Origins of lineage-specific elements via gene duplication, relocation, and regional rearrangement in Neurospora crassa

Abstract

The origin of new genes has long been a central interest of evolutionary biologists. However, their novelty means that they evade reconstruction by the classical tools of evolutionary modelling. This evasion of deep ancestral investigation necessitates intensive study of model species within well-sampled, recently diversified, clades. One such clade is the model genus Neurospora, members of which lack recent gene duplications. Several Neurospora species are comprehensively characterized organisms apt for studying the evolution of lineage-specific genes (LSGs). Using gene synteny, we documented that 78% of Neurospora LSG clusters are located adjacent to the telomeres featuring extensive tracts of non-coding DNA and duplicated genes. Here, we report several instances of LSGs that are likely from regional rearrangements and potentially from gene rebirth. To broadly investigate the functions of LSGs, we assembled transcriptomics data from 68 experimental data points and identified co-regulatory modules using Weighted Gene Correlation Network Analysis, revealing that LSGs are widely but peripherally involved in known regulatory machinery for diverse functions. The ancestral status of the LSG mas-1, a gene with roles in cell-wall integrity and cellular sensitivity to antifungal toxins, was investigated in detail alongside its genomic neighbours, indicating that it arose from an ancient lysophospholipase precursor that is ubiquitous in lineages of the Sordariomycetes. Our discoveries illuminate a "rummage region" in the N. crassa genome that enables the formation of new genes and functions to arise via gene duplication and relocation, followed by fast mutation and recombination facilitated by sequence repeats and unconstrained non-coding sequences.

Keywords: chromosomal rearrangement; de novo origination; fungi; genomics; molecular evolution; orphan gene.

FIGURE 1

Syntenic map near the 3′ end of chromosome V, which is enriched with LSGs clusters and Het‐domain genes in three Neurospora species. To provide a single basis that would clearly illustrate frequent gene duplications, long non‐coding intergenic sequences, chromosomal rearrangement, as well as lineage‐specific genes within the three genomes, links between orthologs (lines between genes; magenta: LSGs in N. crassa and in some cases other Neurospora species, black: non‐LSGs) were drawn from N. discreta whenever possible; otherwise, orthologs were linked between each pair to show the retention of their clustering among the three species. Each colour‐coded block (LSGs in N. crassa and in some cases other Neurospora species, magenta; LSGs only in N. tetrasperma, peach; and LSGs only in N. discreta, green) represents a gene. Blank spaces between genes represent non‐coding intergenic sequences or sequences without annotation. Width of the genes and non‐coding regions is proportional to their length in nucleotide base pairs. Three possible rearrangements that were associated with a long non‐coding region were marked for N. crassa by comparison to the gene orders in N. discreta, with three syntenic chromosomal blocks marked as S1, S2 and S3 (Table S2).

FIGURE 2

Divergent synteny, gene tree, and gene expression during sexual development or asexual spore germination of homologues of hypothetical protein‐coding gene NCU09590. (a) Cladogram of N. discreta, N. crassa and N. tetrasperma and local synteny map of the region surrounding NCU09590 among three Neurospora species featuring six orthologs of NCU09590 (colour coded; boundary genes in N. discreta and N. tetrasperma are in black); (b) Bayesian midpoint‐rooted phylogeny of NCU09590 and its homologues (in parentheses, gene numbers follow the latest genome annotation at FungiDB for N. crassa FGSC2489, mating type A; N. crassa FGSC4200, mating type a; N. discreta FGSC8579, mating type A; and pseudohomothallic N. tetrasperma FGSC2508, mat A genome). Each of the six coloured clades represents a consequence of duplication and divergence (scale bar: 0.2 amino‐acid substitutions per site; *: clade support with Bayesian posterior probability <.95; clades are coloured as their orthologs are coloured in panel a). (c) Divergent relative expression levels of N. crassa homologues (coloured as in panels a and b) during key development stages (whiskers: 95% confidence intervals).

FIGURE 3

Syntenic associations among NCU01135 homologues with two kinds of sequence repeats in three Neurospora species, and their expression profiles during sexual reproduction and asexual spore germination. (a) Cladogram of N. discreta, N. crassa and N. tetrasperma and schematics of homologous gene structures. Three homologues of NCU01135 are clustered in N. discreta (purple, peach, navy). However, the order of the three genes is reversed compared to their orthologous sequences in a single gene in N. crassa and N. tetrasperma, indicating an ancestral gene rearrangement. Short perfect sequence repeats were represented by tildas (~). (b) NCU01135 paralog NCU04998 in N. crassa and Ndisc8579_128717 in N. discreta share sequence similarity with NCU01135 and exhibit two kinds of sequence repeats (black and yellow tildas, ~). Repeats shared between SC013 and NCU01135 indicate a possible (question mark) rearrangement and recombination involving non‐syntenic regions from N. discreta to the most recent common ancestor of N. crassa and N. tetrasperma. (c) Comparative expression levels of homologues of NCU01135 in the three Neurospora species, (d) homologues of NCU04998 during sexual reproduction in N. crassa and N. discreta, and (e) NCU01135 and NCU04998 during asexual spore germination in N. crassa.

FIGURE 4

Reconstruction of the evolutionary history of synteny for the Neurospora crassa LSG mas‐1 (NCU03140) and neighbour genes within the Sordariomycetes, based on an RPB1‐ and RPB2‐based species phylogeny with genome annotations from JGI fungal genomes and FungiDB. (a) Current and ancestral synteny in the Sordariomycetes of homologues of N. crassa mas‐1 and its four neighbouring genes. Ancestral inferences of the likely positions of the homologues of the five genes (colour coded to match the orthologs of the five N. crassa genes as shown in the top with dark green NCU03138, blue NCU03139, orange NCU03140, light green NCU03141, and purple NCU03142) were inferred as in Figures S2 and S3; all nodes were fully resolved with posterior probability equal to one. Gene positions are diagrammatic. Grey bars represent other genes in the region. The lengths of arrows representing coding and non‐coding regions are not drawn proportionate to actual lengths in base pairs. Genes that share the same chromosome or supercontig were linked with the same solid or dashed black lines. Extremely long intervals between genes of interest were broken with (−//−) without tallying genes in the intervening regions. (b) Alignment of amino‐acid sequences translated from non‐coding regions orthologous to mas‐1, exhibiting 16%–52% similarity between Sordaria macrospora (Sm), N. discreta (two pieces of a tandem repeat: Nd1 and Nd2) and the orthologs of mas‐1 in N. crassa (Nc) and N. tetrasperma (Nt). Sequence visualization and analysis used MView (EMBL‐EBI online service) with N. crassa mas‐1 as the reference for coverage (cov) and percentage of identity (pid). (c) Hypothesis consistent with a rummage model of evolution for LSG mas‐1 including chromosomal rearrangement, gene duplication, gene function loss that resulted in a non‐coding region, then origin of de novo function and gene.

FIGURE 5

Expression profiles of the LSG mas‐1 and its neighbour genes during asexual spore germination, from first polar growth of the germ tube, doubling of the germ‐tube length, to the appearance of the first hyphal branch, and evidence of a functional role in response to toxins. Relative expression levels of mas‐1 and four neighbour genes (colour‐coded) with 95% CIs for during asexual spore germination on (a) BM at 25°C and (b) MSM at 25°C. (c) A wild‐type strain on polyoxin D medium for 24 h; (d) a mutant of mas‐1 on polyoxin D medium for 24 h, exhibiting higher tolerance to this fungal cell‐wall biosynthesis inhibitor.