StAR-Related Lipid Transfer (START) Domains Across the Rice Pangenome Reveal How Ontogeny Recapitulated Selection Pressures During Rice Domestication

Abstract

The StAR-related lipid transfer (START) domain containing proteins or START proteins, encoded by a plant amplified family of evolutionary conserved genes, play important roles in lipid binding, transport, signaling, and modulation of transcriptional activity in the plant kingdom, but there is limited information on their evolution, duplication, and associated sub- or neo-functionalization. Here we perform a comprehensive investigation of this family across the rice pangenome, using 10 wild and cultivated varieties. Conservation of START domains across all 10 rice genomes suggests low dispensability and critical functional roles for this family, further supported by chromosomal mapping, duplication and domain structure patterns. Analysis of synteny highlights a preponderance of segmental and dispersed duplication among STARTs, while transcriptomic investigation of the main cultivated variety Oryza sativa var. japonica reveals sub-functionalization amongst genes family members in terms of preferential expression across various developmental stages and anatomical parts, such as flowering. Ka/Ks ratios confirmed strong negative/purifying selection on START family evolution, implying that ontogeny recapitulated selection pressures during rice domestication. Our findings provide evidence for high conservation of START genes across rice varieties in numbers, as well as in their stringent regulation of Ka/Ks ratio, and showed strong functional dependency of plants on START proteins for their growth and reproductive development. We believe that our findings advance the limited knowledge about plant START domain diversity and evolution, and pave the way for more detailed assessment of individual structural classes of START proteins among plants and their domain specific substrate preferences, to complement existing studies in animals and yeast.

Cavity architecture based modulation of ligand binding tunnels in plant START domains

“…Such a detailed residue level comparison is evidence for evolutionary change leading to functionally critical substitutions, thereby causing sub-functionalization to smaller ligands (as in case of minimal STARTs), or neo-functionalization in terms of loss of ligand binding ability altogether (as in case of HD containing STARTs). Supporting evidence for these changes has already been seen in our previous studies in terms of gene expression changes and GO enrichment [20] , [33] , [56] . The evolutionary change of HD START domains into new roles of transcriptional regulation has already been experimentally verified in case of Arabidopsis, and this provides support for the findings reported in case of rice domains [55] .…”

“…Therefore, minimal STARTs appear to be the only class of present day plant STARTs domains that may have retained the capacity to bind and transport large lipid based molecules, similar to mammalian counterparts. This inference is also supported by transcriptomic and gene duplication studies conducted in our group, where we show that minimal STARTs are closest to mammalian homologs, and thus, have the highest potential for being lipid transporters [33] , [55] , [56] . As demonstrated in very recent reports [57] , minimal START domains have played roles in transportation of C 20 fatty acids from endoplasmic reticulum to chloroplast in Marchantia polymorpha .…”

“…In case of several HD containing START domains, our cavity architecture analysis shows a total obliteration of the tunnels, despite conservation of the helix-grip core structural fold, providing further evidence for evolution of new function/s. Furthermore, the paradigm that START domains only play a role in ligand binding and transportation of lipid sterols, does not stand in case of many rice START domains, in which, regardless of very small cavities detected, these were putatively expressed in different anatomical parts and developmental stages as shown in our previous studies [33] . Overall, our findings support earlier reports that majority of plant STARTs have evolved to perform other functions like regulation (through association with other domains like HD-ZIP etc), or at best these may serve as lipid sensors rather than transporters, that signal biological responses.…”

“…A total of 28 START domains proteins were identified in cultivated rice variety (Oryza sativa var. japonica) from our earlier work on evolution of START domains across the rice pangenome [33] . These 28 START domains were classified into four major structural classes based on the arrangement of co-occurring domains (co-domains), namely (i) HD bZIP START MEKHLA; HZSM or Class III HD-Zip (ii) HD START; HS and HD bZIP START; HZS or Class IV HD-Zip (iii) PH START DUF1336; PSD and (iv) minimal START proteins; mS (lacking co-domains) [33] .…”

“…japonica) from our earlier work on evolution of START domains across the rice pangenome [33] . These 28 START domains were classified into four major structural classes based on the arrangement of co-occurring domains (co-domains), namely (i) HD bZIP START MEKHLA; HZSM or Class III HD-Zip (ii) HD START; HS and HD bZIP START; HZS or Class IV HD-Zip (iii) PH START DUF1336; PSD and (iv) minimal START proteins; mS (lacking co-domains) [33] . The START domain regions from all proteins were extracted based on annotated border residues and were fed into the structure prediction pipeline as described below.…”