Discovery and Characterization of Two Folded Intermediates for Outer Membrane Protein TolC Biogenesis.

TolC is the outer membrane protein responsible for antibiotic efflux in E. coli. Compared to other outer membrane proteins it has an unusual fold and has been shown to fold independently of commonly used periplasmic chaperones, SurA and Skp. Here we find that the assembly of TolC involves the formation of two folded intermediates using circular dichroism, gel electrophoresis, site-specific disulfide bond formation and radioactive labeling. First the TolC monomer folds, and then TolC assembles into a trimer both in detergent-free buffer and in the presence of detergent micelles. We find that a TolC trimer also forms in the periplasm and is present in the periplasm before it inserts in the outer membrane. The monomeric and trimeric folding intermediates may be used in the future to develop a new approach to antibiotic efflux pump inhibition by targeting the assembly pathway of TolC.

The energetics and evolution of oxidoreductases in deep time.

The core metabolic reactions of life drive electrons through a class of redox protein enzymes, the oxidoreductases. The energetics of electron flow is determined by the redox potentials of organic and inorganic cofactors as tuned by the protein environment. Understanding how protein structure affects oxidation-reduction energetics is crucial for studying metabolism, creating bioelectronic systems, and tracing the history of biological energy utilization on Earth. We constructed ProtReDox (https://protein-redox-potential.web.app), a manually curated database of experimentally determined redox potentials. With over 500 measurements, we can begin to identify how proteins modulate oxidation-reduction energetics across the tree of life. By mapping redox potentials onto networks of oxidoreductase fold evolution, we can infer the evolution of electron transfer energetics over deep time. ProtReDox is designed to include user-contributed submissions with the intention of making it a valuable resource for researchers in this field.

General features of transmembrane beta barrels from a large database.

Large datasets contribute new insights to subjects formerly investigated by exemplars. We used coevolution data to create a large, high-quality database of transmembrane ß-barrels (TMBB). By applying simple feature detection on generated evolutionary contact maps, our method (IsItABarrel) achieves 95.88% balanced accuracy when discriminating among protein classes. Moreover, comparison with IsItABarrel revealed a high rate of false positives in previous TMBB algorithms. In addition to being more accurate than previous datasets, our database (available online) contains 1,938,936 bacterial TMBB proteins from 38 phyla, respectively, 17 and 2.2 times larger than the previous sets TMBB-DB and OMPdb. We anticipate that due to its quality and size, the database will serve as a useful resource where high-quality TMBB sequence data are required. We found that TMBBs can be divided into 11 types, three of which have not been previously reported. We find tremendous variance in proteome percentage among TMBB-containing organisms with some using 6.79% of their proteome for TMBBs and others using as little as 0.27% of their proteome. The distribution of the lengths of the TMBBs is suggestive of previously hypothesized duplication events. In addition, we find that the C-terminal ß-signal varies among different classes of bacteria though its consensus sequence is LGLGYRF. However, this ß-signal is only characteristic of prototypical TMBBs. The ten non-prototypical barrel types have other C-terminal motifs, and it remains to be determined if these alternative motifs facilitate TMBB insertion or perform any other signaling function.

KIF-Key Interactions Finder: A program to identify the key molecular interactions that regulate protein conformational changes.

Simulation datasets of proteins (e.g., those generated by molecular dynamics simulations) are filled with information about how a non-covalent interaction network within a protein regulates the conformation and, thus, function of the said protein. Most proteins contain thousands of non-covalent interactions, with most of these being largely irrelevant to any single conformational change. The ability to automatically process any protein simulation dataset to identify non-covalent interactions that are strongly associated with a single, defined conformational change would be a highly valuable tool for the community. Furthermore, the insights generated from this tool could be applied to basic research, in order to improve understanding of a mechanism of action, or for protein engineering, to identify candidate mutations to improve/alter the functionality of any given protein. The open-source Python package Key Interactions Finder (KIF) enables users to identify those non-covalent interactions that are strongly associated with any conformational change of interest for any protein simulated. KIF gives the user full control to define the conformational change of interest as either a continuous variable or categorical variable, and methods from statistics or machine learning can be applied to identify and rank the interactions and residues distributed throughout the protein, which are relevant to the conformational change. Finally, KIF has been applied to three diverse model systems (protein tyrosine phosphatase 1B, the PDZ3 domain, and the KE07 series of Kemp eliminases) in order to illustrate its power to identify key features that regulate functionally important conformational dynamics.

