Life History and Developmental Processes in the Basidiomycete Coprinus cinereus

Monokaryons and Dikaryons

Monokaryons arise from germination of the haploid binucleate basidiospores or, alternatively, from germination of the uninucleate haploid oidia. During germination of the basidiospores, a single germ tube displaces the spore pore cap. Similarly, a single germ tube is formed by the oidium. Nuclei within the germ tubes divide, and septa may be formed between the daughter nuclei; alternatively, the germ tubes may stay aseptate through several cycles of nuclear division. Multinucleate side branches emerge, while septum formation occurs in the older portions of the germtubes to yield cell segments with one or only a few nuclei. The tip cells of the main hyphae and those of side branches remain multinucleate during further growth (26, 169). The literature contains reports of the older monokaryotic mycelium having only one nucleus per hyphal cell (452). However, nuclear staining of aerial mycelium of various strains revealed that in most cases up to half of all hyphal cell segments contain not just one but two and sometimes even three nuclei (385; E. Polak, personal communication). Thus, the primary mycelium of C. cinereus does not fulfill the strictest definition of a monokaryon, i.e., a specific homokaryotic mycelium with just one nucleus per cell (4, 121). Nevertheless, we prefer to continue calling the primary mycelium a monokaryon, since this term is traditionally used in C. cinereus (70, 75, 266, 343, 391), and thus follow the less stringent definition given by Hawksworth et al. (164) of a mycelium as having genetically identical haploid nuclei. For the geneticists, the term “monokaryon” nicely targets the most important difference from the dikaryotic mycelium, which typically contains two genetically distinct haploid nuclei in its cells (26, 175a, 187).

Cultures of monokaryons differ greatly in growth rate and also in their morphological appearance, primarily due to variations in the amount and structure of the aerial mycelium produced (Polak et al., submitted). The distance between branches varies between strains but branches usually form in a relatively wide angle of about 70 to 75°. Hyphae of monokaryons are usually thin, with a diameter of about 3 μm (58, 226; Polak et al., submitted). They are generally characterized by simple septa (26, 58, 475) with a dolipore, as is typical for the basidiomycetous fungi (132). The dolipore is barrel shaped due to the swollen edge of the cross-wall around the pore. Parenthosomes, dome-shaped double membranes thought to be a modified part of the endoplasmic reticulum (ER), cover the dolipore swellings and the pore channel to form the septal pore cap. Perforations within the membranes allow cytoplasmic continuity through the parenthosome barrier and between adjacent cellular compartments, giving the hyphae a coenocytic character. Passage of smaller organelles such as mitochondria is possible; however, migration of larger organelles such as nuclei is blocked (299, 348, 437).

Dikaryons arise upon fusion of monokaryons of different mating types (Fig. 2), either by hypha-hypha fusion or by fusion of hyphae with germinating or resting oidia (26, 75, 223). Cellular fusion (anastomosis) may occur between two hyphal tips (tip-to-tip fusion), between a hyphal tip and a lateral hyphal wall (tip-to-side fusion), between a hyphal tip and a lateral swelling of a hypha (tip-to-peg fusion), or between lateral swellings of two neighboring hyphae (peg-to-peg fusion) (59, 410, 440). Upon fusion, nuclei enter the mycelium of the opposite mating type and, with dissolution of the hyphal septa (69, 75, 132), migrate through the hyphae until they reach a hyphal tip cell (26, 58) (Fig. 2). In C. cinereus, the speed of nuclear migration is between 1 and 3 mm h⁻¹, which is up to 20 times faster than that of the hyphal tip growth (58, 236, 460; for a compilation of data for various basidiomycetes, see reference 64). Buller (58) found that migration takes place mainly through the outer part of the colony, and he assumed that migration involved the regular division of the invading nucleus. In Schizophyllum commune, invading nuclei apparently do not densely colonize the established host mycelium and there are areas through which the migrant nuclei obviously pass but in which they cannot be detected (407, 441). Niederpruem (361) monitored nuclear movements over short distances in living hyphae in compatible matings of S. commune and did not observe any mitosis. The observations in S. commune suggest that there may be no or little nuclear division during nuclear migration in basidiomycetes (361, 407, 441). Other authors, however, dispute this conclusion. At least directly after hyphal fusion, at the initial stage of nuclear exchange and migration, there are sometimes nonsynchronous (265, 409) and often also synchronous mitotic divisions of the two nuclei while they are together in the fusion compartment (410, 471). In addition, synchronous nuclear divisions in “migration” hyphae have been visualized by indirect immunofluorescence microscopy (410). The age of the hyphae strongly influences the rate of nuclear migration (425) and might also have an effect on the mitotic activity of the nuclei in the fusing hyphae (M. Raudaskoski, personal communication). Whatever the situation in C. cinereus, once a migrating nucleus reaches a hyphal tip cell, the two types of haploid nuclei pair at average distances of 15 to 20 μm (175a, 211) and the paired nuclei divide synchronously (26, 58, 459) (Fig. 2). Simultaneously, specialized clamp cells (or hook cells) are formed at the position where a new septum will appear (26, 59, 459) (Fig. 2).

During the mitotic prophase, the nucleus localized closer to the tip will enter the developing clamp cell, and in the anaphase, it divides with a short spindle. The second nucleus stays in the hyphal cell beneath the clamp cell and divides with a longer spindle. Due to the different lengths of the spindles, the foremost of the dividing nuclei in the hyphal cell will pass the member of the other pair of dividing nuclei that is leaving the clamp cell. Septa are generated between the dividing nuclei, enclosing one nucleus into the clamp cell, one nucleus of the other type in the newly formed subapical cell, and a nucleus of each type in the newly arising hyphal tip cell. After septum formation, the backward-looped clamp cell fuses with a peg induced at the subapical cell. This fusion releases the enclosed nucleus from the clamp cell (59, 187, 464, 465) (Fig. 2). As a result of this coordinated mechanism of clamp cell formation and synchronized nuclear division, the order of nuclei reverses between the apical and subapical cell with each cellular division (187). The mechanism ensures that every cellular segment within a dikaryon contains one of each of the two distinct nuclei (26, 59, 187) (Fig. 2). However, occasionally, more than two nuclei are found within cells of a C. cinereus dikaryon (175a), and other species are known where the homokaryon also possesses clamp cells (225) or where a dikaryon is perfectly formed in the absence of any clamp cells (224, 459). Consistent with earlier data obtained with S. commune (427), recent analysis of mutant strains indicates that the actin cytoskeleton and microtubules with α₁- and β₁-tubulins play a role both in nuclear migration for dikaryosis and in subsequent nuclear positioning and movements within the established dikaryon (199, 209, 211, 410, 464, 465).

As in the monokaryon, the septa of the dikaryon have a dolipore, irrespectively of being formed between two hyphal segments or between a hyphal cell and a clamp cell (132). Biochemically, the septa of monokaryons and dikaryons are probably somewhat distinct, as suggested by in vitro enzymic dissolution of S. commune septa (69, 75, 519). The more resistant nature of dikaryotic septa had been suggested to be a cause of blockage of nuclear migration through septa within the dikaryon (69, 80). The literature on the cell wall composition of the monokaryotic and the dikaryotic mycelium in C. cinereus is unfortunately poor (42, 43, 195, 296, 298, 432). The innovative work of Marchant (298) and Bottom and Siehr (42, 43) indicates that major differences in chitin content and sugar compositions exist between the cell walls of the two different mycelial stages. In addition, Ásgeirsdóttir et al. (15) recently isolated monokaryon-specific hydrophobins. These are small, hydrophobic, cysteine-rich secreted proteins that assemble at the hyphal wall when hyphae emerge from a submerged culture into the air (116, 231, 521, 522). Visually, there are further differences between monokaryons and dikaryons. Dikaryons of C. cinereus tend to grow faster, with a denser, more protuberant and conspicious aerial mycelium, the hyphae are about 7 μm in diameter, and branches arise with a relatively acute angle of 10 to 45° (58, 266). Due to the vigorous, rapidly propagating character of dikaryotic mycelium, outgrowth of a dikaryon from a cross of compatible monokaryons is easily detected. In the ideal case, there are four positions where outgrowth will be observed. Two dikaryotic sectors will arise at the junction of the growing monokaryotic colonies, where their hyphae met and fuse. Two other dikaryotic mycelia are generated at the outward sides of the two monokaryons after migration of nuclei through the opposite mycelium (58, 158). Since only the nuclei, not the mitochondria, migrate, these latter dikaryons are distinguished from each other by their mitochondrial content. In contrast, the dikaryotic sectors arising from the junction zones of the monokaryotic colonies are expected to be mitochondrial mosaics due to mixed hyphal populations with distinct mitochondrial DNA populations. Usually, such mixed dikaryotic mycelia rapidly segregate for one mitochondrial type. Although mitochondrial inheritance in C. cinereus is basically uniparental, recombined mitochondrial DNA is occasionally found in places of hyphal anastomosis (23, 75, 117, 118, 306; for a recent review of mitochondrial inheritance and recombination, see reference 420).

Following dikaryon formation, there is no habitual restriction to hyphal fusion with other monokaryotic or dikaryotic strains, but, in contrast to monokaryons, dikaryons do not accept invading nuclei (237, 456, 457, 460). However, they are capable of contributing fertilizing nuclei to a haploid monokaryon, resulting in a new dikaryon (57, 58, 69); this reaction has been termed the Buller phenomenon (400). “Legitimate di-mon matings” (for “dikaryon-monokaryon matings”) are those where both kinds of nuclei in the dikaryon are of compatible mating type to the nuclei of the monokaryon. In these cases, heterokaryotization of the monokaryon occurs readily. Either one or both types of the nuclei from the dikaryon migrate into the mycelium of the recipient monokaryotic partner, giving rise to one or two new dikaryotic mycelia. Occasionally, both nuclei from the dikaryon replace the original resident nucleus of the monokaryon (237, 238, 304, 401, 457). In di-mon matings, as in matings of compatible monokaryons, mitochondria are not exchanged but mitochondrial mosaics may arise by a mixed hyphal population (304, 306). In hemicompatible matings, where one nuclear type of the dikaryon is incompatible with the nuclei of the monokaryotic partner, only the compatible nuclei might enter the monokaryon or, alternatively, both kinds of nuclei invade the monokaryon and replace the resident nucleus (304). In illegitimate matings, where all nuclei are incompatible due to differences at either one of the two mating-type loci (see below), heterokaryotization of the monokaryon may occur but only after a delay during which compatible nuclei are generated by somatic recombination (58, 401). Exchange of nuclei between different dikaryons is also possible despite the blockage of nuclear migration, but only in areas where hyphae of both strains intermingle. New dikaryons are thought to form either by outgrowth of fused cells after a fresh sorting of the original pairs of nuclei or by dikaryotization of monokaryotic hyphae that occasionally are observed in the growing front of a dikaryon. Somatic recombination between nuclei has never been observed in dikaryon-dikaryon combinations (456), probably due to the low frequency with which nuclei of the different strains are found together within the same cell.

