Search

Stabilizing proteins is a fundamental challenge in protein engineering and is almost always a prerequisite for the development of industrial and pharmaceutical biotechnologies. Here we present Stability Oracle: a structure-based graph-transformer framework that achieves state-of-the-art performance on predicting the effect of a point mutation on a protein’s thermodynamic stability (ΔΔG). A strength of our model is its ability to identifystabilizingmutations, which often make up a small fraction of a protein’s mutational landscape. Our framework introduces several data and machine learning innovations to overcome well-known challenges in data scarcity and bias, generalization, and computation time. Stability Oracle is first pretrained on over 2M masked microenvironments and then fine-tuned using a novel data augmentation technique, Thermodynamic Permutations (TP), applied to a ∼120K curated subset of the mega-scale cDNA display proteolysis dataset. This technique increases the original 120K mutations to over 2M thermodynamically valid ΔΔG measurements to generate the first structure training set that samples and balances all 380 mutation types. By using the masked microenvironment paradigm, Stability Oracle does not require a second mutant structure and instead uses amino acid structural embeddings to represent a mutation. This architectural design accelerates training and inference times: we can both train on 2M instances with just 119 structures and generate deep mutational scan (DMS) predictions from only the wildtype structure. We benchmark Stability Oracle with both experimental and AlphaFold structures of all proteins on T2837, a test set that aggregates the common test sets (SSym, S669, p53, and Myoglobin) with all additional experimental data from proteins with over a 30% sequence similarity overlap. We used TP augmented T2837 to evaluate performance for engineering protein stability: Stability Oracle correctly identifies 48% of stabilizing mutations (ΔΔG < −0.5 kcal/mol) and 74% of its stabilizing predictions are indeed stabilizing (18% and 8% of predictions were neutral and destabilizing, respectively). For a fair comparison between sequence and structure-based fine-tuned deep learning models, we build on the Prostata framework and fine-tune the sequence embeddings of ESM2 on our training set (Prostata-IFML). A head-to-head comparison demonstrates that Stability Oracle outperforms Prostata-IFML on regression and classification even though the model is 548 times smaller and is pretrained with 4000 times fewer proteins, highlighting the advantages of learning from structures.

Bioengineers increasingly rely on ligand-inducible transcription regulators for chemical-responsive control of gene expression, yet the number of regulators available is limited. Novel regulators can be mined from genomes, but an inadequate understanding of their DNA specificity complicates genetic design. Here we present Snowprint, a simple yet powerful bioinformatic tool for predicting regulator:DNA interactions. Benchmarking results demonstrate that Snowprint predictions are significantly similar for >45% of experimentally validated regulator:operator pairs from organisms across nine phyla and for regulators that span five distinct structural families. We then use Snowprint to design promoters for 33 previously uncharacterized regulators sourced from diverse phylogenies, of which 28 were shown to influence gene expression and 24 produced a >20-fold signal-to-noise ratio. A panel of the newly domesticated regulators were then screened for response to biomanufacturing-relevant compounds, yielding new sensors for a polyketide (olivetolic acid), terpene (geraniol), steroid (ursodiol), and alkaloid (tetrahydropapaverine) with induction ranges up to 10.7-fold. Snowprint represents a unique, generalizable tool that greatly facilitates the discovery of ligand-inducible transcriptional regulators for bioengineering applications.Abstract Figure

A major challenge to achieving industry-scale biomanufacturing of therapeutic alkaloids is the slow process of biocatalyst engineering. Amaryllidaceae alkaloids, such as the Alzheimer’s medication galantamine, are complex plant secondary metabolites with recognized therapeutic value. Due to their difficult synthesis they are regularly sourced by extraction and purification from low-yielding plants, including the wild daffodilNarcissus pseudonarcissus. Engineered biocatalytic methods have the potential to stabilize the supply chain of amaryllidaceae alkaloids. Here, we propose a highly efficient biosensor-AI technology stack for biocatalyst development, which we apply to engineer amaryllidaceae alkaloid production inEscherichia coli. Directed evolution is used to develop a highly sensitive (EC50= 20 uM) and specific biosensor for the key amaryllidaceae alkaloid branchpoint 4-O’Methylnorbelladine. A machine learning model (MutComputeX) was subsequently developed and used to generate activity-enriched variants of a plant methyltransferase, which were rapidly screened with the biosensor. Functional enzyme variants were identified that yielded a 60% improvement in product titer, 17-fold reduced remnant substrate, and 3-fold lower off-product regioisomer formation.

