges spherically at the top and then the biconical zygospore takes shape. At about this time two crosswalls form, one separating the zygospore from the zygosporophore and the other separating the zygosporophore from the conjugated cell. Both are the typical harpellid perforate septa that become occluded by an electron-opaque plug. An especially interesting feature of the entire sexual apparatus in H. melusinae is that at maturity each of the four cells-the two conjugants, the zygosporophore and the zygospore-contains a single nucleus (Moss and Lichtwardt, 1977) (Fig. 7.9). This may indicate that a meiotic division occurs prior to septation and, if true, that the zygospore contains a haploid nucleus. Such a condition would be advantageous, assuming somatic nuclei are haploid, for it would allow the zygospore to be primed for rapid germination, attachment, and thallial development after ingestion by a host, as in trichospores.

In Furculomyces spp., two branches (conjugants) within one thallus develop, usually near the base of the thallus, fuse at their tips, and produce a short zygosporophore from which a bent zygospore (Type II) grows. The conjugants have the appearance of a wishbone (furcula), for which the genus is named (Fig. 11.48, Fig. 11.63, Fig. 11.103).

In Trichozygospora chironomidarum (Type II), the sexual apparatus is somewhat modified. The zygospore is produced some distance from the zone of conjugation and develops at the end of an extended conjugant. The zygosporophore in this species can be identified as a swelling beneath the zygospore, but it is merely a continuation of the conjugant and is not separated from it by a septum as in other harpellids. Moss and Lichtwardt (1977) found in one of their specimens that three of the four nuclei in the sexual apparatus remained in the larger conjugant cell, with the fourth nucleus being located in the zygospore. The numerous appendages (10 or more) in this species develop within the zygosporophore and extend down into the conjugant cell. These appendages do not form in individual appendage sacs, as do the fewer appendages (5-7) of the trichospores of this species. On detachment, a portion of the zygosporophore wall remains with the zygospore to form a collar from within which the numerous appendages trail. In Genistellospora homothallica (Type III) and Zygopolaris ephemeridarum (Type IV), the zygospores contain a single nucleus when first formed, but Moss and Lichtwardt (1977) did not find any nuclei in the zygosporophores of the limited material available for study.

The most distinctive zygospores are Type IV, which are attached at one pole, and therefore pointed at only one end (Fig. 7.23). Of these, two have unique shapes: Lancisporomyces are lance-shaped with a very long extension below the zygospore proper (Fig. 11.161), and zygospores of Plecopteromyces are short and turbinate (Fig. 11.53).

There appear to be several patterns of zygospore release. This process has not been described for all species, and in some cases only a few zygospores have been seen to detach from thalli when kept in water on microscope slides. Therefore, one cannot always be certain that the observed release is not premature or otherwise abnormal. In Genistelloides hibernus and Simuliomyces microspora (Type I), detachment usually takes place at the base of the zygosporophore, so that the zygosporophore and the zygospore release as a unit. All Type II zygospores have a small collar when released due to a circumscissile dehiscence of the zygosporophore wall. In Genistellospora homothallica (Type III), the zygospore is released with the contents of the zygosporophore attached, leaving behind the zygosporophore wall in place on the thallus (Lichtwardt, 1972; Moss and Lichtwardt, 1977). Zygospores of Zygopolaris ephemeridarum (Type IV) have been seen to retain some of the zygosporophore material at their base (Moss and Lichtwardt, 1977), and to produce a skirtlike substance around the zygospore base (Lichtwardt and Williams, 1984).

Before normal release, zygospores begin to develop internal thickenings at their conical ends, and as the zygospores mature the protoplast becomes progressively restricted. In the case of Type IV zygospores the thickenings develop only in the free end of the spores. Completion of this thickening process has been seen to occur also in detached zygospores of many species kept in water mounts. It appears to involve mostly deposition of supplemental wall material, but in addition the protoplasts of some species may develop conical apical caps that aid in the release of the contents on germination. It is believed that normally zygospores are ingested and that germination occurs in the host gut, but germination has been seen only on a few occasions outside the host: in Legeriomyces sp. (Whisler, 1963), Stipella vigilans (Moss, 1970), and Glotzia ephemeridarum (Lichtwardt, 1972). The zygospore contents are apparently under considerable pressure so that on germination the inner structure pushes through one end of the zygospore wall almost explosively. In G. ephemeridarum the inner apical caps are quite sharp, and the internal force is sufficient to perforate the lateral wall of the zygospore in cases where the zygospore protoplast is misaligned within the spore, as sometimes happens in this species (Lichtwardt, 1972).

