Differential Binding Kinetics for Evaluating Immune Checkpoint Inhibitors in Serum

The ability to characterize the binding kinetics of drug-target interactions in a biologically relevant matrix, such as serum or plasma, remains a fundamental challenge in drug discovery. We apply a novel label-based giant magnetoresistance (GMR) biosensor platform to measure protein binding kinetics and affinities of drug-target pairs in buffer and different levels of serum. Specifically, we evaluate three well-established immune checkpoint inhibitors, pembrolizumab, nivolumab and atezolizumab and compare the results with label-free kinetic platforms: surface plasmon resonance (SPR) and bio-layer interferometry (BLI). Labeling of analytes does not affect their association and dissociation rates (on and off rates) from GMR biosensors which enables kinetic measurements in biologically relevant matrices. Only the GMR biossensors is consistently suitable for measuring binding kinetics in up to 80% serum. The faster and different off-rates of the three immune checkpoint inhibitors in the presence of serum should be considered when modeling their pharmacological performance.

The AIRE G228W mutation results in a longer-lasting AIRE-SIRT1 interaction

The in silico and in vitro binding of a peptide covering a part of the autoimmune regulator (AIRE) SAND domain with the SIRT1 protein provides a powerful model system for studying the mechanism of the dominant SAND G228W mutation, which is the causative of APS-1 autoimmune syndrome. It is known that the mutant G228W AIRE protein accumulates more within the nucleus of cells than its wild-type counterpart does. This accumulation is not physiological and is associated with loss of AIRE function. However, the precise molecular mechanism that leads to AIRE accumulation is not yet known. AIRE works as a tetramer and interacts with partner proteins to form the “AIRE complex” that pushes RNA Pol II stalling in the chromatin of medullary thymic epithelial cells. Under normal conditions, the SIRT1 protein temporarily interacts with AIRE and deacetylates Lys residues of the SAND domain. Once AIRE is deacetylated, the binding with SIRT1 is undone, allowing the complex to proceed downstream. Here, we integrate molecular modeling, docking, dynamics, and surface plasmon resonance approaches to compare the structure and energetics of binding/release between AIRE G228 (wild-type) or W228 (mutant) peptides to SIRT1. We find that the proximity of G228W mutation to a K aminoacid residue in the SAND domain promotes a longer-lasting AIRE-SIRT1 interaction. The lasting interaction might cause a delay in the AIRE SAND domain to be released from the SIRT1 catalytic site, which might cause accumulation of the defective AIRE mutant protein in the nuclei of cells.SignificanceThis report reveals the mechanism of the pathogenic and dominant G228W mutation in the AIRE SAND domain. The G228W mutation is found in APS-1 syndrome patients, and it is critical to understand the molecular basis of loss of self-representation, a challenging aspect for immunology. Through modeling, molecular dynamics, and protein binding kinetics, we found that the G228W mutation leads to a stronger physical interaction between the AIRE SAND domain and the SIRT1 protein when compared to the equivalent wild-type segment. The short-term consequence of this stronger interaction is that the release of the AIRE-SIRT1 binding is slower. This might explain the reason that cells carrying the G228W mutation accumulate AIRE protein in their nuclei. This finding reveals with precision the AIRE-SIRT1 binding and the molecular mechanism of the human AIRE G228W mutation.

Measurement of kinetic constant of protein binding using microfluidics and particle diffusometry

