Improving Protein Expression, Stability, and Function with ProteinMPNN.

Natural proteins are highly optimized for function but are often difficult to produce at a scale suitable for biotechnological applications due to poor expression in heterologous systems, limited solubility, and sensitivity to temperature. Thus, a general method that improves the physical properties of native proteins while maintaining function could have wide utility for protein-based technologies. Here, we show that the deep neural network ProteinMPNN, together with evolutionary and structural information, provides a route to increasing protein expression, stability, and function. For both myoglobin and tobacco etch virus (TEV) protease, we generated designs with improved expression, elevated melting temperatures, and improved function. For TEV protease, we identified multiple designs with improved catalytic activity as compared to the parent sequence and previously reported TEV variants. Our approach should be broadly useful for improving the expression, stability, and function of biotechnologically important proteins.

Gut microbiome responds to alteration in female sex hormone status and exacerbates metabolic dysfunction.

Women are at significantly greater risk of metabolic dysfunction after menopause, which subsequently leads to numerous chronic illnesses. The gut microbiome is associated with obesity and metabolic dysfunction, but its interaction with female sex hormone status and the resulting impact on host metabolism remains unclear. Herein, we characterized inflammatory and metabolic phenotypes as well as the gut microbiome associated with ovariectomy and high-fat diet feeding, compared to gonadal intact and low-fat diet controls. We then performed fecal microbiota transplantation (FMT) using gnotobiotic mice to identify the impact of ovariectomy-associated gut microbiome on inflammatory and metabolic outcomes. We demonstrated that ovariectomy led to greater gastrointestinal permeability and inflammation of the gut and metabolic organs, and that a high-fat diet exacerbated these phenotypes. Ovariectomy also led to alteration of the gut microbiome, including greater fecal ß-glucuronidase activity. However, differential changes in the gut microbiome only occurred when fed a low-fat diet, not the high-fat diet. Gnotobiotic mice that received the gut microbiome from ovariectomized mice fed the low-fat diet had greater weight gain and hepatic gene expression related to metabolic dysfunction and inflammation than those that received intact sham control-associated microbiome. These results indicate that the gut microbiome responds to alterations in female sex hormone status and contributes to metabolic dysfunction. Identifying and developing gut microbiome-targeted modulators to regulate sex hormones may be useful therapeutically in remediating menopause-related diseases.

Small-molecule binding and sensing with a designed protein family.

Despite transformative advances in protein design with deep learning, the design of small-molecule-binding proteins and sensors for arbitrary ligands remains a grand challenge. Here we combine deep learning and physics-based methods to generate a family of proteins with diverse and designable pocket geometries, which we employ to computationally design binders for six chemically and structurally distinct small-molecule targets. Biophysical characterization of the designed binders revealed nanomolar to low micromolar binding affinities and atomic-level design accuracy. The bound ligands are exposed at one edge of the binding pocket, enabling the de novo design of chemically induced dimerization (CID) systems; we take advantage of this to create a biosensor with nanomolar sensitivity for cortisol. Our approach provides a general method to design proteins that bind and sense small molecules for a wide range of analytical, environmental, and biomedical applications.

De novo design of protein structure and function with RFdiffusion.

There has been considerable recent progress in designing new proteins using deep-learning methods1-9. Despite this progress, a general deep-learning framework for protein design that enables solution of a wide range of design challenges, including de novo binder design and design of higher-order symmetric architectures, has yet to be described. Diffusion models10,11 have had considerable success in image and language generative modelling but limited success when applied to protein modelling, probably due to the complexity of protein backbone geometry and sequence-structure relationships. Here we show that by fine-tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, we obtain a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomer design, protein binder design, symmetric oligomer design, enzyme active site scaffolding and symmetric motif scaffolding for therapeutic and metal-binding protein design. We demonstrate the power and generality of the method, called RoseTTAFold diffusion (RFdiffusion), by experimentally characterizing the structures and functions of hundreds of designed symmetric assemblies, metal-binding proteins and protein binders. The accuracy of RFdiffusion is confirmed by the cryogenic electron microscopy structure of a designed binder in complex with influenza haemagglutinin that is nearly identical to the design model. In a manner analogous to networks that produce images from user-specified inputs, RFdiffusion enables the design of diverse functional proteins from simple molecular specifications.