The elongate rather than spherical shape of the trichomycete zygospore appears to be adaptive by providing a means to allow rapid germination or expulsion of the protoplast during its short travel time in the gut, and this may be a major reason for the differences in shape one finds between the biconical zygospores of these gut fungi and the spherical ones of other Zygomycota.

ECCRINALES - Over the years there have been a number of reports of putative sexual processes in the Eccrinales. Among them are intracellular fusions in Eccrinidus flexilis and Arundinula capitata (Duboscq et al., 1948); Enterobryus duboscqui (Tuzet and Manier, 1948a), and Palavascia philoscii and P. sphaeromae (Tuzet and Manier, 1948b). These reports have not been substantiated in subsequent studies, and in fact Manier in her 1969 monograph did not acknowledge the presence of sexuality in those species or any other Eccrinales. Maessen (1955) described the fusion of released protoplasts in Enterobryus flavus (nom. dub.) and Trichellopsis schizophylli (nom. dub.), but an occurrence such as she illustrated is not likely. More recently Scheer (1977) described Nodocrinella hylonisci (nom. nud.) and speculated that there were haploid and diploid phases in that species without offering any good evidence to support his view.

A study of marine eccrinids by Hibbits (1978) merits attention and further investigation. Among the various thallial and spore types in Enteropogon sexuale occurring in the anomurid Upogebia pugettensis, she described and illustrated parallel thalli from the anterior hindgut with series of scalariform conjugations (Fig. 11.36). Apparently the uninucleate protoplast of one thallus fused with the adjacent uninucleate protoplast of the other thallus, after which the binucleate protoplast moved into the cell of the larger of the two thalli, rounded up, and karyogamy occurred. No conjugation tubes formed in this process; rather, the cell walls simply dissolved between adjacent cells. Loose uninucleate cells (presumed to be diploid) were seen as well, but further development was not observed, nor was it determined if they were functional in the life cycle of the eccrinid. This process was seen only occasionally and mostly in exuviae. If the events as she described them have been pieced together correctly, they would resemble the type of sexual reproduction characteristic of the algal order Zygnematales. Hibbits also saw thalli possibly undergoing scalariform fusion in Taeniella carcini from the hermit crab Pagurus kennerlyi.

Prior to Hibbits' discovery, Drs. Hiroharu Indoh, Yosio Kobayasi, and the author (Lichtwardt et al., 1987) in May of 1964 collected anomurids [Thalassina sp., possible T. anomale (Herbst.)] from a mud flat at Usa on the island of Shikoku, Japan. Several of the specimens in their stomachs had an unidentified eccrinid (Arundinula sp.?) with a similar type of fusion. In this species, adjacent pairs of cells from two thalli fused so that each pair formed a common chamber containing a binucleate protoplast. Conjugation of two thalli did not necessarily occur at the distal ends of the thalli, and never at their bases. Farther down the series of fused cells, walled protoplasts could be found with a single nucleus. Thus, this is a process, like that described by Hibbits, that is suggestive of karyogamy and sexuality, although possibly vestigial.

Another interpretation of the type of fusion described above is that they merely represent spurious events in a group of eccrinids that often produce aberrant structures. Additional developmental studies are necessary before the presence of sexual reproduction in the Eccrinales is established.
 

Viruses

Virus-like particles (VLPs) have been discovered in two species of Paramoebidium, P. arcuatum (nom. nud.) from the hindgut of a mayfly nymph (Manier et al., 1971) and P. curvum from the rectum of a blackfly larva (Dang and Lichtwardt, 1979). They were essentially hexagonal in sectioned material of both species but differed in several other respects. Those in P. arcuatum measured 105-110 nm; most particles had an electron-opaque circular core of about 70 nm diameter and were seen partly arranged in paracrystalline bodies in the nuclei of a mature thallus beginning to cleave out amoeboid cells. In P. curvum, the VLPs were twice the size (210-220 nm) with an electron-opaque core of about 100 nm in diameter and an outer shell of about 20-30 nm thickness that consisted of two electron-opaque layers separated by an electron-transparent middle layer (Fig. 7.24). These were scattered throughout a highly vacuolated and membranous cytoplasm, but organelles like nuclei and mitochondria appeared normal.

Unfortunately, no species of Paramoebidium has been cultured axenically, and as a consequence it has not been feasible to study the nucleic acid composition of these VLPs or to obtain information on their biological effects on the Paramoebidium species. VLPs have been sought in several axenic cultures of  Amoebidium parasiticum, but were not found, and they have not been reported in any other trichomycete thalli, cultured or not.