Dynamic active-site protection by the M. tuberculosis protein tyrosine phosphatase PtpB lid domain

100th Anniversary of Macromolecular Science Viewpoint: Data-Driven Protein Design

The design of synthetic proteins with the desired function is a long-standing goal in biomolecular science, with broad applications in biochemical engineering, agriculture, medicine, and public health. Rational de novo design and experimental directed evolution have achieved remarkable successes but are challenged by the requirement to find functional "needles" in the vast "haystack" of protein sequence space. Data-driven models for fitness landscapes provide a predictive map between protein sequence and function and can prospectively identify functional candidates for experimental testing to greatly improve the efficiency of this search. This Viewpoint reviews the applications of machine learning and, in particular, deep learning as part of data-driven protein engineering platforms. We highlight recent successes, review promising computational methodologies, and provide an outlook on future challenges and opportunities. The article is written for a broad audience comprising both polymer and protein scientists and computer and data scientists interested in an up-to-date review of recent innovations and opportunities in this rapidly evolving field.

Statistical Learning of Discrete States in Time Series

Time series obtained from time-dependent experiments contain rich information on kinetics and dynamics of the system under investigation. This work describes an unsupervised learning framework, along with the derivation of the necessary analytical expressions, for the analysis of Gaussian-distributed time series that exhibit discrete states. After the time series has been partitioned into segments in a model-free manner using the previously developed change-point (CP) method, this protocol starts with an agglomerative hierarchical clustering algorithm to classify the detected segments into possible states. The initial state clustering is further refined using an expectation-maximization (EM) procedure, and the number of states is determined by a Bayesian information criterion (BIC). Also introduced here is an achievement scalarization function, usually seen in artificial intelligence literature, for quantitatively assessing the performance of state determination. The statistical learning framework, which is comprised of three stages, detection of signal change, clustering, and number-of-state determination, was thoroughly characterized using simulated trajectories with random intensity segments that have no underlying kinetics, and its performance was critically evaluated. The application to experimental data is also demonstrated. The results suggested that this general framework, the implementation of which is based on firm theoretical foundations and does not require the imposition of any kinetics model, is powerful in determining the number of states, the parameters contained in each state, as well as the associated statistical significance.

Compound Molecular Logic in Accessing the Active Site of Mycobacterium tuberculosis Protein Tyrosine Phosphatase B

Protein tyrosine phosphatase B (PtpB) from Mycobacterium tuberculosis (Mtb) extends the bacteria's survival in hosts and hence is a potential target for Mtb-specific drugs. To study how Mtb-specific sequence insertions in PtpB may regulate access to its active site through large-amplitude conformational changes, we performed free-energy calculations using an all-atom explicit solvent model. Corroborated by biochemical assays, the results show that PtpB's active site is controlled via an "either/or" compound conformational gating mechanism, an unexpected discovery that Mtb has evolved to bestow a single enzyme with such intricate logical operations. In addition to providing unprecedented insights for its active-site surroundings, the findings also suggest new ways of inactivating PtpB.

MptpB Promotes Mycobacteria Survival by Inhibiting the Expression of Inflammatory Mediators and Cell Apoptosis in Macrophages

Tuberculosis is a severe contagious disease caused by Mycobacterium tuberculosis (Mtb). To develop new vaccines and medicine against TB, there is an urgent need to provide insights into the mechanisms by which Mtb induces tuberculosis. In this study, we found that secreted Mtb virulence factor MptpB significantly enhanced the survival of H37Rv in macrophages. MptpB suppressed the production of iNOS, the expression of inflammatory factors IL-1β and IL-6, as well as the apoptosis of the macrophage in Mtb infected RAW264.7 cells. Mechanism investigation showed that MptpB simultaneously hampered the NF-κB and MAPK signal pathways, evidenced by its blocking of p65, IKKα, Erk1/2, and p38 phosphorylation induced by Mtb infection. MptpB also inhibited host cell p53 expression. The results demonstrated that MptpB contributed to the survival of H37Rv by inhibiting host inflammatory responses and apoptosis through impeding the NF-κB and MAPK signal pathways and p53 expression in the macrophage.

