Complex evolutionary footprints revealed in an analysis of reused protein segments of diverse lengths

Abstract

SignificanceWe question a central paradigm: namely, that the protein domain is the “atomic unit” of evolution. In conflict with the current textbook view, our results unequivocally show that duplication of protein segments happens both above and below the domain level among amino acid segments of diverse lengths. Indeed, we show that significant evolutionary information is lost when the protein is approached as a string of domains. Our finer-grained approach reveals a far more complicated picture, where reused segments often intertwine and overlap with each other. Our results are consistent with a recursive model of evolution, in which segments of various lengths, typically smaller than domains, “hop” between environments. The fit segments remain, leaving traces that can still be detected.

The Urfold: Structural similarity just above the superfold level?

“…4,16,17 Depending on the threshold of structural overlap, the degree of interconnectedness between distinct folds can range from dense to sparse. First, network representations of FS feature highly interconnected nodes that are bridges or hubs.…”

“…Such hubs have been proposed to contain (sub)structures that are common to many different folds. 4,16,17 Depending on the threshold of structural overlap, the degree of interconnectedness between distinct folds can range from dense to sparse. Second, a highly skewed distribution of folds-in terms of their population by known 3D structures-was first observed long ago, 6,26 and a power-law trend has persisted after many more observations (e.g., poststructural genomics): more than 1,300 folds (as defined by CATH) are currently known, and 10 of these accounts for 50% of all known domain structures.…”

“…Folds are generally viewed as corresponding to the level of structural domains, 6,35 though even for the smallest of folds many subtle and intertwined signals can be detected, such as covariation of amino acid residues that are distant in sequence but near in space. 17 These signals are presumably evolutionary echoes of the physicochemical interactions that stabilize a fold, integrated over millions to billions of years; thus, it may be feasible to detect subtle similarities in patterns within covariance matrices for subsets of proteins lying within a given Urfold (via, e.g., the evolutionary couplings approach). As envisaged here, the Urfold can be a full domain, most likely of relatively small size (e.g., the SBB of Reference 29), or it may comprise a significant fraction of the structural "core" of a larger sized domain (e.g., the β-grasp in the work of Shi et al 36 ).…”

“…14 Interfold similarities, including those which are missed, stem from geometric similarities of structural motifs within the folds. [13][14][15][16][17] Claims as to (a) the extent of structural overlap between two otherwise disparate folds (i.e., the characteristic size of the structural motifs), (b) any conclusions regarding their origin (e.g., convergent vs. divergent evolution), and (c) the potential functional significance of such motifs vary greatly in the literature. 13,[18][19][20] While a detailed and comprehensive treatment of that topic is beyond the scope of this Note, inter-fold relationships clearly exist, and fold space (FS) can be viewed as rather continuous.…”

“…13,[18][19][20] While a detailed and comprehensive treatment of that topic is beyond the scope of this Note, inter-fold relationships clearly exist, and fold space (FS) can be viewed as rather continuous. [13][14][15][16][17][18][19]21,22 In the network view of FS, the degree of connectivity between folds varies, often depending on the precise computational methods. For example, the α/β region of FS appears to be highly interlinked, 4,7,22 and the all-α-helical region may show more connections than other regions 21 ; simultaneously, others have found similar levels of interconnectivity within FS.…”