Lessons from the lysozyme of phage T4

Protein folding as a jamming transition

Proteins fold to a specific functional conformation with a densely packed hydrophobic core that controls their stability. We develop a geometric, yet all-atom model for proteins that explains the universal core packing fraction of ϕ c = 0.55 found in experimental measurements. We show that as the hydrophobic interactions increase relative to the temperature, a novel jamming transition occurs when the core packing fraction exceeds ϕ c . The model also recapitulates the global structure of proteins since it can accurately refold to native-like structures from partially unfolded states.

Enhancing bezlotoxumab binding to C. difficile toxin B2: insights from computational simulations and mutational analyses for antibody design

Clostridioides difficile infection (CDI) is a significant concern caused by widespread antibiotic use, resulting in diarrhea and inflammation from the gram-positive anaerobic bacterium C. difficile. Although bezlotoxumab (Bez), a monoclonal antibody (mAb), was developed to address CDI recurrences, the recurrence rate remains high, partly due to reduced neutralization efficiency against toxin B2. In this study, we aimed to enhance the binding of Bez to C. difficile toxin B2 by combining computational simulations and mutational analyses. We identified specific mutations in Bez, including S28R, S31W/K, Y32R, S56W and G103D/S in the heavy chain (Hc), and S32F/H/R/W/Y in the light chain (Lc), which significantly improved binding to toxin B2 and formed critical protein-protein interactions. Through molecular dynamics simulations, several single mutations, such as HcS28R, LcS32H, LcS32R, LcS32W and LcS32Y, exhibited superior binding affinities to toxin B2 compared to Bez wild-type (WT), primarily attributed to Coulombic interactions. Combining the HcS28R mutation with four different mutations at residue LcS32 led to even greater binding affinities in double mutants (MTs), particularly HcS28R/LcS32H, HcS28R/LcS32R and HcS28R/LcS32Y, reinforcing protein-protein binding. Analysis of per-residue decomposition free energy highlighted key residues contributing significantly to enhanced binding interactions, emphasizing the role of electrostatic interactions. These findings offer insights into rational Bez MT design for improved toxin B2 binding, providing a foundation for developing more effective antibodies to neutralize toxin B2 and combat-related infections.Communicated by Ramaswamy H. Sarma.

Connecting polymer collapse and the onset of jamming

Previous studies have shown that the interiors of proteins are densely packed, reaching packing fractions that are as large as those found for static packings of individual amino-acid-shaped particles. How can the interiors of proteins take on such high packing fractions given that amino acids are connected by peptide bonds and many amino acids are hydrophobic with attractive interactions? We investigate this question by comparing the structural and mechanical properties of collapsed attractive disk-shaped bead-spring polymers to those of three reference systems: static packings of repulsive disks, of attractive disks, and of repulsive disk-shaped bead-spring polymers. We show that the attractive systems quenched to temperatures below the glass transition T≪T_{g} and static packings of both repulsive disks and bead-spring polymers possess similar interior packing fractions. Previous studies have shown that static packings of repulsive disks are isostatic at jamming onset, i.e., the number of interparticle contacts N_{c} matches the number of degrees of freedom, which strongly influences their mechanical properties. We find that repulsive polymer packings are hypostatic at jamming onset (i.e., with fewer contacts than degrees of freedom) but are effectively isostatic when including stabilizing quartic modes, which give rise to quartic scaling of the potential energy with displacements along these modes. While attractive disk and polymer packings are often considered hyperstatic with excess contacts over the isostatic number, we identify a definition for interparticle contacts for which they can also be considered as effectively isostatic. As a result, we show that the mechanical properties (e.g., scaling of the potential energy with excess contact number and low-frequency contribution to the density of vibrational modes) of weakly attractive disk and polymer packings are similar to those of isostatic repulsive disk and polymer packings. Our results demonstrate that static packings generated via attractive collapse or compression of repulsive particles possess similar structural and mechanical properties.

Stepwise introduction of stabilizing mutations reveals nonlinear additive effects in de novo TIM barrels

Over the past decades, the TIM-barrel fold has served as a model system for the exploration of how changes in protein sequences affect their structural, stability, and functional characteristics, and moreover, how this information can be leveraged to design proteins from the ground up. After numerous attempts to design de novo proteins with this specific fold, sTIM11 was the first validated de novo design of an idealized four-fold symmetric TIM barrel. Subsequent efforts to enhance the stability of this initial design resulted in the development of DeNovoTIMs, a family of de novo TIM barrels with various stabilizing mutations. In this study, we present an investigation into the biophysical and thermodynamic effects upon introducing a varying number of stabilizing mutations per quarter along the sequence of a four-fold symmetric TIM barrel. We compared the base design DeNovoTIM0 without any stabilizing mutations with variants containing mutations in one, two, three, and all four quarters-designated TIM1q, TIM2q, TIM3q, and DeNovoTIM6, respectively. This analysis revealed a stepwise and nonlinear change in the thermodynamic properties that correlated with the number of mutated quarters, suggesting positive nonadditive effects. To shed light on the significance of the location of stabilized quarters, we engineered two variants of TIM2q which contain the same number of mutations but positioned in different quarter locations. Characterization of these TIM2q variants revealed that the mutations exhibit varying effects on the overall protein stability, contingent upon the specific region in which they are introduced. These findings emphasize that the amount and location of stabilized interfaces among the four quarters play a crucial role in shaping the conformational stability of these four-fold symmetric TIM barrels. Analysis of de novo proteins, as described in this study, enhances our understanding of how sequence variations can finely modulate stability in both naturally occurring and computationally designed proteins.

Calculation of Protein Folding Thermodynamics Using Molecular Dynamics Simulations

Despite advances in artificial intelligence methods, protein folding remains in many ways an enigma to be solved. Accurate computation of protein folding energetics could help drive fields such as protein and drug design and genetic interpretation. However, the challenge of calculating the state functions governing protein folding from first-principles remains unaddressed. We present here a simple approach that allows us to accurately calculate the energetics of protein folding. It is based on computing the energy of the folded and unfolded states at different temperatures using molecular dynamics simulations. From this, two essential quantities (ΔH and ΔCp) are obtained and used to calculate the conformational stability of the protein (ΔG). With this approach, we have successfully calculated the energetics of two- and three-state proteins, representatives of the major structural classes, as well as small stability differences (ΔΔG) due to changes in solution conditions or variations in an amino acid residue.

ATP induces folding of ALS-causing C71G-hPFN1 and nascent hSOD1

ALS-causing C71G-hPFN1 coexists in both folded and unfolded states, while nascent hSOD1 is unfolded. So far, the mechanisms underlying their ALS-triggering potential remain enigmatic. Here we show by NMR that ATP completely converts C71G-hPFN1 into the folded state at a 1:2 ratio, while inducing nascent hSOD1 into two co-existing states at a 1:8 ratio. Surprisingly, the inducing capacity of ATP comes from its triphosphate, but free triphosphate triggers aggregation. The inducing capacity ranks as: ATP = ATPP = PPP > ADP = AMP-PNP = AMP-PCP = PP, while AMP, adenosine, P, and NaCl show no conversion. Mechanistically, ATP and triphosphate appear to enhance the intrinsic folding capacity encoded in the sequences, as unveiled by comparing conformations and dynamics of ATP- and Zn2+-induced hSOD1 folded states. Our study provides a mechanism for the finding that some single-cell organisms employ polyphosphates as primordial chaperones, and sheds light on the enigma of age-related onset of familial ALS and risk increase of neurodegenerative diseases.

