Biology subjects

Shifman, J. M.

Publications and source records attributed to Shifman, J. M..

6 recordsLinked to original sources

bioRxiv2024

Improving Circulation Half-Life of Therapeutic Candidate N-TIMP2 by Unfolded Peptide Extension

Matrix Metalloproteinases (MMPs) are drivers of many diseases including cancer and are established targets for drug development. Tissue inhibitors of metalloproteinases (TIMPs) are human proteins that inhibit MMPs and are being pursued for the development of anti-MMP therapeutics. TIMPs possess many attractive properties of a drug candidate, such as complete MMP inhibition, low toxicity and immunogenicity, high tissue permeability and others. A major challenge with TIMPs, however, is their formulation and delivery, as these proteins are quickly cleared from the bloodstream due to their small size. In this study, we explore a new method for plasma half-life extension for the N-terminal domain of TIMP2 (N-TIMP2) through appending it with a long intrinsically unfolded tail containing a random combination of Pro, Ala, and Thr (PATylation). We design, produce and explore two PATylated N-TIMP2 constructs with a tail length of 100- and 200-amino acids (N-TIMP2-PAT100 and N-TIMP2-PAT200, respectively). We demonstrate that both PATylated N-TIMP2 constructs possess apparent higher molecular weights compared to the wild-type protein and retain high inhibitory activity against MMP-9. Furthermore, when injected into mice, N-TIMP2-PAT200 exhibited a significant increase in plasma half-life compared to the non-PATylated variant, enhancing the therapeutic potential of the protein. Thus, we establish that PATylation could be successfully applied to TIMP-based therapeutics and offers distinct advantages as an approach for half-life extension, such as fully genetic encoding of the gene construct, mono-dispersion, and biodegradability. Furthermore, PATylation could be easily applied to N-TIMP2 variants engineered to possess high affinity and selectivity toward individual MMP family members, thus creating attractive candidates for drug development against MMP-related diseases.

biochemistry↗

bioRxiv2024

ProBASS: a language model with sequence and structural features for predicting the effect of mutations on binding affinity

Protein-protein interactions (PPIs) govern virtually all cellular processes. Even a single mutation within PPI can significantly influence overall protein functionality and potentially lead to various types of diseases. To date, numerous approaches have emerged for predicting the change in free energy of binding ({Delta}{Delta}Gbind) resulting from mutations, yet the majority of these methods lack precision. In recent years, protein language models (PLMs) have been developed and shown powerful predictive capabilities by leveraging both sequence and structural data from protein-protein complexes. Yet, PLMs have not been optimized specifically for predicting {Delta}{Delta}Gbind. We developed an approach to predict effects of mutations on PPI binding affinity based on two most advanced protein language models ESM2 and ESM-IF1 that incorporate PPI sequence and structural features, respectively. We used the two models to generate embeddings for each PPI mutant and subsequently fine-tuned our model by training on a large dataset of experimental {Delta}{Delta}Gbind values. Our model, ProBASS (Protein Binding Affinity from Structure and Sequence) achieved a correlation with experimental {Delta}{Delta}Gbind values of 0.83 {+/-} 0.05 for single mutations and 0.69 {+/-} 0.04 for double mutations when model training and testing was done on the same PDB. Moreover, ProBASS exhibited very high correlation (0.81 {+/-} 0.02) between prediction and experiment when training and testing was performed on a dataset containing 2325 single mutations in 132 PPIs. ProBASS surpasses the state-of-the-art methods in correlation with experimental data and could be further trained as more experimental data becomes available. Our results demonstrate that the integration of extensive datasets containing {Delta}{Delta}Gbind values across multiple PPIs to refine the pre-trained PLMs represents a successful approach for achieving a precise and broadly applicable model for {Delta}{Delta}Gbind prediction, greatly facilitating future protein engineering and design studies.

bioinformatics↗

bioRxiv2023

Computational design of Matrix Metalloprotenaise-9 (MMP-9) resistant to auto-cleavage

Matrix metalloproteinase-9 (MMP-9) is an endopeptidase that remodels the extracellular matrix and has been implicated as a major driver in cancer metastasis. Hence, there is a high demand for MMP-9 inhibitors for therapeutic purposes. For such drug design efforts, large amounts of MMP-9 are required. Yet, the catalytic domain of MMP-9 (MMP-9Cat) is an intrinsically unstable enzyme that tends to auto-cleave within minutes, making it difficult to use in drug design experiments and other biophysical studies. We set our goal to design MMP-9Cat variant that is active but stable to autocleavage. For this purpose, we first identified potential autocleavage sites on MMP-9Cat using mass spectroscopy and then eliminated the autocleavage site by predicting mutations that minimize autocleavage potential without reducing enzyme stability. Four computationally designed MMP-9Cat variants were experimentally constructed and evaluated for auto-cleavage and enzyme activity. Our best variant, Des2, with 2 mutations, was as active as the wild-type enzyme but did not exhibit auto-cleavage after seven days of incubation at 37{degrees}C. This MMP-9Cat variant, with an identical to MMP- 9Cat WT active site, is an ideal candidate for drug design experiments targeting MMP-9 and enzyme crystallization experiments. The developed strategy for MMP-9CAT stabilization could be applied to redesign of other proteases to improve their stability for various biotechnological applications.

