Short Abstract: Identification of mutations that determine an organism’s phenotype and the distinction from (near-) neutral (‘hitchhikers’) ones is a fundamental challenge in genome research, and is relevant for numerous medical and biotechnological applications. This is usually done by time- and cost-intensive mutation experiments that are inherently low throughput, due to their experimental nature. We present a computational method for the inference of ‘phenotype trees’, i.e. for the inference of phenotypic impact of genotypic changes from homologous genetic sequences and associated pairwise phenotype distances. We demonstrate the accuracy of our method by application to the major viral surface protein hemagglutinin of human influenza A (H3N2) virus. Here, recognizing changes in the antigenic phenotype and a viral strains’ capability to evade pre-existing host immunity is important for the production of efficient vaccines. In our ‘antigenic trees’, antigenic weights are assigned to all branches of a phylogenetic tree using least-squares optimization, which allows us to resolve the antigenic impact of the associated amino acid changes. Antigenic distances are fitted onto the tree with comparable accuracy to antigenic cartography. We identified both known and novel sites, and amino acid changes with antigenic impact in 35 years of evolution of human influenza A (H3N2) viruses. Additionally, we provide further details on individual changes that shape the antigenic evolution of influenza A (H3N2) viruses.

Short Abstract: Mechanistically diverse enzyme superfamilies are composed of evolutionarily related enzymes that share common mechanistic features yet catalyze different chemical reactions. The study of the evolution of these mechanistically diverse enzyme superfamilies provides a unique opportunity to understand how nature has modified specific catalytic scaffolds to catalyze numerous enzymatic reactions. The alkaline phosphatase (AP) superfamily provides an especially compelling system because its founding member, AP, is a prototypic phosphoryl transfer catalyst and its mechanism is well characterized. Homologues of AP catalyze a range of phosphoryl and sulfuryl transfer reactions including phosphatases, sulfatases, phosphodiesterases and phosphomutases.

Protein similarity networks (PSNs) are graphical representations of sequence, structural, and other types of similarities among a group of proteins in which pairwise all-by-all similarity connections are calculated. For example, nodes can be used to represent one or more protein sequences or structures and edges drawn between two nodes represent some measure of their similarity. Mapping biological information to network nodes or edges enables hypothesis creation about sequence-structure-function relationships across entire sets of related proteins.

We present an investigation of the AP superfamily using PSNs to create an evolutionary model for their divergence from a common ancestor, which is supported from phylogenetic analysis. This model demonstrates the elaboration of an ancient minimal precursor structure to produce the various contemporary reactions of known and unknown function. This work has applications in selection of targets for structural genomics, automated function prediction, and enzyme engineering.

Short Abstract: The progress in experimental studies revealed the structure, biochemical function and catalytic mechanisms of many proteins. However, the questions how the first enzymes emerged and what their building blocks were remain unresolved.

Enzymes are complex catalytic machines performing biochemical transformations as sequences of elementary chemical reactions. Therefore, the biochemical functions can be represented as sets of elementary ones. Previously, we explored how the polymer nature of proteins determined the units of protein structure and function and introduced the concept of elementary functional loops. We hypothesized that elementary functional loops (EFLs), 25-30 residue long segments, are the functional units of enzymes, possessing distinct signatures and providing catalytic and binding amino acids to the enzyme's active sites. Some of the EFLs are presumably descendants of primordial catalytic peptides.

We analyzed distant evolutionary connections between protein functions in Archaea based on the EFLs comprising them. We show examples of the EFLs in new functional domains, as well as reutilization of EFLs and functional domains in building multidomain structures and protein complexes.

Our analysis of the archaeal superkingdom yields the dominating mechanisms in different periods of protein evolution, which resulted in several levels of organization of biochemical function. First, functional domains emerged as combinations of prebiotic peptides with the very basic functions, such as nucleotide/phosphate and metal cofactor binding. Second, domain recombination brought to the evolutionary scene the multidomain proteins and complexes. Later, reutilization and de novo design of functional domains and elementary functional loops complemented evolution of protein function.

