Biology subjects

Tian, P.

Publications and source records attributed to Tian, P..

31 records · Page 2Linked to original sources

bioRxiv2021

Kidney organoids: A system to study human basement membrane assembly in health and disease

Basement membranes (BMs) are complex macromolecular networks underlying all continuous layers of cells. Essential components include type IV collagen and laminins, which are affected by human genetic defects leading to a range of debilitating conditions including kidney, muscle, and cerebrovascular phenotypes. We investigated the dynamics of BM assembly in human pluripotent stem cell-derived kidney organoids. We resolved their global BM composition and discovered a conserved temporal sequence in BM assembly that paralleled mammalian fetal kidneys. We identified the emergence of key BM isoforms, which were altered by a pathogenic variant in COL4A5. Integrating organoid, fetal and adult kidney proteomes we found dynamic regulation of BM composition through development to adulthood, and with single-cell transcriptomic analysis we mapped the cellular origins of BM components. Overall, we define the complex and dynamic nature of vertebrate BM assembly and provide a platform for understanding its wider relevance in human development and disease.

cell biology↗

bioRxiv2020

The ribosome modulates folding inside the ribosomal exit tunnel

Proteins commonly fold cotranslationally on the ribosome, while the nascent chain emerges from the ribosomal tunnel. Protein domains that are sufficiently small can even fold while still located inside the tunnel. However, the effect of the tunnel on the folding dynamics of these domains is still not well understood. Here, we combine optical tweezers with single-molecule FRET and molecular dynamics simulations to investigate folding of the small zinc-finger domain ADR1a inside and at the vestibule of the ribosomal tunnel. The tunnel is found to accelerate folding and stabilize the folded state, reminiscent of the effects of chaperonins. However, a simple mechanism involving stabilization by confinement does not reproduce the results. Instead, it appears that electrostatic interactions between the protein and ribosome contribute to the observed folding acceleration and stabilization of ADR1a.Competing Interest StatementThe authors have declared no competing interest.View Full Text

biophysics↗

bioRxiv2020

Exploring the Sequence Fitness Landscape of a Bridge Between Protein Folds

Most foldable protein sequences adopt only a single native fold. Recent protein design studies have, however, created protein sequences which fold into different structures apon changes of environment, or single point mutation, the best characterized example being the switch between the folds of the GA and GB binding domains of streptococcal protein G. To obtain further insight into the design of sequences which can switch folds, we have used a computational model for the fitness landscape of a single fold, built from the observed sequence variation of protein homologues. We have recently shown that such coevolutionary models can be used to design novel foldable sequences. By appropriately combining two of these models to describe the joint fitness landscape of GA and GB, we are able to describe the propensity of a given sequence for each of the two folds. We have successfully tested the combined model against the known series of designed GA/GB hybrids. Using Monte Carlo simulations on this landscape, we are able to identify pathways of mutations connecting the two folds. In the absence of a requirement for domain stability, the most frequent paths go via sequences in which neither domain is stably folded, reminiscent of the propensity for certain intrinsically disordered proteins to fold into different structures according to context. Even if the folded state is required to be stable, we find that there is nonetheless still a wide range of sequences which are close to the transition region and therefore likely fold switches, consistent with recent estimates that fold switching may be more widespread than had been thought. Author SummaryWhile most proteins self-assemble (or "fold") to a unique three-dimensional structure, a few have been identified that can fold into two distinct structures. These so-called "metamorphic" proteins that can switch folds have attracted a lot of recent interest, and it has been suggested that they may be much more widespread than currently appreciated. We have developed a computational model that captures the propensity of a given protein sequence to fold into either one of two specific structures (GA and GB), in order to investigate which sequences are able to fold to both GA and GB ("switch sequences"), versus just one of them. Our model predicts that there is a large number of switch sequences that could fold into both structures, but also that the most likely such sequences are those for which the folded structures have low stability, in agreement with available experimental data. This also suggests that intrinsically disordered proteins which can fold into different structures on binding may provide an evolutionary path in sequence space between protein folds.