Large language models generate functional protein sequences across diverse families

Deep-learning language models have shown promise in various biotechnological applications, including protein design and engineering. Here we describe ProGen, a language model that can generate protein sequences with a predictable function across large protein families, akin to generating grammatically and semantically correct natural language sentences on diverse topics. The model was trained on 280 million protein sequences from >19,000 families and is augmented with control tags specifying protein properties. ProGen can be further fine-tuned to curated sequences and tags to improve controllable generation performance of proteins from families with sufficient homologous samples. Artificial proteins fine-tuned to five distinct lysozyme families showed similar catalytic efficiencies as natural lysozymes, with sequence identity to natural proteins as low as 31.4%. ProGen is readily adapted to diverse protein families, as we demonstrate with chorismate mutase and malate dehydrogenase.

Enhanced stability of the SARS CoV-2 spike glycoprotein following modification of an alanine cavity in the protein core

The spike (S) glycoprotein of SARS CoV-2 is the target of neutralizing antibodies (NAbs) that are crucial for vaccine effectiveness. The S1 subunit binds ACE2 while the S2 subunit mediates virus-cell membrane fusion. S2 is a class I fusion glycoprotein subunit and contains a central coiled coil that acts as a scaffold for the conformational changes associated with fusion function. The coiled coil of S2 is unusual in that the 3-4 repeat of inward-facing positions are mostly occupied by polar residues that mediate few inter-helical contacts in the prefusion trimer. We examined how insertion of bulkier hydrophobic residues (Val, Leu, Ile, Phe) to fill a cavity next to Ala1016 and Ala1020 in the 3-4 repeat affects the stability and antigenicity of S trimers. Substitution of Ala1016 with bulkier hydrophobic residues in the context of a prefusion-stabilized S trimer, S2P-FHA, was associated with increased thermal stability. S glycoprotein membrane fusion function was retained with Ala1016/Ala1020 cavity-filling mutations associated with improved recombinant S2P-FHA thermostability, however 2 mutants, A1016L and A1016V/A1020I, lacked ability to mediate entry of S-HIV-1 pseudoparticles into 293-ACE2 cells. When assessed as immunogens, two thermostable S2P-FHA mutants derived from the ancestral isolate, A1016L (16L) and A1016V/A1020I (VI) elicited neutralizing antibody with 50%-inhibitory dilutions (ID50s) in the range 2,700-5,110 for ancestral and Delta-derived viruses, and 210-1,744 for Omicron BA.1. The antigens elicited antibody specificities directed to the receptor-binding domain (RBD), N-terminal domain (NTD), fusion peptide and stem region of S2. The VI mutation enabled the production of intrinsically stable Omicron BA.1 and Omicron BA.4/5 S2P-FHA-like ectodomain oligomers in the absence of an external trimerization motif (T4 foldon), thus representing an alternative approach for stabilizing oligomeric S glycoprotein vaccines.

Accelerated Enveloping Distribution Sampling (AEDS) Allows for Efficient Sampling of Orthogonal Degrees of Freedom

One of the most challenging aspects in the molecular simulation of proteins is the study of slowly relaxing processes, such as loop rearrangements or ligands that adopt different conformations in the binding site. State-of-the-art methods used to calculate binding free energies rely on performing several short simulations (lambda steps), in which the ligand is slowly transformed into the endstates of interest. This makes capturing the slowly relaxing processes even more difficult, as they would need to be observed in all of the lambda steps. One attractive alternative is the use of a reference state capable of sampling all of the endstates of interest in a single simulation. However, the energy barriers between the states can be too high to overcome, thus hindering the sampling of all of the relevant conformations. Accelerated enveloping distribution sampling (AEDS) is a recently developed reference state technique that circumvents the high-energy-barrier challenge by adding a boosting potential that flattens the energy landscape without distorting the energy minima. In the present work, we apply AEDS to the well-studied benchmark system T4 lysozyme L99A. The T4 lysozyme L99A mutant contains a hydrophobic pocket in which there is a valine (valine 111), whose conformation influences the binding efficiencies of the different ligands. Incorrectly sampling the dihedral angle can lead to errors in predicted binding free energies of up to 16 kJ mol^-1. This protein constitutes an ideal scenario to showcase the advantages and challenges when using AEDS in the presence of slow relaxing processes. We show that AEDS is able to successfully sample the relevant degrees of freedom, providing accurate binding free energies, without the need of previous information of the system in the form of collective variables. Additionally, we showcase the capabilities of AEDS to efficiently screen a set of ligands. These results represent a promising first step toward the development of free-energy methods that can respond to more intricate molecular events.

Genomic characterization of a new phage BUCT541 against Klebsiella pneumoniae K1-ST23 and efficacy assessment in mouse and Galleria mellonella larvae

Over the past decades, the spread of multi-drug-resistant Klebsiella pneumoniae (MDR-KP) is becoming a new threat and new effective therapies against this pathogen are needed. Bacteriophage (phage) therapy is considered to be a promising alternative treatment for MDR-KP infections compared with antibacterial drug usage. Here, we reported a new phage BUCT541 which can lyse MDR-KP ST23. The genome of BUCT541 is a double-stranded linear 46,100-bp long DNA molecule with 48% GC content through the Next generation sequencing (NGS) data. A total of 81 open reading frames and no virulence or antimicrobial resistance genes are annotated in the BUCT541 genome. BUCT541 was able to lyse 7 of the 30 tested MDR-KP according to the host range analysis. And the seven sensitive strains belonged to the K. pneumoniae K1-ST23. BUCT541 exhibited high thermal stability (4–70°C) and broad pH tolerance (pH 3-11) in the stability test. The in vivo results showed that BUCT541 (4 × 10⁵ plaque-forming units (PFU)/each) significantly increased the survival rate of K. pneumoniae infected Galleria mellonella from 5.3% to 83.3% within 48 h. Moreover, in the mouse lung infection model, high doses of BUCT541 (2 × 10⁷ PFU/each) cured 100% of BALB/c mice that were infected with K. pneumoniae. After 30 h of treatment with phage BUCT541 of the multiplicity of infection (MOI) = 10, the K. pneumoniae in the lungs of mice was lower than 10⁴ CFU/mL, compared to the control group 10⁹ CFU/mL. Together, these findings indicate that phage BUCT541 holds great promise as an alternative therapy with excellent stability and a wide lysis range for the treatment of MDR-KP ST23 infection.

Modeling protein structure as a stable static equilibrium

We present evidence that the protein structure can be modeled as a stable static equilibrium, determined mainly by compressive supports in the nonpolar interior. That is, protein structures derive their structural strength through the same mechanical principles as do conventional structures like bridges and buildings. This is based on the observation that the experimentally elucidated structural determinants, the interior nonpolar side chains, are engaged in strong compressions in static terms. At the same time, major substructures in proteins, helices and h-bonded strands, because of their geometry, inherently leave gaps in the space they occupy. Under the compressive force, nonpolar side chains from one substructure can protrude into the gaps of another neighboring substructure and block its motion. As a result, interlocking of substructures can form, which builds up the nonpolar core assembly. The native structure then is the one with the structurally most stable core assembly. While intuitively appealing, this is a radical departure from the prevailing thinking that protein native structure is determined by global energy minimum, which is founded on thermodynamic hypothesis. Furthermore, to develop an effective model for analyzing protein structure with conventional tools, a proper mechanical representation must be established. By proving that the stability of the equilibrium in compressive interactions is conditioned on a form of mechanical energy minimum, we show that our notion of native structure can be equally consistent with the thermodynamic hypothesis. By mathematically treating the blocking action, an interaction, as a bar, a physical object, we succeed in representing and analyzing the core assembly as truss, a conventional structure. In this paper we define and expound step-by-step increasingly integrated interlocking patterns. We then analyze the core assemblies of a large set of diverse protein database structures. A native structure can be distinguished from decoys by comparing the composition and strength of their core assemblies. We show the results for two sets of native structures vs decoys.