Short Abstract: Introduction

A reference tree of life (TOL) including all completely sequenced organisms is an essential component for capitalizing on the rapidly increasing amount of sequence data being produced. Furthermore, providing an ontological infrastructure across the tree enables a shift beyond narrative descriptions of evolutionary relationships among organisms, towards functional insights.

Methods

We report ‘sTOL’, a resource to meet these needs. It is the product of an automated pipeline using existing status quo NCBI taxonomic information in concert with structural domain information from all sequenced genomes that are publicly available. At the core is phylogenomic analysis based on RAxML using weight calibration operating on molecular character data (in the form of SCOP structural superfamilies, domain families, supra-domains and full-length domain architectures).

Results & Conclusions

We quantify the reliability of using the NCBI taxonomy as a partial constraint tree, as a way of representing electronically the current status quo. In doing so we observe that there are potentially significant improvements that can be made to the NCBI taxonomic classification, particularly in the fungi kingdom; we also see that the current state of many animal genome assemblies is inadequate. We illustrate the utility of the sTOL by looking at several examples. When combined domain-centric GO annotations, it allows us to shed evolutionary and functional insights into eukaryotic evolution. The sTOL provides a new analytical platform to study genome evolution and function in the next generation sequencing era. The server hosting the daily-updated sTOL is available as an extension to the SUPERFAMILY database (http://supfam.org/SUPERFAMILY/sTOL).

Short Abstract: Protein evolution can be modeled as a combination of neutral evolution and functional differentiation. Analyzing evolutionary patterns of coding sequences it has been suggested, based on the ratio of rates of evolution of non-synonymous and synonymous codon sites (Ka/Ks), that substitutions between amino acid types of different physico-chemical properties correlate more frequently with events of functional differentiation, whereas substitutions between similar amino acid types more likely represent events of constrained neutral evolution. Here we identify patterns of functional differentiation in the evolution of bacterial proteins based on “switch positions”, sites in a protein family where amino acid types are conserved within individual phylogenetic groups of bacteria, indicating strong lineage-specific constraining negative selection, but differ (“switch”) between different phylogenetic groups, indicating events of functional differentiation that accompany lineage differentiation.
From the analysis of the alignment of 167 protein families conserved across 31 well-defined bacterial groups, we identified several positions where amino acid types were conserved within all groups. Among these we identified a small fraction of sites corresponding to switch positions. Amino acid exchange matrices based on switch positions indicated that, in contrast to inferences from Ka/Ks ratios, amino acid switches ascribable to events of functional differentiation corresponded most frequently to replacements between amino acid types of similar physico-chemical properties.
Being a class of rare genomic changes, switch positions should be appropriate markers to reconstruct phylogenetic relations of bacterial groups. The resulting trees supported the model of biological big-bang in the evolution of bacterial phyla

Short Abstract: The nature of protein fold space is hotly debated. Do the protein folds observed in nature fall into clean, discrete clusters, or is fold space more accurately modeled as a vast continuum, of which only a small sample of proteins has yet been observed? Previous efforts to answer this question have focused on geometric spaces (PCA, multidimensional scaling, locally linear embedding) or network models (conformational space networks). While such schemes may facilitate protein comparison and classification, the choice of a mathematical framework for fold space is arbitrary without a connection to concrete biological processes. To accurately capture the true relationships between protein folds, a model must consider the evolutionary history of those folds.

Here we present a high-level model of protein evolution, which focuses on mutations that preserve the global 3D structure of proteins. We hypothesize that the combination of subtle local changes (e.g. PTMs) and large, but structure-preserving, rearrangements (e.g. duplications) can account for both the continuity of intermediate structures within protein folds and the evolution of seemingly novel folds. Our model categorizes known biological mutation processes, such as DNA replication errors and crossover errors, and places them in a simple theoretical framework.

To test this model, we present evidence from the analysis of a recent systematic comparison of all protein domains from the Protein Data Bank (PDB). We also show that the model is consistent with existing evolutionary models for gene duplication, circular permuted proteins, and proteins with internal symmetry. Future work will focus on explicitly determining evolutionary events relating distantly homologous folds.

Short Abstract:

Short Abstract:

Short Abstract: