Detecting misfolded non-covalent lasso entanglements in protein structures, simulation trajectories, and mass spectrometry data

This paper introduces EntDetect, an open-source computational tool and accompanying webserver designed to identify non-covalent lasso entanglements (NCLEs) in protein structures and simulation trajectories, enabling the detection of misfolded states and integration with mass spectrometry data to advance the study of protein topology and misfolding.

The Gist

The Big Idea: Protein Tangles and the "Lasso" Problem

Imagine your body is a bustling city, and the proteins inside it are the workers. To do their jobs, these workers need to fold themselves into very specific, intricate shapes—like origami.

Sometimes, however, a protein gets confused and folds into the wrong shape. This is called misfolding, and it can lead to diseases.

For a long time, scientists knew about "knots" in proteins (like a shoelace tied in a knot). But this paper introduces a new, overlooked type of tangle called a Non-Covalent Lasso Entanglement (NCLE).

The Lasso Analogy:
Imagine a protein is a long, flexible rope.

• The Loop: Part of the rope folds back on itself and ties a loose loop (held together by weak magnetic forces, not a permanent knot).
• The Thread: The two ends of the rope (the N-terminus and C-terminus) act like a lasso thrower.
• The Entanglement: If one of the rope's ends slips through that loop and gets stuck, you have a Lasso Entanglement.

In a healthy protein, this lasso happens exactly where it's supposed to. But if the protein misfolds, two bad things can happen:

1. The Lost Lasso: The loop forms, but the end of the rope slips out. The protein loses its structural integrity.
2. The Fake Lasso: The rope forms a loop where it shouldn't, and an end gets stuck in it. This creates a "traffic jam" that stops the protein from working.

The Problem: We Didn't Have a "Tangle Detector"

Until now, scientists had tools to find knots, but they didn't have a good way to find these specific "lasso" tangles, especially in complex computer simulations or messy experimental data. It was like trying to find a specific type of knot in a pile of tangled headphones without a flashlight.

Because they couldn't see these tangles, they couldn't fully understand why some proteins were getting stuck in broken shapes.

The Solution: EntDetect (The "Tangle Finder")

The authors built a new software tool called EntDetect. Think of it as a high-tech metal detector for protein tangles.

Here is what EntDetect does, broken down simply:

1. It Maps the Territory (The "Lasso Map")
It scans a protein structure and says, "Okay, here is a loop, and here is the end of the rope threading through it." It creates a map of every lasso in the protein.

2. It Spots the Mistakes (The "Misfold Alarm")
If you feed it a simulation of a protein folding, EntDetect watches the lassos in real-time. If a lasso disappears or a new, weird one appears, it screams, "Alert! This protein is misfolding!"

• Analogy: Imagine watching a origami crane being folded. If a piece of paper suddenly pokes through a loop it shouldn't, EntDetect is the person who yells, "Stop! You folded that wrong!"

3. It Talks to the Lab (The "Reality Check")
Scientists often use Mass Spectrometry (a fancy scale that weighs protein pieces) to see how proteins change shape. But the data is noisy. EntDetect takes the computer simulation of the "tangled" protein and compares it to the real-world lab data.

• Analogy: It's like a detective comparing a suspect's alibi (the computer simulation) with witness testimony (the lab data). If the alibi matches the witness, the suspect is likely guilty (or in this case, the simulation is likely correct).

4. It Finds the "Bad Apples" in the Whole Orchard (The "Proteome Hunt")
Sometimes scientists look at thousands of proteins at once (the whole proteome). It's hard to spot a problem in just one. EntDetect uses a statistical trick (like a Monte Carlo simulation) to group proteins together. It asks, "Do proteins with lassos tend to break more often than those without?" This helps them pick the most dangerous proteins to study further.

Why Does This Matter?

• Better Medicine: If we know exactly how a protein is getting tangled, we can design drugs to untie it or prevent the tangle in the first place.
• Understanding Disease: Many diseases (like Alzheimer's or Parkinson's) involve proteins misfolding. This tool helps us see the specific "lasso" errors causing the trouble.
• Saving Time: Instead of guessing why a protein isn't working, scientists can now look at the tangle map and know exactly what went wrong.

The Bottom Line

This paper is about building a new pair of glasses that lets scientists see a specific, invisible type of knot (the lasso) in proteins. By seeing these tangles, they can finally understand why proteins break, how to fix them, and how to stop diseases caused by these molecular "knots."

The tool is free and open for anyone to use, meaning the whole scientific community can now start hunting for these elusive tangles.

Technical Summary

1. Problem Statement

Protein misfolding is a critical issue in biology, often leading to loss of function or disease. While traditional metrics like Root Mean Square Deviation (RMSD) and the fraction of native contacts ($Q$) are standard for analyzing folding, they often fail to detect specific topological misfolding events.

• The Gap: A specific class of entanglements known as Non-Covalent Lasso Entanglements (NCLEs)—where a protein loop is threaded by a terminus via non-covalent interactions—is prevalent in native structures but difficult to track computationally.
• The Challenge: Misfolding can occur via the loss of a native NCLE or the gain of a non-native NCLE. These events can create kinetically trapped intermediates that conventional order parameters miss. Furthermore, there is a lack of tools to correlate these topological changes with high-throughput experimental data (Mass Spectrometry) or to identify misfolding candidates across entire proteomes.