biophysics↗

bioRxiv2020

Scalable hierarchical clustering by composition rank vector encoding and tree structure

Supervised machine learning, especially deep learning based on a wide variety of neural network architectures, have contributed tremendously to fields such as marketing, computer vision and natural language processing. However, development of un-supervised machine learning algorithms has been a bottleneck of artificial intelligence. Clustering is a fundamental unsupervised task in many different subjects. Unfortunately, no present algorithm is satisfactory for clustering of high dimensional data with strong nonlinear correlations. In this work, we propose a simple and highly efficient hierarchical clustering algorithm based on encoding by composition rank vectors and tree structure, and demonstrate its utility with clustering of protein structural domains. No record comparison, which is an expensive and essential common step to all present clustering algorithms, is involved. Consequently, it achieves linear time and space computational complexity hierarchical clustering, thus applicable to arbitrarily large datasets. The key factor in this algorithm is definition of composition, which is dependent upon physical nature of target data and therefore need to be constructed case by case. Nonetheless, the algorithm is general and applicable to any high dimensional data with strong nonlinear correlations. We hope this algorithm to inspire a rich research field of encoding based clustering well beyond composition rank vector trees.

bioinformatics↗

bioRxiv2020

End to end differentiable protein structure refinement

Free energy is arguably the most important property of molecular systems. Despite great progress in both its efficient estimation by scoring functions/potentials and more rigorous computation based on extensive sampling, we remain far from accurately predicting and manipulating biomolecular structures and their interactions. There are fundamental limitations, including accuracy of interaction description and difficulty of sampling in high dimensional space, to be tackled. Computational graph underlies major artificial intelligence platforms and is proven to facilitate training, optimization and learning. Combining autodifferentiation, coordinates transformation and generalized solvation free energy theory, we construct a computational graph infrastructure to realize seamless integration of fully trainable local free energy landscape with end to end differentiable iterative free energy optimization. This new framework greatly improves efficiency by replacing local sampling with differentiation. Its specific implementation in protein structure refinement achieves superb efficiency and competitive accuracy when compared with state of the art all-atom mainstream methods.

bioinformatics↗

bioRxiv2019

Global atlas tree of natural proteins based on sorted composition vectors

Sequence comparison is the cornerstone of bioinformatics and is traditionally realized by alignment. Unfortunately, exponential computational complexity renders rigorous multiple sequence alignment (MSA) intractable. Approximate algorithms and heuristics provide acceptable performance for relatively small number of sequences but engender prohibitive computational cost and unbounded accumulation of error for massive sequence sets. Alignment free algorithms achieved linear computational cost for sequence pair comparison but the challenge for multiple sequence comparison (MSC) remains. Meanwhile, various number of parameters and procedures need to be empirically adjusted for different MSC tasks with their complex interactions and impact not well understood. Therefore, development of efficient and nonparametric global sequence comparison method is essential for explosive sequencing data. It is shown here that sorted composition vector (SCV), which is based on a physical perspective on sequence composition constraint, is a feasible non-parametric encoding scheme for global protein sequence comparison and classification with linear computational complexity, and provides a global atlas tree for natural protein sequences. This finding renders massive sequence comparison and classification, which is infeasible on supercomputers, routine on a workstation. SCV sets an example of one-way encoding that might revolutionize recognition and classification tasks in general.