Dithiol Based on l-Cysteine and Cysteamine as a Disulfide-Reducing Agent

We report the synthesis, chemical properties, and disulfide bond-reducing performance of a dithiol called NACMEAA, conceived as a hybrid of two biologically relevant thiols: cysteine and cysteamine. NACMEAA is conveniently prepared from inexpensive l-cystine in an efficient manner. As a nonvolatile, highly soluble, and neutral compound at physiological pH with the first thiol pK_a value of 8.0, NACMEAA is reactive and user-friendly. We also demonstrate that NACMEAA reduces disulfide bonds in GSSG and lysozyme.

Creation of Cross-Linked Crystals With Intermolecular Disulfide Bonds Connecting Symmetry-Related Molecules Allows Retention of Tertiary Structure in Different Solvent Conditions

Protein crystals are generally fragile and sensitive to subtle changes such as pH, ionic strength, and/or temperature in their crystallization mother liquor. Here, using T4 phage lysozyme as a model protein, the three-dimensional rigidification of protein crystals was conducted by introducing disulfide cross-links between neighboring molecules in the crystal. The effect of cross-linking on the stability of the crystals was evaluated by microscopic observation and X-ray diffraction. When soaking the obtained cross-linked crystals into a precipitant-free solution, the crystals held their shape without dissolution and diffracted to approximately 1.1 Å resolution, comparable to that of the non-cross-linked crystals. Such cross-linked crystals maintained their diffraction even when immersed in other solutions with pH values from 4 to 10, indicating that the disulfide cross-linking made the packing contacts enforced and resulted in some mechanical strength in response to changes in the preservation conditions. Furthermore, the cross-linked crystals gained stability to permit soaking into solutions containing high concentrations of organic solvents. The results suggest the possibility of obtaining protein crystals for effective drug screening by introducing appropriate cross-linked disulfide bonds.

Addressing Intersite Coupling Unlocks Large Combinatorial Chemical Spaces for Alchemical Free Energy Methods

Alchemical free energy methods are playing a growing role in molecular design, both for computer-aided drug design of small molecules and for computational protein design. Multisite λ dynamics (MSλD) is a uniquely scalable alchemical free energy method that enables more efficient exploration of combinatorial alchemical spaces encountered in molecular design, but simulations have typically been limited to a few hundred ligands or sequences. Here, we focus on coupling between sites to enable scaling to larger alchemical spaces. We first discuss updates to the biasing potentials that facilitate MSλD sampling to include coupling terms and show that this can provide more thorough sampling of alchemical states. We then harness coupling between sites by developing a new free energy estimator based on the Potts models underlying direct coupling analysis, a method for predicting contacts from sequence coevolution, and find it yields more accurate free energies than previous estimators. The sampling requirements of the Potts model estimator scale with the square of the number of sites, a substantial improvement over the exponential scaling of the standard estimator. This opens up exploration of much larger alchemical spaces with MSλD for molecular design.

Site-wise Diversification of Combinatorial Libraries Using Insights from Structure-guided Stability Calculations

Many auspicious clinical and industrial accomplishments have improved the human condition by means of protein engineering. Despite these achievements, our incomplete understanding of the sequence-structure-function relationship prevents rapid innovation. To tackle this problem, we must develop and integrate new and existing technologies. To date, directed evolution and rational design have dominated as protein engineering principles. Even so, prior to screening for novel or improved functions, a large collection of variants, within a protein library, exist along an ambiguous mutational terrain. Complicating things further, the choice of where to initialize investigation along a vast sequence space becomes even more difficult given that the majority of any sequence lacks function entirely. Unfortunately, even when considering functionally relevant positions, random substitutions can prove to be destabilizing, causing a hindrance to an otherwise function-inducing, stability-reliant folding process. To enhance productivity in the field, we seek to address this issue of destabilization, and subsequent disfunction, at protein-protein and protein-ligand interacting regions. Herein, the process of choosing amenable positions - and amino acids at those positions - allows for a refined, knowledge-based approach to combinatorial library design. Using structural data, we perform computational stability prediction with FoldX's PositionScan and Rosetta's ddG_monomer in tandem, allowing for the refinement of our thermodynamic stability data through the comparison of results. In turn, we provide a process for selecting in silico predicted mutually stabilizing positions and avoiding overly destabilizing ones that guides the site-wise diversification of combinatorial libraries.

Recent Insights Into the Prognostic and Therapeutic Applications of Lysozymes

Lysozymes are naturally occurring enzymes present in a variety of biological organisms, such as bacteria, fungi, and animal bodily secretions and tissues. It is also the main ingredient of many ethnomedicines. It is well known that lysozymes and lysozyme-like enzymes can be used as anti-bacterial agents by degrading bacterial cell wall peptidoglycan that leads to cell death, and can also inhibit fungi, yeasts, and viruses. In addition to its direct antimicrobial activity, lysozyme is also an important component of the innate immune system in most mammals. Increasing evidence has shown the immune-modulatory effects of lysozymes against infection and inflammation. More recently, studies have revealed the anti-cancer activities of lysozyme in multiple types of tumors, potentially through its immune-modulatory activities. In this review, we summarized the major functions and underlying mechanisms of lysozymes derived from animal and plant sources. We highlighted the therapeutic applications and recent advances of lysozymes in cancers, hypertension, and viral diseases, aiming toseeking alternative therapies for standard medical treatment bypassing side effects. We also evaluated the role of lysozyme as a promising cancer marker for prognosis to indicate the outcomes recurrence for patients.

High throughput and quantitative enzymology in the genomic era

Accurate predictions from models based on physical principles are the ultimate metric of our biophysical understanding. Although there has been stunning progress toward structure prediction, quantitative prediction of enzyme function has remained challenging. Realizing this goal will require large numbers of quantitative measurements of rate and binding constants and the use of these ground-truth data sets to guide the development and testing of these quantitative models. Ground truth data more closely linked to the underlying physical forces are also desired. Here, we describe technological advances that enable both types of ground truth measurements. These advances allow classic models to be tested, provide novel mechanistic insights, and place us on the path toward a predictive understanding of enzyme structure and function.

Interpretation of Single-Molecule Force Experiments on Proteins Using Normal Mode Analysis

Single-molecule force spectroscopy experiments allow protein folding and unfolding to be explored using mechanical force. Probably the most informative technique for interpreting the results of these experiments at the structural level makes use of steered molecular dynamics (MD) simulations, which can explicitly model the protein under load. Unfortunately, this technique is computationally expensive for many of the most interesting biological molecules. Here, we find that normal mode analysis (NMA), a significantly cheaper technique from a computational perspective, allows at least some of the insights provided by MD simulation to be gathered. We apply this technique to three non-homologous proteins that were previously studied by force spectroscopy: T4 lysozyme (T4L), Hsp70 and the glucocorticoid receptor domain (GCR). The NMA results for T4L and Hsp70 are compared with steered MD simulations conducted previously, and we find that we can recover the main results. For the GCR, which did not undergo MD simulation, our approach identifies substructures that correlate with experimentally identified unfolding intermediates. Overall, we find that NMA can make a valuable addition to the analysis toolkit for the structural analysis of single-molecule force experiments on proteins.