Deep learning models are seeing increased use as methods to predict mutational effects or allowed mutations at various sites in proteins. The models commonly used for these purposes include large language models (LLMs) and 3D Convolutional Neural Networks (CNNs). These two model types have very different architectures and are trained on different representations of proteins. LLMs make use of the transformer architecture and are trained purely on protein sequences whereas 3D CNNs are trained on voxelized representations of local protein structure. While comparable overall prediction accuracies have been reported for both types of models, it is not known to what extent these models make comparable specific predictions and/or generalize protein biochemistry in similar ways. Here, we perform a systematic comparison of two LLMs and one 3D CNN model and show that the different model types have distinct strengths and weaknesses. The overall prediction accuracies are largely uncorrelated between sequence and structure based models. Overall, the 3D CNN model is better at predicting buried aliphatic and hydrophobic residues whereas the LLMs are better at predicting solvent-exposed polar and charged amino acids. A combined model that takes the individual model predictions as input can leverage these individual model strengths and results in significantly improved overall prediction accuracy.

Biosensors can accelerate the engineering of new biosynthetic pathways. Phloroglucinol is a platform chemical of wide utility that can be produced at limited titers inEscherichia coli. Starting from the TetR family repressor RolR that is responsive to the related compound resorcinol, we were able to employ a combined selection and screen to identify variants that had greatly improved activities with phloroglucinol (EC50 for phloroglucinol of 131 uM, relative to an estimated 42 mM for wild-type RolR). The variants obtained were further screened with a panel of similar single ring aromatics, and several were found to be generalists, consistent with the hypothesis that both natural and directed evolution tend to first create semi-specific pockets prior to further optimization for new function

Yeast expression of human G Protein Coupled Receptors (GPCRs) can be used as a biosensor platform for the detection of pharmaceuticals. The Cannabinoid receptors type 1 and 2 (CB1/2R) are of particular interest, given the cornucopia of natural and synthetic cannabinoids being explored as therapeutics. We show for the first time that engineering the N-terminus of CB1R allows for efficient signal transduction in yeast, and that engineering the sterol composition of the yeast membrane optimizes CB2R performance. Using the dual cannabinoid biosensors, large libraries of synthetic cannabinoids and terpenes could be quickly screened to elucidate known and novel structure-activity relationships, including compounds and trends that more selectively target each of the two receptors. The biosensor strains offer a ready platform for evaluating the activity of new synthetic cannabinoids, monitoring drugs of abuse, and developing molecules that target the therapeutically important CB2R receptor while minimizing psychoactive effects.

We report two isothermal nucleic acid amplification assays for detection of monkeypox virus (MPXV) clades 2 and 3 that include the strains responsible for the current global outbreak of monkeypox. The assays use loop-mediated isothermal amplification (LAMP) to amplify two distinct sequences in the MPXV genome. Readout specificity is ensured by oligonucleotide strand displacement (OSD) probes integrated in one-pot LAMP-OSD reactions. OSD probes undergo toehold-mediated strand displacement hybridization to LAMP amplicon loop sequences derived from MPXV clades 2 and 3 resulting in fluorescence readable both in realtime and visually at endpoint. We also perform both assays on two different portable devices, the GeneTiger and the miniPCR, to exemplify compatibility with minimum infrastructure point-of-care (POC) testing in clinical and at-home settings. Both assays could readily detect single digit copies of MPXV synthetic double stranded DNA templates within 30 min.

SummaryThe worldwide spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has led to the repeated emergence of variants of concern. The Omicron variant has two dominant sub-lineages, BA.1 and BA.2, each with unprecedented numbers of nonsynonymous and indel spike protein mutations: 33 and 29, respectively. Some of these mutations individually increase transmissibility and enhance immune evasion, but their interactions within the Omicron mutational background is unknown. We characterize the molecular effects of all Omicron spike mutations on expression, human ACE2 receptor affinity, and neutralizing antibody recognition. We show that key mutations enable escape from neutralizing antibodies at a variety of epitopes. Stabilizing mutations in the N-terminal and S2 domains of the spike protein compensate for destabilizing mutations in the receptor binding domain, thereby enabling the record number of mutations in Omicron sub-lineages. Taken together, our results provide a comprehensive account of the mutational effects in the Omicron spike protein and illuminate previously unknown mechanisms of how the N-terminal domain can compensate for destabilizing mutations within the more evolutionarily constrained RBD.