Extraction of Protein Conformational Modes from Distance Distributions Using Structurally Imputed Bayesian Data Augmentation

Protein conformational changes are known to play important roles in assorted biochemical and biological processes. Driven by thermal motions of surrounding solvent molecules, such a structural remodeling often occurs stochastically. Yet, regardless of how random the conformational reconfiguration may appear, it could in principle be described by a linear combination of a set of orthogonal modes which, in turn, are contained in the intramolecular distance distributions. The central challenge is how to obtain the distribution. This contribution proposes a Bayesian data-augmentation scheme to extract the predominant modes from only few distance distributions, be they from computational sampling or directly from experiments such as single-molecule Förster-type resonance energy transfer (smFRET). The inference of the complete protein structure from insufficient data was recognized as isomorphic to the missing-data problem in Bayesian statistical learning. Using smFRET data as an example, the missing coordinates were deduced, given protein structural constraints and multiple but limited number of smFRET distances; the Boltzmann weighing of each inferred protein structure was then evaluated using computational modeling to numerically construct the posterior density for the global protein conformation. The conformational modes were then determined from the iteratively converged overall conformational distribution using principal component analysis. Two examples were presented to illustrate these basic ideas as well as their practical implementation. The scheme described herein was based on the theory behind the powerful Tanner-Wang algorithm that guarantees convergence to the true posterior density. However, instead of assuming a mathematical model to calculate the likelihood as in conventional statistical inference, here the protein structure was treated as a statistical parameter and was imputed from the numerical likelihood function based on structural information, a probability model-free method. The framework put forth here is anticipated to be generally applicable, offering a new way to articulate protein conformational changes in a quantifiable manner.

Mycobacterium tuberculosis Serine/Threonine Protein Kinases

The Mycobacterium tuberculosis genome encodes 11 serine/threonine protein kinases (STPKs). A similar number of two-component systems are also present, indicating that these two signal transduction mechanisms are both important in the adaptation of this bacterial pathogen to its environment. The M. tuberculosis phosphoproteome includes hundreds of Ser- and Thr-phosphorylated proteins that participate in all aspects of M. tuberculosis biology, supporting a critical role for the STPKs in regulating M. tuberculosis physiology. Nine of the STPKs are receptor type kinases, with an extracytoplasmic sensor domain and an intracellular kinase domain, indicating that these kinases transduce external signals. Two other STPKs are cytoplasmic and have regulatory domains that sense changes within the cell. Structural analysis of some of the STPKs has led to advances in our understanding of the mechanisms by which these STPKs are activated and regulated. Functional analysis has provided insights into the effects of phosphorylation on the activity of several proteins, but for most phosphoproteins the role of phosphorylation in regulating function is unknown. Major future challenges include characterizing the functional effects of phosphorylation for this large number of phosphoproteins, identifying the cognate STPKs for these phosphoproteins, and determining the signals that the STPKs sense. Ultimately, combining these STPK-regulated processes into larger, integrated regulatory networks will provide deeper insight into M. tuberculosis adaptive mechanisms that contribute to tuberculosis pathogenesis. Finally, the STPKs offer attractive targets for inhibitor development that may lead to new therapies for drug-susceptible and drug-resistant tuberculosis.

Protein topology determines cysteine oxidation fate: the case of sulfenyl amide formation among protein families