MAHOMES II: A webserver for predicting if a metal binding site is enzymatic.

Recent advances have enabled high-quality computationally generated structures for proteins with no solved crystal structures. However, protein function data remains largely limited to experimental methods and homology mapping. Since structure determines function, it is natural that methods capable of using computationally generated structures for functional annotations need to be advanced. Our laboratory recently developed a method to distinguish between metalloenzyme and non-enzyme sites. Here we report improvements to this method by upgrading our physicochemical features to alleviate the need for structures with sub-angstrom precision and using machine learning to reduce training data labeling error. Our improved classifier identifies protein bound metal sites as enzymatic or non-enzymatic with 94% precision and 92% recall. We demonstrate that both adjustments increased predictive performance and reliability on sites with sub-angstrom variations. We constructed a set of predicted metalloprotein structures with no solved crystal structures and no detectable homology to our training data. Our model had an accuracy of 90 - 97.5% depending on the quality of the predicted structures included in our test. Finally, we found the physicochemical trends that drove this model's successful performance were local protein density, second shell ionizable residue burial, and the pocket's accessibility to the site. We anticipate that our model's ability to correctly identify catalytic metal sites could enable identification of new enzymatic mechanisms and improve de novo metalloenzyme design success rates. Significance statement: Identification of enzyme active sites on proteins with unsolved crystallographic structures can accelerate discovery of novel biochemical reactions, which can impact healthcare, industrial processes, and environmental remediation. Our lab has developed an ML tool for predicting sites on computationally generated protein structures as enzymatic and non-enzymatic. We have made our tool available on a webserver, allowing the scientific community to rapidly search previously unknown protein function space.

MAHOMES II: A webserver for predicting if a metal binding site is enzymatic.

Recent advances have enabled high-quality computationally generated structures for proteins with no solved crystal structures. However, protein function data remains largely limited to experimental methods and homology mapping. Since structure determines function, it is natural that methods capable of using computationally generated structures for functional annotations need to be advanced. Our laboratory recently developed a method to distinguish between metalloenzyme and nonenzyme sites. Here we report improvements to this method by upgrading our physicochemical features to alleviate the need for structures with sub-angstrom precision and using machine learning to reduce training data labeling error. Our improved classifier identifies protein bound metal sites as enzymatic or nonenzymatic with 94% precision and 92% recall. We demonstrate that both adjustments increased predictive performance and reliability on sites with sub-angstrom variations. We constructed a set of predicted metalloprotein structures with no solved crystal structures and no detectable homology to our training data. Our model had an accuracy of 90%-97.5% depending on the quality of the predicted structures included in our test. Finally, we found the physicochemical trends that drove this model's successful performance were local protein density, second shell ionizable residue burial, and the pocket's accessibility to the site. We anticipate that our model's ability to correctly identify catalytic metal sites could enable identification of new enzymatic mechanisms and improve de novo metalloenzyme design success rates.

Detergent headgroups control TolC folding in vitro.

TolC is the trimeric outer membrane component of the efflux pump system in Escherichia coli that is responsible for antibiotic efflux from bacterial cells. Overexpression of efflux pumps has been reported to decrease susceptibility to antibiotics in a variety of bacterial pathogens. Reliable production of membrane proteins allows for the biophysical and structural characterization needed to better understand efflux and for the development of therapeutics. Preparation of recombinant protein for biochemical/structural studies often involves the production of proteins as inclusion body aggregates from which active proteins are recovered. Here, we find that the in vitro folding of TolC into its functional trimeric state from inclusion bodies is dependent on the headgroup composition of detergent micelles used. Nonionic detergent favors the formation of functional trimeric TolC, whereas zwitterionic detergents induce the formation of a non-native, oligomeric TolC fold. We also find that nonionic detergents with shorter alkyl lengths facilitate TolC folding. It remains to be seen whether the charges in lipid headgroups have similar effects on membrane insertion and folding in biological systems.

Colicin E1 opens its hinge to plug TolC.