Background The construction and application of synthetic genetic circuits is frequently improved if gene expression can be orthogonally controlled, relative to the host. In plants, orthogonality can be achieved via the use of CRISPR-based transcription factors that are programmed to act on natural or synthetic promoters. The construction of complex gene circuits can require multiple, orthogonal regulatory interactions, and this in turn requires that the full programmability of CRISPR elements be adapted to non-natural and non-standard promoters that have few constraints on their design. Therefore, we have developed synthetic promoter elements in which regions upstream of the minimal 35S CaMV promoter are designed from scratch to interact via programmed gRNAs with dCas9 fusions that allow activation of gene expression. Results A panel of three, mutually orthogonal promoters that can be acted on by artificial gRNAs bound by CRISPR regulators were designed. Guide RNA expression targeting these promoters was in turn controlled by either Pol III (U6) or ethylene-inducible Pol II promoters, implementing for the first time a fully artificial Orthogonal Control System (OCS). Following demonstration of the complete orthogonality of the designs, the OCS was tied to cellular metabolism by putting gRNA expression under the control of an endogenous plant signaling molecule, ethylene. The ability to form complex circuitry was demonstrated via the ethylene-driven, ratiometric expression of fluorescent proteins in single plants. Conclusions The design of synthetic promoters is highly generalizable to large tracts of sequence space, allowing Orthogonal Control Systems of increasing complexity to potentially be generated at will. The ability to tie in several different basal features of plant molecular biology (Pol II and Pol III promoters, ethylene regulation) to the OCS demonstrates multiple opportunities for engineering at the system level. Moreover, given the fungibility of the core 35S CaMV promoter elements, the derived synthetic promoters can potentially be utilized across a variety of plant species.

Aptamers have been the subject of more than 144,000 papers to date. However, there has been a growing concern that errors in reporting aptamer research limit the reliability of these reagents for research and other applications. These observations noting inconsistencies in the use of our RNA anti-lysozyme aptamer served as an impetus for our systematic review of the reporting of aptamer sequences in the literature. Our detailed examination of literature citing the RNA anti-lysozyme aptamer revealed that 93% of the 61 publications reviewed reported unexplained altered sequences with 86% of those using DNA variants. The ten most cited aptamers were examined using a standardized methodology in order to categorize the extent to which the sequences themselves were apparently improperly reproduced, both in the literature and presumably in experiments beyond their discovery. Our review of 800 aptamer publications spanned decades, multiple journals, and research groups, and revealed that 44% of the papers reported unexplained sequence alterations. We identified ten common categories of sequence alterations including deletions, substitutions, additions, among others. The robust data set we have produced elucidates a source of irreproducibility and unreliability in our field and can be used as a starting point for building evidence-based best practices in publication standards to elevate the rigor and reproducibility of aptamer research.

Plastic waste poses an ecological challenge1. While current plastic waste management largely relies on unsustainable, energy-intensive, or even hazardous physicochemical and mechanical processes, enzymatic degradation offers a green and sustainable route for plastic waste recycling2. Poly(ethylene terephthalate) (PET) has been extensively used in packaging and for the manufacture of fabrics and single-used containers, accounting for 12% of global solid waste3. The practical application of PET hydrolases has been hampered by their lack of robustness and the requirement for high processing temperatures. Here, we use a structure-based, deep learning algorithm to engineer an extremely robust and highly active PET hydrolase. Our best resulting mutant (FAST-PETase: Functional, Active, Stable, and Tolerant PETase) exhibits superior PET-hydrolytic activity relative to both wild-type and engineered alternatives, (including a leaf-branch compost cutinase and its mutant4) and possesses enhanced thermostability and pH tolerance. We demonstrate that whole, untreated, post-consumer PET from 51 different plastic products can all be completely degraded by FAST-PETase within one week, and in as little as 24 hours at 50 °C. Finally, we demonstrate two paths for closed-loop PET recycling and valorization. First, we re-synthesize virgin PET from the monomers recovered after enzymatic depolymerization. Second, we enable in situ microbially-enabled valorization using a Pseudomonas strain together with FAST-PETase to degrade PET and utilize the evolved monomers as a carbon source for growth and polyhydroxyalkanoate production. Collectively, our results demonstrate the substantial improvements enabled by deep learning and a viable route for enzymatic plastic recycling at the industrial scale.

InEscherichia coli, DNA replication termination is orchestrated by two clusters ofTersites forming a DNA replication fork trap when bound by Tus proteins. The formation of a ‘locked’ Tus-Tercomplex is essential for halting incoming DNA replication forks. However, the absence of replication fork arrest at someTersites raised questions about their significance. In this study, we examined the genome-wide distribution of Tus and found that only the six innermostTersites (TerA-EandG) were significantly bound by Tus. We also found that a single ectopic insertion ofTerBin its non-permissive orientation could not be achieved, advocating against a need for ‘back-up’Tersites. Finally, examination of the genomes of a variety of Enterobacterales revealed a new replication fork trap architecture exclusively found outside the Enterobacteriaceae family. Taken together, our data enabled the delineation of a narrow prototypical Tus-dependent DNA replication fork trap consisting of only twoTersites.Graphical Abstract

A key bottleneck in the microbial production of therapeutic plant metabolites is identifying enzymes that can greatly improve yield. The facile identification of genetically encoded biosensors can overcome this limitation and become part of a general method for engineering scaled production. We have developed a unique combined screening and selection approach that quickly refines the affinities and specificities of generalist transcription factors, and using RamR as a starting point we evolve highly specific (>100-fold preference) and sensitive (EC50

The yeast Saccharomyces cerevisiae is a powerful tool for studying G protein-coupled receptors (GPCRs) as they can be functionally coupled to its pheromone response pathway. However, some exogenous GPCRs, including the mu opioid receptor, are non-functional in yeast, which may be due to the presence of the fungal sterol ergosterol instead of the animal sterol cholesterol. We engineered yeast to produce cholesterol and introduced the human mu opioid receptor, creating an opioid biosensor capable of detecting the peptide DAMGO at an EC50 of 62 nM and the opiate morphine at an EC50 of 882 nM. Furthermore, introducing mu, delta, and kappa opioid receptors from diverse vertebrates consistently yielded active opioid biosensors that both recapitulated expected agonist binding profiles with EC50s as low as 2.5 nM and were inhibited by the antagonist naltrexone. Additionally, clinically relevant human mu opioid receptor alleles, or variants with terminal mutations, resulted in biosensors that largely displayed the expected changes in activity. We also tested mu opioid receptor-based biosensors with systematically adjusted biosynthetic intermediates of cholesterol, enabling us to relate sterol profiles with biosensor sensitivity. Finally, cholesterol-producing and sterol intermediate biosensor backgrounds were applied to other human GPCRs, resulting in SSTR5, 5-HTR4, FPR1 and NPY1R signaling with varying degrees of cholesterol dependence. Our sterol-optimized platform will be a valuable tool in generating human GPCR-based biosensors, aiding in ongoing receptor deorphanization efforts, and providing a framework for high-throughput screening of receptors and effectors.

DNA polymerase fromGeobacillus stearothermophilus, BstDNA polymerase (BstDNAP), is a versatile enzyme with robust strand-displacing activity that enables loop-mediated isothermal amplification (LAMP). Despite its exclusive usage in LAMP assay, its properties remain open to improvement. Here, we describe logical redesign ofBstDNAP by using multimodal application of several independent and orthogonal rational engineering methods such as domain addition, supercharging, and machine learning predictions of amino acid substitutions. The resulting Br512g3 enzyme is not only thermostable and extremely robust but it also displays improved reverse transcription activity and the ability to carry out ultrafast LAMP at 74 °C. Our study illustrates a new enzyme engineering strategy as well as contributes a novel engineered strand displacing DNA polymerase of high value to diagnostics and other fields.

