7.9 Cytoplasmic segregations: mitochondria, plasmids, viruses and prions

There are a number of cytoplasmic factors that affect the fungal phenotype and which depend for their transmission upon some of the features of heterokaryosis we have described so far. Chief among these are the mitochondria. Mitochondrial genomes are independent of and quite distinct from the nuclear genome. Loss of function mutations in mitochondrial genes result in characteristic phenotypes in both Neurospora crassa and Saccharomyces cerevisiae. Mitochondrial DNA (mtDNA) of yeast usually makes up 18% of the total DNA, but it has a distinctive very high AT content (82%) so is relatively easy to separate from chromosomal DNA. The mtDNA is circular, of 25 nm circumference, and comprises about 7.5 × 105 base pairs. In yeast, mtDNA codes for three of seven polypeptides of the cytochrome c complex (the rest derive from nuclear genes), four polypeptide components of a mitochondrial ATPase, and one component of cytochrome b. Mutations in these genes can produce recognisable respiratory deficiency phenotypes (for example, mutants named petite in S. cerevisiae, poky in Neurospora crassa) and thereby provide mitochondrial mutants. Also, although chromosomal genes code mitochondrial ribosomal proteins, mtDNA determines mitochondrial ribosomal RNA (rRNA) and transfer RNAs (tRNAs).

Mitochondrial ribosomes are similar in size to prokaryotic ribosomes and share some other prokaryotic properties; in particular, protein synthesis on mitochondrial ribosomes is inhibited by chloramphenicol, erythromycin and several other antibacterials that have no effect on cytoplasmic (eukaryotic) ribosomes. Consequently, another kind of mutant phenotype due to mitochondrial mutation is resistance to inhibition by mitochondrion-specific drugs. Mitochondrial gene sequences are also similar to equivalent genes in prokaryotes. These features encouraged the endosymbiont theory of mitochondrial origin, which envisages mitochondria to be relics of ancient bacteria-like organisms that formed a symbiotic association that resulted in the ancestral ‘eukaryotic’ cell.

S. cerevisiae can have over 100 genomes per mitochondrion, corresponding to about 6500 in each yeast cell; some of the mitochondrial genomes are circular, others are linear. Oddly enough, when segregation of organelle genes is followed in genetic crosses, the segregation patterns are consistent with there being only one copy of the mitochondrial genome in the cell. The fact that this is clearly not the case indicates that we do not fully understand how organelle genomes are transmitted from parent to offspring. The basic procedure for making mitochondrial crosses involves making heterokaryons (diploids in yeast) between haploids carrying mitochondrial markers. After some vegetative growth of the heterokaryon or diploid, diploid daughter cells (of yeast) or spores or hyphal fragments (of filamentous fungi) are plated out and the resulting colonies are scored for the mitochondrial markers present in the original haploids.

In Neurospora crassa, phenotypes controlled by mitochondrial genes are generally transmitted through the female, that is the protoperithecial parent. This is vertical mitochondrial transmission, from one generation to the next. Horizontal transmission, which is between individuals of the same generation, occurs as a result of hyphal fusion. Although complementation and recombination can be detected between mitochondrial genomes, mycelia containing genetically different mitochondria (called heteroplasmons or heteroplasmic mixtures) tend to segregate the different mitochondria into different cells.

Uniparental inheritance of mitochondrial phenotypes has been observed in yeast, which is isogamous (does not show a male/female differentiation). The mechanism is unknown, but mitochondrial genomes can segregate in association with mitosis and the consequential bud formation. In filamentous fungi, mitochondria are not closely associated with the mitotic spindle, so vegetative segregation may simply be a matter of random physical sorting, though nuclear and mitochondrial genes do influence mitochondrial genome transmission and it is also affected by membrane chemistry. In some Ascomycota sub-cultured for a long time in the laboratory, altered mitochondrial DNAs due to molecular rearrangements have been associated with modifications in mycelial growth in N. crassa and N. intermedia, and in growth of yeast.

Amongst the Basidiomycota, mitochondrial inheritance has been studied in matings of Schizophyllum commune, Agaricus bitorquis, A. brunnescens, Coprinopsis cinerea, Lentinula edodes, Pleurotus ostreatus, plant pathogenic Armillaria species and Ustilago violacea, and in all these cases, hyphal anastomoses result in the production of mycelial colonies composed of sectors containing different mitochondrial DNAs (mitochondrial mosaics). In Agaricus bitorquis, A. bisporus, Armillaria bulbosa, P. ostreatus and U. violacea, dikaryons had mixed populations of mitochondria, which sometimes resulted in recombination between mitochondrial genomes; recombinant mitochondrial DNAs have also been recovered from dikaryons and dikaryotic protoplasts of C. cinerea.