bioinformatics↗

bioRxiv2019

Cotranslational folding cooperativity of contiguous domains of α-spectrin

Proteins synthesized in the cell can begin to fold during translation before the entire polypeptide has been produced, which may be particularly relevant to the folding of multidomain proteins. Here, we study the cotranslational folding of adjacent domains from the cytoskeletal protein -spectrin using Force Profile Analysis (FPA). Specifically, we investigate how the cotranslational folding behavior of the R15 and R16 domains are affected by their neighboring R14 and R16, and R15 and R17 domains, respectively. Our results show that the domains impact each others folding in distinct ways that may be important for the efficient assembly of -spectrin, and may reduce its dependence on chaperones. Furthermore, we directly relate the experimentally observed yield of full-length protein in the FPA assay to the force exerted by the folding protein in pN. By combining pulse-chase experiments to measure the rate at which the arrested protein is converted into full-length protein with a Bell model of force-induced rupture, we estimate that the R16 domain exerts a maximal force on the nascent chain of [~]15 pN during cotranslational folding. SignificanceIn living cells, proteins are produced in a sequential way by ribosomes. This vectoral process allows the growing protein chain to start to fold before translation has been completed. Thereby, cotranslational protein folding can be significantly different than the folding of a full-length protein in isolation. Here we show how structurally similar repeat domains, normally produced as parts of a single long polypeptide, affect the cotranslational folding of their neighbors. This provides insight into how the cell may efficiently produce multidomain proteins, paving the way for future studies in vivo or with chaperones. We also provide an estimated magnitude of the mechanical force on the nascent chain generated by cotranslational folding, calculated from biochemical measurements and molecular dynamics simulations.

biochemistry↗

bioRxiv2019

Protein secondary structure prediction with context convolutional neural network

Protein secondary structure (SS) prediction is important for studying protein structure and function. Both traditional machine learning methods and deep learning neural networks have been utilized and great progress has been achieved in approaching the theoretical limit. Convolutional and recurrent neural networks are two major types of deep leaning architectures with comparable prediction accuracy but different training procedures to achieve optimal performance. We are interested in seeking novel architectural style with competitive performance and in understanding performance of different architectures with similar training procedures.\n\nResultsWe constructed a context convolutional neural network (Contextnet) and compared its performance with popular models (e.g. convolutional neural network, recurrent neural network, conditional neural fields ...) under similar training procedures on Jpred dataset. the Contextnet was proven to be highly competitive. Additionally, we retrained the network with the Cullpdb data set and compared with Jpred, ReportX and Spider3 server, the Contextnet was found to be more accurate on CASP13 dataset. Training procedures were found to have significant impact on the accuracy of the Contextnet.\n\nAvailabilityThe full source code and dataset have been uploaded at https://github.com/qzlshy/second_structure_model and https://github.com/qzlshy/ss_pssm_hhm.\n\nContacttianpu@jlu.edu.cn

bioinformatics↗

bioRxiv2018

Monte Carlo Sampling of Protein Folding by Combining an All-Atom Physics-Based Model with a Native State Bias

Energy landscape theory suggests that native interactions are a major determinant of the folding mechanism of a protein. Thus, structure-based (G[o]) models have, aided by coarse-graining techniques, shown great success in capturing the mechanisms of protein folding and conformational changes. In certain cases, however, non-native interactions and atomic details are also essential to describe the protein dynamics, prompting the development of a variety of structure-based models which include non-native interactions, and differentiate between different types of attractive potentials. Here, we describe an all-protein-atom hybrid model, termed ProfasiGo, that integrates an implicit solvent all-atom physics-based model (called Profasi) and a structure-based G[o] potential, and its implementation in two software packages (PHAISTOS and ProFASi) that are developed for Monte Carlo sampling of protein molecules. We apply the ProfasiGo model to study the folding free energy landscapes of four topologically similar proteins, one of which can be folded by the simplified potential Profasi, and two that have been folded by explicit solvent, all-atom molecular dynamics simulations with the CHARMM22* force field. Our results reveal that the hybrid ProfasiGo model is able to capture many of the details present in the physics-based potentials, while retaining the advantages of G[o] models for sampling and guiding to the native state. We expect that the model will be widely applicable to study the folding of more complex proteins, or to study conformational dynamics and integration with experimental data.

biophysics↗

bioRxiv2018

A small single-domain protein folds through the same pathway on- and off- the ribosome