De novo design of luciferases using deep learning.

De novo enzyme design has sought to introduce active sites and substrate-binding pockets that are predicted to catalyse a reaction of interest into geometrically compatible native scaffolds1,2, but has been limited by a lack of suitable protein structures and the complexity of native protein sequence-structure relationships. Here we describe a deep-learning-based 'family-wide hallucination' approach that generates large numbers of idealized protein structures containing diverse pocket shapes and designed sequences that encode them. We use these scaffolds to design artificial luciferases that selectively catalyse the oxidative chemiluminescence of the synthetic luciferin substrates diphenylterazine3 and 2-deoxycoelenterazine. The designed active sites position an arginine guanidinium group adjacent to an anion that develops during the reaction in a binding pocket with high shape complementarity. For both luciferin substrates, we obtain designed luciferases with high selectivity; the most active of these is a small (13.9 kDa) and thermostable (with a melting temperature higher than 95 °C) enzyme that has a catalytic efficiency on diphenylterazine (kcat/Km = 106 M-1 s-1) comparable to that of native luciferases, but a much higher substrate specificity. The creation of highly active and specific biocatalysts from scratch with broad applications in biomedicine is a key milestone for computational enzyme design, and our approach should enable generation of a wide range of luciferases and other enzymes.

De novo protein design by deep network hallucination.

There has been considerable recent progress in protein structure prediction using deep neural networks to predict inter-residue distances from amino acid sequences1-3. Here we investigate whether the information captured by such networks is sufficiently rich to generate new folded proteins with sequences unrelated to those of the naturally occurring proteins used in training the models. We generate random amino acid sequences, and input them into the trRosetta structure prediction network to predict starting residue-residue distance maps, which, as expected, are quite featureless. We then carry out Monte Carlo sampling in amino acid sequence space, optimizing the contrast (Kullback-Leibler divergence) between the inter-residue distance distributions predicted by the network and background distributions averaged over all proteins. Optimization from different random starting points resulted in novel proteins spanning a wide range of sequences and predicted structures. We obtained synthetic genes encoding 129 of the network-'hallucinated' sequences, and expressed and purified the proteins in Escherichia coli; 27 of the proteins yielded monodisperse species with circular dichroism spectra consistent with the hallucinated structures. We determined the three-dimensional structures of three of the hallucinated proteins, two by X-ray crystallography and one by NMR, and these closely matched the hallucinated models. Thus, deep networks trained to predict native protein structures from their sequences can be inverted to design new proteins, and such networks and methods should contribute alongside traditional physics-based models to the de novo design of proteins with new functions.

Garcinoic Acid Is a Natural and Selective Agonist of Pregnane X Receptor.

Pregnane X receptor (PXR) is a master xenobiotic-sensing transcription factor and a validated target for immune and inflammatory diseases. The identification of chemical probes to investigate the therapeutic relevance of the receptor is still highly desired. In fact, currently available PXR ligands are not highly selective and can exhibit toxicity and/or potential off-target effects. In this study, we have identified garcinoic acid as a selective and efficient PXR agonist. The properties of this natural molecule as a specific PXR agonist were demonstrated by the screening on a panel of nuclear receptors, the assessment of the physical and thermodynamic binding affinity, and the determination of the PXR-garcinoic acid complex crystal structure. Cytotoxicity, transcriptional, and functional properties were investigated in human liver cells, and compound activity and target engagement were confirmed in vivo in mouse liver and gut tissue. In conclusion, garcinoic acid is a selective natural agonist of PXR and a promising lead compound toward the development of new PXR-regulating modulators.

Targeted inhibition of gut bacterial ß-glucuronidase activity enhances anticancer drug efficacy.

