Deep learning methods for designing proteins scaffolding functional sites

Current approaches to de novo design of proteins harboring a desired binding or catalytic motif require pre-specification of an overall fold or secondary structure composition, and hence considerable trial and error can be required to identify protein structures capable of scaffolding an arbitrary functional site. Here we describe two complementary approaches to the general functional site design problem that employ the RosettaFold and AlphaFold neural networks which map input sequences to predicted structures. In the first "constrained hallucination" approach, we carry out gradient descent in sequence space to optimize a loss function which simultaneously rewards recapitulation of the desired functional site and the ideality of the surrounding scaffold, supplemented with problem-specific interaction terms, to design candidate immunogens presenting epitopes recognized by neutralizing antibodies, receptor traps for escape-resistant viral inhibition, metalloproteins and enzymes, and target binding proteins with designed interfaces expanding around known binding motifs. In the second "missing information recovery" approach, we start from the desired functional site and jointly fill in the missing sequence and structure information needed to complete the protein in a single forward pass through an updated RoseTTAFold trained to recover sequence from structure in addition to structure from sequence. We show that the two approaches have considerable synergy, and AlphaFold2 structure prediction calculations suggest that the approaches can accurately generate proteins containing a very wide array of functional sites.

Stage-specific action of Runx1 and GATA3 controls silencing of PU.1 expression in mouse pro–T cells

PU.1 (encoded by Spi1), an ETS-family transcription factor with many hematopoietic roles, is highly expressed in the earliest intrathymic T cell progenitors but must be down-regulated during T lineage commitment. The transcription factors Runx1 and GATA3 have been implicated in this Spi1 repression, but the basis of the timing was unknown. We show that increasing Runx1 and/or GATA3 down-regulates Spi1 expression in pro–T cells, while deletion of these factors after Spi1 down-regulation reactivates its expression. Leveraging the stage specificities of repression and transcription factor binding revealed an unconventional but functional site in Spi1 intron 2. Acute Cas9-mediated deletion or disruption of the Runx and GATA motifs in this element reactivates silenced Spi1 expression in a pro–T cell line, substantially more than disruption of other candidate elements, and counteracts the repression of Spi1 in primary pro–T cells during commitment. Thus, Runx1 and GATA3 work stage specifically through an intronic silencing element in mouse Spi1 to control strength and maintenance of Spi1 repression during T lineage commitment.

Circ_0008542 in osteoblast exosomes promotes osteoclast-induced bone resorption through m6A methylation

With an increasing aging society, China is the world’s fastest growing markets for oral implants. Compared with traditional oral implants, immediate implants cause marginal bone resorption and increase the failure rate of osseointegration, but the mechanism is still unknown. Therefore, it is important to further study mechanisms of tension stimulus on osteoblasts and osteoclasts at the early stage of osseointegration to promote rapid osseointegration around oral implants. The results showed that exosomes containing circ_0008542 from MC3T3-E1 cells with prolonged tensile stimulation promoted osteoclast differentiation and bone resorption. Circ_0008542 upregulated Tnfrsf11a (RANK) gene expression by acting as a miR-185-5p sponge. Meanwhile, the circ_0008542 1916-1992 bp segment exhibited increased m6A methylation levels. Inhibiting the RNA methyltransferase METTL3 or overexpressing the RNA demethylase ALKBH5 reversed osteoclast differentiation and bone resorption induced by circ_0008542. Injection of circ_0008542 + ALKBH5 into the tail vein of mice reversed the same effects in vivo. Site-directed mutagenesis study demonstrated that 1956 bp on circ_0008542 is the m6A functional site with the abovementioned biological functions. In conclusion, the RNA methylase METTL3 acts on the m6A functional site of 1956 bp in circ_0008542, promoting competitive binding of miRNA-185-5p by circ_0008542, and leading to an increase in the target gene RANK and the initiation of osteoclast bone absorption. In contrast, the RNA demethylase ALKBH5 inhibits the binding of circ_0008542 with miRNA-185-5p to correct the bone resorption process. The potential value of this study provides methods to enhance the resistance of immediate implants through use of exosomes releasing ALKBH5.

Design of proteins presenting discontinuous functional sites using deep learning

An outstanding challenge in protein design is the design of binders against therapeutically relevant target proteins via scaffolding the discontinuous binding interfaces present in their often large and complex binding partners. There is currently no method for sampling through the almost unlimited number of possible protein structures for those capable of scaffolding a specified discontinuous functional site; instead, current approaches make the sampling problem tractable by restricting search to structures composed of pre-defined secondary structural elements. Such restriction of search has the disadvantage that considerable trial and error can be required to identify architectures capable of scaffolding an arbitrary discontinuous functional site, and only a tiny fraction of possible architectures can be explored. Here we build on recent advances in de novo protein design by deep network hallucination to develop a solution to this problem which eliminates the need to pre-specify the structure of the scaffolding in any way. We use the trRosetta residual neural network, which maps input sequences to predicted inter-residue distances and orientations, to compute a loss function which simultaneously rewards recapitulation of a desired structural motif and the ideality of the surrounding scaffold, and generate diverse structures harboring the desired binding interface by optimizing this loss function by gradient descent. We illustrate the power and versatility of the method by scaffolding binding sites from proteins involved in key signaling pathways with a wide range of secondary structure compositions and geometries. The method should be broadly useful for designing small stable proteins containing complex functional sites.