The double membrane architecture of Gram-negative bacteria forms a barrier that is impermeable to most extracellular threats. Bacteriocin proteins evolved to exploit the accessible, surface-exposed proteins embedded in the outer membrane to deliver cytotoxic cargo. Colicin E1 is a bacteriocin produced by, and lethal to, Escherichia coli that hijacks the outer membrane proteins (OMPs) TolC and BtuB to enter the cell. Here, we capture the colicin E1 translocation domain inside its membrane receptor, TolC, by high-resolution cryo-electron microscopy to obtain the first reported structure of a bacteriocin bound to TolC. Colicin E1 binds stably to TolC as an open hinge through the TolC pore-an architectural rearrangement from colicin E1's unbound conformation. This binding is stable in live E. coli cells as indicated by single-molecule fluorescence microscopy. Finally, colicin E1 fragments binding to TolC plug the channel, inhibiting its native efflux function as an antibiotic efflux pump, and heightening susceptibility to three antibiotic classes. In addition to demonstrating that these protein fragments are useful starting points for developing novel antibiotic potentiators, this method could be expanded to other colicins to inhibit other OMP functions.

Bacteria are constantly warring with each other for space and resources. As a result, they have developed a range of molecular weapons to poison, damage or disable other cells. For instance, bacteriocins are proteins that can latch onto structures at the surface of enemy bacteria and push toxins through their outer membrane. Bacteria are increasingly resistant to antibiotics, representing a growing concern for modern healthcare. One way that they are able to survive is by using 'efflux pumps' studded through their external membranes to expel harmful drugs before these can cause damage. Budiardjo et al. wanted to test whether bacteriocins could interfere with this defence mechanism by blocking efflux pumps. Bacteriocins are usually formed of binding elements (which recognise specific target proteins) and of a 'killer tail' that can stab the cell. Experiments showed that the binding parts of a bacteriocin could effectively 'plug' efflux pumps in Escherichia coli bacteria: high-resolution molecular microscopy revealed how the bacteriocin fragment binds to the pump, while fluorescent markers showed that it attached to the surface of E. coli and stopped the efflux pumps from working. As a result, lower amounts of antibiotics were necessary to kill the bacteria when bacteriocins were present. The work by Budiardjo et al. could lead to new ways to combat bacteria that will reduce the need for current antibiotics. In the future, bacteriocins could also be harnessed to target other proteins than efflux pumps, allowing scientists to manipulate a range of bacterial processes.

Machine learning for enzyme engineering, selection and design.

Machine learning is a useful computational tool for large and complex tasks such as those in the field of enzyme engineering, selection and design. In this review, we examine enzyme-related applications of machine learning. We start by comparing tools that can identify the function of an enzyme and the site responsible for that function. Then we detail methods for optimizing important experimental properties, such as the enzyme environment and enzyme reactants. We describe recent advances in enzyme systems design and enzyme design itself. Throughout we compare and contrast the data and algorithms used for these tasks to illustrate how the algorithms and data can be best used by future designers.

Machine learning differentiates enzymatic and non-enzymatic metals in proteins.

Metalloenzymes are 40% of all enzymes and can perform all seven classes of enzyme reactions. Because of the physicochemical similarities between the active sites of metalloenzymes and inactive metal binding sites, it is challenging to differentiate between them. Yet distinguishing these two classes is critical for the identification of both native and designed enzymes. Because of similarities between catalytic and non-catalytic  metal binding sites, finding physicochemical features that distinguish these two types of metal sites can indicate aspects that are critical to enzyme function. In this work, we develop the largest structural dataset of enzymatic and non-enzymatic metalloprotein sites to date. We then use a decision-tree ensemble machine learning model to classify metals bound to proteins as enzymatic or non-enzymatic with 92.2% precision and 90.1% recall. Our model scores electrostatic and pocket lining features as more important than pocket volume, despite the fact that volume is the most quantitatively different feature between enzyme and non-enzymatic sites. Finally, we find our model has overall better performance in a side-to-side comparison against other methods that differentiate enzymatic from non-enzymatic sequences. We anticipate that our model's ability to correctly identify which metal sites are responsible for enzymatic activity could enable identification of new enzymatic mechanisms and de novo enzyme design.

Membrane Barrels Are Taller, Fatter, Inside-Out Soluble Barrels.

Up-and-down ß-barrel topology exists in both the membrane and soluble environment. By comparing features of these structurally similar proteins, we can determine what features are particular to the environment rather than the fold. Here we compare structures of membrane ß-barrels to soluble ß-barrels and evaluate their relative size, shape, amino acid composition, hydrophobicity, and periodicity. We find that membrane ß-barrels are generally larger than soluble ß-barrels, with more strands per barrel and more amino acids per strand, making them wider and taller. We also find that membrane ß-barrels are inside-out soluble ß-barrels. The inward region of membrane ß-barrels has similar hydrophobicity to the outward region of soluble ß-barrels, and the outward region of membrane ß-barrels has similar hydrophobicity to the inward region of the soluble ß-barrels. Moreover, even though both types of ß-barrel have been assumed to have strands with amino acids that alternate in direction and hydrophobicity, we find that the membrane ß-barrels have more regular alternation than soluble ß-barrels. These features give insight into how membrane barrels maintain their fold and function in the membrane.

High-Yield Preparation of Outer Membrane Protein Efflux Pumps by in Vitro Refolding is Concentration Dependent.

Overexpression of tripartite efflux pump systems in gram-negative bacteria is a principal component of antibiotic resistance. High-yield purification of the outer membrane component of these systems will enable biochemical and structural interrogation of their mechanisms of action and allow testing of compounds that target them. However, preparation of these proteins is typically hampered by low yields, requiring laborious large-scale efforts. If refolding conditions can be found, refolding these proteins from inclusion bodies can lead to increased yields as compared to membrane isolations. A classical method for refolding outer membrane proteins involves unfolding inclusion bodies in urea followed by refolding in lipid or detergent micelles. However, that method has not yet been successful in refolding tripartite efflux pump TolC. Here, we find that refolding TolC from inclusion bodies requires an additional oligomerization enhancing step of sample concentration. We show that by our method of refolding, homotrimeric TolC remains folded in SDS-PAGE, retains binding to an endogenous ligand, and recapitulates the known crystal structure by single particle cryoEM analysis. We find that TolC refolding is concentration dependent. We then extended our method to refolding CmeC, a homologous protein from Campylobacter jejuni, and find that concentration-dependent oligomerization is a general feature of these systems. Because outer membrane efflux pump components are ubiquitous across gram-negative species, we anticipate that incorporating a concentration step in refolding protocols will promote correct refolding allowing for reliable, high-yield preparation of this family of proteins.

Modulating Integrin αIIbß3 Activity through Mutagenesis of Allosterically Regulated Intersubunit Contacts.

Integrin αIIbß3, a transmembrane heterodimer, mediates platelet aggregation when it switches from an inactive to an active ligand-binding conformation following platelet stimulation. Central to regulating αIIbß3 activity is the interaction between the αIIb and ß3 extracellular stalks, which form a tight heterodimer in the inactive state and dissociate in the active state. Here, we demonstrate that alanine replacements of sensitive positions in the heterodimer stalk interface destabilize the inactive conformation sufficiently to cause constitutive αIIbß3 activation. To determine the structural basis for this effect, we performed a structural bioinformatics analysis and found that perturbing intersubunit contacts with favorable interaction geometry through substitutions to alanine quantitatively accounted for the degree of constitutive αIIbß3 activation. This mutational study directly assesses the relationship between favorable interaction geometry at mutation-sensitive positions and the functional activity of those mutants, giving rise to a simple model that highlights the importance of interaction geometry in contributing to the stability between protein-protein interactions.

Efflux Pumps Represent Possible Evolutionary Convergence onto the ß-Barrel Fold.

There are around 100 varieties of outer membrane proteins in each Gram-negative bacteria. All of these proteins have the same fold-an up-down ß-barrel. It has been suggested that all membrane ß-barrels excluding lysins are homologous. Here we suggest that ß-barrels of efflux pumps have converged on this fold as well. By grouping structurally solved outer membrane ß-barrels (OMBBs) by sequence we find that the membrane environment may have led to convergent evolution of the barrel fold. Specifically, the lack of sequence linkage to other barrels coupled with distinctive structural differences, such as differences in strand tilt and barrel radius, suggest that the outer membrane factor of efflux pumps evolutionarily converged on the barrel. Rather than being related to other OMBBs, sequence and structural similarity in the periplasmic region of the outer membrane factor of efflux pumps suggests an evolutionary link to the periplasmic subunit of the same pump complex.

Tight Turns of Outer Membrane Proteins: An Analysis of Sequence, Structure, and Hydrogen Bonding.