Irinotecan treats a range of solid tumors, but its effectiveness is severely limited by gastrointestinal (GI) tract toxicity caused by gut bacterial ß-glucuronidase (GUS) enzymes. Targeted bacterial GUS inhibitors have been shown to partially alleviate irinotecan-induced GI tract damage and resultant diarrhea in mice. Here, we unravel the mechanistic basis for GI protection by gut microbial GUS inhibitors using in vivo models. We use in vitro, in fimo, and in vivo models to determine whether GUS inhibition alters the anticancer efficacy of irinotecan. We demonstrate that a single dose of irinotecan increases GI bacterial GUS activity in 1 d and reduces intestinal epithelial cell proliferation in 5 d, both blocked by a single dose of a GUS inhibitor. In a tumor xenograft model, GUS inhibition prevents intestinal toxicity and maintains the antitumor efficacy of irinotecan. Remarkably, GUS inhibitor also effectively blocks the striking irinotecan-induced bloom of Enterobacteriaceae in immune-deficient mice. In a genetically engineered mouse model of cancer, GUS inhibition alleviates gut damage, improves survival, and does not alter gut microbial composition; however, by allowing dose intensification, it dramatically improves irinotecan's effectiveness, reducing tumors to a fraction of that achieved by irinotecan alone, while simultaneously promoting epithelial regeneration. These results indicate that targeted gut microbial enzyme inhibitors can improve cancer chemotherapeutic outcomes by protecting the gut epithelium from microbial dysbiosis and proliferative crypt damage.

Discovering the Microbial Enzymes Driving Drug Toxicity with Activity-Based Protein Profiling.

It is increasingly clear that interindividual variability in human gut microbial composition contributes to differential drug responses. For example, gastrointestinal (GI) toxicity is not observed in all patients treated with the anticancer drug irinotecan, and it has been suggested that this variability is a result of differences in the types and levels of gut bacterial ß-glucuronidases (GUSs). GUS enzymes promote drug toxicity by hydrolyzing the inactive drug-glucuronide conjugate back to the active drug, which damages the GI epithelium. Proteomics-based identification of the exact GUS enzymes responsible for drug reactivation from the complexity of the human microbiota has not been accomplished, however. Here, we discover the specific bacterial GUS enzymes that generate SN-38, the active and toxic metabolite of irinotecan, from human fecal samples using a unique activity-based protein profiling (ABPP) platform. We identify and quantify gut bacterial GUS enzymes from human feces with an ABPP-enabled proteomics pipeline and then integrate this information with ex vivo kinetics to pinpoint the specific GUS enzymes responsible for SN-38 reactivation. Furthermore, the same approach also reveals the molecular basis for differential gut bacterial GUS inhibition observed between human fecal samples. Taken together, this work provides an unprecedented technical and bioinformatics pipeline to discover the microbial enzymes responsible for specific reactions from the complexity of human feces. Identifying such microbial enzymes may lead to precision biomarkers and novel drug targets to advance the promise of personalized medicine.

Selecting a Single Stereocenter: The Molecular Nuances That Differentiate ß-Hexuronidases in the Human Gut Microbiome.

The human gut microbiome is a ripe space for the discovery of new proteins and novel functions. Many genes in the gut microbiome encode glycoside hydrolases that help bacteria scavenge sugars present in the human gut. Glycoside hydrolase family 2 (GH2) is one group of sugar-scavenging proteins, which includes ß-glucuronidases (GUS) and ß-galacturonidases (GalAses), enzymes that cleave the sugar conjugates of the epimers glucuronate and galacturonate. Here we structurally and functionally characterize a GH2 GalAse and a hybrid GUS/GalAse, which reveal the molecular details that enable these GHs to differentiate a single stereocenter. First, we characterized a previously annotated GUS from Eisenbergiella tayi and demonstrated that it is, in fact, a GalAse. We determined the crystal structure of this GalAse, identified the key residue that confers GalAse activity, and convert this GalAse into a GUS by mutating a single residue. We performed bioinformatic analysis of 279 putative GUS enzymes from the human gut microbiome and identified 12 additional putative GH2 GalAses, one of which we characterized and confirmed is a GalAse. Lastly, we report the structure of a hybrid GUS/GalAse from Fusicatenibacter saccharivorans, which revealed a unique hexamer that positions the N-terminus of adjacent protomers in the aglycone binding site. Taken together, these data reveal a new class of bacterial GalAses in the human gut microbiome and unravel the structural details that differentiate GH2 GUSs and GalAses.