The ongoing evolution of SARS-CoV-2 into more easily transmissible and infectious variants has sparked concern over the continued effectiveness of existing therapeutic antibodies and vaccines. Hence, together with increased genomic surveillance, methods to rapidly develop and assess effective interventions are critically needed. Here we report the discovery of SARS-CoV-2 neutralizing antibodies isolated from COVID-19 patients using a high-throughput platform. Antibodies were identified from unpaired donor B-cell and serum repertoires using yeast surface display, proteomics, and public light chain screening. Cryo-EM and functional characterization of the antibodies identified N3-1, an antibody that binds avidly (Kd,app = 68 pM) to the receptor binding domain (RBD) of the spike protein and robustly neutralizes the virus in vitro. This antibody likely binds all three RBDs of the trimeric spike protein with a single IgG. Importantly, N3-1 equivalently binds spike proteins from emerging SARS-CoV-2 variants of concern, neutralizes UK variant B.1.1.7, and binds SARS-CoV spike with nanomolar affinity. Taken together, the strategies described herein will prove broadly applicable in interrogating adaptive immunity and developing rapid response biological countermeasures to emerging pathogens.

Despite the fact that strand-displacing activity is of great utility for a variety of applications, including isothermal amplification assays, there are relatively few strand-displacing DNA polymerases. In particular, the thermotolerant DNA polymerase from Geobacillus stearothermophilus (previously Bacillus stearothermophilus), Bst DNA polymerase (Bst DNAP), is used in a variety of assays, including loop-mediated isothermal amplification. However, despite its wide use, its properties remain open to improvement, as has been demonstrated by a variety of engineering efforts, including the identification of point mutations that impact its robustness, strand-displacement capabilities, and nascent reverse transcriptase activity.Interestingly, a strategy that has been commonly used to alter the capabilities of DNA polymerases, the addition of additional DNA- or RNA-binding domains, has yet to be applied to Bst DNAP. To this end, we now show that by adding fusion domains the performance of Bst DNAP in isothermal amplification assays, including its nascent RT activity, can be greatly improved. The impact of these improvements on the development of LAMP assays for the detection of SARS-CoV-2 is fully explored.

Since the emergence of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, there have been demands on the testing infrastructure that have strained testing capacity. As a simplification of method, we confirm the efficacy of RNA extraction-free RT-qPCR and saline as an alternative patient sample storage buffer. In addition, amongst potential reagent shortages, it has sometimes been difficult to obtain inactivated viral particles. We have therefore also characterized armored SARS-CoV-2 RNA from Asuragen as an alternative diagnostic standard to ATCC genomic SARS-CoV-2 RNA and heat inactivated virions and provide guidelines for its use in RT-qPCR.

Taq DNA polymerase, one of the first thermostable DNA polymerases to be discovered, has been typecast as a DNA-dependent DNA polymerase commonly employed for PCR. However, Taq polymerase belongs to the same DNA polymerase superfamily as the Molony murine leukemia virus reverse transcriptase and has in the past been shown to possess reverse transcriptase activity. We report optimized buffer and salt compositions that promote the reverse transcriptase activity of Taq DNA polymerase, and thereby allow it to be used as the sole enzyme in TaqMan RT-qPCR reactions. We demonstrate the utility of Taq-alone RT-qPCR reactions by executing CDC SARS-CoV-2 N1, N2, and N3 TaqMan RT-qPCR assays that could detect as few as 2 copies/µL of input viral genomic RNA.

Isothermal nucleic acid amplification tests (iNAT), such as loop-mediated isothermal amplification (LAMP), are good alternatives to polymerase chain reaction (PCR)-based amplification assays, especially for point-of-care and low resource use, in part because they can be carried out with relatively simple instrumentation. However, iNATs can generate spurious amplicons, especially in the absence of target sequences, resulting in false positive results. This is especially true if signals are based on non-sequence-specific probes, such as intercalating dyes or pH changes. In addition, pathogens often prove to be moving, evolving targets, and can accumulate mutations that will lead to inefficient primer binding and thus false negative results. Internally redundant assays targeting different regions of the target sequence can help to reduce such false negatives. Here we describe rapid conversion of three previously described SARS-CoV-2 LAMP assays that relied on non-sequence-specific readout into assays that can be visually read using sequence-specific fluorogenic oligonucleotide strand exchange (OSD) probes. We evaluate one-pot operation of both individual and multiplex LAMP-OSD assays and demonstrate detection of SARS-CoV-2 virions in crude human saliva.