As a structural class, tight turns can control molecular recognition, enzymatic activity, and nucleation of folding. They have been extensively characterized in soluble proteins but have not been characterized in outer membrane proteins (OMPs), where they also support critical functions. We clustered the 4 to 6 residue tight turns of 110 OMPs to characterize the phi/psi angles, sequence, and hydrogen bonding of these structures. We find significant differences between reports of soluble protein tight turns and OMP tight turns. Since OMP strands are less twisted than soluble strands, they favor different turn structures types. Moreover, the membrane localization of OMPs yields different sequence hallmarks for their tight turns relative to soluble protein turns. We also characterize the differences in phi/psi angles, sequence, and hydrogen bonding between OMP extracellular loops and OMP periplasmic turns. As previously noted, the extracellular loops tend to be much longer than the periplasmic turns. We find that this difference in length is due to the broader distribution of lengths of the extracellular loops not a large difference in the median length. Extracellular loops also tend to have more charged residues as predicted by the charge-out rule. Finally, in all OMP tight turns, hydrogen bonding between the side chain and backbone 2 to 4 residues away from that side chain plays an important role. These bonds preferentially use an Asp, Asn, Ser, or Thr residue in a beta or pro phi/psi conformation. We anticipate that this study will be applicable to future design and structure prediction of OMPs.

Why have small multidrug resistance proteins not evolved into fused, internally duplicated structures?

The increasing number of solved membrane protein structures has led to the recognition of a common feature in a large fraction of the small-molecule transporters: inverted repeat structures, formed by two fused homologous membrane domains with opposite orientation in the membrane. An evolutionary pathway in which the ancestral state is a single gene encoding a dual-topology membrane protein capable of forming antiparallel homodimers has been posited. A gene duplication event enables the evolution of two oppositely orientated proteins that form antiparallel heterodimers. Finally, fusion of the two genes generates an internally duplicated transporter with two oppositely orientated membrane domains. Strikingly, however, in the small multidrug resistance (SMR) family of transporters, no fused, internally duplicated proteins have been found to date. Here, we have analyzed fused versions of the dual-topology transporter EmrE, a member of the SMR family, by blue-native PAGE and in vivo activity measurements. We find that fused constructs give rise to both intramolecular inverted repeat structures and competing intermolecular dimers of varying activity. The formation of several intramolecularly and intermolecularly paired species indicates that a gene fusion event may lower the overall amount of active protein, possibly explaining the apparent absence of fused SMR proteins in nature.

In vivo trp scanning of the small multidrug resistance protein EmrE confirms 3D structure models'.

The quaternary structure of the homodimeric small multidrug resistance protein EmrE has been studied intensely over the past decade. Structural models derived from both two- and three-dimensional crystals show EmrE as an anti-parallel homodimer. However, the resolution of the structures is rather low and their relevance for the in vivo situation has been questioned. Here, we have challenged the available structural models by a comprehensive in vivo Trp scanning of all four transmembrane helices in EmrE. The results are in close agreement with the degree of lipid exposure of individual residues predicted from coarse-grained molecular dynamics simulations of the anti-parallel dimeric structure obtained by X-ray crystallography, strongly suggesting that the X-ray structure provides a good representation of the active in vivo form of EmrE.

Charge asymmetry in the proteins of the outer membrane.

MOTIVATION: Outer membrane beta-barrels (OMBBs) are the proteins found in the outer membrane of bacteria, mitochondria and chloroplasts. There are thousands of beta-barrels reported in genomic databases with ∼2-3% of the genes in gram-negative bacteria encoding these proteins. These proteins have a wide variety of biological functions including active and passive transport, cell adhesion, catalysis and structural anchoring. Of the non-redundant OMBB structures in the Protein Data Bank, half have been solved during the past 5 years. This influx of information provides new opportunities for understanding the chemistry of these proteins. The distribution of charges in proteins in the outer membrane has implications for how the mechanism of outer membrane protein insertion is understood. Understanding the distribution of charges might also assist in organism selection for the heterologous expression of mitochondrial OMBBs. RESULTS: We find a strong asymmetry in the charge distribution of these proteins. For the outward-facing residues of the beta-barrel within regions of similar amino acid density for both membrane leaflets, the external side of the outer membrane contains almost three times the number of charged residues as the internal side of the outer membrane. Moreover, the lipid bilayer of the outer membrane is asymmetric, and the overall preference for amino acid types to be in the external leaflet of the membrane correlates roughly with the hydrophobicity of the membrane lipids. This preference is demonstrably related to the difference in lipid composition of the external and internal leaflets of the membrane.