Description and recognition of regular and distorted secondary structures in proteins using the automated protein structure analysis method

SSnet: A Deep Learning Approach for Protein-Ligand Interaction Prediction

Computational prediction of Protein-Ligand Interaction (PLI) is an important step in the modern drug discovery pipeline as it mitigates the cost, time, and resources required to screen novel therapeutics. Deep Neural Networks (DNN) have recently shown excellent performance in PLI prediction. However, the performance is highly dependent on protein and ligand features utilized for the DNN model. Moreover, in current models, the deciphering of how protein features determine the underlying principles that govern PLI is not trivial. In this work, we developed a DNN framework named SSnet that utilizes secondary structure information of proteins extracted as the curvature and torsion of the protein backbone to predict PLI. We demonstrate the performance of SSnet by comparing against a variety of currently popular machine and non-Machine Learning (ML) models using various metrics. We visualize the intermediate layers of SSnet to show a potential latent space for proteins, in particular to extract structural elements in a protein that the model finds influential for ligand binding, which is one of the key features of SSnet. We observed in our study that SSnet learns information about locations in a protein where a ligand can bind, including binding sites, allosteric sites and cryptic sites, regardless of the conformation used. We further observed that SSnet is not biased to any specific molecular interaction and extracts the protein fold information critical for PLI prediction. Our work forms an important gateway to the general exploration of secondary structure-based Deep Learning (DL), which is not just confined to protein-ligand interactions, and as such will have a large impact on protein research, while being readily accessible for de novo drug designers as a standalone package.

N-alkylacylamides in thin films display infrared spectra of 3₁₀-, α-, and π-helices with visible static and dynamic growth phases

A peptide model is a physical system containing a CONH group, the simplest being HCONHCH3 , N-methylformamide (NMF). We have discovered that NMF and N-methylacetamide (NMA), which form hydrogen-bonded oligomers in thin films on a planar AgX fiber, display infrared (IR) spectra with peaks like those of polypeptide helices. Structures can be assigned by their amide I maxima near 1672 (3(10)), 1655 (3(10)), 1653 (α), 1655 (π), and 1635 cm(-1) (π), which are the first IR data for the π-helix. Sharp peaks are an outcome of immobilization of polar species on the polar surface of silver halides. We report the first use of expanded thin-film IR spectroscopy, in which plots of every spectrum over the amide I-II range show pauses or slow stages in the increase or decrease of absorption. These are identified as static phases followed by dynamic phases, with the incremental gain or loss of a helix turn. A general theory can be stated for such processes. Density functional calculations show that the NMA α-helix pentamer (crystal structure geometry) is transformed into a π-helix-like form. For the first time, an entire sequence (3(10)-helix, α-helix, π-helix, quasiplanar species) of spectra has been recorded for NMA.

N-methyl-trimethylacetamide in thin films displays infrared spectra of π-helices, with visible static and dynamic growth phases, and then a β-sheet

The simplest (minimal) peptide model is HCONHCH3. An increase in the π-helix content with increased substitution in the acyl portion suggested the examination of N-methyl-trimethylacetamide) (NMT). NMT displays spectra, in which there is evolution of a set of helices defined by their amide I maxima near 1686 (3(10)), 1655 (first π), and, most importantly, at 1637 cm(-1) (π). Expanded thin-film infrared spectroscopy (XTFIS) shows pauses or slow stages, which are identified as static phases followed by dynamic phases with the incremental gain or loss of a helix turn. In addition, absorbance at 1637 cm(-1) suddenly increases at 82.1 s (30% over 0.3 s), indicating a phase change and crystallization of the π-helix, along with a coincidental decrease in the absorbance for the first π-helix. A sharp peak occurs at the maximum of the phase change at 82.5 s, representing a pure NMT π-helix. The spectra then undergo a decreasing general absorption loss over 150 s, with the π-helix evolving further to an antiparallel β-sheet fragment. The spectral quality arises from the immobilization of polar molecules on polar surfaces. The crystal structure is that of an antiparallel β-sheet.

