Publications:

Hallen HE, H Luo, JS Scott-Craig, JD Walton (2007) Gene family encoding the toxins of lethal Amanita mushrooms. Proceedings of the National Academy of Sciences USA doi/10.1073/pnas.0707340104. Abstract.

Hallen HE, M Huebner, S-H Shiu, U Gueldener, F Trail (2007) Gene expression shifts during perithecium development in Gibberella zeae (anamorph Fusarium graminearum), with particular emphasis on ion transport proteins. Fungal Genetics and Biology 44:1146-1156. Abstract.

Adams RI, HE Hallen, A Pringle (2006) Primer Note: Using the incomplete genome of the ectomycorrhizal fungus Amanita bisporigera to identify molecular polymorphisms in the related Amanita phalloides. Molecular Ecology Notes 6:218-220. Abstract. PDF.

Trail F, I Gaffoor, JC Guenther, HE Hallen (2005) Using genomics to understand the disease cycle of Fusarium graminearum. Canadian Journal of Plant Pathology 27:486-498. Abstract.

Hallen HE, R Watling, GC Adams (2003) Taxonomy and toxicity of Conocybe lactea and related species. Mycological Research 107:969-979. Abstract. PDF.

Hallen HE, GC Adams, A Eicker (2002) Amatoxins and phallotoxins in indigenous and introduced South African Amanita species. South African Journal of Botany 68:322-326. Abstract. PDF.

Hallen HE, GC Adams (2002) Don't Pick Poison! - When gathering mushrooms for food in Michigan. Michigan State University Extension Bulletin E-2777.

Amanita and amatoxins

Research: background, questions and approaches
Amatoxins are bicyclic octapeptides, produced by certain species of mushrooms in the genera Amanita, Galerina, Lepiota and Conocybe. Alpha- and beta-amanitin are the primary amatoxins in Amanita, and differ by the presence of the amino acid asparagine in alpha-amanitin while beta-amanitin contains aspartic acid in the same position. Alpha- and gamma-amanitin predominate in Lepiota and Galerina; these toxins have the same amino acid composition, but differ in the degree of hydroxylation.  Alpha-, beta- and gamma-amanitin, as well as some less prevalent amatoxins, bind tightly to RNA polymerase II, the enzyme that transcribes DNA into messenger RNA (mRNA), which is in turn translated into proteins. Amatoxins are extremely toxic, with a human LD₅₀ estimated at 0.1 mg/kg body weight. One bite of the Death Cap (Amanita phalloides) or Destroying Angel (A. bisporigera, A. ocreata, A. virosa and allies) can contain a lethal dose.

Among the most interesting questions raised by amatoxins (interesting to me, at any rate) are those dealing with evolution and ecology. Amatoxins are by nature slow poisons. In order for poisoning to occur, the toxins are actively imported into cells in the gastrointestinal tract and the liver (and any other cells; GI tract and liver are hardest hit because they tend to be the first organs encountering an ingested toxin). Once inside the cells, the toxins must bind to RNA polymerase II in sufficient quantities to incapacitate that enzyme, and existing mRNA and protein pools in the cells must be used up. All this takes time; a minimum of six hours in a human before the initial symptoms of severe gastrointestinal distress are felt. However, many toxins act much more rapidly: insect and amphibian toxins taste bad, and the animals are spit out promptly, ideally before a predator has done them too much damage. Dumbcane (Diffenbachia), a common houseplant, contains crystals of toxic oxalic acid that instantly cause pain and swelling in the mouth of any creature that tries to eat the plant; again, the toxin acts before anyone can do the plant serious damage. In either case, would-be predators or herbivores quickly learn to associate the poisonous plant or animal with discomfort, and learn not to eat it again. In a tasty* mushroom that takes at least six hours (more often 12-24; rarely as long as 36) to produce any effects, is anybody even going to associate the mushroom with the future distress? Can a delayed-action toxin serve as a defense mechanism?

*Poisonous Amanita species have been reported to taste very good - by those who survived to report.