One of the most remarkable aspects of the life cycle of most homobasidiomycetes is the temporal and spatial separation of cellular fusion from nuclear fusion. Karyogamy normally occurs in the fruiting bodies within the specialized cells, the basidia. It is directly followed by meiosis, thereby restricting the diplophase to a single nuclear generation (283, 469) (see below). The extended dikaryotic growth phase, together with the unusual mating system developed in Coprinus and other basidiomycetes (see below), has special consequences for the understanding of the individual fungus. The free fusion between any mono- and dikaryotic mycelia and the variety of donating, accepting, and expelling of nuclei allows the formation of genetic mosaic colonies composed of genetically different sectors. However, somatic incompatibilities (mycelial antagonisms) lead to processes that rapidly separate newly formed and original dikaryons, despite their history of hyphal fusions and a common nucleus, ensuring that different dikaryons can represent discrete individuals. The vegetative self-nonself recognition of dikaryons is associated with pigmented zones, sparse hyphae or, less commonly, tight hyphal knitting, and an increased production of sclerotia at the interface of colonies. In most cases, the nuclear genomes but occasionally also the mitochondria seem to determine somatic incompatibility. Usually, mitochondrial differences are tolerated and a mitochondrial mosaic behaves as one individual (304; for a further discussion of somatic incompatibility and the problem of fungal individualism, see references 122, 227, 412, and 511).

Other Heterokaryons and Homokaryons

Dikaryons are only one type of heterokaryon found in C. cinereus. Dikaryon formation is governed by the two independent mating type loci, A and B. For a mating to be successful, monokaryons must be different at both loci (for recent reviews, see references 79, 172, and 245) (see below). However, monokaryons can fuse independently of whether different mating-type specificities are present (438, 440). If monokaryons differ in the B but not in the A mating-type loci, a so-called common A heterokaryon is formed that is not distinct in hyphal morphology from the monokaryon, although such strains usually grow less vigorously and more slowly than their component monokaryons. If strains differ within the A but not the B loci, a common B heterokaryon results which grows as vigorously as a dikaryon and which is characterized by formation of unfused clamp cells at the hyphal septa. Common B heterokaryons may form fruiting bodies but not as readily as does the dikaryon. In contrast, fruiting body formation in common A heterokaryons is the exception (399, 460, 459; for further review, see references 75, 76, and 407). Usually, common A heterokaryons are not stable and can be recognized and kept as such only under strong selection pressure (forced heterokaryon), e.g., by selection for complementation of auxotrophies. Similarly, unstable clampless common AB heterokaryons can be isolated by a forced mating of monokaryotic strains of the same mating type but with different auxotrophies. By contrast, common B heterokaryons are normally quite stable. Like dikaryons, common B heterokaryons have been reported to act as donors but not as acceptors of nuclei upon fusion with suitable monokaryons. In contrast, common A and common AB heterokaryons will usually act both as donors and as acceptors of nuclei, probably also because they frequently produce hyphae with only one type of nucleus (455, 459, 460).

Normally, it is the dikaryon that produces the fruiting bodies (see below). However, under specific conditions, monokaryons may also form fruiting bodies, e.g., under total nutritional depletion (159, 503, 508a). Certain mutations outside of the mating-type loci (su-A [103]), fis^c [324, 373, 480, 491], pcc1, and CopD5 [356, 357]) also lead to formation of fruiting bodies, karyogamy, and meiosis on monokaryons. These mutations are all accompanied by the formation of unfused clamp cells at hyphal septa (356, 357), and the su-A mutation is known to block nuclear acceptance (103). A specific class of mutants comprises homokaryons with defects in the mating-type loci that overcome the self-incompatibility of monokaryotic strains. Such strains, commonly called Amut Bmut homokaryons, are fully self-compatible and will initiate fruiting-body development under certain environmental conditions (452; see below). Depending on the general genetic constitution of the strain, fruiting-body initials and primordia will mature and karyogamy, meiosis, and the production of four basidiospores will occur within the gills of the fruiting body (165, 452) (Fig. 3). Amut Bmut homokaryons resemble dikaryons in their vigorous growth and hyphal morphology. They have fused clamp cells at the hyphal septa, but their distribution is usually not as regular as in dikaryons, and septa with unfused or no clamps are frequently been detected (452; U. Kües, unpublished data). The nuclei in aerial cells of Amut Bmut homokaryons are distributed in similar patterns to those of related monokaryons (385), but in submerged mycelium a fairly strict binucleate distribution has been reported (452). Although fully self-compatible, Amut Bmut homokaryons might mate with any strain in the junction zones of a cross, irrespective of which wild-type (self or nonself) or mutant mating types are present in the mating partner. However, heterokaryons formed with strains of related mating types appear to be short-lived and unstable. Nuclear acceptance of Amut Bmut homokaryons might be restricted but donation by nuclei into monokaryons is possible (269; Kües, unpublished).

To generate Amut Bmut homokaryons, independent mutations were selected in the A and B mating-type loci and combined into the same nucleus by crosses (165, 452). Amut homokaryons with only a mutated A locus mimic common B heterokaryons and are characterized by mycelia with unfused (“false”) clamp cells at some but usually not all septa. Amut homokaryons may have one, two or occasionally even three nuclei per cell (103, 165, 452) at frequencies comparable to what we have found in monokaryons (384, 385). Amut homokaryons are hemicompatible and form stable dikaryons with any monokaryotic strain as long it has a different B specificity (103, 165, 452). In mating, Amut homokaryons may act as both donors and acceptors of nuclei (Kües, unpublished). Bmut homokaryons are also hemicompatible. They mate bilaterally and accept and donate nuclei in crosses with monokaryons as long as these differ in the A mating-type locus. Bmut homokaryons have simple septa like normal monokaryons (165, 452). In contrast to the analogous Bmut homokaryons of S. commune (243, 361), the dolipore septa are not disrupted in the C. cinereus Bmut homokaryons and the nuclear distribution within the cells is even over the entire hyphal length as in monokaryons (165, 452).

Somatic Diploids

Although the diplophase is normally restricted to the basidia in the life cycle of C. cinereus, nuclear fusion occasionally occurs within the vegetative mycelium of heterokaryons. Diploid mycelia can be selected from common A and common AB heterokaryons by plating and germinating uninucleate oidia at frequencies ranging from approximately 10⁻³ to 10⁻⁴ (68, 187, 355). There are no reports in the literature on the rescue of diploids from common B heterokaryons or from dikaryons, but indirect evidence for the infrequent existence of diploid nuclei comes from the observations of somatic recombinations (98, 237, 401, 457, 458).

Diploid oidia of common A and of common AB heterokaryons have increased cell size, with a single enlarged nucleus that is twice the size of a haploid nucleus. Germinated hyphae are significantly broader than those of the monokaryotic parents (68, 187, 355, 364). Diploid common A and common AB mycelia readily mate with monokaryons and with other diploid common A strains, provided that these strains contain other A mating-type genes. However, common A diploids serve only as nuclear donors in matings, whereas common AB diploids both donate and accept nuclei (77, 80, 98, 364). In agreement with their different behavior in nuclear acceptance, septal dissolution was not detected in common A diploids, in contrast to common A heterokaryons, indicating that the cytological difference between the two nuclear conditions has an impact on the mating behavior of the mycelium (364). Possibly due to uneven nuclear distribution within colonies of common A heterokaryons (78, 455), common A diploids and common A heterokaryons also differ in complementation efficiency of metabolic functions whereas common A diploids and ordinary dikaryons usually behave similarly (78, 105, 367, 472). However, bringing genes together in a single nucleus in common A diploids or separating them into two haploid nuclei within dikaryotic cells can also cause changes in expression, as observed with certain resistance genes (267, 365).

Diploid mycelia are often fairly stable—only 0.5% of oidia produced by such mycelia were found to be haploid (68, 78)—but treatment with the fungicide griseofulvin can induce haploidization by a gradual loss of chromosomes (364). Most interestingly, diploid-haploid and diploid-diploid pairings produce dikaryons in which the diploid nuclear components are unstable in the vegetative growth phase, indicating that the dikaryotic situation is preferred over the diploid. Loss of chromosomes in such dikaryons is also progressive as aneuploid and haploid nuclei are recovered from these colonies (68, 78, 397). Finally, it is interesting that common A diploids are not fully compatible even if they are distinct in their A mating-type loci. In dikaryons formed by such common A diploids, the produced clamp cells at first fail to fuse, but progressive haploidization, eliminating one of the two B chromosomes from the diploid nuclei, eventually results in a compatible situation and clamp cell fusion (78).

Oidiophores and Oidia

Monokaryons of C. cinereus have for a long time been known to constitutively produce abundant uninucleate asexual spores on specialized aerial structures (oidiophores). These spores are rod shaped, and because of this shape, they are called oidia (26, 47). Until recently, dikaryons were believed not to form these mitotic spores (48, 70, 234), but their production has been observed after light induction (232, 452) (Fig. 4). Likewise, formation of oidia on Amut and Amut Bmut homokaryons is under light control, whereas Bmut homokaryons and common A heterokaryons behave like monokaryons and constitutively produce numerous oidia (232, 385, 452). Common B heterokaryons are also reported to produce abundant oidia in the light (459); due to the presence of different A mating-type genes within these strains, repression of oidiation in the dark is expected (232, 261) (see below) but was not tested. Common A and common AB diploids also give rise to oidia, and spore production is presumably constitutive (68, 78, 364). Extensive use of oidia has been made in single-cell purification of cultures, in generation of uninucleate protoplasts for transformation, and in providing isolated uninucleate cells in classical and modern mutagenesis techniques (30, 31, 68, 145, 366, 386, 388). Regardless of their importance in genetic studies, attention was only recently drawn to the cytological process of formation of oidia (385; Polak et al., submitted). Under certain genetic conditions, oidiation is light induced, and light acts at the level of oidiophore formation, which has enabled the whole process of sporulation to be monitored over time in microslide cultures (385).