Protein-protein interaction is widely used in biological science and biomedical engineering research Moreira et al. (2007). Accurate measurement of binding kinetics is essential for understanding protein-protein interactions. Current gold standard assays, such as surface plasmon resonance (SPR), bio-layer interferometry (BLI) and quartz crystal microbalance (QCM), can generate precise and real-time kinetics data. However, these methods usually require expensive instruments housed in core facilities and high-level expertise, which is not convenient for most labs to implement. We developed a new method based on microfluidics and particle diffusometry (PD) to measure protein binding kinetics, which only needs very general lab equipment including a fluorescent microscope to take photos, a syringe pump to inject solutions, capillary tubing, a simple chip made on a glass slide and a computer to process images. To measure the binding rate of a protein pair, both proteins are conjugated with beads of different sizes, respectively. The bead solutions are diluted to appropriate concentrations and injected into a Y-junction channel by a syringe pump. In the microchannel, the two kinds of beads will meet at the interface and bind due to surface protein interactions. Therefore, the size of the beads in solution gradually increases and the Brownian motion will be less and less drastic until the reaction is saturated. Taking photos recording this dynamic process, the apparent change in size of the beads can be measured by particle diffusometry and used for extracting binding kinetics. Particle diffusometry is a correlation-based and non-intrusive optical detection method to analyze properties of fluid and particle such as viscosity, temperature and particle diameter Chamarthy et al. (2009); Clayton et al. (2016, 2017b,a); Hohreiter et al. (2002). It was initially developed to determine errors caused by thermal noise in particle image velocimetry (PIV). PD always analyzes image pairs. A single image of a particle laden flow is first used to do auto-correlation, correlating with itself, which will generate a high and sharp peak. Then it is cross-correlated with a successive image with a known time interval ∆t. Because particles slightly deviate away from the initial positions after time ∆t, due to Brownian motion, the correlation peak is lower and broader than that of auto-correlation. Auto- and cross-correlation peaks are fit into Gaussian function to find peak widths, by which the particle size can be computed as long as the viscosity and temperature do not change. Processing the image sets of the protein-conjugated particles’ binding process, we acquire the relation of particle size and time, which can be used to solve protein binding kinetics. An equation of protein interaction and particle volume is derived to work out association rate from particle diameter data acquired by PD. In this study, we measured streptavidin-biotin binding rate. Streptavidin is conjugated with 20nm beads and biotin is immobilized onto 200nm beads. Proteins on the two kinds of beads bind rapidly after mixing in the main channel. It is necessary to choose a narrow area at the interface of particle streams that diffusion does not limit the reaction. Since the liquid is flowing, there is both Brownian motion and advection in particle images. We used EDPIV, a software package developed by Prof. Steven Wereley’s lab, to measure advection velocity. When doing PD analysis, images are shifted following the PIV data to catch up with the flow. The photos are taken at the center layer in the middle of the channel, where there is no velocity gradient. Measuring a series of photo sets along the main channel at several points with known distances to each other, the relation of complex bead size and time can be acquired. Solving for the association constant, the measured value is 1.74 × 107M−1s−1, which is close to that of current gold standard assays. This novel PD-based method is accurate and requires only general lab facilities, making protein binding kinetics measurements accessible and practical for biological and biomedical labs.

Devil in the details: Mechanistic variations impact information transfer across models of transcriptional cascades

The transcriptional network determines a cell’s internal state by regulating protein expression in response to changes in the local environment. Due to the interconnected nature of this network, information encoded in the abundance of various proteins will often propagate across chains of noisy intermediate signaling events. The data-processing inequality (DPI) leads us to expect that this intracellular game of “telephone” should degrade this type of signal, with longer chains losing successively more information to noise. However, a previous modeling effort predicted that because the steps of these signaling cascades do not truly represent independent stages of data processing, the limits of the DPI could seemingly be surpassed, and the amount of transmitted information could actually increase with chain length. What that work did not examine was whether this regime of growing information transmission was attainable by a signaling system constrained by the mechanistic details of more complex protein-binding kinetics. Here we address this knowledge gap through the lens of information theory by examining a model that explicitly accounts for the binding of each transcription factor to DNA. We analyze this model by comparing stochastic simulations of the fully nonlinear kinetics to simulations constrained by the linear response approximations that displayed a regime of growing information. Our simulations show that even when molecular binding is considered, there remains a regime wherein the transmitted information can grow with cascade length, but ends after a critical number of links determined by the kinetic parameter values. This inflection point marks where correlations decay in response to an oversaturation of binding sites, screening informative transcription factor fluctuations from further propagation down the chain where they eventually become indistinguishable from the surrounding levels of noise.