Below: Representatives of the four amatoxin-containing genera of fungi, and the structural formula of a-amanitin, one of the principle amatoxins. A) Amanita bisporigera; B) Lepiota subincarnata; C) Conocybe filaris (photo courtesy Mike Wood); D) Structural formula of a-amanitin; E) Galerina marginata.


 
    The answer to that question is a definite "maybe". Mushrooms have many enemies, not all of them mammals. Insects and insect larvae (fungus gnats, beetle larvae, etc.) colonize many mushrooms, growing to maturity inside the fungus. In a case like this, where there is prolonged contact between the animal and the fungus, amatoxin could play a role in discouraging predation. It so happens that amatoxin-producing Amanita species are occasionally colonized by insects - but some insects, such as fruit flies, are known to develop amatoxin-resistance mutations when reared in the presence of amatoxin. Amatoxin could be a biological control that succeeds in keeping out most unwanted insect tenants, but has been overcome by a few strains that have developed resistance. Another intriguing possibility is that amatoxin plays a role in discouraging parasitic fungi. Several Amanita species in subgenus Lepidella, section Validae are attacked by an ascomycete parasite, Hypomyces hyalinus. The amatoxin-producing Amanita belong to subgenus Lepidella, section Phalloideae, which is the sister group to Validae. Is amatoxin protecting its producers from parasites?

    In searching for the gene(s) responsible for amatoxin synthesis, we had to start with the known chemical structure of amatoxins. Small, cyclic peptides with unusual amino acids (dihydroxy isoleucine, for example) that occur as part of a family of related molecules often suggest a non-ribosomal biosynthetic pathway. Non-ribosomal peptide synthetases (NRPSs) make several fungal secondary metabolites, including many antibiotics, and plant and animal toxins. NRPS genes have received a lot of attention in recent years, and have been characterized from a number of fungi and bacteria. Some characteristics of these genes are 1) Modular construction, with one module in the NRPS gene corresponding to one module in the peptide end product (so an amatoxin synthetase would be expected to have eight modules, one for each amino acid); 2) Large size, with each module approaching 4 kb in size (so an amatoxin synthetase would be on the order of 30 kb); 3) Few or no introns; and 4) Multiple conserved motifs in each module.

    Much of my research proceeded on the theory that amatoxins are, in fact, synthesized by a non-ribosomal peptide synthetase. I approached this from a variety of angles: A PCR-based approach involving the design of degenerate primers based on conserved NRPS motifs; a biochemical approach based on ATP-pyrophosphate exchange assays; and, most recently, a shotgun genome sequencing project of Amanita bisporigera. "Amatoxins are produced by a non-ribosomal peptide synthetase" is a testable hypothesis, and we performed numerous tests. As it happens, we were wrong; amatoxins are encoded ribosomally, as part of a larger protein which appears to be cleaved twice, at conserved proline residues, to release the amatoxin precursor. Interestingly, phallotoxins, actin-binding bicyclic heptapeptides (remember, amatoxins are bicyclic octapeptides) turn out to be encoded by very closely related genes.

Fusarium graminearum genome project

Background
Fusarium graminearum (more appropriately known by its holomorph name Gibberella zeae - Fusarium is a name specific to the anamorph, or asexual, stage of the fungus) is a serious pathogen of a number of cultivated crops: maize, wheat, barley, rice... A principle disease, head blight of wheat, is estimated to have caused more than $3 billion in losses in the US alone over the past decade. In addition to the reduction in yield (damage to the plant in the field), Gibberella zeae is a serious storage problem, producing mycotoxins in stored grain that render the grain unusable. Zearelanone is a polyketide toxin that mimics estrogen. It can cause feminization of male animals, or spontaneous abortions (along with other problems) in females. Deoxynivalenol (DON) is also known as vomitoxins, which may provide some clue as to its effects. DON is considered an antifeedant; animals that eat grain which promptly makes them sick are understandably reluctant to eat any more.