Oidiophores are formed predominantly in the aerial mycelium (47, 385), although in most strains small numbers of oidiophores can regularly be detected in the submerged mycelium of agar cultures (Polak et al., submitted). Not every cell of the aerial mycelium becomes an oidiophore foot cell, i.e., the hyphal cell that will give rise to the oidiophore. Unlike the oidiophore foot cells, undifferentiated aerial cells vary greatly in length, indicating that the former are physiologically predetermined (Polak et al., submitted). Development starts with the protrusion of a young oidiophore from its foot cell (385) (Fig. 5). Subsequently, a nucleus in the foot cell divides, one daughter nucleus migrates into the bulging young oidiophore, and a septum is formed which separates the first oidiophore stemcell from its foot cell. The stem cell elongates and may divide once or twice, and side stems may branch from the cells of the main stalk. From the tip of stem cells of both main and side stalks, short, oblong oidial hyphae bud off one after the other. Successive nuclear divisions within the stem cells ensure that each oidial hypha is provided with a nucleus before it is separated from the stem cell by septation. Within an oidial hypha, a further nuclear division occurs followed by formation of a septum that separates the oidial hypha into two equally sized cells. Eventually, these cells are released by schizolysis of the septa between the oidial hypha and between the oidial hypha and its stem cell. A few hundred free spores are successively collected in a sticky liquid droplet secreted at the tips of the oidiophores (385) (Fig. 5). Since their mode of generation involves budding and septum schizolysis, oidia are classified as arthroconidia (228, 385). Mature haploid spores average about 2 by 4 to 6 μm in size, and mature diploid oidia measure about 3 by 7 μm (68, 364, 385; Polak et al., submitted). Oidia are enclosed by mucilage and possess a double-layered primary cell wall with hair-like structures, except at former sites of cell attachments, where a single-layer secondary cell wall is present (168, 169, 385). These secondary walls are later involved in the germination of the spores (168). Oidia are distinguished as “wet” spores due to their strong hydrophilic character (15, 222). Consistent with this, proteins with the characteristic properties of hydrophobins are not detected on oidia (15), unlike the “dry” hydrophobic spores of other fungal organisms, which are coated with hydrophobins (116, 231, 518, 521, 522). Oidia of C. cinereus are not wind-borne spores. Instead, they stick to flies and other insects attracted by the horse dung substrate and are thereby distributed to new substrates (47). Independently of mating types, oidia attract hyphae of colonies—a process called oidial homing—and act as spermatia in matings with compatible C. cinereus cells (223) (see above). Most interestingly, they also attract and fuse to hyphae of other related coprophilous fungi. Upon fusion, somatic incompatibility reactions are initiated, leading to the death of the foreign hypha. Oidia thereby act as a killing agent of opponents competing for the same limited substrate (223).

Analysis of various C. cinereus strains showed that oidiophore formation is not a fixed cytological process. Oidiophores may have simple stems or may be branched, and the stem cells may not elongate or may totally be absent (26; Polak et al., submitted), indicating that the different steps in oidiophore development, as defined in Fig. 5, are independent of each other and that some may be omitted without terminating development. To denote the main structural differences, four main types of oidiophores have been defined. Type 1 oidiophores have unbranched elongated stems and produce oidial hyphae at only their tips. Type 2 oidiophores form oidial hyphae at the tip of a main stem and also at the tip of side branches; oidial hyphae may also bud directly from the uppermost stem cell (type 2A) and from stem cells beneath (type 2B). In contrast to these more advanced structures, type 3 oidiophores do not elongate their stem cells. Type 4 oidiophores are those lacking a stem cell, where oidial hyphae bud off either singly (type 4A) or in bundles (type 4B) from cells of the aerial hyphae. These various types of oidiophores occur within the same aerial mycelium of all strains so far analyzed, although at strain-specific frequencies (Polak et al., submitted).

Chlamydospores

Chlamydospores are large, thick-walled mitospores of variable forms and with condensed cytoplasm. They are found in brown patches present on the mycelial mating of older cultures of dikaryons (6, 71, 121, 266) and, occasionally, on certain monokaryons (26, 261) (Fig. 6). The generation mode of this type of spore in C. cinereus is not well documented. In general, chlamydospores may arise endogenously in cells of the vegetative hyphae following compression of the cytoplasm (chlamydospores in the strictest sense) or by transfer of the compressed cytoplasm from a hyphal cell into a bud (blastocyst) (95). Diagrams found in some publications (6, 121, 266), combined by occasional observations in our laboratory (U. Kües and E. Polak, unpublished data), suggest that both generation modes exist in C. cinereus.

Despite their poor cytological description, chlamydospores have been used to collect component monokaryons from dikaryotic mycelia. Chlamydospores from dikaryons generally contain one nucleus of each parental type. Dikaryotic chlamydospores germinate with either one germ tube or two germ tubes, one at each end. When two germ tubes are present, each will receive one nuclear type. Monokaryotic hyphal cells can be isolated by microsurgery using a sharp scalpel before the germ tubes newly dikaryotize each other (121, 266). However, the fact that oidia production has now been observed in dikaryotic cultures after light illumination simplifies the process of monokaryotic culture isolation, but there is normally an uneven recovery of the two nuclear types (175a, 383).

Developmental Course of Fruiting

Fruiting-body formation is the most complex developmental process in the life cycle of C. cinereus. Fruiting-body development involves a dramatic change of the growth pattern from a relatively regular three-dimensional loose mesh of free, undifferentiated hyphae to a compact multihyphal structure composed of many different cell types that associate with each other in distinct hypha-hypha interactions (335, 337, 338). Fruiting-body formation in Coprinus is a rapid process. From the first sign of fruiting (formation of initials from hyphal knots) to maturation (autolysis of the cap of the mature fruiting body), it takes about 4 to 5 days (263, 277, 344, 353, 450) (Fig. 7). To initiate and continue fruiting-body development, several environmental signals are called for, of which dark and light periods are the most important (see below). Indeed, fruiting-body development is perfectly synchronized to the light-dark rhythm fixed by the normal day-night cycle (21, 40, 184, 203, 277, 350, 351, 353, 476) (Fig. 7). Hyphal knots are formed in the dark (261, 330), and light then induces the formation of globose fruiting-body initials. Within the initials, cellular differentiation begins, but a further light signal is needed to develop primordia with distinctive cellular tissues (21, 184, 203, 277, 350, 351), otherwise the initials will grow into elongated structures with underdeveloped caps (see below), variously called pseudorhizas (32, 56, 60) dark stipes (353, 476) or etiolated stipes (40, 120). Following another dark period, light is needed to complete meiosis within the basidia, while stipe elongation and cap maturation take place in parallel. Meiosis will be finished by midnight, and basidiospore formation and cap opening occur in the early-morning hours. Within a few more hours, the cap will autolyze and the black basidiospores will be released within a brown liquid (283, 286, 353) (for further details of light regulation, see below).

Early Stages

Fruiting-body development in C. cinereus is monocentric: a hyphal knot (also referred to as a hyphal tuft [335] or nodulus [93, 94]) forms as a single independent organ from which a fruiting-body initial may arise (302, 303). In the literature, there is no sharp distinction between a hyphal knot and an initial, and the difference between the initial and the primordium is not easy to understand due to the different uses of these terms by different authors (see, e.g., references 93, 94, 303, 335, 338, and 344). Hyphal knots as defined by Reijnders (413, 415) occur in various tissues of basidiomycetous fruiting bodies and their primordia (e.g., in the stem, veil, and cap). Reijnders (413, 415) defines hyphal knots as aggregates of hyphae that form small communities consisting of an induction hypha with the surrounding hyphae brought under its influence (for further discussion, see reference 338).

Work in our laboratory focuses on fruiting-body initiation. This has made it necessary to define the initial hyphal knot in fruiting-body formation of C. cinereus more specifically as areas of intense localized branching of undifferentiated hyphae within the vegetative mycelium (R. P. Boulianne, Y. Liu, M. Aebi, B. C. Lu, and U. Kües, submitted for publication). (Fig. 8, left). Mature hyphal knots are about 0.2 mm in size (201, 302, 344; Y. Liu, personal communcation). Microscopic studies revealed that the simplest hyphal knot arises from just a single hypha that will intensely ramify with short branches, which in turn give rise to further generations of branches with restricted tip growth (46; Liu, personal communication). In the more common case, hyphal knots originate from more than one generative hypha. Branches of neighboring aerial hyphae will grow toward and alongside each other and possibly merge by anastomosis through the lateral hyphal walls to form an intricate and easily distinguishable lattice. Such locally restricted interlacings will serve as the active center for the intense production of short hyphal branches. Hyphal cells within the resulting bunches of branches are often of a globose or inflated morphology (201, 302; Liu, personal communication). The process as described so far occurs in the dark and is repressed by illumination with blue light (261, 330). However, continuation of development toward fruiting-body formation is light dependent (277, 350) (Fig. 8, middle), indicating a major change in regulation and the beginning of a new, distinct developmental phase. This new phase includes a switch from a (purely) ramifying to an aggregating mode of hyphal growth and results in the formation of a round, aggregated body of tightly interwoven hyphae (95, 335, 338; Boulianne et al., submitted). For this reason, we restrict the definition of a hyphal knot to the locally restricted bunch of hyphal branches which is formed in the dark and which does not necessarily involve hyphal aggregation (Boulianne et al., submitted). These dark-formed structures are not fruiting-body specific but also serve as precursors in the dark-dependent development of sclerotia (330) (Fig. 8, right) (see below). Thus, the formation of globose hyphal aggregates from hyphal knots in the light is the first fruiting-body-specific stage (330, 338), and these structures are referred to as the fruiting-body initials (see also references 263, 344, and 450). This distinction between hyphal knots (i.e., the primary nodules of Clémençon [93, 94]) and initials (in their early appearance, the secondary nodules of Clémençon [93, 94]) has recently been confirmed by using molecular markers. Two different galectins, Cgl1 and Cgl2, belonging to a β-galactoside sugar binding subgroup of the lectin family, had been isolated from Coprinus mushrooms of a wild-type dikaryon and an Amut Bmut homokaryon (44, 96; Boulianne et al., submitted). Expression of Cgl2 in the Amut Bmut homokaryon has been shown to correlate with hyphal knot formation and to be repressed by light, whereas Cgl1 expression starts with the development of initials and is light-induced (Boulianne et al., submitted). Mutants have served before to elucidate fungal development (95), and we generated a collection of mutants from an Amut Bmut homokaryon by restriction enzyme-mediated DNA integration (REMI) and UV mutagenesis (145; Kües, Granado, et al., unpublished). Among these mutants, we identified a number of strains with specific defects in fruiting-body initiation, having a block either before or directly after hyphal knot formation (knt and prm, respectively). Most interestingly, we found one mutant that formed hyphal knots but produced neither Cgl1 nor Cgl2, some knotless strains that produced only Cgl2, and mutants that formed knots but no initials and produced either only Cgl2, or, in rare cases, both galectins at a low level. These classes of mutants correlate well with the distinction of hyphal knot and initial and a two-phase regulation of development (Boulianne et al., submitted; Liu, unpublished). Galectins are thought to be involved in hypha-hypha aggregation (44, 96), and the results obtained with the mutants indicate further that galectins are not essential for hyphal knot formation (Y. Liu, M. Aebi, and U. Kües, unpublished data), consistent with the notation that aggregation is not essential for hyphal knot formation.