Discovery and Characterization of FMN-Binding ß-Glucuronidases in the Human Gut Microbiome.

The human gut microbiota encodes ß-glucuronidases (GUSs) that play key roles in health and disease via the metabolism of glucuronate-containing carbohydrates and drugs. Hundreds of putative bacterial GUS enzymes have been identified by metagenomic analysis of the human gut microbiome, but less than 10% have characterized structures and functions. Here we describe a set of unique gut microbial GUS enzymes that bind flavin mononucleotide (FMN). First, we show using mass spectrometry, isothermal titration calorimetry, and x-ray crystallography that a purified GUS from the gut commensal microbe Faecalibacterium prausnitzii binds to FMN on a surface groove located 30 Šaway from the active site. Second, utilizing structural and functional data from this FMN-binding GUS, we analyzed the 279 unique GUS sequences from the Human Microbiome Project database and identified 14 putative FMN-binding GUSs. We characterized four of these hits and solved the structure of two, the GUSs from Ruminococcus gnavus and Roseburia hominis, which confirmed that these are FMN binders. Third, binding and kinetic analysis of the FMN-binding site mutants of these five GUSs show that they utilize a conserved site to bind FMN that is not essential for GUS activity, but can affect KM. Lastly, a comprehensive structural review of the PDB reveals that the FMN-binding site employed by these enzymes is unlike any structurally characterized FMN binders to date. These findings reveal the first instance of an FMN-binding glycoside hydrolase and suggest a potential link between FMN and carbohydrate metabolism in the human gut microbiota.

Structure, function, and inhibition of drug reactivating human gut microbial ß-glucuronidases.

Bacterial ß-glucuronidase (GUS) enzymes cause drug toxicity by reversing Phase II glucuronidation in the gastrointestinal tract. While many human gut microbial GUS enzymes have been examined with model glucuronide substrates like p-nitrophenol-ß-D-glucuronide (pNPG), the GUS orthologs that are most efficient at processing drug-glucuronides remain unclear. Here we present the crystal structures of GUS enzymes from human gut commensals Lactobacillus rhamnosus, Ruminococcus gnavus, and Faecalibacterium prausnitzii that possess an active site loop (Loop 1; L1) analogous to that found in E. coli GUS, which processes drug substrates. We also resolve the structure of the No Loop GUS from Bacteroides dorei. We then compare the pNPG and diclofenac glucuronide processing abilities of a panel of twelve structurally diverse GUS proteins, and find that the new L1 GUS enzymes presented here process small glucuronide substrates inefficiently compared to previously characterized L1 GUS enzymes like E. coli GUS. We further demonstrate that our GUS inhibitors, which are effective against some L1 enzymes, are not potent towards all. Our findings pinpoint active site structural features necessary for the processing of drug-glucuronide substrates and the inhibition of such processing.

Three structurally and functionally distinct ß-glucuronidases from the human gut microbe Bacteroides uniformis.

The glycoside hydrolases encoded by the human gut microbiome play an integral role in processing a variety of exogenous and endogenous glycoconjugates. Here we present three structurally and functionally distinct ß-glucuronidase (GUS) glycoside hydrolases from a single human gut commensal microbe, Bacteroides uniformis We show using nine crystal structures, biochemical, and biophysical data that whereas these three proteins share similar overall folds, they exhibit different structural features that create three structurally and functionally unique enzyme active sites. Notably, quaternary structure plays an important role in creating distinct active site features that are hard to predict via structural modeling methods. The enzymes display differential processing capabilities toward glucuronic acid-containing polysaccharides and SN-38-glucuronide, a metabolite of the cancer drug irinotecan. We also demonstrate that GUS-specific and nonselective inhibitors exhibit varying potencies toward each enzyme. Together, these data highlight the diversity of GUS enzymes within a single Bacteroides gut commensal and advance our understanding of how structural details impact the specific roles microbial enzymes play in processing drug-glucuronide and glycan substrates.