Research
To the scientist, Gibberella zeae possesses some good qualities as well: it can serve as an excellent model organism. G. zeae is homothallic (self-fertile) and can undergo its sexual reproductive cycle without needing to be mated with a different strain. Furthermore, it can be induced very readily to undergo sexual development in the lab. It grows relatively quickly. It is amenable to genetic manipulations; genes of interest can be removed or modified without undue difficulty. The entire genome has been sequenced by Broad Institute and the USDA - a crucial asset to identifying genes of interest for the aforementioned genetic manipulations. And an Affymetrix GeneChip exists for microarray analysis (at present, the GeneChip is in its first generation, and is only available to the three collaborating labs responsible for its development).
Why are these good and useful features? - It all ties together. Our lab is interested in sexual development and spore discharge in ascomycete fungi. Ascomycetes, such as G. zeae, normally shoot their spores off forcibly. In the field, plants are primarily infected by G. zeae in the Spring, when the ascospores (products of sexual reproduction) are shot off. Conidia (asexual spores) are produced abundantly year-round, but are not as important in helping the fungus spread from plant to plant. So, if you want to minimize disease spread, it would be handy to reduce the production and dispersal of ascospores. With a genome sequence, we can look for sequences similar to those for genes that have been implicated in spore discharge in other fungi, or for genes that we otherwise deduce might be involved in spore discharge. With a genetically-tractible organism, we can use the genome information to very specifically disrupt or remove only those genes we are interested in and see what happens. I, personally, have produced two distinct mutants incapable of shooting off their ascospores using these methods, and others in the Trail lab have produced other discharge-minus mutants. Does this stop Gibberella from infecting plants? No. But by understanding how Gibberella launches its spores and spreads, we are potentially generating targets for new generations of fungicides, chemicals that could be used very specifically against Giberrella zeae without harming innocent fungi. And we are certainly increasing our knowledge and understanding of a large number of understudied organisms.

Another branch of our research into Gibberella zeae is the Affymetrix microarray project. Very briefly, a microarray is a small area onto which pieces of DNA from an organism have been spotted. This can be done in several ways. For the Fusarium GeneChip, Affymetrix took the genomic sequence produced by Broad Institute, the gene annotations produced by Broad and the Munich Information Centre for Protein Sequences (MIPS), and any additional annotations provided by the project PIs (i.e. EST data). Affymetrix then synthesized probes 25 bases in length, corresponding exactly to the DNA sequence. As a control, for each perfect match probe there is a mismatch counterpart, in which the base in the middle (the 13th base) differs from the actual sequence of the organism. Each gene, predicted gene or EST gets between ten and twenty probes. These ten-to-twenty probes, plus their mismatch counterparts, are one probe set. There are 18,030 probe sets on the Fusarium gene chip, all spotted onto an area one square centimeter in size.
My part in the GeneChip project is the sexual development time course. I grow Gibberella zeae out on a petri plate and induce it to undergo its sexual developmental cycle. The moment of induction is considered the zero hour (0H) time point. I collect fungal tissue at 0H (vegetative hyphae), 24H (wide, dikaryotic hyphae are present), 48H (perithecial initials are forming), 96H (immature perithecia are present, with paraphyses) and 144H (mature perithecia are present, with multiseptate ascospores; paraphyses have mostly collapsed).

From this tissue, I harvest RNA and label it with a fluorescent label (biotin). The labeled cRNA is hybridized overnight to the gene chip. Any probe set that lights up when viewed with the proper imager is lighting up because RNA hybridized to it. Since messenger RNA (the only RNA that will hybridize to the GeneChip) has a relatively short lifespan - it is presumed to be only floating around in the cell when the gene to which it corresponds is actively being transcribed, or in the time between transcription and the end of translation - genes that light up can be assumed to be genes playing an active role in the cell at the time of harvest. Some genes - the so-called housekeeping genes - are always active; we say they are constituitively expressed. These are generally extremely important genes to the organism, but are of relatively little interest to us here. What we are searching for are genes that show obvious differences in expression during sexual development. A gene that is highly expressed at 96H and during no other time point might very well be involved in the production of perithecia, asci or ascospores. A gene that is expressed at all times, at a similar expression level, may also play a role in sexual development, but it would be more difficult to predict. Ultimately, the mere presence or absence of a gene product at a given time point does not definitively implicate that gene in a particular developmental role - it merely narrows down the search for interesting genes, and allows us to focus our efforts in gene knockouts.


Updated November 20, 2007