In the hyphal knot, there is a noticeable reduction in the chitin content with respect to that in the vegetative hyphae (201). However, the globose hyphal aggregate of the initial, the “primordial bud” (277), is the first structure that shows clear histological differentiation. It is enclosed within a coat of large, vacuolated, and mainly outwardly directed hyphae covered with various amounts of amorphous material that protect a compact core of prosenchymal tissue. In the young initial (about 0.2 mm in diameter), this core of heavily branched short cells is only loosely grouped, but in older initials (up to 0.5 mm in diameter), the cells are tightly twisted and fixed together (303, 500). The densely packed cells in the core are rich in glycogen, and between the cells, a large amount of mucilaginous material is present which is probably involved in aggregation. Anastomosis and presumably exchange of cytoplasm occur frequently between adjacent cells. Septa between core cells are usually surrounded at one or both sides by a parenthosome and an additional hemispherical cap also composed of ER membranes. In contrast, cellular fusions do not occur between the large hyphae of the outer layer and the septa are dolipores with normal parenthosomes (500). The current lack of suitable molecular markers makes it difficult to give any information about the possible mechanisms involved in defining the further morphogenesis of the structures which eventually lead to distinctly differentiated tissues within the core area. Since the fruiting-body embryo is initiated within a hyphal agglomeration and is protected by outer layers of large cells (veil cells), which rupture later in development, fruiting-body formation in C. cinereus is classified as hemiangiocarp (95, 335, 512) (Fig. 9). Most important is the change in polarity: the lower and the upper parts of the core of the initial give rise to the fruiting-body stipe and cap, respectively (303, 362). It is not clear when the change from undirected hyphal growth to a defined polarity is determined, and the responsible physiological determinants are also unknown. As a step forward in elucidating this problem, a promising method of vital staining with Janus green has recently been developed to identify initials easily and to study the physiological processes involved in fruiting-body initiation (430).

Following the definitions of Moore and coworkers (263, 344, 450), up to a maximal size of 2 mm we will call the developing globular, later more oval aggregation the initial. Staining of fixed material indicated that the first cytological differentiation within the aggregate becomes obvious at a size of about 0.4 to 0.7 mm. Differentiation starts visibly in the upper third of the aggregate, the prepileus region. This ventral area is still prosenchymal and is the base for the upper peripheral cells (pileus trama) and, underneath, the ring of cells in which polysaccharides (glycogen) accumulate. Cells in the middle part of the initial line up to give rise to the stipe tissue, whereas cells in the lower third are randomly orientated to form the basal plectenchyma, representing a tissue of tightly packed hyphal cells not belonging to the stipe. A distinct mushroom shape emerges at a size of 0.8 to 1.0 mm, when hyphal tissue grows from the subapical region in an outward direction to give the annular boundaries of the pileus (303, 335, 344, 362, 414).

The whole complex is surrounded by a marginal veil of hyphae which originate from the top of the pileus (303, 344, 353, 362, 414, 422, 424). Veil cells are large, multinucleate, and septate, but they lack clamp connections (277). With an increasing size of 1.5 to 2 mm, a dense band develops which clearly defines the boundary of the hymenium and stipe (277, 333, 344, 362, 414, 422, 424). To suppress the formation of etiolated stipes and to continue hymenium development (referred to as day 0 by Lu [277]), light is needed at this stage (277). In the lateral part of the cap, between the trama of the cap (pileus) and the differentiated veil cells, a zone of cell division (meristemoid) of strictly parallel hyphae appears (413, 414). Glycogen deposits can be detected in the basal plectenchyma and in the hymenium (89, 197, 303, 344). The hymenium differentiates further into dome-shaped rudiments, which eventually will become the gills (lamella), which carry the basidia in their outer hymenial cell layer (89, 277, 333, 413, 414) (Fig. 9). Primary gill formation proceeds from the edge closest to the stipe toward the cap and from the margin to the apex of the cap (277, 333, 422, 424). Autodigestion of the cells in areas of the gill cavities has been proposed to be the origin of gill construction (277, 282), although the cell debris and the multivesicular and membraneous residual bodies seen in electron micrographs and taken as remnants of degenerating cells possibly represent fixation artifacts (335, 337, 338). Support for the occurrence of morphogenetic cell death as part of the programmed process of primordia development comes from studies with Agaricus bisporus and other mushrooms. In Agaricus, programmed cell death repeatedly occurs during establishment of the hymenial chamber within the primordia (339, 478, 479). It has been suggested that this cellular autolysis contributes to the formation of the primary extracellular matrix (ECM) (479).

Light is essential to enter the next stage of development (203, 277, 351) (Fig. 7), termed stage 1 primordia by Moore and coworkers (263, 344, 450), day 1 primordia by Lu (277), and stage 0 by Kamada et al. (203). Light induces hymenium differentiation and ultimately karyogamy (203, 277, 351) (see below), but the nuclear prefusion phase is long (taking about 24 h) (479). During this time, stage 1 primordia grow from about 2–3 mm to 7–9 mm in diameter. They are characterized by well-developed gills containing basidia of about 12 μm and cystidia (large sterile cells sized between 20 and 25 μm) at their surface (263, 344, 402, 450). Gills are essentially vertical plates arranged radially around the stipe. Primary gills are connected with their tramal tissue of their apex to the outer layers of the stipe; consequently, they have a hymenium at both their sides but not at their edge. With increasing primordium size, space for further gills becomes available within the cap. Secondary and lesser ranked gills are formed by bifurcation of the primary gills. Secondary gills grow radially outward, away from the stipe, by a replication fork-like movement of the gill organizer (a hypothetical formative element in the tissue at the extreme end of the gill cavity) into the trama of the primary gills. As the gill cavity moves outward, it expands tangentially, making room for a new gill organizer to appear, which initiates further gill splitting (89, 333, 414, 422, 424). When space becomes available, further gills may arise from prosenchymal tissue as folds of the cap. Gills of this kind appear as convoluted plates usually in the most apical regions of the cap (89, 422; for further reading on the process of gill formation, see references 95, 335, 337, and 338). Due to their mode of formation, secondary gills are not attached to the stipe. Therefore, their hymenium is continuous over the gill edge. At the edge of the secondary gills, there is an increased number of cystidia (333, 414, 422, 424). The hymenium develops as a layer of similar-sized palisade-shaped cells by branching from sister hyphae of the prosenchymal subhymenium. This hymenial layer of cells is occasionally interrupted by larger cystidia (90, 179, 344, 422–424). Hymenial cells have clamp cells, and the nuclei stay paired, indicating that the dikaryotic stage is preserved (277, 344). During further development of stage 1 primordia, the basidia change from a cylindrical to a club-shaped structure (402). Glycogen accumulates in the subhymenium and cystidia and in the basal plectenchyma of the stipe (263, 344, 450). The timing of this stage is difficult to determine, and this makes it particularly difficult to compare the observations on primordium development made by different researchers (21, 203, 277, 344, 351). Karyogamy is therefore used as a reference point (277). Karyogamy occurs toward the end of this phase and is considered, according to Lu's definition, to have begun when 5% of the basidia have undergone nuclear fusion (277, 402) (see below). At this point, by the beginning of the following day, tissue development in the cap is completed (277).

Stage 2 (day 3) primordia (Fig. 7) are 6 to 10 mm (occasionally up to 15 mm) tall. Meiosis starts within the basidia (see below), which are now 15 to 18 μm long. The cystidia enlarge up to 45 μm. Primary gills are still connected to the stipe, but the gills begin to separate. Initially, the veil of stage 2 primordia is still intact, but it becomes more and more free. In the cap, more glycogen accumulates, while the glycogen content in the basal plectenchyma decreases (263, 277, 344, 450). Galectin distribution has been analyzed in stage 2 primordia. Galectin antibodies localize preferentially to the veil cells, to the outer layer of narrow hyphae of the stipe, and to the apex of the primary gills connected to the stipe (Boulianne et al., submitted). Significantly, these different tissues (called in specialist terms lemmablem [the external tissue of the cap] and lipsanoblem [the external tissue of the stipe], respectively) are functionally related in protection of the inner primordium tissues, and at least the veil cells of the cap are dispensible at later stages and degenerate or are discarded (95, 155). To a lesser extent, galectins are also present in the gill trama and the inner tissues of the stipe. Galectins are found attached to cell walls and in the extracellular matrix and, within the cells, localized in vesicles and in the dolipores (Boulianne et al., submitted). Consistent with the immunolocalization studies, Northern blot analysis indicates high expression of galectin genes in stage 1 and stage 2 primordia, but at later meiotic stages (late pachytene/diplotene, metaphase I) at the transition to the next developmental stage (fruiting-body maturation) galectin expression is halted in both the cap and the stipe (85).

