A novel pattern recognition algorithm to classify membrane protein unfolding pathways with high-throughput single-molecule force spectroscopy

Iterative Machine Learning for Classification and Discovery of Single-Molecule Unfolding Trajectories from Force Spectroscopy Data

We report the application of machine learning techniques to expedite classification and analysis of protein unfolding trajectories from force spectroscopy data. Using kernel methods, logistic regression, and triplet loss, we developed a workflow called Forced Unfolding and Supervised Iterative Online (FUSION) learning where a user classifies a small number of repeatable unfolding patterns encoded as images, and a machine is tasked with identifying similar images to classify the remaining data. We tested the workflow using two case studies on a multidomain XMod-Dockerin/Cohesin complex, validating the approach first using synthetic data generated with a Monte Carlo algorithm and then deploying the method on experimental atomic force spectroscopy data. FUSION efficiently separated traces that passed quality filters from unusable ones, classified curves with high accuracy, and identified unfolding pathways that were undetected by the user. This study demonstrates the potential of machine learning to accelerate data analysis and generate new insights in protein biophysics.

Unfolding and identification of membrane proteins in situ

Single-molecule force spectroscopy (SMFS) uses the cantilever tip of an AFM to apply a force able to unfold a single protein. The obtained force-distance curve encodes the unfolding pathway, and from its analysis it is possible to characterize the folded domains. SMFS has been mostly used to study the unfolding of purified proteins, in solution or reconstituted in a lipid bilayer. Here, we describe a pipeline for analyzing membrane proteins based on SMFS, that involves the isolation of the plasma membrane of single cells and the harvesting of force-distance curves directly from it. We characterized and identified the embedded membrane proteins combining, within a Bayesian framework, the information of the shape of the obtained curves, with the information from Mass Spectrometry and proteomic databases. The pipeline was tested with purified/reconstituted proteins and applied to five cell types where we classified the unfolding of their most abundant membrane proteins. We validated our pipeline by overexpressing 4 constructs, and this allowed us to gather structural insights of the identified proteins, revealing variable elements in the loop regions. Our results set the basis for the investigation of the unfolding of membrane proteins in situ, and for performing proteomics from a membrane fragment.

Automatic classification of single-molecule force spectroscopy traces from heterogeneous samples

MOTIVATION

Single-molecule force spectroscopy (SMFS) experiments pose the challenge of analysing protein unfolding data (traces) coming from preparations with heterogeneous composition (e.g. where different proteins are present in the sample). An automatic procedure able to distinguish the unfolding patterns of the proteins is needed. Here, we introduce a data analysis pipeline able to recognize in such datasets traces with recurrent patterns (clusters).

RESULTS

We illustrate the performance of our method on two prototypical datasets: ∼50 000 traces from a sample containing tandem GB1 and ∼400 000 traces from a native rod membrane. Despite a daunting signal-to-noise ratio in the data, we are able to identify several unfolding clusters. This work demonstrates how an automatic pattern classification can extract relevant information from SMFS traces from heterogeneous samples without prior knowledge of the sample composition.

AVAILABILITY AND IMPLEMENTATION

https://github.com/ninailieva/SMFS_clustering.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Fodis: Software for Protein Unfolding Analysis

The folding dynamics of proteins at the single-molecule level has been studied with single-molecule force spectroscopy experiments for 20 years, but a common standardized method for the analysis of the collected data and for sharing among the scientific community members is still not available. We have developed a new open-source tool-Fodis-for the analysis of the force-distance curves obtained in single-molecule force spectroscopy experiments, providing almost automatic processing, analysis, and classification of the obtained data. Our method provides also a classification of the possible unfolding pathways and the structural heterogeneity present during the unfolding of proteins.

Evolutionary Influenced Interaction Pattern as Indicator for the Investigation of Natural Variants Causing Nephrogenic Diabetes Insipidus

The importance of short membrane sequence motifs has been shown in many works and emphasizes the related sequence motif analysis. Together with specific transmembrane helix-helix interactions, the analysis of interacting sequence parts is helpful for understanding the process during membrane protein folding and in retaining the three-dimensional fold. Here we present a simple high-throughput analysis method for deriving mutational information of interacting sequence parts. Applied on aquaporin water channel proteins, our approach supports the analysis of mutational variants within different interacting subsequences and finally the investigation of natural variants which cause diseases like, for example, nephrogenic diabetes insipidus. In this work we demonstrate a simple method for massive membrane protein data analysis. As shown, the presented in silico analyses provide information about interacting sequence parts which are constrained by protein evolution. We present a simple graphical visualization medium for the representation of evolutionary influenced interaction pattern pairs (EIPPs) adapted to mutagen investigations of aquaporin-2, a protein whose mutants are involved in the rare endocrine disorder known as nephrogenic diabetes insipidus, and membrane proteins in general. Furthermore, we present a new method to derive new evolutionary variations within EIPPs which can be used for further mutagen laboratory investigations.