A high-throughput predictive method for sequence-similar fold switchers

Although most experimentally characterized proteins with similar sequences assume the same folds and perform similar functions, an increasing number of exceptions is emerging. One class of exceptions comprises sequence-similar fold switchers, whose secondary structures shift from α-helix <-> β-sheet through a small number of mutations, a sequence insertion, or a deletion. Predictive methods for identifying sequence-similar fold switchers are desirable because some are associated with disease and/or can perform different functions in cells. Here, we use homology-based secondary structure predictions to identify sequence-similar fold switchers from their amino acid sequences alone. To do this, we predicted the secondary structures of sequence-similar fold switchers using three different homology-based secondary structure predictors: PSIPRED, JPred4, and SPIDER3. We found that α-helix <-> β-strand prediction discrepancies from JPred4 discriminated between the different conformations of sequence-similar fold switchers with high statistical significance (P < 1.8*10-19 ). Thus, we used these discrepancies as a classifier and found that they can often robustly discriminate between sequence-similar fold switchers and sequence-similar proteins that maintain the same folds (Matthews Correlation Coefficient of 0.82). We found that JPred4 is a more robust predictor of sequence-similar fold switchers because of (a) the curated sequence database it uses to produce multiple sequence alignments and (b) its use of sequence profiles based on Hidden Markov Models. Our results indicate that inconsistencies between JPred4 secondary structure predictions can be used to identify some sequence-similar fold switchers from their sequences alone. Thus, the negative information from inconsistent secondary structure predictions can potentially be leveraged to identify sequence-similar fold switchers from the broad base of genomic sequences.

Enhanced bone regeneration via spatiotemporal and controlled delivery of a genetically engineered BMP-2 in a composite Hydrogel

Scaffolds functionalized with bone morphogenetic protein-2 (BMP-2) have shown great potential for bone regeneration. However, structural instability and the necessity for supra-physiological dose have thus far limited practical applications for BMP-2. Protein modification and site-specific covalent immobilization of BMP-2 to carrier materials might be optimal strategies to overcome these problems. Here, we report a broadly applicable strategy where the polyhistidine tag-T4 Lysozyme (His6-T4L) was genetically fused at the N-terminus of BMP-2 and used as a protein spacer, which on one hand enhanced protein solubility and stability, and on the other hand mediated site-specific covalent anchoring of BMP-2 upon binding to nickel-chelated nitrilotriacetic acid (Ni-NTA) microparticles (denoted as MPs-His6-T4L-BMP2) to further maximize its rescued activity. We also constructed a novel gelatin-based hydrogel that was crosslinked by transglutaminase (TG) and tannic acid (TA). This hydrogel, when incorporated with MPs-His6-T4L-BMP2, displayed excellent in-situ injectability, thermosensitivity, adhesiveness and improved mechanical properties. The effective loading mode led to a controlled and long-term sustained release of His6-T4L-BMP2, thereby resulting in enhancement of bone regeneration in a critical-sized bone defect. We believe that the protein modification strategy proposed here opens up new route not only for BMP-2 applications, but can be used to inform novel uses for other macromolecules.

Energy penalties enhance flexible receptor docking in a model cavity

Protein flexibility remains a major challenge in library docking because of difficulties in sampling conformational ensembles with accurate probabilities. Here, we use the model cavity site of T4 lysozyme L99A to test flexible receptor docking with energy penalties from molecular dynamics (MD) simulations. Crystallography with larger and smaller ligands indicates that this cavity can adopt three major conformations: open, intermediate, and closed. Since smaller ligands typically bind better to the cavity site, we anticipate an energy penalty for the cavity opening. To estimate its magnitude, we calculate conformational preferences from MD simulations. We find that including a penalty term is essential for retrospective ligand enrichment; otherwise, high-energy states dominate the docking. We then prospectively docked a library of over 900,000 compounds for new molecules binding to each conformational state. Absent a penalty term, the open conformation dominated the docking results; inclusion of this term led to a balanced sampling of ligands against each state. High ranked molecules were experimentally tested by Tm upshift and X-ray crystallography. From 33 selected molecules, we identified 18 ligands and determined 13 crystal structures. Most interesting were those bound to the open cavity, where the buried site opens to bulk solvent. Here, highly unusual ligands for this cavity had been predicted, including large ligands with polar tails; these were confirmed both by binding and by crystallography. In docking, incorporating protein flexibility with thermodynamic weightings may thus access new ligand chemotypes. The MD approach to accessing and, crucially, weighting such alternative states may find general applicability.

Temperature artifacts in protein structures bias ligand-binding predictions

X-ray crystallography is the gold standard to resolve conformational ensembles that are significant for protein function, ligand discovery, and computational methods development. However, relevant conformational states may be missed at common cryogenic (cryo) data-collection temperatures but can be populated at room temperature. To assess the impact of temperature on making structural and computational discoveries, we systematically investigated protein conformational changes in response to temperature and ligand binding in a structural and computational workhorse, the T4 lysozyme L99A cavity. Despite decades of work on this protein, shifting to RT reveals new global and local structural changes. These include uncovering an apo helix conformation that is hidden at cryo but relevant for ligand binding, and altered side chain and ligand conformations. To evaluate the impact of temperature-induced protein and ligand changes on the utility of structural information in computation, we evaluated how temperature can mislead computational methods that employ cryo structures for validation. We find that when comparing simulated structures just to experimental cryo structures, hidden successes and failures often go unnoticed. When using structural information in ligand binding predictions, both coarse docking and rigorous binding free energy calculations are influenced by temperature effects. The trend that cryo artifacts limit the utility of structures for computation holds across five distinct protein classes. Our results suggest caution when consulting cryogenic structural data alone, as temperature artifacts can conceal errors and prevent successful computational predictions, which can mislead the development and application of computational methods in discovering bioactive molecules.

Adenosine triphosphate energy-independently controls protein homeostasis with unique structure and diverse mechanisms

Proteins function in the crowded cellular environments with high salt concentrations, thus facing tremendous challenges of misfolding/aggregation which represents a pathological hallmark of aging and an increasing spectrum of human diseases. Recently, intrinsically disordered regions (IDRs) were recognized to drive liquid-liquid phase separation (LLPS), a common principle for organizing cellular membraneless organelles (MLOs). ATP, the universal energy currency for all living cells, mysteriously has concentrations of 2-12 mM, much higher than required for its previously-known functions. Only recently, ATP was decoded to behave as a biological hydrotrope to inhibit protein LLPS and aggregation at mM. We further revealed that ATP also acts as a bivalent binder, which not only biphasically modulates LLPS driven by IDRs of human and viral proteins, but also bind to the conserved nucleic-acid-binding surfaces of the folded proteins. Most unexpectedly, ATP appears to act as a hydration mediator to antagonize the crowding-induced destabilization as well as to enhance folding of proteins without significant binding. Here, this review focuses on summarizing the results of these biophysical studies and discussing their implications in an evolutionary context. By linking triphosphate with unique hydration property to adenosine, ATP appears to couple the ability for establishing hydrophobic, π-π, π-cation and electrostatic interactions to the capacity in mediating hydration of proteins, which is at the heart of folding, dynamics, stability, phase separation and aggregation. Consequently, ATP acquired a category of functions at ~mM to energy-independently control protein homeostasis with diverse mechanisms, thus implying a link between cellular ATP concentrations and protein-aggregation diseases.

