Genomic Origin and Diversification of the Glucosinolate MAM Locus

Abstract

Glucosinolates are a diverse group of plant metabolites that characterize the order Brassicales. The MAM locus is one of the most significant QTLs for glucosinolate diversity. However, most of what we understand about evolution at the locus is focused on only a few species and not within a phylogenetic context. In this study, we utilize a micro-synteny network and phylogenetic inference to investigate the origin and diversification of the MAM/IPMS gene family. We uncover unique MAM-like genes found at the orthologous locus in the Cleomaceae that shed light on the transition from IPMS to MAM. In the Brassicaceae, we identify six distinct MAM clades across Lineages I, II, and III. We characterize the evolutionary impact and consequences of local duplications, transpositions, whole genome duplications, and gene fusion events, generating several new hypothesizes on the function and diversity of the MAM locus.

Keywords: brassicaceae; gene duplication; gene family; gene fusion; glucosinolates; polyploidy.

FIGURE 1

Synteny clusters and gene tree phylogeny of identified IPMS and MAM genes consisting of 262 total. For (A,B) the bar along the tips represent species lineage where black bars indicate genes from out group genomes, the pink bar indicates genes from Cleomaceae genomes, and the gray bar indicates genes from Brassicaceae genomes. (A) Syntenic cluster analysis identified three distinct gene clusters, each representing a different conserved genomic location. The IPMS cluster in blue, the MAM-Ancestral cluster in green, and the novel lineage specific MAM-Transposed cluster in orange. Gray lines here indicate connections between the IPMS and MAM clusters. (B) Emphasizes those connections between MAM-like genes in the Cleomaceae that exhibit both IPMS & MAM cluster membership (Clevi.0004s0713 and tha_Th2v24105) despite being physically located at the MAM-Ancestral locus in their respective genomes. [For Bootstrap scores: Supplementary Figure S6; Online interactive trees: (A) http://bit.ly/2tHVgYK; (B) http://bit.ly/2Svu8Vf].

FIGURE 2

Inferred evolutionary trajectory of MAM and IPMS loci in the cleomaceae based on genomic synteny and phylogenetic information. (A) The MAM-Ancestral locus originated from the β whole genome duplication event and is characterized by MAM–like genes that experience local duplication and have retained their LeuA domain. (B) In the genome of Cleome violacea there is a gene deletion followed by a novel tandem duplication at the MAM locus. (C) Following the Th-α whole genome triplication the IPMS locus is duplicated and the MAM locus experiences compensatory gene loss and is reflected in the Gyanandropsis gynandra genome. (D) In the Tarenaya hassleriana genome, the IPMS locus experiences a gene conversion event that maintains sequence similarity between the two copies. There is also a novel transposition of the MAM-like gene from the MAM-Ancestral locus that does not maintain the LeuA domain. As the placement of the Th-α whole genome duplication event is not confirmed to be fully shared by both lineages, an alternative reconstruction is also possible.

FIGURE 3

Clade comparison between Brassicaceae MAM domain and full sequence gene trees highlighting incongruence. HMGL-like sequences were used for the domain tree and resolved six clades of MAM (MAMa-f) but could not infer branching order. In the full sequence tree there is a breakdown of MAMa and MAMb that is correlated with species lineage. In both trees, the placement of MAM1 and MAM3 from the Arabidopsis thaliana Col genome are indicated. [For Bootstrap scores: Supplementary Figure S7 – Domain Tree; Supplementary Figure S8 – Full Sequence; Online interactive trees: Domain – http://bit.ly/2Hb5jIS; Full Sequence – http://bit.ly/37btHEZ].

FIGURE 4

MAM clade and genomic context diversity within the Brassicaceae based on a subsample of analyzed genomes. Each square represents a MAM gene with an indicated HMGL-like domain type. Connected squares are found at the same physical location in the genome and not connected squares represent separate MAM loci (i.e., the MAM-Ancestral locus, a syntenic duplicate of the of the MAM-Ancestral locus, or MAM-Transposed). Non-syntenic gene transpositions were not included. (A) We estimate that the shared ancestor of Lineages I, II, and III maintained both MAMb and MAMd domain types. In the Lineage III genomes sampled MAMb genes were not located at the MAM-Ancestral locus, but at transposed loci. (B) At the ancestor of Lineage I and II, MAMa and MAMe appear, while the MAMc innovation occurs within a sub clade of Lineage I. (C) MAMf originates at the ancestor of Lineage II. The MAMet transposition that creates the MAM-Transposed locus occurs following the split from Lunaria annua, with all MAMet genes being closely related to a MAMe gene at the MAM-Ancestral locus. Lunaria annua also contains a context duplication of the MAM-Ancestral locus* that does not appear to be associated with whole genome duplication. (D) The unnamed whole genome duplication found in the tribe Brassiceae of Lineage II has resulted in multiple context duplications of the MAM-Ancestral locus. Full comparison is found in Supplementary Figure S3.