Gut Microbial ß-Glucuronidase Inhibition via Catalytic Cycle Interception.

Microbial ß-glucuronidases (GUSs) cause severe gut toxicities that limit the efficacy of cancer drugs and other therapeutics. Selective inhibitors of bacterial GUS have been shown to alleviate these side effects. Using structural and chemical biology, mass spectrometry, and cell-based assays, we establish that piperazine-containing GUS inhibitors intercept the glycosyl-enzyme catalytic intermediate of these retaining glycosyl hydrolases. We demonstrate that piperazine-based compounds are substrate-dependent GUS inhibitors that bind to the GUS-GlcA catalytic intermediate as a piperazine-linked glucuronide (GlcA, glucuronic acid). We confirm the GUS-dependent formation of inhibitor-glucuronide conjugates by LC-MS and show that methylated piperazine analogs display significantly reduced potencies. We further demonstrate that a range of approved piperazine- and piperidine-containing drugs from many classes, including those for the treatment of depression, infection, and cancer, function by the same mechanism, and we confirm through gene editing that these compounds selectively inhibit GUS in living bacterial cells. Together, these data reveal a unique mechanism of GUS inhibition and show that a range of therapeutics may impact GUS activities in the human gut.

Structural basis for the regulation of ß-glucuronidase expression by human gut Enterobacteriaceae.

The gut microbiota harbor diverse ß-glucuronidase (GUS) enzymes that liberate glucuronic acid (GlcA) sugars from small-molecule conjugates and complex carbohydrates. However, only the Enterobacteriaceae family of human gut-associated Proteobacteria maintain a GUS operon under the transcriptional control of a glucuronide repressor, GusR. Despite its potential importance in Escherichia, Salmonella, Klebsiella, Shigella, and Yersinia opportunistic pathogens, the structure of GusR has not been examined. Here, we explore the molecular basis for GusR-mediated regulation of GUS expression in response to small-molecule glucuronides. Presented are 2.1-Å-resolution crystal structures of GusRs from Escherichia coli and Salmonella enterica in complexes with a glucuronide ligand. The GusR-specific DNA operator site in the regulatory region of the E. coli GUS operon is identified, and structure-guided GusR mutants pinpoint the residues essential for DNA binding and glucuronide recognition. Interestingly, the endobiotic estradiol-17-glucuronide and the xenobiotic indomethacin-acyl-glucuronide are found to exhibit markedly differential binding to these GusR orthologs. Using structure-guided mutations, we are able to transfer E. coli GusR's preferential DNA and glucuronide binding affinity to S. enterica GusR. Structures of putative GusR orthologs from GUS-encoding Firmicutes species also reveal functionally unique features of the Enterobacteriaceae GusRs. Finally, dominant-negative GusR variants are validated in cell-based studies. These data provide a molecular framework toward understanding the control of glucuronide utilization by opportunistic pathogens in the human gut.

An Atlas of ß-Glucuronidases in the Human Intestinal Microbiome.

Microbiome-encoded ß-glucuronidase (GUS) enzymes play important roles in human health by metabolizing drugs in the gastrointestinal (GI) tract. The numbers, types, and diversity of these proteins in the human GI microbiome, however, remain undefined. We present an atlas of GUS enzymes comprehensive for the Human Microbiome Project GI database. We identify 3,013 total and 279 unique microbiome-encoded GUS proteins clustered into six unique structural categories. We assign their taxonomy, assess cellular localization, reveal the inter-individual variability within the 139 individuals sampled, and discover 112 novel microbial GUS enzymes. A representative in vitro panel of the most common GUS proteins by read abundances highlights structural and functional variabilities within the family, including their differential processing of smaller glucuronides and larger carbohydrates. These data provide a sequencing-to-molecular roadmap for examining microbiome-encoded enzymes essential to human health.

Glucuronides in the gut: Sugar-driven symbioses between microbe and host.