The further development of stage 2 primordia into mature basidiocarps takes another 24 h (263, 277, 344, 353, 394, 450). Maturation consists of at least four processes: stipe elongation, basidium maturation (including meiotic divisions and spore formation), pileus expansion, and autolysis of the pileus (see below).

Structure and Development of the Stipe

The mature stipe is a thin-walled hollow cylinder with two types of hyphae, i.e., narrow and wide hyphae. It has been found that 23 to 54% of all hyphae are narrow. Most of these narrow hyphae are concentrated at the exterior of the stipe, but they are also found at the lumen and in between the wide, inflated, and highly vacuolated hyphae that make up most of the interior tissues (39, 156). The narrow hyphae resemble vegetative hyphae, since they clearly have clamp cells and form interconnections independent of the inflated hyphae (39). This network of narrow hyphae has been suggested to act in nutrient translocation from stipe to cap (188, 335, 338). Cytologically, the cell walls of the narrow hyphae at the different locations differ from each other and from those of the inflated hyphae, as indicated in various staining experiments (39, 139, 156) and by a different immunohistochemical response with antigalectin antibodies (Boulianne et al., submitted). On the whole, the composition of cell walls of the stipe seems very much like that of monokaryotic hyphae and is very different from that of vegetative dikaryotic hyphae (298). However, compared to the monokaryotic mycelium, the overall chitin content is increased in the stipe (298), consistent with the facts that chitin synthesis is especially active in stipe formation (106, 137, 140, 143) and that the chitin synthesis inhibitor polyoxin D suppresses stipe elongation (144). Chitin microfibrils, 7 to 25 nm in diameter, are indeed a major component of the walls of stipe cells. They are arranged as shallow helices transverse to the long axis of the cell. Two-thirds of these helices are left-handed, and one-third are right handed (140, 195, 210). In contrast, chitin microfibrils are randomly orientated in the vegetative mycelium; the shift to a parallel transverse arrangement can occur in hyphal knots of only 0.1 to 0.2 mm in diameter, although at this point it is still incomplete (201).

Stipe growth is mainly the result of manifold cell elongation rather than cell division (39, 99, 142, 196). However, in the very young primordium, stipe cells are cuboid or polyhedral, and proliferation still occurs in the apex meristematic region of the stipe (39, 155, 302). At the stage of karyogamy, many of the cells inflate and become cylindrical, and it is obvious that these enlarged cells lack clamp connections, in contrast to cells found in the central yet closed region of the stipe (39, 155, 156, 277). During stipe elongation, inflated hyphae inflate further but only vertically; the width of the stipe cells remains almost constant (155, 156, 335, 337, 338). However, the proportion of narrow hyphae declines as the stem grows from 45 to 70 mm, indicating that normal vertical stem extension involves both an increase in the cross-sectional area of inflated hyphae and recruitment of narrow hyphae into the inflated population (156, 335, 337, 338).

The enlarged stipe cells are usually multinucleate, and more nuclei are progressively produced by mitosis as the cells grow by vertical elongation (277, 449). Indeed, there is a linear correlation between the length of the stipe and the numbers of nuclei found within the elongated cells. There are between 2 and 8 nuclei in stipe cells of 2- to 4.5-mm primordia, about 16 at a primordial size of 5 to 7.5 mm, and more than 150 when the stipes are longer than 20 mm. In 29% of stipe cells, nuclei are present in uneven numbers and there is no indication of synchronized nuclear division (142, 449). The increase in the number of nuclei is indicated by the dramatic increase in DNA content 3 h before the onset of rapid stipe elongation (142, 202). In contrast, the RNA content increases continuously with stipe maturation and the protein content is constant (141, 202, 344). The young stipe contains large amounts of glycogen inclusions, whereas glucans, including glycogens, are almost absent in the fully expanded stipe, probably because they serve as a source of cell wall precursor material (32, 195, 197, 198). Apart from glycogen inclusions, insoluble protein inclusions were also detected which first increase but then decrease in size with progressive stipe development (32). This is most interesting, since, apart from the intercellular compartments and cell walls, the galectins have been localized to vesicles within stipe cells of stage 2 primordia and young elongating fruiting bodies (Boulianne et al., submitted).

Decapitation of the stipe before meiosis arrests stipe expansion, showing that the cap is needed for the early development of the stipe (39, 99, 138, 150, 195). Consistent with this, grafting experiments revealed that during meiosis, the cap produces a diffusible substance acting on basidiocarp maturation, including stipe elongation (200). Rapid stipe elongation correlates with the end of meiosis 8 h after the nuclei in the bulk of basidia have fused (155, 202). Although the cap still influences the absolute length a stipe can achieve, in the postmeiotic basidiomes the stipe and cap develop independently from each other. Also, the stipe does not need to be connected to the vegetative mycelium, indicating that stipe elongation is an autonomous endotrophic event requiring no exogenous nutrients, growth factors, or water (39, 99, 138, 142, 155). Elongation is not evenly distributed over the length of a stipe. The ultimate apex of the stipe does not extend much, although in general the cells in the apical half elongate most, in contrast to the cells of the basal half, which show reduced stretching, and the cells in the base, which show no stretching at all (39, 99, 156, 196, 198, 344). Stipe cells elongate in an intercalary fashion by diffuse extension in which the extension of the cell surface is not confined to the hyphal tip but occurs throughout the whole cell (139, 141, 156, 195). The stipe has to be formed quickly and as strongly and economically as possible. Therefore, cell walls have to be plastic but also rigid to withstand the high turgor pressure. Plasticity is given by the ordered arrangement of chitin helices, and rigidity is given by cross-linking between chitin and glucans (195, 196, 198). Throughout elongation, the chitin content stays constant and insertion of new chitin microfilaments occurs over the whole length of the cell in a uniform intercalary fashion, probably mediated by a constant unfastening and resealing of linkages between chitin and glucan (139, 141, 197, 210). Consistent with active cell wall hydrolysis playing a role in the mechanism of stipe elongation, stipes are rich in chitinases and glucanases (204, 205, 207).

An average stipe may elongate 80 mm in less than 12 h (195, 341, 449). It is very strong in its longitudinal axis and withstands pressures of 7 × 10⁴ N m⁻², a fact impressively shown by the occasional documentations of fungal breaks through asphalt (58, 142, 327). Rapid stipe elongation requires high turgor pressures (142, 327), and throughout elongation the osmotic value stays almost constant at 0.45 to 0.5 M (99, 195). This high turgor must be actively maintained, probably by organic metabolites which have not yet been identified (39, 141, 142, 338). Trehalose and polyols have been excluded as the sole osmotic stabilizers, since the concentration of the former decreases during elongation whereas the latter are hardly present in stipes at all (141, 142, 338, 406). Amino nitrogen compounds and urea have also been implicated in osmostabilization, but their contents are low and decline with development (123, 142, 344). Regularly in fruiting-body development, shortly before rapid stipe elongation, clear or sometimes yellowish droplets are excreted at the region where the cap with the gills is attached to the stipe (241; Kües, unpublished). These droplets are rich in potassium oxalacetate, but it is not known whether this contributes to the high turgor pressure of the stipe (241). Cox and Niederpruem (99) reported that a brown gel was present within the lumen of young stipes but disappeared during elongation, and this has also been discussed as a source of an organic osmotic stabilizer (195). Since no sugars or other substances have been found to be of prime importance, cocktails of simple carbohydrates together with inorganic ions are anticipated to control the osmotic pressure in stipe elongation (338, 344).

Developmental Processes in the Basidia

Cap Maturation

With progressing meiosis, the hymenium within the cap differentiates fully. In the immature stage, the hymenium lacks the close packing of sterile hyphal elements (paraphyses), which may provide some stability in later stages (277, 344, 394, 402, 422). In late stage 2 primordia, at the end of prophase I, binucleate paraphyses emerge as outgrowths of binucleate sub-basidial cells, increasing the width of the gill from 100 μm to 300–400 μm (90, 394, 395, 422–424). In the postmeiotic cap of the immature fruiting body (10 to 20 mm in size; stage 3 immature fruiting bodies [263, 344, 450] [Fig. 7]), the paraphyses expand from 3–4 μm to 8 μm, incorporate glycogen, vacuolize, and take on their characteristic rectangular shape. Inflation of paraphyses makes a major contribution to cap extension, which starts within the 1.5-h period between the ending of meiosis and the beginning of sterigma production (123, 155, 344, 423). Basidia begin to form sterigmata and spores when they measure around 20 μm (344, 394) (Fig. 10). Each basidium is surrounded by about five paraphyses, and consequently the basidia are displaced from each other (123, 312, 394, 422). Although comparably low in numbers (8% of all the hyphal tips of the juvenile hymenium become cystidia [423]), the solitary cystidia are very prominent on the gill surfaces due to their final length of about 60 μm (193, 344). The cystidia also accumulate glycogen (312). Cystidia are commonly believed to function as spacing organs to keep the hymenial surfaces apart during the development of basidia and basidiospores (54, 55, 193), to aid in evaporation of moisture and other volatile substances (439), and to enable ballistic spore release (55). Moore and coworkers suggest another role for cystidia in anchoring the gills to each other in order to provide a robust stabilization of the loosely arranged gill tissues and pull them into the typically radial-parallel arrangement due to the tension exerted by cap expansion (89, 179, 335, 337, 338). Cystidia are regularly spaced over the hymenium surface at both sides of the gills. As they grow, cystidia come into contact with the opposing gill face. Collision of the tips of the cystidia with cells of the opposite hymenium induces the latter to vacuolize and to adhere firmly to the cystidia. Cells whose differentiation is determined by attaching a cystidium have been named cystesia (179). Mutants lacking cystidium-cystesium connections have a distorted hymenium, and the cap rolls up prematurely (89). Initially in cap expansion, gills are still physically connected to the young stipe, but eventually, after onset of rapid stipe elongation and by the time sterigmata appear, the stipe-hymenium boundary breaks up, separating the stipe from the gills by a clear zone (338, 344, 394, 422). When primary gills pull away from the central stipe, marginal, often multinucleate cystidia arise from the multinucleate gill trama, probably to protect and stabilize the apex of the primary gills (90, 277). The cystidia on the edge of a gill are more oval than the solitary facial cystidia, which are mostly binucleate and only occasionally uninucleate or multinucleate (90, 193).