Given the scale of the ongoing COVID-19 pandemic, the need for reliable, scalable testing, and the likelihood of reagent shortages, especially in resource-poor settings, we have developed a RT-qPCR assay that relies on an alternative to conventional viral reverse transcriptases, a thermostable reverse transcriptase / DNA polymerase (RTX)1. Here we show that RTX performs comparably to the other assays sanctioned by the CDC and validated in kit format. We demonstrate two modes of RTX use – (i) dye-based RT-qPCR assays that require only RTX polymerase, and (ii) TaqMan RT-qPCR assays that use a combination of RTX and Taq DNA polymerases (as the RTX exonuclease does not degrade a TaqMan probe). We also provide straightforward recipes for the purification of this alternative reagent. We anticipate that in low resource or point-of-need settings researchers could obtain the available constructs from Addgene or our lab and begin to develop their own assays, within whatever regulatory framework exists for them.We lay out protocols for dye-based and TaqMan probe-based assays, in order to best compare with ‘gold standard’ reagents. These protocols should form the basis of further modifications that can simplify the assay to the use of overexpressing cells themselves as reagents.Developing dye-based and TaqMan probe-based RT-qPCR assays with RTX

Natural selection acting on synonymous mutations in protein-coding genes influences genome composition and evolution. In viruses, introducing synonymous mutations in genes encoding structural proteins can drastically reduce viral growth, providing a means to generate potent, live attenuated vaccine candidates. However, an improved understanding of what compositional features are under selection and how combinations of synonymous mutations affect viral growth is needed to predictably attenuate viruses and make them resistant to reversion. We systematically recoded all non-overlapping genes of the bacteriophage ΦX174 with codons rarely used in itsE. colihost. The fitness of recombinant viruses decreases as additional deoptimizing mutations are made to the genome, although not always linearly, and not consistently across genes. Combining deoptimizing mutations may reduce viral fitness more or less than expected from the effect size of the constituent mutations and we point out difficulties in untangling correlated compositional features. We test our model by optimizing the same genes and find that the relationship between codon usage and fitness does not hold for optimization, suggesting that wild-type ΦX174 is at a fitness optimum. This work highlights the need to better understand how selection acts on patterns of synonymous codon usage across the genome and provides a convenient system to investigate the genetic determinants of virulence.

While deep learning methods exist to guide protein optimization, examples of novel proteins generated with these techniques require a priori mutational data. Here we report a 3D convolutional neural network that associates amino acids with neighboring chemical microenvironments at state-of-the-art accuracy. This algorithm enables identification of novel gain-of-function mutations, and subsequent experiments confirm substantive phenotypic improvements in stability-associated phenotypes in vivo across three diverse proteins.

We extend the concept of DNA proximity ligation from a single readout per oligonucleotide pair to multiple reversible, iterative ligations re-using the same oligonucleotide molecules. Using iterative proximity ligation (IPL), we can in principle capture multiple ligation events between each oligonucleotide and its various neighbors and thus recover a far richer knowledge about their relative positions than single, irreversible ligation events. IPL would thus act to sample and record local molecular neighborhoods. By integrating a unique DNA barcode into each participating oligonucleotide, we can catalog the individual ligation events and thus capture the positional information contained therein in a high throughput manner using next-generation DNA sequencing. We propose that by interpreting IPL sequencing results in the context of graph theory and by applying spring layout algorithms, we can recover geometric patterns of objects labeled by DNA. Using simulations, we demonstrate that we can in principle recover letter patterns photolithographed onto slide surfaces using only IPL sequencing data, illustrating how our technique maps complex spatial configurations into DNA sequences and then – using only this sequence information – recovers them. We complement our theoretical work with an experimental proof-of-concept of iterative proximity ligation on an oligonucleotide population.

