bioRxiv preprint doi: https://doi.org/10.1101/184242; this version posted September 4, 2017. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1 Article-Discoveries 2 Specialized plant biochemistry drives gene clustering in fungi 3 Emile Gluck-Thaler1, Jason C. Slot1* 4 1 Department of Plant Pathology, The Ohio State University, Columbus, Ohio, 5 United States of America 6 *Corresponding author 7 E-mail: [email protected] 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1 bioRxiv preprint doi: https://doi.org/10.1101/184242; this version posted September 4, 2017. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 23 Abstract 24 The fitness and evolution of both prokaryotes and eukaryotes are affected 25 by the organization of their genomes. In particular, the physical clustering of 26 functionally related genes can facilitate coordinated gene expression and can 27 prevent the breakup of co-adapted alleles in recombining populations. While 28 clustering may thus result from selection for phenotype optimization and 29 persistence, the extent to which eukaryotic gene organization in particular is 30 driven by specific environmental selection pressures has rarely been 31 systematically explored. Here, we investigated the genetic architecture of fungal 32 genes involved in the degradation of phenylpropanoids, a class of plant-produced 33 secondary metabolites that mediate many ecological interactions between plants 34 and fungi. Using a novel gene cluster detection method, we identified over one 35 thousand gene clusters, as well as many conserved combinations of clusters, in 36 a phylogenetically and ecologically diverse set of fungal genomes. We 37 demonstrate that congruence in gene organization over small spatial scales in 38 fungal genomes is often associated with similarities in ecological lifestyle. 39 Additionally, we find that while clusters are often structured as independent 40 modules with little overlap in content, certain gene families merge multiple 41 modules in a common network, suggesting they are important components of 42 phenylpropanoid degradation strategies. Together, our results suggest that 43 phenylpropanoids have repeatedly selected for gene clustering in fungi, and 2 bioRxiv preprint doi: https://doi.org/10.1101/184242; this version posted September 4, 2017. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 44 highlight the interplay between gene organization and ecological evolution in this 45 ancient eukaryotic lineage. 46 47 Introduction 48 Genome organization is intimately linked to the trajectory of organismal 49 evolution. The impacts of genome organization on prokaryotic evolution in 50 particular have been extensively studied (Baquero 2004), and increasingly, the 51 organization of genes in eukaryotic genomes is also recognized to affect 52 organismal fitness and evolution. For example, the spatial clustering of 53 functionally related genes may enable the compartmentalization and optimization 54 of phenotypes through coordinated gene expression (Hurst et al. 2002; Al- 55 Shahrour et al. 2010; McGary et al. 2013). Similarly, the formation of loci 56 composed of co-adapted alleles can facilitate the inheritance of locally adapted 57 ecotypes within recombining populations over short time periods (Yeaman 2013; 58 Holliday et al. 2016). Rather than resulting from non-adaptive processes, such as 59 genetic hitchhiking, the persistence of such organizational patterns in eukaryotic 60 genomes suggests they may instead result from natural selection (Hurst et al. 61 2002; Lynch 2007). However, the extent to which eukaryotic genome 62 organization is driven by specific environmental selection pressures, especially 63 over macroevolutionary timescales, is not clear. 64 Organized genome structure is particularly apparent in fungi, a lineage of 65 eukaryotic microorganisms that perform critical ecosystem services and biomass 3 bioRxiv preprint doi: https://doi.org/10.1101/184242; this version posted September 4, 2017. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 66 transformation, and also impact plant and animal health (Peay et al. 2016). 67 Fungal genomes are more or less replete with metabolic gene clusters (MGCs) 68 composed of genes encoding enzymes, transporters and regulators that 69 participate in specialized metabolic processes such as nutrient acquisition, 70 competition and defense (Wisecaver et al. 2014). Although MGCs are far more 71 rare in eukaryotes compared with bacteria, fungal MGCs exhibit similarly sparse 72 and disjunct phylogenetic distributions among distantly related species with 73 overlapping niches (Greene et al. 2014; Dhillon et al. 2015; Glenn et al. 2016). 