Neurospora fmf-1: lure and lore.

The Neurospora crassa fmf-1 mutant has a unique phenotype. It arrests sexual development when the fruiting bodies (perithecia) attain only 40% of their normal diameter, regardless of whether the mutant participates in a cross with the wild type (fmf-1 x fmf-1⁺) as the male or female parent. I first learnt about fmf-1 when this journal invited me to review 'The Neurospora compendium: chromosomal loci' by D. D. Perkins, A. Radford and M. S. Sachs (J. Genet. 80: 53-54, 2001). The compendium also informed me that the first Neurospora genetic map was published here (J. Genet. 32, 243-256, 1936). The mutant was discovered and characterized by T. E. Johnson, who also localized the mutation to a chromosome 1 segment that spanned more than 3.3 Mb DNA (Genetics 92, 1107-1120, 1979). The second fmf-1 paper came 30 years later from my laboratory. We mapped the mutation to a single base pair, a T:A to A:T transversion mutation, and thus identified the altered gene (J. Genet. 88: 33-39, 2009). To map fmf-1, we leveraged our expertise in making strains bearing chromosome segment duplications. The Dp strains were generated in crosses of the wild type with translocation strains (WT x T). A translocation transfers a segment of one chromosome into another. Mapping with Dps localized fmf-1 to a 330 kbp segment. Conventional mapping with crossovers and selection against noncrossovers subsequently localized it to a 33 kbp segment. This interval was small enough to pick up the mutation by sequencing its DNA. The Fmf-1 protein activates genes required for mating pheromone signalling. The fmf-1 male gametes (conidia) fail to secrete the pheromone that attracts receptors on the fmf-1⁺ female sexual structures (protoperithecia). Conversely, fmf-1 protoperithecia do not express the cognate receptor for the pheromone from the fmf-1⁺ conidia. Consequently, the fmf-1⁺ x fmf-1 cross fails to fertilize protoperithecia and arrests their maturation into perithecia. Genetic mapping, especially Dp mapping, fails to impress many nongeneticists these days. How do WT x T crosses produce Dp progeny? Why are Dps and crossovers even needed? Why select against noncrossovers? Why not just sequence the genomes of the wild type and mutant, identify genes whose DNA is altered in the mutant, and then test them one by one? Many forget that DNA sequencing, especially of 'hard to access' centromeric sequences, was not as easy and inexpensive then. Isolating fmf-1 offered us the possibility of enriching for RIP-defective mutants. RIP is a mutational process that occurs during a sexual cross and induces multiple G:C to A:T transition mutations in all copies of any DNA sequences duplicated in the otherwise haploid Neurospora genome. It is the most mutagenic process known in biology. Reputedly, linked duplications were 'mutated at frequencies of 95% or more' (J. Genet. 75: 313-324, 1996). My student, Srividhya Iyer, created a linked duplication of fmf-1 by inserting a second copy of it within 5 kbp of the endogenous gene. Most progeny from duplication-homozygous crosses would inherit a RIP-mutated fmf-1 allele, rendering them infertile. If the f₁ progeny are germinated en masse, and allowed to randomly inter-cross, then only crosses between the minor fraction of non-RIPed progeny can generate the f₂. Likewise, for the f₃, f₄, etc. Later generations, hence, become progressively enriched for RIP-defective mutants. In the f₁ progeny examined by Iyer, the RIP-induced fmf-1 mutant fraction was not 95%, but 'merely' 85%, a lesser enrichment efficiency than we desired. Therefore, the enrichment attempt was abandoned. This is not for the first time, nor the last, that a beautiful strategy was killed by an ugly fact.