Polyphony: superposition independent methods for ensemble-based drug discovery

Background

Structure-based drug design is an iterative process, following cycles of structural biology, computer-aided design, synthetic chemistry and bioassay. In favorable circumstances, this process can lead to the structures of hundreds of protein-ligand crystal structures. In addition, molecular dynamics simulations are increasingly being used to further explore the conformational landscape of these complexes. Currently, methods capable of the analysis of ensembles of crystal structures and MD trajectories are limited and usually rely upon least squares superposition of coordinates.

Results

Novel methodologies are described for the analysis of multiple structures of a protein. Statistical approaches that rely upon residue equivalence, but not superposition, are developed. Tasks that can be performed include the identification of hinge regions, allosteric conformational changes and transient binding sites. The approaches are tested on crystal structures of CDK2 and other CMGC protein kinases and a simulation of p38α. Known interaction - conformational change relationships are highlighted but also new ones are revealed. A transient but druggable allosteric pocket in CDK2 is predicted to occur under the CMGC insert. Furthermore, an evolutionarily-conserved conformational link from the location of this pocket, via the αEF-αF loop, to phosphorylation sites on the activation loop is discovered.

Conclusions

New methodologies are described and validated for the superimposition independent conformational analysis of large collections of structures or simulation snapshots of the same protein. The methodologies are encoded in a Python package called Polyphony, which is released as open source to accompany this paper [http://wrpitt.bitbucket.org/polyphony/].

Differential geometric analysis of alterations in MH α-helices

Antigen presenting cells present processed peptides via their major histocompatibility (MH) complex to the T cell receptors (TRs) of T cells. If a peptide is immunogenic, a signaling cascade can be triggered within the T cell. However, the binding of different peptides and/or different TRs to MH is also known to influence the spatial arrangement of the MH α-helices which could itself be an additional level of T cell regulation. In this study, we introduce a new methodology based on differential geometric parameters to describe MH deformations in a detailed and comparable way. For this purpose, we represent MH α-helices by curves. On the basis of these curves, we calculate in a first step the curvature and torsion to describe each α-helix independently. In a second step, we calculate the distribution parameter and the conical curvature of the ruled surface to describe the relative orientation of the two α-helices. On the basis of four different test sets, we show how these differential geometric parameters can be used to describe changes in the spatial arrangement of the MH α-helices for different biological challenges. In the first test set, we illustrate on the basis of all available crystal structures for (TR)/pMH complexes how the binding of TRs influences the MH helices. In the second test set, we show a cross evaluation of different MH alleles with the same peptide and the same MH allele with different peptides. In the third test set, we present the spatial effects of different TRs on the same peptide/MH complex. In the fourth test set, we illustrate how a severe conformational change in an α-helix can be described quantitatively. Taken together, we provide a novel structural methodology to numerically describe subtle and severe alterations in MH α-helices for a broad range of applications.

Description of local and global shape properties of protein helices

A new method, dubbed "HAXIS" is introduced to describe local and global shape properties of a protein helix via its axis. HAXIS is based on coarse-graining and spline-fitting of the helix backbone. At each Cα anchor point of the backbone, a Frenet frame is calculated, which directly provides the local vector presentation of the helix. After cubic spline-fitting of the axis line, its curvature and torsion are calculated. This makes a rapid comparison of different helix forms and the determination of helix similarity possible. Distortions of the helix caused by individual residues are projected onto the helix axis and presented either by the rise parameter per residue or by the local curvature of the axis. From a non-redundant set of 2,017 proteins, 15,068 helices were investigated in this way. Helix start and helix end as well as bending and kinking of the helix are accurately described. The global properties of the helix are assessed by a polynomial fit of the helix axis and the determination of its overall curving and twisting. Long helices are more regular shaped and linear whereas short helices are often strongly bent and twisted. The distribution of different helix forms as a function of helix length is analyzed.