74 This ecological pattern of distribution suggests that conserved combinations of 75 genes may be signatures of ecological selection in fungal genomes. 76 Fungal MGCs encoding the production of specialized, or secondary, 77 metabolites (SMs) have been studied extensively (Hoffmeister and Keller 2007) 78 and more recently, several reports suggest that adaptations to degrade plant 79 SMs are also encoded in MGCs (Shanmugam et al. 2010; Greene et al. 2014; 80 Wang et al. 2014; Kettle et al. 2015; Glenn et al. 2016). Plant SMs mediate 81 important biotic and abiotic interactions, including the exclusion of fungal 82 pathogens, the establishment of mutualisms, and the rates of nutrient cycling 83 long after the plant has died. The largest group of plant SMs are the 84 phenylpropanoids, which not only contribute to constitutive and inducible 85 chemical defenses, but are also the main barriers to wood decay in terrestrial 86 ecosystems (Floudas et al. 2012), and costly inhibitors of lignocellulose biofuel 87 production (Jönsson and Martín 2016). As the primary colonizers of plant 4 bioRxiv preprint doi: https://doi.org/10.1101/184242; this version posted September 4, 2017. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 88 material in natural and artificial environments, fungi are frequently in contact with 89 phenylpropanoids, and must often mitigate their inhibitory effects in order to 90 grow. Common fungal adaptations to phenylpropanoid toxicity include 91 detoxification through sequestration, excretion and degradation (Mäkelä et al. 92 2015). Despite the characterization of many phenylpropanoid degradation 93 pathways in fungi, the genomic bases of these pathways are largely unknown 94 (Mäkelä et al. 2015), precluding the use of currently available algorithms to 95 investigate whether or not these metabolic processes are encoded in MGCs 96 (Wisecaver et al. 2014; Weber et al. 2015). 97 Here, we developed a novel algorithm based on empirically derived 98 models of fungal genome evolution in order to systematically identify clusters of 99 genes putatively involved in phenylpropanoid degradation, enabling us to test the 100 hypothesis that selection pressures from plant SMs impact genome organization 101 across disparate fungal lineages. Using a database of 556 fungal genomes 102 representing 481 species, we detected 1168 MGCs and many conserved 103 combinations of MGCs that putatively degrade a broad array of 104 phenylpropanoids. We then tested for associations between MGCs and various 105 fungal ecological lifestyles, and found that the presence of certain MGCs was 106 enriched in plant pathotrophs and saprotrophs. While many clusters appear to 107 have evolved independently, we identified several gene families that are 108 commonly associated with diverse MGCs, suggesting they play important roles in 109 phenylpropanoid catabolism. These results suggest that phenylpropanoids are 5 bioRxiv preprint doi: https://doi.org/10.1101/184242; this version posted September 4, 2017. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 110 drivers of gene organization in plant-associated fungi, and that MGCs may in turn 111 determine patterns of fungal community assembly on both living and decaying 112 plant tissues. 113 114 Results 115 116 Diverse candidate gene clusters are associated with phenylpropanoid 117 degradation 118 119 Using 27 different “anchor” gene families involved in phenylpropanoid 120 degradation as separate queries (Methods; Supplementary Table 1), we 121 searched 556 genomes from 481 fungal species for clusters associated with 122 each anchor gene family (Methods; Supplementary Figure 1, Supplementary 123 Table 2). After removing those clusters containing genes known to exclusively 124 participate in fungal secondary metabolite biosynthesis (Supplementary Table 3; 125 Supplementary Table 4), as well as gene clusters with fewer than 4 genes, we 126 found evidence of unexpected clustering in 13 anchor gene families, which we 127 defined as separate cluster classes. We identified a total of 1168 clusters 128 distributed across 363 fungal genomes (Figure 1, Supplementary Figure 2, 129 Supplementary Table 5). 31 of these clusters belonged to multiple clusters 130 classes