The Molecular Basis for Escherichia coli O157:H7 Phage FAHEc1 Endolysin Function and Protein Engineering to Increase Thermal Stability

Bacteriophage-encoded endolysins have been identified as antibacterial candidates. However, the development of endolysins as mainstream antibacterial agents first requires a comprehensive biochemical understanding. This study defines the atomic structure and enzymatic function of Escherichia coli O157:H7 phage FAHEc1 endolysin, LysF1. Bioinformatic analysis suggests this endolysin belongs to the T4 Lysozyme (T4L)-like family of proteins and contains a highly conserved catalytic triad. We then solved the structure of LysF1 with x-ray crystallography to 1.71 Å. LysF1 was confirmed to exist as a monomer in solution by sedimentation velocity experiments. The protein architecture of LysF1 is conserved between T4L and related endolysins. Comparative analysis with related endolysins shows that the spatial orientation of the catalytic triad is conserved, suggesting the catalytic mechanism of peptidoglycan degradation is the same as that of T4L. Differences in the sequence illustrate the role coevolution may have in the evolution of this fold. We also demonstrate that by mutating a single residue within the hydrophobic core, the thermal stability of LysF1 can be increased by 9.4 °C without compromising enzymatic activity. Overall, the characterization of LysF1 provides further insight into the T4L-like class of endolysins. Our study will help advance the development of related endolysins as antibacterial agents, as rational engineering will rely on understanding mutable positions within this protein fold.

A strategy for proline and glycine mutations to proteins with alchemical free energy calculations

Computation of the thermodynamic consequences of protein mutations holds great promise in protein biophysics and design. Alchemical free energy methods can give improved estimates of mutational free energies, and are already widely used in calculations of relative and absolute binding free energies in small molecule design problems. In principle, alchemical methods can address any amino acid mutation with an appropriate alchemical pathway, but identifying a strategy that produces such a path for proline and glycine mutations is an ongoing challenge. Most current strategies perturb only side chain atoms, while proline and glycine mutations also alter the backbone parameters and backbone ring topology. Some strategies also perturb backbone parameters and enable glycine mutations. This work presents a strategy that enables both proline and glycine mutations and comprises two key elements: a dual backbone with restraints and scaling of bonded terms, facilitating backbone parameter changes, and a soft bond in the proline ring, enabling ring topology changes in proline mutations. These elements also have utility for core hopping and macrocycle studies in computer-aided drug design. This new strategy shows slight improvements over an alternative side chain perturbation strategy for a set T4 lysozyme mutations lacking proline and glycine, and yields good agreement with experiment for a set of T4 lysozyme proline and glycine mutations not previously studied. To our knowledge this is the first report comparing alchemical predictions of proline mutations with experiment. With this strategy in hand, alchemical methods now have access to the full palette of amino acid mutations.

Pressure and Chemical Unfolding of an α-Helical Bundle Protein: The GH2 Domain of the Protein Adaptor GIPC1

When combined with NMR spectroscopy, high hydrostatic pressure is an alternative perturbation method used to destabilize globular proteins that has proven to be particularly well suited for exploring the unfolding energy landscape of small single-domain proteins. To date, investigations of the unfolding landscape of all-β or mixed-α/β protein scaffolds are well documented, whereas such data are lacking for all-α protein domains. Here we report the NMR study of the unfolding pathways of GIPC1-GH2, a small α-helical bundle domain made of four antiparallel α-helices. High-pressure perturbation was combined with NMR spectroscopy to unravel the unfolding landscape at three different temperatures. The results were compared to those obtained from classical chemical denaturation. Whatever the perturbation used, the loss of secondary and tertiary contacts within the protein scaffold is almost simultaneous. The unfolding transition appeared very cooperative when using high pressure at high temperature, as was the case for chemical denaturation, whereas it was found more progressive at low temperature, suggesting the existence of a complex folding pathway.

Beyond history and "on a roll": The list of the most well-studied human protein structures and overall trends in the protein data bank

Of the roughly 20,000 canonical human protein sequences, as of January 20, 2021, 7,077 proteins have had their full or partial, medium- to high-resolution structures determined by x-ray crystallography or other methods. Which of these proteins dominate the protein data bank (the PDB) and why? In this paper, we list the 273 top human protein structures based on the number of their PDB entries. This set of proteins accounts for more than 40% of all available human PDB entries and represent past trends as well as current status for protein structural biology. We briefly discuss the relationship which some of the prominent protein structures have with protein research as a whole and mention their relevance to human diseases. The top-10 soluble and membrane proteins are all well-known (most of their first structures being deposited more than 30 years ago). Overall, there is no dramatic change in recent trends in the PDB. Remarkably, the number of structure depositions has grown nearly exponentially over the last 10 or more years (with a doubling time of 7 years for proteins, obtained from any organism). Growth in human protein structures is slightly faster (at 5.9 years). The information in this paper may be informative to senior scientists but also inspire researchers who are new to protein science, providing the year 2021 snap-shot for the state of protein structural biology.

Wepy: A Flexible Software Framework for Simulating Rare Events with Weighted Ensemble Resampling

Here, we introduce the open-source software framework wepy (https://github.com/ADicksonLab/wepy) which is a toolkit for running and analyzing weighted ensemble (WE) simulations. The wepy toolkit is in pure Python and as such is highly portable and extensible, making it an excellent platform to develop and use new WE resampling algorithms such as WExplore, REVO, and others while leveraging the entire Python ecosystem. In addition, wepy simplifies WE-specific analyses by defining out-of-core tree-like data structures using the cross-platform HDF5 file format. In this paper, we discuss the motivations and challenges for simulating rare events in biomolecular systems. As has previously been shown, high-dimensional WE resampling algorithms such as WExplore and REVO have been successful at these tasks, especially for rare events that are difficult to describe by one or two collective variables. We explain in detail how wepy facilitates implementation of these algorithms, as well as aids in analyzing the unique structure of WE simulation results. To explain how wepy and WE work in general, we describe the mathematical formalism of WE, an overview of the architecture of wepy, and provide code examples of how to construct, run, and analyze simulation results for a protein-ligand system (T4 Lysozyme in an implicit solvent). This paper is written with a variety of readers in mind, including (1) those curious about how to leverage WE rare-event simulations for their domain, (2) current WE users who want to begin using new high-dimensional resamplers such as WExplore and REVO, and (3) expert users who would like to prototype or implement their own algorithms that can be easily adopted by others.

SAXS studies of X-ray induced disulfide bond damage: Engineering high-resolution insight from a low-resolution technique

A significant problem in biological X-ray crystallography is the radiation chemistry caused by the incident X-ray beam. This produces both global and site-specific damage. Site specific damage can misdirect the biological interpretation of the structural models produced. Cryo-cooling crystals has been successful in mitigating damage but not eliminating it altogether; however, cryo-cooling can be difficult in some cases and has also been shown to limit functionally relevant protein conformations. The doses used for X-ray crystallography are typically in the kilo-gray to mega-gray range. While disulfide bonds are among the most significantly affected species in proteins in the crystalline state at both cryogenic and higher temperatures, there is limited information on their response to low X-ray doses in solution, the details of which might inform biomedical applications of X-rays. In this work we engineered a protein that dimerizes through a susceptible disulfide bond to relate the radiation damage processes seen in cryo-cooled crystals to those closer to physiologic conditions. This approach enables a low-resolution technique, small angle X-ray scattering (SAXS), to detect and monitor a residue specific process. A dose dependent fragmentation of the engineered protein was seen that can be explained by a dimer to monomer transition through disulfide bond cleavage. This supports the crystallographically derived mechanism and demonstrates that results obtained crystallographically can be usefully extrapolated to physiologic conditions. Fragmentation was influenced by pH and the conformation of the dimer, providing information on mechanism and pointing to future routes for investigation and potential mitigation. The novel engineered protein approach to generate a large-scale change through a site-specific interaction represents a promising tool for advancing radiation damage studies under solution conditions.