Graph representation of high-dimensional alpha-helical membrane protein data

Background

In genomics and proteomics, membrane protein analysis have shown that such analyses are very important to support the understanding of complex biological processes. In Genome-wide investigations of membrane proteins a large number of short, distinct sequence motifs has been revealed. Such motifs found so far support the understanding of the folded membrane protein in the membrane environment. They provide important information about functional or stabilizing properties. Recently several integrative approaches have been proposed to extract meaningful information out of the membrane environment. However, many information based approaches deliver results having deficits of visualisation outputs. Outgoing from high-throughput protein data analysis, these outputs play an important role in the evaluation of high-dimensional protein data, to establish a biological relationship and ultimately to provide useful information for research.

Results

We have evaluated different resulting graphs generated from statistical analysis of consecutive motifs in helical structures of the membrane environment. Our results show that representative motifs with high occurrence in all investigated protein families are responsible for the general importance in alpha-helical membrane structure formation. Further, motifs which often occur with others in their function as so called “hubs” lead to the assumption, that these motifs constitute as important components in helical structures within the membrane. Otherwise, consecutive motifs and hubs which show a high occurrence in certain families only can be classified as important for family-specific functional characteristics. Summarized, we are able to bridge our graphical results from high-throughput analysis of membrane proteins over networking with databases to a biological context.

Conclusions

Our results and the corresponding graphical visualisation support the understanding and interpretation of structure forming and functional motifs of membrane proteins. Our results are useful to interpret and refine results of common developed approaches. At last we show a simple way to visualise high-dimensional protein data in context to biological relevant information.

Supersecondary structure prediction of transmembrane beta-barrel proteins

We introduce a graph-theoretic model for predicting the supersecondary structure of transmembrane β-barrel proteins--a particular class of proteins that performs diverse important functions but it is difficult to determine their structure with experimental methods. This ab initio model resolves the protein folding problem based on pseudo-energy minimization with the aid of a simple probabilistic filter. It also allows for determining structures whose barrel follows a given permutation on the arrangement of β-strands, and allows for rapidly discriminating the transmembrane β-barrels from other kinds of proteins. The model is fairly accurate, robust and can be run very efficiently on PC-like computers, thus proving useful for genome screening.

Reference-free alignment and sorting of single-molecule force spectroscopy data

Single-molecule force spectroscopy has become a versatile tool for investigating the (un)folding of proteins and other polymeric molecules. Like other single-molecule techniques, single-molecule force spectroscopy requires recording and analysis of large data sets to extract statistically meaningful conclusions. Here, we present a data analysis tool that provides efficient filtering of heterogeneous data sets, brings spectra into register based on a reference-free alignment algorithm, and determines automatically the location of unfolding barriers. Furthermore, it groups spectra according to the number of unfolding events, subclassifies the spectra using cross correlation-based sorting, and extracts unfolding pathways by principal component analysis and clustering methods to extracted peak positions. Our approach has been tested on a data set obtained through mechanical unfolding of bacteriorhodopsin (bR), which contained a significant number of spectra that did not show the well-known bR fingerprint. In addition, we have tested the performance of the data analysis tool on unfolding data of the soluble multidomain (Ig27)(8) protein.

Membrane protein stability analyses by means of protein energy profiles in case of nephrogenic diabetes insipidus

Diabetes insipidus (DI) is a rare endocrine, inheritable disorder with low incidences in an estimated one per 25,000-30,000 live births. This disease is characterized by polyuria and compensatory polydypsia. The diverse underlying causes of DI can be central defects, in which no functional arginine vasopressin (AVP) is released from the pituitary or can be a result of defects in the kidney (nephrogenic DI, NDI). NDI is a disorder in which patients are unable to concentrate their urine despite the presence of AVP. This antidiuretic hormone regulates the process of water reabsorption from the prourine that is formed in the kidney. It binds to its type-2 receptor (V2R) in the kidney induces a cAMP-driven cascade, which leads to the insertion of aquaporin-2 water channels into the apical membrane. Mutations in the genes of V2R and aquaporin-2 often lead to NDI. We investigated a structure model of V2R in its bound and unbound state regarding protein stability using a novel protein energy profile approach. Furthermore, these techniques were applied to the wild-type and selected mutations of aquaporin-2. We show that our results correspond well to experimental water ux analysis, which confirms the applicability of our theoretical approach to equivalent problems.