We have developed a generalizable ‘smart molecular diagnostic’ capable of accurate point-of-care (POC) detection of variable nucleic acid targets. Our one-pot isothermal assay relies on multiplex execution of four loop-mediated isothermal amplification reactions, with primers that are degenerate and redundant, thereby increasing the breadth of targets while reducing the probability of amplification failure. An easy-to-read visual answer is computed directly by a multi-input Boolean OR gate signal transducer that uses degenerate strand exchange probes to assess any combination of amplicons. We demonstrate our platform by using the same assay to detect divergent Asian and African lineages of the evolving Zika virus (ZIKV), while maintaining selectivity against non-target viruses. Direct analysis of biological specimens proved possible, with 20 virions / µl being directly detected in human saliva within 90 minutes, and crudely macerated ZIKV-infected Aedes aegypti mosquitoes being identified with 100% specificity and sensitivity. The ease-of-use with minimal instrumentation, broad programmability, and built-in fail-safe reliability make our smart molecular diagnostic attractive for POC use.

There are few methods for the assembly of defined protein oligomers and higher order structures that could serve as novel biomaterials. Using fluorescent proteins as a model system, we have engineered novel oligomerization states by combining oppositely supercharged variants. A well-defined, highly symmetrical 16-mer (two stacked, circular octamers) can be formed from alternating charged proteins; higher order structures then form in a hierarchical fashion from this discrete protomer. During SUpercharged PRotein Assembly (SuPrA), electrostatic attraction between oppositely charged variants drives interaction, while shape and patchy physicochemical interactions lead to spatial organization along specific interfaces, ultimately resulting in protein assemblies never before seen in nature.

A number of proteins containing arginine-rich motifs (ARMs) are known to bind RNA and are involved in regulating RNA processing in viruses and cells. Using automated selection methods we have generated a number of aptamers against ARM peptides from various natural proteins. Aptamers bind tightly to their cognate ARMs, with Kd values in the nanomolar range, and frequently show no propensity to bind to other ARMs or even to single amino acid variants of the cognate ARM. However, at least some anti-ARM aptamers can cross-recognize a limited set of other ARMs, just as natural RNA-binding sites have been shown to exhibit so-called “chameleonism.” We expand upon the number of examples of cross-recognition and, using mutational and circular dichroism (CD) analyses, demonstrate that there are multiple mechanisms by which RNA ligands can cross-recognize ARMs. These studies support a model in which individual arginine residues govern binding to an RNA ligand, and the inherent flexibility of the peptide backbone may make it possible for “semi-specific” recognition of a discrete set of RNAs by a discrete set of ARM peptides and proteins.

We have used a compartmentalized in vitro selection method to directly select for ligase ribozymes that are capable of acting on and turning over separable oligonucleotide substrates. Starting from a degenerate pool, we selected a trans-acting variant of the Bartel class I ligase which statistically may have been the only active variant in the starting pool. The isolation of this sequence from the population suggests that this selection method is extremely robust at selecting optimal ribozymes and should, therefore, prove useful for the selection and optimization of other trans-acting nucleic acid catalysts capable of multiple turnover catalysis.

A peptide-dependent ribozyme ligase (aptazyme ligase) has been selected from a random sequence population based on the small L1 ligase. The aptazyme ligase is activated > 18,000-fold by its cognate peptide effector, the HIV-1 Rev arginine-rich motif (ARM), and specifically recognizes the Rev ARM relative to other peptides containing arginine-rich motifs. Moreover, the aptazyme ligase can preferentially recognize the Rev ARM in the context of the full-length HIV-1 Rev protein. The only cross-reactivity exhibited by the aptazyme is toward the Tat ARM. Reselection of peptide- and protein-dependent aptazymes from a partially randomized population yielded aptazymes that could readily discriminate against the Tat ARM. These results have important implications for the development of aptazymes that can be used in arrays for the detection and quantitation of multiple cellular proteins (proteome arrays).

A small ribozyme ligase (L1) selected from a random sequence population appears to utilize non-Watson-Crick base pairs at its ligation junction. Mutational and selection analyses confirmed the presence of these base pairings. Randomization of the L1 core and selection of active ligases yielded highly active variants whose rates were on the order of 1 min(-1). Base-pairing covariations confirmed the general secondary structure of the ligase, and the most active ligases contained a novel pentuple sequence covariation. The optimized L1 ligases may be optimal within their sequence spaces, and minimal ligases that span less than 60 nt in length have been engineered based on these results.