2. Methodology: The EntDetect Protocol

The authors present EntDetect, an open-source Python-based software suite and protocol designed to detect, characterize, and analyze NCLEs. The workflow is divided into four main components:

A. Detection and Characterization of NCLEs

• Definition: An NCLE is defined by a loop closed by a non-covalent contact (residues $i, j$) and threaded by a terminus segment.
• Mathematical Core: The protocol uses a discrete version of the Gaussian Linking Number (GLN) to determine if a loop is threaded.
  • A threshold of $|g| \geq 0.6$ is used to distinguish true entanglements from noise (a modification from the standard 0.5 threshold to reduce false positives).
  • For higher accuracy, the Topoly package is used to calculate the Topoly Linking Number (TLN) and identify specific crossing residues via surface tessellation.
• Clustering: To handle degeneracy (multiple similar entanglements), a clustering algorithm groups entanglements based on spatial proximity of crossing residues and loop locations, outputting a set of non-redundant representative NCLEs.

B. Analysis of Simulation Trajectories (MD)

• Order Parameters:
  • $Q$: Fraction of native contacts.
  • $G$: The fraction of native contacts that exhibit a change in entanglement status (linking number) relative to a reference structure.
  • $K$: A chirality order parameter to detect and filter "mirror image" artifacts in simulations.
• Metastable State Identification: The protocol clusters simulation frames using $Q$ and $G$ to define microstates, which are then grouped into metastable states using Markov State Modeling (MSM). This allows for the mapping of folding pathways and the identification of kinetically trapped misfolded states.
• Hysteresis Analysis: The Jensen-Shannon Divergence (JSD) is used to compare structural ensembles from different initial conditions (e.g., fast vs. slow translation) to quantify divergence.

C. Integration with Mass Spectrometry (LiP-MS and XL-MS)

The protocol validates simulated ensembles against experimental data:

• Limited Proteolysis (LiP-MS): Compares experimentally observed protease cut-sites with changes in Solvent Accessible Surface Area (SASA) in simulated metastable states.
• Cross-Linking (XL-MS): Compares observed cross-linking frequencies with simulated Cross-Linking Propensity (XP) scores, calculated using a modified MNXL scoring function based on SASD (Solvent Accessible Surface Distance).
• Consistency Test: A statistical permutation test identifies which metastable states in the simulation ensemble are most consistent with the experimental signals.

D. Proteome-Wide Analysis

For large-scale datasets where single-protein statistical power is low:

• Logistic Regression: Models the odds of a protein misfolding as a function of the presence of native NCLEs, controlling for confounders (amino acid composition, protein length, SASA).
• Monte Carlo Selection: A Monte Carlo algorithm (Metropolis criteria) is used to randomly partition proteins into groups and optimize an objective function. This identifies sub-populations of proteins with the strongest statistical association between native entanglements and observed misfolding signals, selecting candidates for further study.

3. Key Contributions

1. EntDetect Software: A comprehensive, open-source tool for identifying NCLEs, calculating topological order parameters ($G$), and performing MSM analysis.
2. NCLEweb: A webserver allowing users to upload protein structures to identify NCLEs without coding.
3. Novel Order Parameter ($G$): Introduces a topology-aware metric specifically designed to detect misfolding events that preserve native contacts but alter entanglement topology.
4. Experimental Integration: A rigorous, post-hoc statistical framework to validate simulation ensembles against LiP-MS and XL-MS data without biasing the simulation force fields.
5. Proteome-Scale Screening: A statistical pipeline to overcome low statistical power in high-throughput MS data to identify proteins prone to entanglement-driven misfolding.

4. Results and Validation

• Case Study (Phosphoglycerate Kinase - PGK): The protocol was applied to a coarse-grained simulation ensemble of E. coli PGK.
  • Misfolding Detection: The analysis successfully identified metastable states with high $G$ values (altered topology) that were near-native in $Q$ but kinetically trapped.
  • Experimental Consistency: The identified metastable states were cross-referenced with LiP-MS and XL-MS data. The protocol successfully pinpointed specific misfolded ensembles that matched the experimental protease accessibility and cross-linking patterns, validating the topological hypothesis.
  • Pathway Analysis: The MSM revealed specific folding pathways and transitions into kinetically trapped states, demonstrating how translation speed (simulated as initial conditions) influences misfolding.
• Proteome-Wide Application: Applied to a proteome-wide LiP-MS dataset, the logistic regression and Monte Carlo selection identified specific subsets of proteins where native entanglements significantly correlated with conformational changes upon refolding, providing a list of high-priority candidates for experimental validation.

5. Significance

• Mechanistic Insight: This work provides a mechanistic explanation for protein misfolding that was previously invisible to standard structural metrics, linking topological errors to kinetic traps.
• Therapeutic Design: By identifying specific non-native entangled states, the protocol offers targets for small-molecule therapeutics designed to correct misfolding or stabilize native states.
• Bridging Simulation and Experiment: It establishes a robust workflow for interpreting complex mass spectrometry data through the lens of protein topology, moving beyond simple structural averaging.
• Accessibility: By providing open-source tools and a webserver, the authors democratize the study of protein entanglements, enabling the broader community to investigate the role of NCLEs in disease, aging, and enzyme function.

In summary, this paper presents a transformative approach to protein structural biology, shifting the focus from purely geometric distance metrics to topological connectivity, thereby uncovering a hidden layer of protein misfolding mechanisms.