Efficient unfolding pattern recognition in single molecule force spectroscopy data

Background

Single-molecule force spectroscopy (SMFS) is a technique that measures the force necessary to unfold a protein. SMFS experiments generate Force-Distance (F-D) curves. A statistical analysis of a set of F-D curves reveals different unfolding pathways. Information on protein structure, conformation, functional states, and inter- and intra-molecular interactions can be derived.

Results

In the present work, we propose a pattern recognition algorithm and apply our algorithm to datasets from SMFS experiments on the membrane protein bacterioRhodopsin (bR). We discuss the unfolding pathways found in bR, which are characterised by main peaks and side peaks. A main peak is the result of the pairwise unfolding of the transmembrane helices. In contrast, a side peak is an unfolding event in the alpha-helix or other secondary structural element. The algorithm is capable of detecting side peaks along with main peaks.

Therefore, we can detect the individual unfolding pathway as the sequence of events labeled with their occurrences and co-occurrences special to bR's unfolding pathway. We find that side peaks do not co-occur with one another in curves as frequently as main peaks do, which may imply a synergistic effect occurring between helices. While main peaks co-occur as pairs in at least 50% of curves, the side peaks co-occur with one another in less than 10% of curves. Moreover, the algorithm runtime scales well as the dataset size increases.

Conclusions

Our algorithm satisfies the requirements of an automated methodology that combines high accuracy with efficiency in analyzing SMFS datasets. The algorithm tackles the force spectroscopy analysis bottleneck leading to more consistent and reproducible results.

Atomic force microscopy force mapping in the study of supported lipid bilayers

Investigating the structural and mechanical properties of lipid bilayer membrane systems is vital in elucidating their biological function. One route to directly correlate the morphology of phase-segregated membranes with their indentation and rupture mechanics is the collection of atomic force microscopy (AFM) force maps. These force maps, while containing rich mechanical information, require lengthy processing time due to the large number of force curves needed to attain a high spatial resolution. A force curve analysis toolset was created to perform data extraction, calculation and reporting specifically in studying lipid membrane morphology and mechanical stability. The procedure was automated to allow for high-throughput processing of force maps with greatly reduced processing time. The resulting program was successfully used in systematically analyzing a number of supported lipid membrane systems in the investigation of their structure and nanomechanics.

Hooke: an open software platform for force spectroscopy

SUMMARY

Hooke is an open source, extensible software intended for analysis of atomic force microscope (AFM)-based single molecule force spectroscopy (SMFS) data. We propose it as a platform on which published and new algorithms for SMFS analysis can be integrated in a standard, open fashion, as a general solution to the current lack of a standard software for SMFS data analysis. Specific features and support for file formats are coded as independent plugins. Any user can code new plugins, extending the software capabilities. Basic automated dataset filtering and semi-automatic analysis facilities are included.

AVAILABILITY

Software and documentation are available at (http://code.google.com/p/hooke). Hooke is a free software under the GNU Lesser General Public License.

Fully automated single-molecule force spectroscopy for screening applications

With the introduction of single-molecule force spectroscopy (SMFS) it has become possible to directly access the interactions of various molecular systems. A bottleneck in conventional SMFS is collecting the large amount of data required for statistically meaningful analysis. Currently, atomic force microscopy (AFM)-based SMFS requires the user to tediously 'fish' for single molecules. In addition, most experimental and environmental conditions must be manually adjusted. Here, we developed a fully automated single-molecule force spectroscope. The instrument is able to perform SMFS while monitoring and regulating experimental conditions such as buffer composition and temperature. Cantilever alignment and calibration can also be automatically performed during experiments. This, combined with in-line data analysis, enables the instrument, once set up, to perform complete SMFS experiments autonomously.