Cysteine residues have a rich chemistry and play a critical role in the catalytic activity of a plethora of enzymes. However, cysteines are susceptible to oxidation by Reactive Oxygen and Nitrogen Species, leading to a loss of their catalytic function. Therefore, cysteine oxidation is emerging as a relevant physiological regulatory mechanism. Formation of a cyclic sulfenyl amide residue at the active site of redox-regulated proteins has been proposed as a protection mechanism against irreversible oxidation as the sulfenyl amide intermediate has been identified in several proteins. However, how and why only some specific cysteine residues in particular proteins react to form this intermediate is still unknown. In the present work using in-silico based tools, we have identified a constrained conformation that accelerates sulfenyl amide formation. By means of combined MD and QM/MM calculation we show that this conformation positions the NH backbone towards the sulfenic acid and promotes the reaction to yield the sulfenyl amide intermediate, in one step with the concomitant release of a water molecule. Moreover, in a large subset of the proteins we found a conserved beta sheet-loop-helix motif, which is present across different protein folds, that is key for sulfenyl amide production as it promotes the previous formation of sulfenic acid. For catalytic activity, in several cases, proteins need the Cysteine to be in the cysteinate form, i.e. a low pK_a Cys. We found that the conserved motif stabilizes the cysteinate by hydrogen bonding to several NH backbone moieties. As cysteinate is also more reactive toward ROS we propose that the sheet-loop-helix motif and the constraint conformation have been selected by evolution for proteins that need a reactive Cys protected from irreversible oxidation. Our results also highlight how fold conservation can be correlated to redox chemistry regulation of protein function.

Cysteine oxidation is emerging as a relevant regulatory mechanism of enzymatic function in the cell. Many proteins are protected from over oxidation by reactive oxygen species by the formation of a cyclic sulfenyl amide. Understanding how cyclic sulfenyl amide is formed and its dependence on protein structure is not only a basic question but necessary to predict which proteins may auto protect from over oxidation We describe a structural motif, which includes cysteine residues with a constrained conformation in a “forbidden” region of the Ramachandran plot plus a Beta-Cys-loop-helix motif, which has a reactive low pK_a Cysteine and also enables to form the cyclic sulfenyl amide with a low activation barrier. Our QM/MM computations show that the cyclization reaction only occurs if the “forbidden” conformation is acquired by the Cysteine residue. This structural motif was identified at least in 7 PFAM families and 145 proteins with solved structure, showing that a large number of proteins could have the ability to go through such cyclic product preventing irreversible oxidation.

Picosecond-resolved fluorescent probes at functionally distinct tryptophans within a thermophilic alcohol dehydrogenase: relationship of temperature-dependent changes in fluorescence to catalysis

Two single-tryptophan variants were generated in a thermophilic alcohol dehydrogenase with the goal of correlating temperature-dependent changes in local fluorescence with the previously demonstrated catalytic break at ca. 30 °C (Kohen et al., Nature 1999, 399, 496). One tryptophan variant, W87in, resides at the active site within van der Waals contact of bound alcohol substrate; the other variant, W167in, is a remote-site surface reporter located >25 Å from the active site. Picosecond-resolved fluorescence measurements were used to analyze fluorescence lifetimes, time-dependent Stokes shifts, and the extent of collisional quenching at Trp87 and Trp167 as a function of temperature. A subnanosecond fluorescence decay rate constant has been detected for W87in that is ascribed to the proximity of the active site Zn(2+) and shows a break in behavior at 30 °C. For the remainder of the reported lifetime measurements, there is no detectable break between 10 and 50 °C, in contrast with previously reported hydrogen/deuterium exchange experiments that revealed a temperature-dependent break analogous to catalysis (Liang et al., Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9556). We conclude that the motions that lead to the rigidification of ht-ADH below 30 °C are likely to be dominated by global processes slower than the picosecond to nanosecond motions measured herein. In the case of collisional quenching of fluorescence by acrylamide, W87in and W167in behave in a similar manner that resembles free tryptophan in water. Stokes shift measurements, by contrast, show distinctive behaviors in which the active-site tryptophan relaxation is highly temperature-dependent, whereas the solvent-exposed tryptophan's dynamics are temperature-independent. These data are concluded to reflect a significantly constrained environment surrounding the active site Trp87 that both increases the magnitude of the Stokes shift and its temperature-dependence. The results are discussed in the context of spatially distinct differences in enthalpic barriers for protein conformational sampling that may be related to catalysis.

Reactive nitrogen species and hydrogen sulfide as regulators of protein tyrosine phosphatase activity

SIGNIFICANCE

Redox modifications of thiols serve as a molecular code enabling precise and complex regulation of protein tyrosine phosphatases (PTPs) and other proteins. Particular gasotransmitters and even the redox modifications themselves affect each other, of which a typical example is S-nitrosylation-mediated protection against the further oxidation of protein thiols.