The Effects of p-Azidophenylalanine Incorporation on Protein Structure and Stability

Functionalization is often needed to harness the power of proteins for beneficial use but can cause losses to stability and/or activity. State of the art methods to limit these deleterious effects accomplish this by substituting an amino acid in the wild-type molecule into an unnatural amino acid, such as p-azidophenylalanine (pAz), but selecting the residue for substitution a priori remains an elusive goal of protein engineering. The results of this work indicate that all-atom molecular dynamics simulation can be used to determine whether substituting pAz for a natural amino acid will be detrimental to experimentally determined protein stability. These results offer significant hope that local deviations from wild-type structure caused by pAz incorporation observed in simulations can be a predictive metric used to reduce the number of costly experiments that must be done to find active proteins upon substitution with pAz and subsequent functionalization.

Structure and Function of the T4 Spackle Protein Gp61.3

The bacteriophage T4 genome contains two genes that code for proteins with lysozyme activity—e and 5. Gene e encodes the well-known T4 lysozyme (commonly called T4L) that functions to break the peptidoglycan layer late in the infection cycle, which is required for liberating newly assembled phage progeny. Gene product 5 (gp5) is the tail-associated lysozyme, a component of the phage particle. It forms a spike at the tip of the tail tube and functions to pierce the outer membrane of the Escherichia coli host cell after the phage has attached to the cell surface. Gp5 contains a T4L-like lysozyme domain that locally digests the peptidoglycan layer upon infection. The T4 Spackle protein (encoded by gene 61.3) has been thought to play a role in the inhibition of gp5 lysozyme activity and, as a consequence, in making cells infected by bacteriophage T4 resistant to later infection by T4 and closely related phages. Here we show that (1) gp61.3 is secreted into the periplasm where its N-terminal periplasm-targeting peptide is cleaved off; (2) gp61.3 forms a 1:1 complex with the lysozyme domain of gp5 (gp5Lys); (3) gp61.3 selectively inhibits the activity of gp5, but not that of T4L; (4) overexpression of gp5 causes cell lysis. We also report a crystal structure of the gp61.3-gp5Lys complex that demonstrates that unlike other known lysozyme inhibitors, gp61.3 does not interact with the active site cleft. Instead, it forms a “wall” that blocks access of an extended polysaccharide substrate to the cleft and, possibly, locks the enzyme in an “open-jaw”-like conformation making catalysis impossible.

Alteration of enzymes and their application to nucleic acid amplification (Review)

Since the discovery of polymerase chain reaction (PCR) in 1985, several methods have been developed to achieve nucleic acid amplification, and are currently used in various fields including clinical diagnosis and life science research. Thus, a wealth of information has accumulated regarding nucleic acid-related enzymes. In this review, some nucleic acid-related enzymes were selected and the recent advances in their modification along with their application to nucleic acid amplification were described. The discussion also focused on optimization of the corresponding reaction conditions. Using newly developed enzymes under well-optimized reaction conditions, the sensitivity, specificity, and fidelity of nucleic acid tests can be improved successfully.

Atomic Details of Carbon-Based Nanomolecules Interacting with Proteins

Since the discovery of fullerene, carbon-based nanomolecules sparked a wealth of research across biological, medical and material sciences. Understanding the interactions of these materials with biological samples at the atomic level is crucial for improving the applications of nanomolecules and address safety aspects concerning their use in medicine. Protein crystallography provides the interface view between proteins and carbon-based nanomolecules. We review forefront structural studies of nanomolecules interacting with proteins and the mechanism underlying these interactions. We provide a systematic analysis of approaches used to select proteins interacting with carbon-based nanomolecules explored from the worldwide Protein Data Bank (wwPDB) and scientific literature. The analysis of van der Waals interactions from available data provides important aspects of interactions between proteins and nanomolecules with implications on functional consequences. Carbon-based nanomolecules modulate protein surface electrostatic and, by forming ordered clusters, could modify protein quaternary structures. Lessons learned from structural studies are exemplary and will guide new projects for bioimaging tools, tuning of intrinsically disordered proteins, and design assembly of precise hybrid materials.

Single-Molecule Studies of Protein Folding with Optical Tweezers

Manipulation of individual molecules with optical tweezers provides a powerful means of interrogating the structure and folding of proteins. Mechanical force is not only a relevant quantity in cellular protein folding and function, but also a convenient parameter for biophysical folding studies. Optical tweezers offer precise control in the force range relevant for protein folding and unfolding, from which single-molecule kinetic and thermodynamic information about these processes can be extracted. In this review, we describe both physical principles and practical aspects of optical tweezers measurements and discuss recent advances in the use of this technique for the study of protein folding. In particular, we describe the characterization of folding energy landscapes at high resolution, studies of structurally complex multidomain proteins, folding in the presence of chaperones, and the ability to investigate real-time cotranslational folding of a polypeptide.

Pan-retroviral Nucleocapsid-Mediated Phase Separation Regulates Genomic RNA Positioning and Trafficking

The duality of liquid-liquid phase separation (LLPS) of cellular components into membraneless organelles defines the nucleation of both normal and disease processes including stress granule (SG) assembly. From mounting evidence of LLPS utility by viruses, we discover that HIV-1 nucleocapsid (NC) protein condenses into zinc-finger (ZnF)-dependent LLPSs that are dynamically influenced by cytosolic factors. ZnF-dependent and Zinc (Zn2+)-chelation-sensitive NC-LLPS are formed in live cells. NC-Zn2+ ejection reverses the HIV-1 blockade on SG assembly, inhibits NC-SG assembly, disrupts NC/Gag-genomic RNA (vRNA) ribonucleoprotein complexes, and causes nuclear sequestration of NC and the vRNA, inhibiting Gag expression and virus release. NC ZnF mutagenesis eliminates the HIV-1 blockade of SG assembly and repositions vRNA to SGs. We find that NC-mediated, Zn2+-coordinated phase separation is conserved among diverse retrovirus subfamilies, illustrating that this exquisitely evolved Zn2+-dependent feature of virus replication represents a critical target for pan-antiretroviral therapies.

How evolution shapes enzyme selectivity - lessons from aminoacyl-tRNA synthetases and other amino acid utilizing enzymes

Aminoacyl-tRNA synthetases (AARSs) charge tRNA with their cognate amino acids. Many other enzymes use amino acids as substrates, yet discrimination against noncognate amino acids that threaten the accuracy of protein translation is a hallmark of AARSs. Comparing AARSs to these other enzymes allowed us to recognize patterns in molecular recognition and strategies used by evolution for exercising selectivity. Overall, AARSs are 2-3 orders of magnitude more selective than most other amino acid utilizing enzymes. AARSs also reveal the physicochemical limits of molecular discrimination. For example, amino acids smaller by a single methyl moiety present a discrimination ceiling of ~200, while larger ones can be discriminated by up to 10⁵ -fold. In contrast, substrates larger by a hydroxyl group challenge AARS selectivity, due to promiscuous H-bonding with polar active site groups. This 'hydroxyl paradox' is resolved by editing. Indeed, when the physicochemical discrimination limits are reached, post-transfer editing - hydrolysis of tRNAs charged with noncognate amino acids, evolved. The editing site often selectively recognizes the edited noncognate substrate using the very same feature that the synthetic site could not efficiently discriminate against. Finally, the comparison to other enzymes also reveals that the selectivity of AARSs is an explicitly evolved trait, showing some clear examples of how selection acted not only to optimize catalytic efficiency with the target substrate, but also to abolish activity with noncognate threat substrates ('negative selection').