High-throughput single-molecule force spectroscopy for membrane proteins

Atomic force microscopy-based single-molecule force spectroscopy (SMFS) is a powerful tool for studying the mechanical properties, intermolecular and intramolecular interactions, unfolding pathways, and energy landscapes of membrane proteins. One limiting factor for the large-scale applicability of SMFS on membrane proteins is its low efficiency in data acquisition. We have developed a semi-automated high-throughput SMFS (HT-SMFS) procedure for efficient data acquisition. In addition, we present a coarse filter to efficiently extract protein unfolding events from large data sets. The HT-SMFS procedure and the coarse filter were validated using the proton pump bacteriorhodopsin (BR) from Halobacterium salinarum and the L-arginine/agmatine antiporter AdiC from the bacterium Escherichia coli. To screen for molecular interactions between AdiC and its substrates, we recorded data sets in the absence and in the presence of L-arginine, D-arginine, and agmatine. Altogether ∼400 000 force-distance curves were recorded. Application of coarse filtering to this wealth of data yielded six data sets with ∼200 (AdiC) and ∼400 (BR) force-distance spectra in each. Importantly, the raw data for most of these data sets were acquired in one to two days, opening new perspectives for HT-SMFS applications.

Comparing proteins by their unfolding pattern

Single molecule force spectroscopy has evolved into an important and extremely powerful technique for investigating the folding potentials of biomolecules. Mechanical tension is applied to individual molecules, and the subsequent, often stepwise unfolding is recorded in force extension traces. However, because the energy barriers of the folding potentials are often close to the thermal energy, both the extensions and the forces at which these barriers are overcome are subject to marked fluctuations. Therefore, force extension traces are an inadequate representation despite widespread use particularly when large populations of proteins need to be compared and analyzed. We show in this article that contour length, which is independent of fluctuations and alterable experimental parameters, is a more appropriate variable than extension. By transforming force extension traces into contour length space, histograms are obtained that directly represent the energy barriers. In contrast to force extension traces, such barrier position histograms can be averaged to investigate details of the unfolding potential. The cross-superposition of barrier position histograms allows us to detect and visualize the order of unfolding events. We show with this approach that in contrast to the sequential unfolding of bacteriorhodopsin, two main steps in the unfolding of the enzyme titin kinase are independent of each other. The potential of this new method for accurate and automated analysis of force spectroscopy data and for novel automated screening techniques is shown with bacteriorhodopsin and with protein constructs containing GFP and titin kinase.

From valleys to ridges: exploring the dynamic energy landscape of single membrane proteins

Membrane proteins are involved in essential biological processes such as energy conversion, signal transduction, solute transport and secretion. All biological processes, also those involving membrane proteins, are steered by molecular interactions. Molecular interactions guide the folding and stability of membrane proteins, determine their assembly, switch their functional states or mediate signal transduction. The sequential steps of molecular interactions driving these processes can be described by dynamic energy landscapes. The conceptual energy landscape allows to follow the complex reaction pathways of membrane proteins while its modifications describe why and how pathways are changed. Single-molecule force spectroscopy (SMFS) detects, quantifies and locates interactions within and between membrane proteins. SMFS helps to determine how these interactions change with temperature, point mutations, oligomerization and the functional states of membrane proteins. Applied in different modes, SMFS explores the co-existence and population of reaction pathways in the energy landscape of the protein and thus reveals detailed insights into local mechanisms, determining its structural and functional relationships. Here we review how SMFS extracts the defining parameters of an energy landscape such as the barrier position, reaction kinetics and roughness with high precision.

Vertebrate membrane proteins: structure, function, and insights from biophysical approaches

Membrane proteins are key targets for pharmacological intervention because they are vital for cellular function. Here, we analyze recent progress made in the understanding of the structure and function of membrane proteins with a focus on rhodopsin and development of atomic force microscopy techniques to study biological membranes. Membrane proteins are compartmentalized to carry out extra- and intracellular processes. Biological membranes are densely populated with membrane proteins that occupy approximately 50% of their volume. In most cases membranes contain lipid rafts, protein patches, or paracrystalline formations that lack the higher-order symmetry that would allow them to be characterized by diffraction methods. Despite many technical difficulties, several crystal structures of membrane proteins that illustrate their internal structural organization have been determined. Moreover, high-resolution atomic force microscopy, near-field scanning optical microscopy, and other lower resolution techniques have been used to investigate these structures. Single-molecule force spectroscopy tracks interactions that stabilize membrane proteins and those that switch their functional state; this spectroscopy can be applied to locate a ligand-binding site. Recent development of this technique also reveals the energy landscape of a membrane protein, defining its folding, reaction pathways, and kinetics. Future development and application of novel approaches during the coming years should provide even greater insights to the understanding of biological membrane organization and function.