CATH functional families predict functional sites in proteins

Motivation Identification of functional sites in proteins is essential for functional characterization, variant interpretation and drug design. Several methods are available for predicting either a generic functional site, or specific types of functional site. Here, we present FunSite, a machine learning predictor that identifies catalytic, ligand-binding and protein–protein interaction functional sites using features derived from protein sequence and structure, and evolutionary data from CATH functional families (FunFams). Results FunSite’s prediction performance was rigorously benchmarked using cross-validation and a holdout dataset. FunSite outperformed other publicly available functional site prediction methods. We show that conserved residues in FunFams are enriched in functional sites. We found FunSite’s performance depends greatly on the quality of functional site annotations and the information content of FunFams in the training data. Finally, we analyze which structural and evolutionary features are most predictive for functional sites. Availabilityand implementation https://github.com/UCL/cath-funsite-predictor. Contact [email protected] or [email protected] Supplementary information Supplementary data are available at Bioinformatics online.

ResiRole: residue-level functional site predictions to gauge the accuracies of protein structure prediction techniques

Motivation Methods to assess the quality of protein structure models are needed for user applications. To aid with the selection of structure models and further inform the development of structure prediction techniques, we describe the ResiRole method for the assessment of the quality of structure models. Results Structure prediction techniques are ranked according to the results of round-robin, head-to-head comparisons using difference scores. Each difference score was defined as the absolute value of the cumulative probability for a functional site prediction made with the FEATURE program for the reference structure minus that for the structure model. Overall, the difference scores correlate well with other model quality metrics; and based on benchmarking studies with NaïveBLAST, they are found to detect additional local structural similarities between the structure models and reference structures. Availabilityand implementation Automated analyses of models addressed in CAMEO are available via the ResiRole server, URL http://protein.som.geisinger.edu/ResiRole/. Interactive analyses with user-provided models and reference structures are also enabled. Code is available at github.com/wamclaughlin/ResiRole. Supplementary information Supplementary data are available at Bioinformatics online.

Allostery and Epistasis: Emergent Properties of Anisotropic Networks

Understanding the underlying mechanisms behind protein allostery and non-additivity of substitution outcomes (i.e., epistasis) is critical when attempting to predict the functional impact of mutations, particularly at non-conserved sites. In an effort to model these two biological properties, we extend the framework of our metric to calculate dynamic coupling between residues, the Dynamic Coupling Index (DCI) to two new metrics: (i) EpiScore, which quantifies the difference between the residue fluctuation response of a functional site when two other positions are perturbed with random Brownian kicks simultaneously versus individually to capture the degree of cooperativity of these two other positions in modulating the dynamics of the functional site and (ii) DCIasym, which measures the degree of asymmetry between the residue fluctuation response of two sites when one or the other is perturbed with a random force. Applied to four independent systems, we successfully show that EpiScore and DCIasym can capture important biophysical properties in dual mutant substitution outcomes. We propose that allosteric regulation and the mechanisms underlying non-additive amino acid substitution outcomes (i.e., epistasis) can be understood as emergent properties of an anisotropic network of interactions where the inclusion of the full network of interactions is critical for accurate modeling. Consequently, mutations which drive towards a new function may require a fine balance between functional site asymmetry and strength of dynamic coupling with the functional sites. These two tools will provide mechanistic insight into both understanding and predicting the outcome of dual mutations.

CATH functional families predict protein functional sites

MotivationIdentification of functional sites in proteins is essential for functional characterisation, variant interpretation and drug design. Several methods are available for predicting either a generic functional site, or specific types of functional site. Here, we present FunSite, a machine learning predictor that identifies catalytic, ligand-binding and protein-protein interaction functional sites using features derived from protein sequence and structure, and evolutionary data from CATH functional families (FunFams).ResultsFunSite’s prediction performance was rigorously benchmarked using cross-validation and a holdout dataset. FunSite outperformed all publicly-available functional site prediction methods. We show that conserved residues in FunFams are enriched in functional sites. We found FunSite’s performance depends greatly on the quality of functional site annotations and the information content of FunFams in the training data. Finally, we analyse which structural and evolutionary features are most predictive for functional sites.AvailabilityThe datasets and prediction models are available on [email protected] informationSupplementary data are available at Bioinformatics online.