During fruiting-body maturation, there are no further cell divisions and most changes in the shape of the cap depend on cell expansion (338, 344). Cap expansion correlates temporally and proportionally with stipe elongation, which lifts up the cap for more effective spore dispersal (155, 344). At a size of 15 to 45 mm (stage 4 immature fruiting bodies [263, 344, 450] [Fig. 7]), about 3.5 h after sterigma production at the time of basidiospore pigmentation, hymenial cells become vacuolated and glycogen disperses from the subhymenial tissue (155, 310, 312, 344). Within a further 8 h, the fruiting body adopts its final size and the cap fully expands, with outward upward and curling of the initially vertical oriented gills (58, 263, 338, 344, 450) (Fig. 7). Cap expansion is associated with a considerable cellular uptake of water, and this is probably osmotically driven (123, 338, 344) (see below). Cap expansion remains constant and continues in the still intact tissues even in the final autodigestive phase, possibly by utilizing substrates provided by the already digested cells (338, 344). Autolysis is clearly part of the developmental program of fruiting-body formation (55, 58, 344) and starts at the edge of the gill closest to the stipe (422), like karyogamy, meiosis, spore formation, and pigmentation (338, 402, 424). Autolysis of the mature cap is caused by cell wall degradation by chitinases, which are released into the wall from vacuoles in senescing gills about 8 h after the spores begin to blacken and 7 h after rapid elongation and opening of the cap start. Autolysis affects all cap tissues, the lamellae, and the interior trama but not the outer veil layers. The chitinase is produced 2 h in advance and is also not effective on stipes or the vegetative mycelium (155, 184, 185). Apart from the chitinases, proteases and glucanases have also been identified in extracts and autolysates of C. cinereus caps (184, 323, 338).

Mutant Analysis

Variants in fruiting-body development were first isolated by UV and chemical mutagenesis in dikaryons. The yield of variants was surprisingly high, with about 15% of a total of ∼11,000 isolates displaying a change in pattern of development (461–463), of which a large proportion were temperature sensitive (466). Component nuclei of dikaryotic isolates have been found to carry naturally recessive alleles acting negatively on fruiting. This suggests that the mutations induced in dikaryons are not all dominant but, rather, that phenotypes are due to a mixed genetic background with naturally recessive and newly mutated alleles. The alleles present in the unmutated dikaryons obviously complement each other for fruiting (269, 330, 353). Amut Bmut homokaryons are able to fruit, and their use in isolating developmental mutants has a number of advantages over the use of dikaryons. First, it is clear that all genes essential for fruiting are present in an active form. Second, because of the presence of a single haploid genome, all possible mutations will be detected, regardless of whether they are dominant or recessive. Third, to ease the mutagenesis procedure, oidia with a single haploid nucleus can be used (388, 389). Because of these advantages, it is nowadays rather common to use Amut Bmut homokaryons and other fruiting homokaryons in UV and in REMI mutagenesis (45, 89, 101, 145, 213–215, 354, 389, 543; Kües, Granado, et al. unpublished) unless a phenotype can also be expected in the monokaryotic growth phase (206, 208, 209, 498, 538, 542). Following UV irradiation of Amut Bmut or CopD5 oidia (with a 10% survival rate), 30 to 40% of all germinated clones have a defect in fruiting, and similar percentages have been determined after REMI mutagenesis (145, 354; Kües, Granado, et al., unpublished). In comparison, auxotrophic mutations are rare, with ≤1% (354; Kües, Granado, et al., unpublished). The large number of developmental mutants in Amut Bmut homokaryons and in dikaryons suggests that development is affected by defects in various housekeeping genes in addition to fruiting-specific genes, although some of the mutants may also carry more than one mutation (354, 463; Kües, Granado, et al., unpublished). One-third to one-half of the mutants with defects in fruiting are also defective in mycelial growth and/or oidiation, and two-thirds of clones abnormal in oidiation are also defective in fruiting, reflecting the fact that regulation and execution of fruiting and oidiation are probably closely related (383; Kües, Granado, et al., unpublished).

Mutants with mutations in fruiting-body development have been categorized into eight major groups: mutants without hyphal knots (knt), mutants blocked after hyphal knot formation (prm), mutants with primordia that do not form mature fruiting bodies (mat), mutants with a short fruiting body where the stipes do not elongate (eln), mutants with fruiting bodies where the cap does not expand (exp), mutants with white fruiting bodies that do not form basidiospores (spo, sometimes also called bad), mutants with combined defects in expansion, elongation and/or sporulation, and other mutants that do not fit into these categories (214, 215, 354, 389, 463; Kües, Granado, et al., unpublished). The unclassified group of strains include such interesting mutants as those forming well-proportioned miniature fruiting bodies of 1 to 1.5 cm (dwarfs or bonsais) and mutants forming etiolated stipes in both the dark and the light (45, 215; Granado, personal communication). Our group has concentrated on analyzing knt and prim mutants (269; Kües, Granado, et al., unpublished). The defect in fruiting-body initiation of the first prm mutant has now been overcome by introduction of a gene cluster that seems to act in initial induction and primordium maturation (Liu, personal communication). Sequencing data suggest that the cluster contains a gene for a cyclopropane fatty acid synthase (a specific methyltransferase for membrane-bound unsaturated fatty acids postulated to exist only in bacteria [149]), a gene for a putative ADP-ribosylation factor (a member of the RAS superfamily of small GTPases involved in vesicle formation and trafficking [4, 33, 109, 426]), and a gene for a kinesin that may be linked to transport of vesicles and other organelles (157, 448). These putative gene functions promise to become very interesting, since the galectins have been localized in vesicles within the primordia and the young fruiting body and in the heterologous host Saccharomyces cerevisiae, galectins are secreted by a nonclassical pathway (Boulianne et al., submitted).

Temperature-sensitive mutants with defects in hyphal growth often also exhibit a defect in stipe elongation (323). In one temperature-sensitive hyphal growth mutant, the mutated gene hyt1, encoding a novel protein, has been identified, but in this case it is not known whether stipe elongation is also affected (206). However, a specific defect in stipe elongation has recently been complemented in an eln UV mutant. The gene isolated, eln2, is constitutively expressed independently of fruiting and was found to encode a cytochrome P450, with 31% amino acid identity to enzymes of the CYP64 family (354a). This family includes the oxidoreductases of the filamentous ascomycetes Aspergillus parasiticus and Aspergillus flavus that convert O-methylsterigmatocystin to aflatoxin (221, 387, 531). Another gene, ich1, has been cloned by complementation of a natural mat mutant with deformed “fig-shaped” primordia that are lacking a differentiated pileus. The ich1 gene is specifically expressed in the pileus. The protein is 1,353 amino acids (aa) long, but the last 590 aa, including a six-times-repeated sequence, are dispensable for function. The remaining part of the protein contains three positively charged sequences suggested to be nuclear localization signals (NLSs), and the protein is therefore believed to act in the nucleus, for example in the regulation of expression of other genes required for the development of the pileus (353). However, another interesting motif is present in the protein, with a sequence of IVDVGGGIA...15...E...94...ADFV at positions 483 to 605 (cf. the Ich1 protein sequence given in reference 353), which is typical of the S-adenosylmethionine binding domain in non-nucleic acid methyltransferases (154, 229). This motif therefore suggests a totally different enzymatic function. Most interestingly, a Blast search revealed that Ich1 has low identity (ca. 20%) to O-methyltransferases from A. parasiticus and A. flavus, which are involved in the conversion of sterigmatocystin to O-methylsterigmatocystin and of dihydrosterigmatocystin to dihydro-O-methylsterigmatocystin (220, 473, 529, 530). Aflatoxin has recently been connected to the development of sclerotia and conidiospores in Aspergillus; when aflatoxin production is missing, development is blocked, and when aflatoxin is overexpressed, development is increased (49, 152, 473, 474). It is therefore likely that aflatoxin plays a biological role in development and that one possible function would be in signaling. In C. cinereus, it is probably not the polyketide aflatoxin that is produced within the tissues of the primordia but some other, probably analogous substances. It is striking that the ich1 mutant fails to develop a differentiated pileus at the apex of the primordial shaft whereas in the eln2 mutant the tissues of the stipes are composed of meandering cells and are disturbed in their organization (353, 354a). ich1 dikaryons still form normal “etiolated” stipes in the dark (353), indicating that the mutant can reach the developmental stage at which a darkness-induced factor from the cap can bring about maturation (200). It is thus possible that Eln2 produces a metabolite that will signal the stipe to organize its tissues properly and, when further modified by Ich1 in the pileus, will determine further development of the cap and stipe. With this in mind, it will be interesting to test the behavior of an ich1 eln2 double mutant.

Much attention has been given in mutant analysis to the stages of karyogamy, meiosis, and basidiospore formation in fruiting-body development. Mutants with defects in meiosis and spore formation are easy to recognize by their white caps, since it is the mature basidiospores that color the cap of Coprinus black. Kanda et al. (214) identified at least six different genes acting at meta-anaphase I in meiosis in a collection of white cap mutants. Two other genes act in sterigma formation following meiosis, 13 genes are involved in prespore formation, and 4 genes contribute to nuclear migration into the spores (214). Other studies report defects in premeiotic DNA replication and meiotic prophase I, but the majority of defects appear to be localized at the stage of meta-anaphase I (216, 389, 396, 397, 405, 435, 466, 498, 502, 538, 539, 542). From γ-radiation-sensitive mutants exhibiting defects in chromosome condensation and synapsis in meiosis, an epistasis group of rad3-9-11-12 has been defined (405, 498, 538, 542). The first genes involved in meiosis and DNA repair (rad9, rad11/mre11, and spo11) have been isolated by mutant complementation and by a two-step, semirandom PCR recovering DNA inserts of a REMI mutant ((101, 130, 539, 542). Mre11 has a well-studied homolog in the yeast Saccharomyces cerevisiae which acts in double-strand break repair (83, 153, 349, 495). Spo11 of yeast is a topoisomerase II-like protein that is also involved in catalyzing meiotic double-strand breaks (219). Rad9 of C. cinereus, however is a novel, proline-rich protein required for meiotic chromosome condensation and synapsis (435), indicating that studying meiosis and recombination in C. cinereus will add new aspects to the cytological and mechanistic knowledge we have from studies with S. cerevisiae (for further information on meiotic recombination, see reviews in references 100, 380, 417–419, and 534). Homologous sequences to Rad9 from human (AB019494) and Drosophila (AF114160) are now deposited in GenBank. The latter, being a chromosomal adherin homolog, had been implicated in gene activation by remote enhancers (421). Rad9 is homologous to Mis4 of Schizosaccharomoyces pombe and Scc2 of S. cerevisiae, which are sister chromatid cohesion molecules (adherins) that prevent premature separation of sister chromatids in anaphase of mitosis (127, 321, 421). Rad51, homologous to RecA of Escherichia coli, has a central place in meiotic homologous recombination and functions in sequence identification, strand exchange, and branch migration (28, 151, 359, 467, 532). An orthologous gene of the S. cerevisiae RAD51 gene, Ccrad51, has been isolated from C. cinereus by a PCR approach using degenerate primers. The gene does not complement any of the rad3, rad9, rad11, and rad12 mutants and maps to chromosome 1, unlike these other genes, indicating that Ccrad51 represents a new gene (446, 542).