Approaching protein design with multisite λ dynamics: Accurate and scalable mutational folding free energies in T4 lysozyme

The estimation of changes in free energy upon mutation is central to the problem of protein design. Modern protein design methods have had remarkable success over a wide range of design targets, but are reaching their limits in ligand binding and enzyme design due to insufficient accuracy in mutational free energies. Alchemical free energy calculations have the potential to supplement modern design methods through more accurate molecular dynamics based prediction of free energy changes, but suffer from high computational cost. Multisite λ dynamics (MSλD) is a particularly efficient and scalable free energy method with potential to explore combinatorially large sequence spaces inaccessible with other free energy methods. This work aims to quantify the accuracy of MSλD and demonstrate its scalability. We apply MSλD to the classic problem of calculating folding free energies in T4 lysozyme, a system with a wealth of experimental measurements. Single site mutants considering 32 mutations show remarkable agreement with experiment with a Pearson correlation of 0.914 and mean unsigned error of 1.19 kcal/mol. Multisite mutants in systems with up to five concurrent mutations spanning 240 different sequences show comparable agreement with experiment. These results demonstrate the promise of MSλD in exploring large sequence spaces for protein design.

How internal cavities destabilize a protein

Although many proteins possess a distinct folded structure lying at a minimum in a funneled free energy landscape, thermal energy causes any protein to continuously access lowly populated excited states. The existence of excited states is an integral part of biological function. Although transitions into the excited states may lead to protein misfolding and aggregation, little structural information is currently available for them. Here, we show how NMR spectroscopy, coupled with pressure perturbation, brings these elusive species to light. As pressure acts to favor states with lower partial molar volume, NMR follows the ensuing change in the equilibrium spectroscopically, with residue-specific resolution. For T4 lysozyme L99A, relaxation dispersion NMR was used to follow the increase in population of a previously identified "invisible" folded state with pressure, as this is driven by the reduction in cavity volume by the flipping-in of a surface aromatic group. Furthermore, multiple partly disordered excited states were detected at equilibrium using pressure-dependent H/D exchange NMR spectroscopy. Here, unfolding reduced partial molar volume by the removal of empty internal cavities and packing imperfections through subglobal and global unfolding. A close correspondence was found for the distinct pressure sensitivities of various parts of the protein and the amount of internal cavity volume that was lost in each unfolding event. The free energies and populations of excited states allowed us to determine the energetic penalty of empty internal protein cavities to be 36 cal⋅Å^-3.

A Second Backbone: The Contribution of a Buried Asparagine Ladder to the Global and Local Stability of a Leucine-Rich Repeat Protein

Parallel β-sheet-containing repeat proteins often display a structural motif in which conserved asparagines form a continuous ladder buried within the hydrophobic core. In such "asparagine ladders", the asparagine side-chain amides form a repetitive pattern of hydrogen bonds with neighboring main-chain NH and CO groups. Although asparagine ladders have been thought to be important for stability, there is little experimental evidence to support such speculation. Here we test the contribution of a minimal asparagine ladder from the leucine-rich repeat protein pp32 to stability and investigate lattice rigidity and hydrogen bond character using solution nuclear magnetic resonance (NMR) spectroscopy. Point substitutions of the two ladder asparagines of pp32 are strongly destabilizing and decrease the cooperativity of unfolding. The chemical shifts of the ladder side-chain HZ protons are shifted significantly downfield in the NMR spectrum and have low temperature coefficients, indicative of strong hydrogen bonding. In contrast, the HE protons are shifted upfield and have temperature coefficients close to zero, suggesting an asymmetry in hydrogen bond strength along the ladder. Ladder NH2 groups have weak 1H-15N cross-peak intensities; 1H-15N nuclear Overhauser effect and 15N CPMG experiments show this to be the result of high rigidity. Hydrogen exchange measurements demonstrate that the ladder NH2 groups exchange very slowly, with rates approaching the global exchange limit. Overall, these results show that the asparagine side chains are held in a very rigid, nondynamic structure, making a significant contribution to the overall stability. In this regard, buried asparagine ladders can be considered "second backbones" within the cores of their elongated β-sheet host proteins.

Protein stability engineering insights revealed by domain-wide comprehensive mutagenesis

The accurate prediction of protein stability upon sequence mutation is an important but unsolved challenge in protein engineering. Large mutational datasets are required to train computational predictors, but traditional methods for collecting stability data are either low-throughput or measure protein stability indirectly. Here, we develop an automated method to generate thermodynamic stability data for nearly every single mutant in a small 56-residue protein. Analysis reveals that most single mutants have a neutral effect on stability, mutational sensitivity is largely governed by residue burial, and unexpectedly, hydrophobics are the best tolerated amino acid type. Correlating the output of various stability-prediction algorithms against our data shows that nearly all perform better on boundary and surface positions than for those in the core and are better at predicting large-to-small mutations than small-to-large ones. We show that the most stable variants in the single-mutant landscape are better identified using combinations of 2 prediction algorithms and including more algorithms can provide diminishing returns. In most cases, poor in silico predictions were tied to compositional differences between the data being analyzed and the datasets used to train the algorithm. Finally, we find that strategies to extract stabilities from high-throughput fitness data such as deep mutational scanning are promising and that data produced by these methods may be applicable toward training future stability-prediction tools.

A critical review of five machine learning-based algorithms for predicting protein stability changes upon mutation

A number of machine learning (ML)-based algorithms have been proposed for predicting mutation-induced stability changes in proteins. In this critical review, we used hypothetical reverse mutations to evaluate the performance of five representative algorithms and found all of them suffer from the problem of overfitting. This approach is based on the fact that if a wild-type protein is more stable than a mutant protein, then the same mutant is less stable than the wild-type protein. We analyzed the underlying issues and suggest that the main causes of the overfitting problem include that the numbers of training cases were too small, and the features used in the models were not sufficiently informative for the task. We make recommendations on how to avoid overfitting in this important research area and improve the reliability and robustness of ML-based algorithms in general.

Exhaustive Search of Ligand Binding Pathways via Volume-Based Metadynamics

Determining the complete set of ligands' binding-unbinding pathways is important for drug discovery and for rational interpretation of mutation data. Here we have developed a metadynamics-based technique that addresses this issue and allows estimating affinities in the presence of multiple escape pathways. Our approach is shown on a lysozyme T4 variant in complex with a benzene molecule. The calculated binding free energy is in agreement with experimental data. Remarkably, not only were we able to find all the previously identified ligand binding pathways, but also we identified three pathways previously not identified as such. These results were obtained at a small computational cost, making this approach valuable for practical applications, such as screening of small compound libraries.

A Double-Resonance CEST Experiment To Study Multistate Protein Conformational Exchange: An Application to Protein Folding

Despite the importance of protein dynamics to function, studying exchange between multiple conformational states remains a challenge because sparsely populated states are invisible to conventional techniques. CEST NMR experiments can detect minor states with lifetimes between 5 and 200 ms populated to a level of just ∼1%. However, CEST often cannot provide the exchange mechanism for processes involving three or more states, leaving the role of the detected minor states unknown. Here a double-resonance CEST experiment to determine the kinetics of multistate exchange is presented. The approach that involves irradiating resonances from two minor states simultaneously is used to study the exchange of T4 lysozyme (T4L) between the dominant native state and two minor states, the unfolded state and a second minor state (B), each populated to only ∼4%. Regular CEST does not provide the folding mechanism, but double-resonance CEST clearly shows that T4L can fold directly without going through B.