In vivo, proteins fold and function in a complex environment where they are subject to many stresses that can modulate protein energy landscapes. One aspect of the environment pertinent to protein folding is the ribosome, since proteins have the opportunity to fold while still bound to the ribosome during translation. We use a combination of force and chemical denaturant (chemo-mechanical unfolding), as well as point mutations, to characterize the folding mechanism of the src SH3 domain both as a stalled ribosome nascent chain and free in solution. Our results indicate that src SH3 folds through the same pathway on and off the ribosome. Molecular simulations also indicate that the ribosome does not affect the folding pathway for this small protein. Taken together, we conclude that the ribosome does not alter the folding mechanism of this small protein, which appears to fold at the mouth of the ribosome as the protein emerges from the exit tunnel. These results, if general, suggest the ribosome may exert a bigger influence on the folding of multi-domain proteins or protein domains that can partially fold before the entire domain sequence is outside the ribosome exit tunnel.

biophysics↗

bioRxiv2018

The Folding Pathway of an Ig Domain is Conserved On and Off the Ribosome

Proteins that fold cotranslationally may do so in a restricted configurational space, due to the volume occupied by the ribosome. How does this environment, coupled with the close proximity of the ribosome, affect the folding pathway of a protein? Previous studies have shown that the cotranslational folding process for many proteins, including small, single domains, is directly affected by the ribosome. Here, we investigate the cotranslational folding of an all-b immunoglobulin domain, titin I27. Using an arrest peptide-based assay and structural studies by cryo-EM, we show that I27 folds in the mouth of the ribosome exit tunnel. Simulations that use a kinetic model for the force-dependence of escape from arrest, accurately predict the fraction of folded protein as a function of length. We used these simulations to probe the folding pathway on and off the ribosome. Our simulations - which also reproduce experiments on mutant forms of I27 - show that I27 folds, while still sequestered in the mouth of the ribosome exit tunnel, by essentially the same pathway as free I27, with only subtle shifts of critical contacts from the C to the N terminus.\n\nSignificance StatementMost proteins need to fold into a specific three-dimensional structure in order to function. The mechanism by which isolated proteins fold has been thoroughly studied by experiment and theory. However, in the cell proteins do not fold in isolation, but are synthesized as linear chains by the ribosome during translation. It is therefore natural to ask at which point during synthesis proteins fold, and whether this differs from the folding of isolated protein molecules. By studying folding of a well characterized protein domain, titin I27, stalled at different points during translation, we show that it already folds in the mouth of the ribosome exit tunnel, and that the mechanism is almost identical to that of the isolated protein.

biophysics↗

bioRxiv2018

The universal mechanism of intermediate filament transport

Intermediate filaments (IFs) are a major component of the cytoskeleton that regulates a wide range of physiological properties in eukaryotic cells. In motile cells, the IF network has to adapt to constant changes of cell shape and tension. In this study, we used two cell lines that express vimentin and keratins 8/18 to study the dynamic behavior of these IFs. We demonstrated that both IF types undergo extensive transport along microtubules. This was an unexpected result as keratin filament remodeling has been described to depend on actin dynamics. We established the role of kinesin-1 in vimentin and keratin IF transport by knocking out KIF5B, the ubiquitous isoform of kinesin-1. Futhermore, we demonstrated that unlike typical membrane cargoes, transport of both types of IFs does not involve kinesin light chains, but requires the presence of the same region of the kinesin-1 tail, suggesting a unified mechanism of IF transport.

cell biology↗

bioRxiv2017

Peptide Terminus Tilting: an Unusual conformational transition of MHC Class I Revealed by Molecular Dynamics Simulations

A conventional picture for major histocompatibility complex class I (MHCI) antigen presentation is that the terminal anchor residues of the antigenic peptide bind to the pockets at the bottom of the MHC cleft, leaving the central peptide residues exposed for T cell antigen receptor (TCR) recognition. However, in the present study, we show that in canonical or accelerated molecular dynamics (MD) simulations, the peptide terminus in some immunodominant peptide-MHCI (pMHCI) complexes can detach from their binding pockets and stretch outside the MHC cleft. These pMHCI complexes include the complex of the H-2Kb and the lymphocytic choriomeningitis virus (LCMV) gp33 peptide, and the complex of the HLA-A*0201 and the influenza A virus M1 peptide. The detached peptide terminus becomes the most prominent spot at the pMHC interface, and so can serves as a novel TCR recognition target. Thus, peptide terminus detaching may be a novel mechanism for MHC antigen presentation.

immunology↗