The intestinal milieu is astonishingly complex and home to a constantly changing mixture of small and large molecules, along with an abundance of bacteria, viral particles, and eukaryotic cells. Such complexity makes it difficult to develop testable molecular hypotheses regarding host-microbe interactions. Fortunately, mammals and their associated gastrointestinal (GI) microbes contain complementary systems that are ideally suited for mechanistic studies. Mammalian systems inactivate endobiotic and xenobiotic compounds by linking them to a glucuronic acid sugar for GI excretion. In the GI tract, the microbiota express ß-glucuronidase enzymes that remove the glucuronic acid as a carbon source, effectively reversing the actions of mammalian inactivation. Thus, by probing the actions of microbial ß-glucuronidases, and by understanding which substrate glucuronides they process, molecular insights into mammalian-microbial symbioses may be revealed amid the complexity of the intestinal tract. Here, we focus on glucuronides in the gut and the microbial proteins that process them.

Structure and Inhibition of Microbiome ß-Glucuronidases Essential to the Alleviation of Cancer Drug Toxicity.

The selective inhibition of bacterial ß-glucuronidases was recently shown to alleviate drug-induced gastrointestinal toxicity in mice, including the damage caused by the widely used anticancer drug irinotecan. Here, we report crystal structures of representative ß-glucuronidases from the Firmicutes Streptococcus agalactiae and Clostridium perfringens and the Proteobacterium Escherichia coli, and the characterization of a ß-glucuronidase from the Bacteroidetes Bacteroides fragilis. While largely similar in structure, these enzymes exhibit marked differences in catalytic properties and propensities for inhibition, indicating that the microbiome maintains functional diversity in orthologous enzymes. Small changes in the structure of designed inhibitors can induce significant conformational changes in the ß-glucuronidase active site. Finally, we establish that ß-glucuronidase inhibition does not alter the serum pharmacokinetics of irinotecan or its metabolites in mice. Together, the data presented advance our in vitro and in vivo understanding of the microbial ß-glucuronidases, a promising new set of targets for controlling drug-induced gastrointestinal toxicity.

Combining content and elements of communication into an upper-level biochemistry course.

This report describes how a science communication module was incorporated into an advanced biochemistry course. Elements of communication were taught synergistically with biochemistry content in this course in an effort to expose students to a variety of effective oral communication strategies. Students were trained to use these established techniques and incorporated them into various presentations throughout the course. Three students describe their use of specific resources and how the skills learned relate to their future career. The importance and relevance of science communication are receiving unprecedented national attention. The academic scientific community must respond by incorporating more communication-centered instruction and opportunities in the classroom and laboratory.

Examination of tyrosine/adenine stacking interactions in protein complexes.

The π-stacking interactions between tyrosine amino acid side chains and adenine-bearing ligands are examined. Crystalline protein structures from the protein data bank (PDB) exhibiting face-to-face tyrosine/adenine arrangements were used to construct 20 unique 4-methylphenol/N9-methyladenine (p-cresol/9MeA) model systems. Full geometry optimization of the 20 crystal structures with the M06-2X density functional theory method identified 11 unique low-energy conformations. CCSD(T) complete basis set (CBS) limit interaction energies were estimated for all of the structures to determine the magnitude of the interaction between the two ring systems. CCSD(T) computations with double-ζ basis sets (e.g., 6-31G*(0.25) and aug-cc-pVDZ) indicate that the MP2 method overbinds by as much as 3.07 kcal mol(-1) for the crystal structures and 3.90 kcal mol(-1) for the optimized structures. In the 20 crystal structures, the estimated CCSD(T) CBS limit interaction energy ranges from -4.00 to -6.83 kcal mol(-1), with an average interaction energy of -5.47 kcal mol(-1), values remarkably similar to the corresponding data for phenylalanine/adenine stacking interactions. Geometry optimization significantly increases the interaction energies of the p-cresol/9MeA model systems. The average estimated CCSD(T) CBS limit interaction energy of the 11 optimized structures is 3.23 kcal mol(-1) larger than that for the 20 crystal structures.