An increasingly important aspect of mitochondrial transmission is their content of plasmids. Isolates of Neurospora from nature commonly contain both linear and circular mitochondrial plasmids. Most are cryptic (that is, neutral) passengers, but some linear plasmids (notably of Podospora anserina) insert into mitochondrial DNA and cause mycelial senescence. Most linear plasmids exhibit typical virus characteristics as far as structure, replication and function are concerned even to the extent that plasmid‑free strains may contain plasmid remnants integrated into their mitochondrial DNA. Plasmid DNA sequences generally encode an RNA polymerase and DNA polymerase, or reverse transcriptase which are used to maintain and propagate the plasmid.

However, plasmid DNA is responsible for the killer phenomenon in the yeast Kluyveromyces lactis by coding for a killer toxin, which kills cells lacking the plasmid (cells hosting the killer plasmid are immune to the toxin). These plasmids reside in the cytoplasm and have an expression system independent of both nucleus and mitochondrion. K. lactis plasmids can be transferred to other yeasts (including Saccharomyces cerevisiae), conferring the killer/immunity phenotype. This shows that the plasmids are autonomous replicons, which can be expressed, in a wide range of host yeasts. The K. lactis killer plasmid toxin is chemically and functionally different from a killer toxin produced in S. cerevisiae, which is encoded by a double-stranded RNA (dsRNA) virus.

Virus‑like particles (VLPs) have been observed in electronmicrographs of many fungi. They are very similar in appearance to small spherical RNA viruses, but there is little evidence that these particles are effective in hypha‑to‑hypha infection. Many of the observed VLPs are presumably degenerate or defective viruses that can only be transmitted by hyphal fusions. No vectors are known for fungal viruses; transmission seems to depend on hyphal fusions. Unexpectedly, virus infections of fungi usually cause no recognisable phenotype. The exception is a mycovirus of Agaricus bisporus, which causes La France disease and ruins the crop. Diseased crops contain three virus particles and require up to 10 different RNA molecules to produce infective particles; as though some are defective viruses and others are helper viruses, or perhaps different viruses perform complementary, but essential, functions.

S. cerevisiae also harbours five retrovirus‑like elements, Ty1 to Ty5 as transposons able to integrate into the nuclear genome by targeting particular chromatin structures. The first cytoplasmic plasmid to be observed is the so-called two‑micron DNA of Saccharomyces (Ty1). The name refers to the contour length of the circular DNA molecules in electronmicrographs; it has a base composition similar to nuclear DNA and quite different from mtDNA. There can be 50 to 100 two‑micron DNA molecules per diploid cell, amounting to something like 3% of the nuclear DNA. The two-micron DNA molecules are transmitted to buds independently of both nuclei and mitochondria. The two‑micron circular DNA carries inverted repeat sequences at either end of two different unique sequence segments; this structure implies that it inserts itself as a whole into the yeast chromosome.

So far we have described nucleic acid molecules that encode features segregating in the cytoplasm. In the final decade of the twentieth century, however, great attention was given (and continues to be given) to a proteinaceous hereditary element, called a prion protein. The attention devoted to prions derives from their ability to cause diseases in mammals: scrapie in sheep, bovine spongiform encephalopathy (BSE) in cattle; and in humans, kuru and, most important recently, new variant Creutzfeldt-Jakob disease (nvCJD). In these cases, the pathogenic agent is a variant of a normal membrane protein (the prion protein) that is encoded in the mammalian genome. The variant prion protein folds abnormally and in addition causes normal prion proteins to fold abnormally so that the proteins aggregate in the central nervous system and cause the encephalopathy. Fungi also have prion proteins.
We have already referred briefly to the infectious het-s phenotype of Podospora anserina as a prion protein. Another candidate is the PSI+ form of the Sup35p protein in Saccharomyces cerevisiae. Sup35p is an essential yeast protein involved in the termination of translation. In the PSI+ state, Sup35p adopts a structural conformation that causes it to direct the refolding of native molecules into a form that can aggregate into filaments of discarded nonfunctional protein. This depletes the cytoplasm of functional translation terminator and results in translation errors. This is the PSI+ phenotype, which is inherited by daughter cells following budding and is infectious following cell fusion, in which case it propagates by autocatalytic conversion of the normal form of the protein. The part of the Sup35p protein that makes it a prion (the prion determining domain) is a glutamine/asparagine-rich amino-terminal region that contains several oligopeptide repeats. Removal of these repeats eliminates the ability of Sup35p to propagate PSI+, and expanding the repeat region increases the spontaneous occurrence of PSI+. Although deleting the analogous repeats from BSE prion protein does not prevent prion propagation and transmission in experimental mice, expansion of the repeat region does increase the spontaneous appearance of spongiform encephalopathies by several orders of magnitude in humans. It is likely that the oligopeptide repeats give the prion protein the intrinsic tendency to acquire a conformation that enables the protein to refold and effectively polymerise with sister molecules. Database searches for regions with amino acid content comparable to the yeast prions has revealed numerous such domains in eukaryotes, but these are lacking from prokaryotes (Silar & Daboussi, 1999; Serio & Lindquist, 2000; Pál, 2001).

Updated December 17, 2016