biochemistry↗

bioRxiv2022

A broad matrix metalloproteinase inhibitor with designed loop extension exhibits ultrahigh specificity for MMP-14

Matrix metalloproteinases (MMPs) are key drivers of various diseases, including cancer. While several antibodies against MMPs are in development, our goal is to construct therapeutic anti-MMP inhibitors based on a natural broad MMP inhibitor, tissue inhibitor of metalloproteinases-2 (N-TIMP2). To confer high binding specificity toward one MMP type, we extend one of the N-TIMP2 loops, allowing it to interact with the non-conserved MMP surface. Multiple computational designs of the loop were used to design a focused library for yeast surface display, which was sorted for high binding to the target MMP-14 and low binding to off-target MMP-3. Deep sequencing of the two selected populations followed by comparative data analysis was used to identify the most promising variants, which were expressed, purified, and tested for inhibition of MMP-14 and off-target MMPs. Our best N-TIMP2 variant exhibited 29 pM binding affinity to MMP-14 and 2.4 {micro}M affinity to MMP-3, 7500-fold more specific than WT N-TIMP2. Furthermore, the variant inhibited cell invasion with increased potency relative to WT N-TIMP2 in two breast cancer cell lines. We obtained the engineered variant high-accuracy model by including NGS data as input to AlphaFold multiple sequence alignment (MSA). Modeling results together with experimental mutagenesis demonstrate that the loop packs tightly against non-conserved residues on MMP-14 and clashes with MMP-3. This study demonstrates that introduction of loop extensions into inhibitors to stretch to the non-conserved surface of the target proteins is an attractive strategy for conferring high binding specificity in design of MMP inhibitors and other therapeutic proteins.

biochemistry↗

bioRxiv2020

Climbing up and down binding landscapes: a high-throughput study of mutational effects in homologous protein-protein complexes

Each protein-protein interaction (PPI) has evolved to possess binding affinity that is compatible with its cellular function. As such, cognate enzyme/inhibitor interactions frequently exhibit very high binding affinities, while structurally similar non-cognate PPIs possess substantially weaker binding affinities. To understand how slight differences in sequence and structure could lead to drastic changes in PPI binding free energy ({Delta}{Delta}Gbind), we study three homologous PPIs that span nine orders of magnitude in binding affinity and involve a serine protease interacting with an inhibitor BPTI. Using state-of-the-art methodology that combines protein randomization and affinity sorting coupled to next-generation sequencing and data normalization, we report quantitative binding landscapes consisting of {Delta}{Delta}Gbind values for the three PPIs, gleaned from tens of thousands of single and double mutations in the BPTI binding interface. We demonstrate that the three homologous PPIs possess drastically different binding landscapes and lie at different points in respect to the landscape maximum. Furthermore, the three PPIs demonstrate distinct patterns of coupling energies between two simultaneous mutations that depend not only on positions involved but also on the nature of the mutation. Interestingly, we find that in all three PPIs positive epistasis is frequently observed at hot-spot positions where mutations lead to loss of high affinity, while conversely negative epistasis is observed at cold-spot positions, where mutations lead to affinity enhancement. The new insights on PPI evolution revealed in this study will be invaluable in understanding evolution of other biological complexes and can greatly facilitate design of novel high-affinity protein inhibitors. SignificanceProtein-protein interactions (PPIs) have evolved to display binding affinities that can support their function. As such, cognate and non-cognate PPIs could be highly similar structurally but exhibit huge differences in binding affinities. To understand this phenomenon, we studied the effect of tens of thousands of single and double mutations on binding affinity of three homologous protease-inhibitor complexes. We show that binding landscapes of the three complexes are strikingly different and depend on the PPI evolutionary optimality. We observe different patterns of couplings between mutations for the three PPIs with negative and positive epistasis appearing most frequently at hot-spot and cold-spot positions, respectively. The evolutionary trends observed here are likely to be universal to all biological complexes in the cell.

biophysics↗

bioRxiv2017

Converting a broad matrix metalloproteinase family inhibitor into a specific inhibitor of MMP-9 and MMP-14

MMP-14 and MMP-9 are two well established cancer targets for which no specific clinically relevant inhibitor is available. Using a powerful combination of computational design and yeast surface display technology, we engineered such an inhibitor starting from a non-specific MMP inhibitor, N-TIMP2. The engineered purified N-TIMP2 variants showed enhanced specificity towards MMP-14 and MMP-9 relative to a panel of off-target MMPs. MMP-specific N-TIMP2 sequence signatures were obtained that could be understood from the structural perspective of MMP/N-TIMP2 interactions. Our MMP-9 inhibitor exhibited 1000-fold preference for MMP-9 vs. MMP-14, which is likely to translate into significant differences under physiological conditions. Our results provide new insights regarding evolution of promiscuous proteins and optimization strategies for design of inhibitors with single-target specificities.

biophysics↗