From phenotype to structure

Much insight has come from structures of macromolecular complexes determined by methods such as crystallography or cryoelectron microscopy. However, looking at transient complexes remains challenging, as does determining structures in the context of the cellular environment. Braberg et al. used an integrative approach in which they mapped the phenotypic profiles of a comprehensive set of mutants in a protein complex in the context of gene deletions or environmental perturbations (see the Perspective by Wang). By associating the similarity between phenotypic profiles with the distance between residues, they determined structures for the yeast histone H3-H4 complex, subunits Rpb1-Rpb2 of yeast RNA polymerase II, and subunits RpoB-RpoC of bacterial RNA polymerase. Comparison with known structures shows that the accuracy is comparable to structures determined based on chemical cross-links.

Science, this issue p. eaaz4910; see also p. 1269

Determining the structures of protein complexes is crucial for understanding cellular functions. Here, we describe an integrative structure determination approach that relies on in vivo quantitative measurements of genetic interactions. Genetic interactions report on how the effect of one mutation is altered by the presence of a second mutation and have proven effective for identifying groups of genes or residues that function in the same pathway. The point mutant epistatic miniarray profile (pE-MAP) platform allows for rapid measurement of genetic interactions between sets of point mutations and deletion libraries. A pE-MAP is made up of phenotypic profiles, each of which contains all genetic interactions between a single point mutant and the entire deletion library.

We observe a statistical association between the distance spanned by two mutated residues in a protein complex and the similarity of their phenotypic profiles (phenotypic similarity) in a pE-MAP. This observation is in agreement with the expectation that mutations within the same functional region (e.g., active, allosteric, and binding sites) are likely to share more similar phenotypes than those that are distant in space. Here, we explore how to use these associations for determining in vivo structures of protein complexes using integrative modeling.

We generated a large pE-MAP by crossing 350 mutations in yeast histones H3 and H4 against 1370 gene deletions (or hypomorphic alleles of essential genes). The phenotypic similarities were then used to generate spatial restraints for integrative modeling of the H3-H4 complex structure. The resulting ensemble of H3-H4 configurations is accurate and precise, as evidenced by its close similarity to the crystal structure. This finding indicates the utility of the pE-MAP data for integrative structure determination. Furthermore, we show that the pE-MAP provides a wealth of biological insight into the function of the nucleosome and can connect individual histone residues and regions to associated complexes and processes. For example, we observe very high phenotypic similarities between modifiable histone residues and their cognate enzymes, such as H3K4 and COMPASS, or H3K36 and members of the Set2 pathway. Furthermore, the pE-MAP reveals several residues involved in DNA repair and others that function in cryptic transcription.

We demonstrate that the approach is transferable to other complexes and other types of phenotypic profiles by determining the structures of two complexes of known structure: (i) subunits Rpb1 and Rpb2 of yeast RNA polymerase II, using a pE-MAP of 53 point mutants crossed against 1200 deletions and hypomorphic alleles; and (ii) subunits RpoB and RpoC of bacterial RNA polymerase, using a chemical genetics map of 44 point mutants subjected to 83 environmental stresses. The accuracy and precision of the models are comparable to those based on chemical cross-linking, which is commonly used to determine protein complex structures. Moreover, the accuracy and precision improve when using pE-MAP and cross-linking data together, indicating complementarity between these methods and demonstrating a premise of integrative structure determination.

We show that the architectures of protein complexes can be determined using quantitative genetic interaction maps. Because pE-MAPs contain purely phenotypic measurements, collected in living cells, they generate spatial restraints that are orthogonal to other commonly used data for integrative modeling. The pE-MAP data may also enable the characterization of complexes that are difficult to isolate and purify, or those that are only transiently stable. Recent advances in CRISPR-Cas9 genome editing provide a means for extending our platform to human cells, allowing for identification and characterization of functionally relevant structural changes that take place in disease alleles. Expanding this analysis to look at structural changes in host-pathogen complexes and how they affect infection will also be feasible by introducing specific mutations into the pathogenic genome and studying the phenotypic consequences using genetic interaction profiling of relevant host genes.

pE-MAPs are generated by measuring the growth of yeast colonies (left) and visualized as a heatmap (background). We present an application of pE-MAPs to determine protein complex structures, using integrative modeling, and apply it to histones H3 and H4 (right) and other complexes. H3 (purple) and H4 (teal) are highlighted in the context of the nucleosome [gray, modified Protein Data Bank (PDB) 1ID3].

Determining structures of protein complexes is crucial for understanding cellular functions. Here, we describe an integrative structure determination approach that relies on in vivo measurements of genetic interactions. We construct phenotypic profiles for point mutations crossed against gene deletions or exposed to environmental perturbations, followed by converting similarities between two profiles into an upper bound on the distance between the mutated residues. We determine the structure of the yeast histone H3-H4 complex based on ~500,000 genetic interactions of 350 mutants. We then apply the method to subunits Rpb1-Rpb2 of yeast RNA polymerase II and subunits RpoB-RpoC of bacterial RNA polymerase. The accuracy is comparable to that based on chemical cross-links; using restraints from both genetic interactions and cross-links further improves model accuracy and precision. The approach provides an efficient means to augment integrative structure determination with in vivo observations.

Read more here:
Genetic interaction mapping informs integrative structure determination of protein complexes - Science Magazine