Mating-Type Loci

C. cinereus, like many other homobasidiomycetes, is an outbreeding organism with a tetrapolar incompatibility system determined by the two mating-type loci A and B. “Tetrapolar” refers to the fact that four different types of basidiospores arise from meiosis: AxBx, AxBy, AyBx, and AyBy, where x and y represent the different mating-type specificities of the parental haploid nuclei. Since the two mating-type loci must be different for a successful mating (see above), each spore has a 25% probability of finding a fully compatible partner within the tetrad it comes from. This reduction of outcrossing per tetrad is compensated for in the whole natural Coprinus population by acquisition of multiple specificities for each of the two mating-type loci. From C. cinereus isolates from the wild (102, 104), Raper (407) calculated that there are about 160 different A and 80 different B specificities, resulting in a total of 12,800 different mating types. However, estimates obtained from other samples are even greater, with 240 different specificities for each of the two loci, suggesting 57,600 different mating types (50, 64, 235). Even using the lower of the two estimates, a general mating probability of >98% and thus inbreeding repression is achieved (for extensive discussions, see the papers by Hiscock et al. [172–174]).

Other well-understood multiallelic mating-type systems are present in the tetrapolar basidiomycetes Ustilago maydis and S. commune. The A mating type locus of C. cinereus is homologous to the b mating type locus of U. maydis and to the A mating type locus of S. commune, whereas the B mating type locus of C. cinereus is homologous to the a locus of U. maydis and the B locus of S. commune (79, 172, 194, 245, 246, 496). It has been proposed to combine the A mating loci of C. cinereus and S. commune and the b locus of U. maydis under the unifying name matT and the B mating type loci of C. cinereus and S. commune and the a locus of U. maydis under the name matP (172).

Physical Factors

Environmental conditions including light, temperature, humidity and nutrients play a decisive role in determining the developmental pathways that Coprinus enters. On a superficial level, alternative developmental choices are possible. At each point of choice, environmental signals act, and light is probably the most important of these (Fig. 16). Light conditions will determine whether a dikaryon will form hyphal knots or oidiophores with oidia and whether the hyphal knots will mature into sclerotia or fruiting-body initials. Higher temperatures favor the development of oidia, and lower temperatures lead to fruiting-body initiation; the application of a cold shock can raise the absolute number of fruiting bodies per culture. Fruiting-body initiation and maturation need high humidity, whereas sclerotia and chlamydospores will also appear at low humidity (39, 40, 232, 261, 277, 293, 330, 352, 476, 506; Kües, unpublished; Kües, Granado, et al., unpublished). After light induction of initials, light continues to be the most important regulator: it suppresses the development of dark stipes and induces proper gill development of the cap tissues. At 48 h later in development, when gill formation is completed, light is required for induction of karyogamy at the stage of DNA synthesis prior to nuclear fusion (21, 184, 203, 277, 286, 350, 351, 476, 492) (Fig. 7). However, the situation is not as simple as this, since strains differ in their responses to light and since environmental signals other than light may influence the process. Light induction of fruiting bodies takes place at 20 to 37°C (275, 276, 476), but complementation of fruiting-body development needs more defined conditions. At 25 to 28°C, most but not all strains perform the whole program of fruiting in continuous light (21, 275, 276, 350, 506; Granado, personal communication), but at 35°C, a 2-h dark period is required before meiotic DNA replication can proceed and complete within the following 6 h (275, 286). Certain strains will form initials in the dark, depending on the amount of ammonium in the media (352, 476) or the age of the culture (292, 293). For some specific strains, an extra, dark-dependent 2.5- to 3-h step has been observed at prophase I before they enter metaphase of meiosis, where strong light and long exposure time lead to arrest of development and the formation of abortive basidiocarps (B. C. Lu, submitted for publication). Different workers used different stage identification and different starting points, making it difficult to recognize how many different regulated steps indeed occur during fruiting-body morphogenesis (21, 203, 277, 344, 351). It is thus likely that the extra light-sensitive phase at prophase I corresponds to a light-sensitive phase at a later stage of development mentioned before in other papers (203, 324; Lu, submitted). Altogether, this yields at least five light-sensitive phases in the course of developing a mature fruiting body: the negative effect on hyphal knot formation, at least three positive effects to induce the formation of initials, maturation of primordia, and karyogamy, and another negative effect on completion of meiosis. However, light illumination before the extra-sensitive dark period at prophase I slows development and illumination after this phase accelerates development, indicating the possible presence of more light-sensitive steps (203).

Light in C. cinereus acts only on already established mycelium, and light response signals are not transmitted through the mycelia. Light effects are therefore local and not systemic events (232, 293). In an Amut Bmut homokaryon, a brief illumination of 1 min was found sufficient to induce the development of oidia and only 0.1 μE m² s⁻¹ was needed for maximum production of oidia (232). Similarly, light requirements in fruiting are low. A 1-s light exposure and 1 ft-candle h (0.98 lx) is enough to induce fruiting (277, 293), and a 30-s exposure to 500 lx has been shown to induce karyogamy and basidiocarp maturation (203, 350). The active wavelengths in all light-controlled reactions, irrespective of whether they are negatively or positively affected, are in the blue region (400 to 500 nm) of the spectrum (40, 232, 261, 277, 278, 350, 476). The identity of the light receptor(s) in C. cinereus is not known, but from action spectra determined for other basidiomycetes (111, 112, 527, 528), blue light receptors are predicted to be flavoproteins. A number of blue light flavin-binding receptors (cryptochromes) have been found in plants and animals that have particular amino acid sequence identities to subgroups of bacterial and eukaryotic photolyases (67). Attempts to isolate a homologous Coprinus photoreceptor using primers deduced from plant cryptochromes and the closely related bacterial type II photolyases unfortunately failed (Bottoli, personal communication). However, comparison of plant and animal cryptochromes and the three major subgroups of photolyases suggested that there was repeated evolution of cryptochromes, in which different lineages arose independently from orthologous genes (67). It is thus possible that in Coprinus cryptochromes from other ancestries exist, possibly more closely related to the animal cryptochromes, since fungi are closer to animals than to plants (18). Moreover, other flavoproteins with photoreceptor properties are being discovered in plants (92), and the red/far-red photoreceptors (phytochromes) can also function in blue light responses (370, 398). Other candidates for blue light receptors have been proposed for the filamentous ascomycete N. crassa with the white collar Zn finger transcription factors WC1 and WC2. Both proteins contain PAS (PER-ARNT-SIM) dimerization domains that have identity to the xanthopsin-containing blue light photoreceptor PYP (photoactive yellow protein) from the halophilic purple bacterium Ectothiorodospira halophila (19, 20, 268). A further potential photoreceptor in N. crassa is a recently discovered rhodopsin-like protein (29).

Another fungal model system for blue light photoreception is the zygomycete Phycomyces blakesleanus, in which phototropism of sporangiophores (reproductive structures containing mitotic endospores) is used to monitor light response (128, 129). Mutant analysis in this fungus indicates that the photoreceptor system is also required for gravitropism (66). In the Coprini, the growing primordia and young fruiting bodies show a positive phototrophism before a negative geotrophic response takes over (53, 60, 241). Gravitropism is a postmeiotic event. It develops after an initial period of light-seeking growth in the early stages of fruiting-body development at the point of completion of meiosis (53, 55, 60, 233, 233, 347), but it is not clear whether the two types of tropism are also linked genetically in Coprinus. Clinostat experiments with C. cinereus and space laboratory experiments with other mushrooms (Flammulina velutipes and Polyporus brumalis) revealed that the basic form of a mushroom, i.e., the overall tissue arrangement of stem, cap, gills, hymenium, and veil is established independently from a gravity vector (162, 230, 336), although without gravity, stipes can be abnormally helically twisted (230). In the absence of a gravity vector, all C. cinereus tissues remain underdeveloped and fruiting-body development stops prior to meiosis, indicating that commitment to the meiosis-sporulation pathway requires the normal gravity vector (336); however, this is not the case in all fungi (229).