Role of Disulfide Bonds and Topological Frustration in the Kinetic Partitioning of Lysozyme Folding Pathways

Disulfide bonds in proteins can strongly influence the folding pathways by constraining the conformational space. Lysozyme has four disulfide bonds and is widely studied for its antibacterial properties. Experiments on lysozyme infer that the protein folds through a fast and a slow pathway. However, the reasons for the kinetic partitioning in the folding pathways are not completely clear. Using a coarse-grained protein model and simulations, we show that two out of the four disulfide bonds, which are present in the α-domain of lysozyme, are responsible for the slow folding pathway. In this pathway, a kinetically trapped intermediate state, which is close to the native state, is populated. In this state, the orientations of α-helices present in the α-domain are misaligned relative to each other. The protein in this state has to partially unfold by breaking down the interhelical contacts between the misaligned helices to fold to the native state. However, the topological constraints due to the two disulfide bonds present in the α-domain make the protein less flexible, and it is trapped in this conformation for hundreds of milliseconds. On disabling these disulfide bonds, we find that the kinetically trapped intermediate state and the slow folding pathway disappear. Simulations mimicking the folding of protein without disulfide bonds under oxidative conditions show that the native disulfide bonds are formed as the protein folds, indicating that folding guides the formation of disulfide bonds. The sequence of formation of the disulfide bonds is Cys64-Cys80 → Cys76-Cys94 → Cys30-Cys115 → Cys6-Cys127. Any disulfide bond that forms before its precursor in the sequence has to break and follow the sequence for the protein to fold. These results show that lysozyme also serves as a very good model system to probe the role of disulfide bonds and topological frustration in protein folding. The predictions from the simulations can be verified by single-molecule fluorescence resonance energy transfer or single-molecule pulling experiments, which can probe heterogeneity in the folding pathways.

A protein interaction free energy model based on amino acid residue contributions: Assessment of point mutation stability of T4 lysozyme

Here we present a model to estimate the interaction free energy contribution of each amino acid residue of a given protein. Protein interaction energy is described in terms of per-residue interaction factors, μ. Multibody interactions are implicitly captured in μ through the combination of amino acid terms (γ) guided by local conformation indices (σ). The model enables construction of an interaction factor heat map for a protein in a given fold, allows prima facie assessment of the degree of residue-residue interaction, and facilitates a qualitative and quantitative evaluation of protein association properties. The model was used to compute thermal stability of T4 bacteriophage lysozyme mutants across seven sites. Qualitative assessment of mutational effects provides a straightforward rationale regarding whether a particular site primarily perturbs native or non-native states, or both. The presented model was found to be in good agreement with experimental mutational data (R 2 = 0.73) and suggests an approach by which to convert structure space into energy space.

Escape of a Small Molecule from Inside T4 Lysozyme by Multiple Pathways

The T4 lysozyme L99A mutant is often used as a model system to study small-molecule binding to proteins, but pathways for ligand entry and exit from the buried binding site and the associated protein conformational changes have not been fully resolved. Here, molecular dynamics simulations were employed to model benzene exit from its binding cavity using the weighted ensemble (WE) approach to enhance sampling of low-probability unbinding trajectories. Independent WE simulations revealed four pathways for benzene exit, which correspond to transient tunnels spontaneously formed in previous simulations of apo T4 lysozyme. Thus, benzene unbinding occurs through multiple pathways partially created by intrinsic protein structural fluctuations. Motions of several α-helices and side chains were involved in ligand escape from metastable microstates. WE simulations also provided preliminary estimates of rate constants for each exit pathway. These results complement previous works and provide a semiquantitative characterization of pathway heterogeneity for binding of small molecules to proteins.

Mechanisms for Benzene Dissociation through the Excited State of T4 Lysozyme L99A Mutant

The atomic-level mechanisms that coordinate ligand release from protein pockets are only known for a handful of proteins. Here, we report results from accelerated molecular dynamics simulations for benzene dissociation from the buried cavity of the T4 lysozyme Leu99Ala mutant (L99A). In these simulations, benzene is released through a previously characterized, sparsely populated room-temperature excited state of the mutant, explaining the coincidence for experimentally measured benzene off rate and apo protein slow-timescale NMR relaxation rates between ground and excited states. The path observed for benzene egress is a multistep ligand migration from the buried cavity to ultimate release through an opening between the F/G-, H-, and I-helices and requires a number of cooperative multiresidue and secondary-structure rearrangements within the C-terminal domain of L99A. These rearrangements are identical to those observed along the ground state to excited state transitions characterized by molecular dynamic simulations run on the Anton supercomputer. Analyses of the molecular properties of the residues lining the egress path suggest that protein surface electrostatic potential may play a role in the release mechanism. Simulations of wild-type T4 lysozyme also reveal that benzene-egress-associated dynamics in the L99A mutant are potentially exaggerations of the substrate-processivity-related dynamics of the wild type.

Pressure-Temperature Analysis of the Stability of the CTL9 Domain Reveals Hidden Intermediates

The observation of two-state unfolding for many small single-domain proteins by denaturants has led to speculation that protein sequences may have evolved to limit the population of partially folded states that could be detrimental to fitness. How such strong cooperativity arises from a multitude of individual interactions is not well understood. Here, we investigate the stability and folding cooperativity of the C-terminal domain of the ribosomal protein L9 in the pressure-temperature plane using site-specific NMR. In contrast to apparent cooperative unfolding detected with denaturant-induced and thermal-induced unfolding experiments and stopped-flow refolding studies at ambient pressure, NMR-detected pressure unfolding revealed significant deviation from two-state behavior, with a core region that was selectively destabilized by increasing temperature. Comparison of pressure-dependent NMR signals from both the folded and unfolded states revealed the population of at least one invisible excited state at atmospheric pressure. The core destabilizing cavity-creating I98A mutation apparently increased the cooperativity of the loss of folded-state peak intensity while also increasing the population of this invisible excited state present at atmospheric pressure. These observations highlight how local stability is subtly modulated by sequence to tune protein conformational landscapes and illustrate the ability of pressure- and temperature-dependent studies to reveal otherwise hidden states.

The pKa values of the catalytic residues in the retaining glycoside hydrolase T26H mutant of T4 lysozyme

T4 phage lysozyme (T4L) is an enzyme that cleaves bacterial cell wall peptidoglycan. Remarkably, the single substitution of the active site Thr26 to a His (T26H) converts T4L from an inverting to a retaining glycoside hydrolase with transglycosylase activity. It has been proposed that T26H-T4L follows a double displacement mechanism with His26 serving as a nucleophile to form a covalent glycosyl-enzyme intermediate (Kuroki et al., PNAS 1999; 96:8949-8954). To gain further insights into this or alternative mechanisms, we used NMR spectroscopy to measure the acid dissociation constants (pK_a values) and/or define the ionization states of the Asp, Glu, His, and Arg residues in the T4L mutant. Most notably, the pK_a value of the putative nucleophile His26 is 6.8 ± 0.1, whereas that of the general acid Glu11 is 4.7 ± 0.1. If the proposed mechanism holds true, then T26H-T4L follows a reverse protonation pathway in which only a minor population of the free enzyme is in its catalytically competent ionization state with His26 deprotonated and Glu11 protonated. Our studies also confirm that all arginines in T26H-T4L, including the active site Arg145, are positively charged under neutral pH conditions. BRIEF STATEMENT: The replacement of a single amino acid changes T4 lysozyme from an inverting to a retaining glycoside hydrolase. Using NMR spectroscopy, we measured the pK_a values of the ionizable residues in the active site of this mutant enzyme. Along with previously reported data, these results provide important constraints for understanding the catalytic mechanisms by which the wild-type and mutant form of T4 lysozyme cleave bacterial peptidoglycan.