RECENT ADVANCES

For a long time, PTPs were considered constitutively active housekeeping enzymes. This view has changed substantially over the last two decades, and the PTP family is now recognized as a group of tightly and flexibly regulated fundamental enzymes. In addition to the conventional ways in which they are regulated, including noncovalent interactions, phosphorylation, and oxidation, the evidence that has accumulated during the past two decades suggests that many of these enzymes are also modulated by gasotransmitters, namely by nitric oxide (NO) and hydrogen sulfide (H2S).

CRITICAL ISSUES

The specificity and selectivity of the methods used to detect nitrosylation and sulfhydration remains to be corroborated, because several researchers raised the issue of false-positive results, particularly when using the most widespread biotin switch method. Further development of robust and straightforward proteomic methods is needed to further improve our knowledge of the full extent of the gasotransmitters-mediated changes in PTP activity, selectivity, and specificity. FURTHER DIRECTIONS: Results of the hitherto performed studies on gasotransmitter-mediated PTP signaling await translation into clinical medicine and pharmacotherapeutics. In addition to directly affecting the activity of particular PTPs, the use of reversible S-nitrosylation as a protective mechanism against oxidative stress should be of high interest.

Expectation-maximization of the potential of mean force and diffusion coefficient in Langevin dynamics from single molecule FRET data photon by photon

The dynamics of a protein along a well-defined coordinate can be formally projected onto the form of an overdamped Lagevin equation. Here, we present a comprehensive statistical-learning framework for simultaneously quantifying the deterministic force (the potential of mean force, PMF) and the stochastic force (characterized by the diffusion coefficient, D) from single-molecule Förster-type resonance energy transfer (smFRET) experiments. The likelihood functional of the Langevin parameters, PMF and D, is expressed by a path integral of the latent smFRET distance that follows Langevin dynamics and realized by the donor and the acceptor photon emissions. The solution is made possible by an eigen decomposition of the time-symmetrized form of the corresponding Fokker-Planck equation coupled with photon statistics. To extract the Langevin parameters from photon arrival time data, we advance the expectation-maximization algorithm in statistical learning, originally developed for and mostly used in discrete-state systems, to a general form in the continuous space that allows for a variational calculus on the continuous PMF function. We also introduce the regularization of the solution space in this Bayesian inference based on a maximum trajectory-entropy principle. We use a highly nontrivial example with realistically simulated smFRET data to illustrate the application of this new method.

Functional roles of slow enzyme conformational changes in network dynamics

Extensive studies from different fields reveal that many macromolecules, especially enzymes, show slow transitions among different conformations. This phenomenon is named such things as dynamic disorder, heterogeneity, hysteretic or mnemonic enzymes across these different fields, and has been directly demonstrated by single molecule enzymology and NMR studies recently. We analyzed enzyme slow conformational changes in the context of regulatory networks. A single enzymatic reaction with slow conformational changes can filter upstream network noises, and can either resonantly respond to the system stimulus at certain frequencies or respond adaptively for sustained input signals of the network fluctuations. It thus can serve as a basic functional motif with properties that are normally for larger intermolecular networks in the field of systems biology. We further analyzed examples including enzymes functioning against pH fluctuations, metabolic state change of Artemia embryos, and kinetic insulation of fluctuations in metabolic networks. The study also suggests that hysteretic enzymes may be building blocks of synthetic networks with various properties such as narrow-banded filtering. The work fills the missing gap between studies on enzyme biophysics and network level dynamics, and reveals that the coupling between the two is functionally important; it also suggests that the conformational dynamics of some enzymes may be evolutionally selected.