Fungus-like plant graviresponse systems divide into reception, transduction, and realization (320a, 347, 451). As in other agarics (53), the negative gravity response of the C. cinereus stipe is very quick. It takes a 7- to 10-min exposure to note a change in the gravity vector and about a 25-min reaction time for the first visible response (bending of the stem), which is strong (51° h⁻¹) within the first 30 to 60 min and completed in 2 to 3 h. Reaction times do not depend on the exposure time, and exposures are not additive, indicating that graviresponse is an “all-or-nothing” reaction. After reaching the maximum curvature, bending is still reversible, but if the tropic stimulus maintains, bending is rigid, indicating that gravitropic bending is a two-stage process (146, 147, 162, 163, 233, 320, 347). It is not exactly known when, where, and how gravity is sensed, but perception and response probably occur in the same tissue (163, 336). The response in horizontally placed stipes is most intense just behind the apex of the stipe and is almost absent at the apex and the base. More than half of the upper stipe can be removed without affecting the ability of the remaining stem to show gravitropic bending or its ability to compensate for curvature so induced and adjust to the vertical. However, the greater the part eliminated, the more time is needed to respond, suggesting that graviperception is gradually reduced in zones further from the apex (146, 147, 319). Gravisensing in fungi is thought to be simple and to occur, for example, just by a static gravistimulus mediated by the fungal mass or by a dynamic stimulus caused by alterations of the rate of cytoplasmic streaming and the slow sedimentation of organelles (such as the nucleus) acting as statoliths (24, 347). In C. cinereus, extracellular mass and/or mechanical stress have been excluded as the cause of the gravitrophic response (147). Geotropism of the stipe has been suggested to relate to a displaced distribution of the vacuole in relation to the cytoplasm, caused by a rearrangement of the cytoskeleton (142), but so far this has not been substantiated (347). Delayed and diminished gravitropic bending of stipes has been observed after treatment with the cytoskeletal inhibitor cytochalasin B, suggesting that the cytoskeleton indeed contributes to gravity perception. Since inhibitors of microtubule polymerization had no effect on gravitropism, the membrane connections of the actin microfilaments may mediate communication (146, 347, 369). In contrast, experiments with antagonists of Ca²⁺ transport and inhibitors of calmodulin have led to the conclusion that, unlike gravitropic responses in plant roots, Ca²⁺ acts in C. cinereus not in perception but in signal transduction (336, 368, 369). Ca²⁺-mediated signal transduction may be involved in directing growth differentials, since the gravitropic bending of the stipe is the result of an asymmetric distribution of longitudinal cell expansion. Neither cell cross-sectional area nor cell size population structure change during bending, indicating that the growth mechanism of stipe bending is different from that of normal vertical stipe growth (148, 336, 347, 369).

Physiological Factors

C. cinereus initiates fruiting when nutrients are depleted (338, 406, 450), but only the youngest mycelium maintains fruiting competence (293, 350, 352; Kües, Granado, et al., unpublished); when a culture is too old at the point of light exposure, mycelial growth and consequently fruiting can be induced by injuring the established mycelium (145; Kües, Granado, et al., unpublished). Mycelial age also determines the execution of light-induced oidiation (232, 262), indicating a need for physiological activity to perform developmental programs upon exposure to environmental signals. Increasing concentrations of C and N sources act negatively on both fruiting-body formation and sclerotium maturation (292, 342, 476). This is also observed in the expression of galectins (Boulianne et al., submitted) and, to a lesser extent, in the production of oidia. In oidiation, absolute spore numbers are influenced by nutritional conditions (508a) but, as tested in an Amut Bmut homokaryon, light perception per se is not affected by nutrients (232, 508a). In principle, it is not so much an overexcess that is inhibitory, but development is particularly sensitive to balanced C/N ratios (292, 338, 352, 476). Ammonium is released by the fungus during the early stages of fruiting-body development (450), possibly to delay sclerotium maturation, which is sensitive to ammonium (476). However, the addition of free ammonium to competent mycelia also stimulates fruiting (352). Analysis of physiological effects on development have concentrated on fruiting-body formation (338). Fruiting-body initiation is inhibited by glucose and glucose analogues (302, 372, 476, 484, 485, 493). cAMP production is glucose repressed (484, 485, 493), but addition of cAMP overcomes a block in hyphal knot formation (302), initial formation, and fruiting-body maturation (480, 481, 491). High activities of adenylate cyclase and low levels of phosphodiesterase and, consequently, accumulation of cAMP were detected in light-incubated mycelia of fis^c strains and of dikaryons that form fruiting bodies but not in mycelia that do not fruit (262, 482, 491, 492). cAMP production by adenylate cyclase follows a typical pattern. The cAMP concentration in the mycelium rises slightly in the light before or during the onset of fruiting and stays high for 3 to 4 days. It then decreases rapidly due to a fall in adenylate cyclase activity and an increase in phosphodiesterase activity during primordium maturation (262, 484, 492), just before or at the time of karyogamy and early meiosis (344, 480, 482, 485). Isolated fruiting-body initials have a high cAMP content, but with primordium formation, the level of cAMP decreases. Although the levels are not as high as in the initials, the cAMP level rises again in both cap and stipe as meiosis proceeds, and it drastically drops once more in the mature fruiting body (263, 344). Accordingly, treatment of isolated basidia with cAMP allows karyogamy but blocks the process of meiosis (89). Generally, cAMP stimulates protein phosphorylation and activates a number of protein kinases (483, 484, 486, 488, 489, 491) but other protein kinases present in the mycelium are inhibited (483, 490). To really understand the actions of cAMP in developmental processes, it is necessary to use molecular biological techniques. A first step in this direction is the analysis of the adenylase cyclase gene cac (41). Expression of the gene is hardly influenced at all by external and internal signals (Bottoli, personal communication), indicating that enzyme activity is regulated by posttranslational activation. One possible candidate for such a regulator of adenylase cyclase is Ras (136, 322, 468). A ras gene has also been isolated from C. cinereus (41, 183), and its effects on development are being studied (Bottoli, personal communication).

Despite the relatively extensive studies on cAMP and glucose regulation, there are few other individual observations linked to induction of fruiting-body formation. Phenol-oxidase activities have repeatedly been related to fruiting-body initiation and maturation in basidiomycetes, but the relevance of this is not clear (417, 518). In C. cinereus, a phenol-oxidase (tyrosinase) is present in the mycelium without induction and another phenol-oxidase is specifically induced when fruiting is inhibited (372). A third enzyme, a laccase, is induced at fruiting-body initiation, and its level rises dramatically in activity during meiosis and spore formation and then decreases after spore maturation. Analysis of enzyme activity in spo and exp mutants suggests that this laccase contributes to the black pigmation of basidiospores (505). Six different laccase genes have been found in C. cinereus (Fig. 17). mRNA expression studies might show whether one of these is correlated with fruiting-body initiation or later stages in fruiting.

Certain cerebrosides produced in the fruiting bodies of, for example, S. commune and Ganoderma lucidum will cause fruiting-body formation when applied to dikaryotic mycelium of C. cinereus (325). The cerebrosides, sphingosine-based glycolipids, have been shown to repress replicative DNA polymerases (325), but how this can relate to fruiting-body induction is not obvious. Implanting from stipes, regardless of whether they are from light- or dark-grown initials, causes renewed fruiting. This response is a very rapid one, occurring 2 to 3 days earlier than on vegetative mycelium, and a new generation of basidiomes is formed at the implants (51, 86, 476, 513). Thus, there must be a memory function for fruiting within the stipe tissues; cerebrosides may play a role in this. The potential for renewed fruiting resides especially in the parts of the fruitbody with accumulations of intracellular glycogen. This may reflect the need to attain a certain threshold level within the tissue before fruiting can be expressed (51). Nutritional transport in the vegetative mycelium of C. cinereus is very slow (375) and, in nature, probably even mostly blocked over longer distances by incompatibility borders in mycelial communities (520). It is therefore very possible that at least the initial stages of fruiting-body formation rely on energy and carbon sources stored in place rather than on substance transport over longer distances. Accordingly, accumulation of glycogen in inflated cells of the vegetative mycelium is also associated with fruiting-body production (294). However, only specific primordia of a culture will mature whereas most others fail, probably because young fruiting bodies initiate a flow of nutritions in their direction in preference to others (292, 335, 338). As an alternative to fruiting-body formation, glycogen in the swollen vegetative cells acts as a mobilizable source of carbon reserve for sclerotium development (342, 511).

Glycogen phosphorylase activity is stimulated (483, 484, 486, 488–491) and glycogen synthetase is inhibited (483, 487, 489, 491, 492) by cAMP. At early stages of fruiting-body development, degradation of glycogen and translocation from the mycelium to the developing fruiting-bodies have been demonstrated and are correlated with the high cAMP levels (188, 292, 294, 338). Glycogen first accumulates in the stipe tissue and then relocates to the cap of the primordium, especially into the subhymenial tissues, and is almost totally utilized in the mature fruiting body (23, 142, 312, 332, 338, 344). Glycogen phosphorylase activity in the cap increases after karyogamy at the stage of sterigma formation; the onset of large-scale utilization of accumulated glycogen is a postmeiotic event. However, inhibition studies with ammonium and analysis of white cap mutants indicated that glycogen degradation and sporulation are not closely coupled (188, 345). Glycogen synthetase activity is low during spore production but increases during basidiospore maturation (188), obviously to produce the glycogen that is found in the mature spores (37, 169, 312).

Maturation of the cap is accompanied by a specific pattern of changes in enzyme activities and metabolite levels. C. cinereus has two forms of glutamate dehydrogenase (5, 345). The NAD-linked enzyme is most active during vegetative growth but is also present in the stipe and cap of the fruiting body. In contrast, the NADP-dependent glutamate dehydrogenase occurs at high specific activity in cap cells of fruiting bodies but is barely detectable in stipe cells and the vegetative mycelium (450). Production of the NADP-dependent enzyme is initiated at or very soon after karyogamy at the early stage of meiosis (323, 345). Urea and urea cycle intermediates activate the NAD-linked enzyme and inactivate the NADP-linked dehydrogenase, indicating a link to the urea cycle (5, 332, 450). Glutamate dehydrogenase, glutamine synthase, ornithine acetyltransferase, and ornithine carbamyltransferase are other enzymes associated with the urea cycle and ammonium metabolism that are highly expressed in the cap but not in the stipe. In contrast, urease activity is present within the stipe and the vegetative mycelium but not in the cap (123, 332, 341). During cap development, synthesis of urea is amplified and water is driven osmotically in cells in which urea accumulates. The urea concentration in cap cells therefore remains unchanged, and it is thought that this mechanism contributes to the inflation of hymenial cells (332). The cap always contains less free ammonium than does the stipe (123, 332); this is especially important for meiosis and spore formation, which are inhibited by free ammonium and also by glutamine (87, 89, 346). It is possible that the NADP-dependent glutamate dehydrogenase and glutamine synthase act together as an ammonium-scavenging system to protect the meiotic apparatus from ammonium inhibition (119, 123, 331, 333). When ammonia is limiting, acetyl coenzyme A accumulates, and this appears to be necessary for induction of NADP-dependent glutamate dehydrogenase. Consistent with acetyl coenzyme A accumulation, the tricarboxylic acid cycle is highly active in the cap, as indicated by a strong increase in succinate dehydrogenase and isocitrate dehydrogenase activities (332). However, there is no evidence for 2-oxoglutarate dehydrogenase, and it is assumed that the 2-oxoglutarate dehydrogenase step is bypassed by a glutamate decarboxylation loop (332, 338, 341).