Structural distributions from single-molecule measurements as a tool for molecular mechanics

A mechanical view provides an attractive alternative for predicting the behavior of complex systems since it circumvents the resource-intensive requirements of atomistic models; however, it remains extremely challenging to characterize the mechanical responses of a system at the molecular level. Here, the structural distribution is proposed to be an effective means to extracting the molecular mechanical properties. End-to-end distance distributions for a series of short poly-L-proline peptides with the sequence P(n)CG(3)K-biotin (n = 8, 12, 15 and 24) were used to experimentally illustrate this new approach. High-resolution single-molecule Förster-type resonance energy transfer (FRET) experiments were carried out and the conformation-resolving power was characterized and discussed in the context of the conventional constant-time binning procedure for FRET data analysis. It was shown that the commonly adopted theoretical polymer models-including the worm-like chain, the freely jointed chain, and the self-avoiding chain-could not be distinguished by the averaged end-to-end distances, but could be ruled out using the molecular details gained by conformational distribution analysis because similar polymers of different sizes could respond to external forces differently. Specifically, by fitting the molecular conformational distribution to a semi-flexible polymer model, the effective persistence lengths for the series of short poly-L-proline peptides were found to be size-dependent with values of ~190 Å, ~67 Å, ~51 Å, and ~76 Å for n = 8, 12, 15, and 24, respectively. A comprehensive computational modeling was carried out to gain further insights for this surprising discovery. It was found that P(8) exists as the extended all-trans isomaer whereas P(12) and P(15) predominantly contained one proline residue in the cis conformation. P(24) exists as a mixture of one-cis (75%) and two-cis (25%) isomers where each isomer contributes to an experimentally resolvable conformational mode. This work demonstrates the resolving power of the distribution-based approach, and the capacity of integrating high-resolution single-molecule FRET experiments with molecular modeling to reveal detailed structural information about the conformation of molecules on the length scales relevant to the study of biological molecules.

LipA, a tyrosine and lipid phosphatase involved in the virulence of Listeria monocytogenes

Intracellular bacterial pathogens manipulate host cell functions by producing enzymes that stimulate or antagonize signal transduction. The Listeria monocytogenes genome contains a gene, lmo1800, encoding a protein with a conserved motif of conventional tyrosine phosphatases. Here, we report that the lmo1800-encoded protein LipA is secreted by Listeria and displays tyrosine as well as lipid phosphatase activity in vitro. Bacteria lacking LipA are severely attenuated in virulence in vivo, thus revealing a so-far-undescribed enzymatic activity involved in Listeria infection.

Structural landscape of isolated agonist-binding domains from single AMPA receptors

α-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) receptors mediate fast excitatory neurotransmission by converting chemical signals into electrical signals. Thus, it is important to understand the relationship between their chemical biology and their function. Single molecule fluorescence resonance energy transfer (smFRET) was used to examine the conformations explored by the agonist binding domain of the AMPA receptor for wild type and T686 mutant proteins. Each form of the agonist binding domain exhibited a dynamic, multi-state sequential equilibrium, which could only be identified using wavelet shrinkage, a signal processing technique that removes experimental shot-noise. These results illustrate that the extent of activation is dependent not on a rigid closed cleft, but instead on the probability that a given subunit will occupy a closed cleft conformation, which in turn is not only determined by the lowest energy state but by the range of states that the protein explores.

Seeing the forest for the trees: fluorescence studies of single enzymes in the context of ensemble experiments

Enzymes are remarkable molecular machines that make many difficult biochemical reactions possible under mild biological conditions with incredible precision and efficiency. Our understanding of the working principles of enzymes, however, has not reached the level where one can readily deduce the mechanism and the catalytic rates from an enzyme's structure. Resolving the dynamics that relate the three-dimensional structure of an enzyme to its function has been identified as a key issue. While still challenging to implement, single-molecule techniques have emerged as one of the most useful methods for studying enzymes. We review enzymes studied using single-molecule fluorescent methods but placing them in the context of results from other complementary experimental work done on bulk samples. This review primarily covers three enzyme systems--flavoenzymes, dehydrofolate reductase, and adenylate kinase--with additional enzymes mentioned where appropriate. When the single-molecule experiments are discussed together with other methods aiming at the same scientific question, the weakness, strength, and unique contributions become clear.