Tapasin-mediated editing of the MHC I immunopeptidome is epitope specific and dependent on peptide off-rate, abundance, and level of tapasin expression.

Tapasin, a component of the major histocompatibility complex (MHC) I peptide loading complex, edits the repertoire of peptides that is presented at the cell surface by MHC I and thereby plays a key role in shaping the hierarchy of CD8+ T-cell responses to tumors and pathogens. We have developed a system that allows us to tune the level of tapasin expression and independently regulate the expression of competing peptides of different off-rates. By quantifying the relative surface expression of peptides presented by MHC I molecules, we show that peptide editing by tapasin can be measured in terms of "tapasin bonus," which is dependent on both peptide kinetic stability (off-rate) and peptide abundance (peptide supply). Each peptide has therefore an individual tapasin bonus fingerprint. We also show that there is an optimal level of tapasin expression for each peptide in the immunopeptidome, dependent on its off-rate and abundance. This is important, as the level of tapasin expression can vary widely during different stages of the immune response against pathogens or cancer and is often the target for immune escape.

Programming Molecular Systems To Emulate a Learning Spiking Neuron.

Hebbian theory seeks to explain how the neurons in the brain adapt to stimuli to enable learning. An interesting feature of Hebbian learning is that it is an unsupervised method and, as such, does not require feedback, making it suitable in contexts where systems have to learn autonomously. This paper explores how molecular systems can be designed to show such protointelligent behaviors and proposes the first chemical reaction network (CRN) that can exhibit autonomous Hebbian learning across arbitrarily many input channels. The system emulates a spiking neuron, and we demonstrate that it can learn statistical biases of incoming inputs. The basic CRN is a minimal, thermodynamically plausible set of microreversible chemical equations that can be analyzed with respect to their energy requirements. However, to explore how such chemical systems might be engineered de novo, we also propose an extended version based on enzyme-driven compartmentalized reactions. Finally, we show how a purely DNA system, built upon the paradigm of DNA strand displacement, can realize neuronal dynamics. Our analysis provides a compelling blueprint for exploring autonomous learning in biological settings, bringing us closer to realizing real synthetic biological intelligence.

Decentralizing Cell-Free RNA Sensing With the Use of Low-Cost Cell Extracts.

Cell-free gene expression systems have emerged as a promising platform for field-deployed biosensing and diagnostics. When combined with programmable toehold switch-based RNA sensors, these systems can be used to detect arbitrary RNAs and freeze-dried for room temperature transport to the point-of-need. These sensors, however, have been mainly implemented using reconstituted PURE cell-free protein expression systems that are difficult to source in the Global South due to their high commercial cost and cold-chain shipping requirements. Based on preliminary demonstrations of toehold sensors working on lysates, we describe the fast prototyping of RNA toehold switch-based sensors that can be produced locally and reduce the cost of sensors by two orders of magnitude. We demonstrate that these in-house cell lysates provide sensor performance comparable to commercial PURE cell-free systems. We further optimize these lysates with a CRISPRi strategy to enhance the stability of linear DNAs by knocking-down genes responsible for linear DNA degradation. This enables the direct use of PCR products for fast screening of new designs. As a proof-of-concept, we develop novel toehold sensors for the plant pathogen Potato Virus Y (PVY), which dramatically reduces the yield of this important staple crop. The local implementation of low-cost cell-free toehold sensors could enable biosensing capacity at the regional level and lead to more decentralized models for global surveillance of infectious disease.

A deep learning model for predicting next-generation sequencing depth from DNA sequence.

Targeted high-throughput DNA sequencing is a primary approach for genomics and molecular diagnostics, and more recently as a readout for DNA information storage. Oligonucleotide probes used to enrich gene loci of interest have different hybridization kinetics, resulting in non-uniform coverage that increases sequencing costs and decreases sequencing sensitivities. Here, we present a deep learning model (DLM) for predicting Next-Generation Sequencing (NGS) depth from DNA probe sequences. Our DLM includes a bidirectional recurrent neural network that takes as input both DNA nucleotide identities as well as the calculated probability of the nucleotide being unpaired. We apply our DLM to three different NGS panels: a 39,145-plex panel for human single nucleotide polymorphisms (SNP), a 2000-plex panel for human long non-coding RNA (lncRNA), and a 7373-plex panel targeting non-human sequences for DNA information storage. In cross-validation, our DLM predicts sequencing depth to within a factor of 3 with 93% accuracy for the SNP panel, and 99% accuracy for the non-human panel. In independent testing, the DLM predicts the lncRNA panel with 89% accuracy when trained on the SNP panel. The same model is also effective at predicting the measured single-plex kinetic rate constants of DNA hybridization and strand displacement.

Conditional sanctioning in a legume-Rhizobium mutualism.

Legumes are high in protein and form a valuable part of human diets due to their interaction with symbiotic nitrogen-fixing bacteria known as rhizobia. Plants house rhizobia in specialized root nodules and provide the rhizobia with carbon in return for nitrogen. However, plants usually house multiple rhizobial strains that vary in their fixation ability, so the plant faces an investment dilemma. Plants are known to sanction strains that do not fix nitrogen, but nonfixers are rare in field settings, while intermediate fixers are common. Here, we modeled how plants should respond to an intermediate fixer that was otherwise isogenic and tested model predictions using pea plants. Intermediate fixers were only tolerated when a better strain was not available. In agreement with model predictions, nodules containing the intermediate-fixing strain were large and healthy when the only alternative was a nonfixer, but nodules of the intermediate-fixing strain were small and white when the plant was coinoculated with a more effective strain. The reduction in nodule size was preceded by a lower carbon supply to the nodule even before differences in nodule size could be observed. Sanctioned nodules had reduced rates of nitrogen fixation, and in later developmental stages, sanctioned nodules contained fewer viable bacteria than nonsanctioned nodules. This indicates that legumes can make conditional decisions, most likely by comparing a local nodule-dependent cue of nitrogen output with a global cue, giving them remarkable control over their symbiotic partners.

A systematic approach to inserting split inteins for Boolean logic gate engineering and basal activity reduction.

Split inteins are powerful tools for seamless ligation of synthetic split proteins. Yet, their use remains limited because the already intricate split site identification problem is often complicated by the requirement of extein junction sequences. To address this, we augment a mini-Mu transposon-based screening approach and devise the intein-assisted bisection mapping (IBM) method. IBM robustly reveals clusters of split sites on five proteins, converting them into AND or NAND logic gates. We further show that the use of inteins expands functional sequence space for splitting a protein. We also demonstrate the utility of our approach over rational inference of split sites from secondary structure alignment of homologous proteins, and that basal activities of highly active proteins can be mitigated by splitting them. Our work offers a generalizable and systematic route towards creating split protein-intein fusions for synthetic biology.

Interpretation of morphogen gradients by a synthetic bistable circuit.

During development, cells gain positional information through the interpretation of dynamic morphogen gradients. A proposed mechanism for interpreting opposing morphogen gradients is mutual inhibition of downstream transcription factors, but isolating the role of this specific motif within a natural network remains a challenge. Here, we engineer a synthetic morphogen-induced mutual inhibition circuit in E. coli populations and show that mutual inhibition alone is sufficient to produce stable domains of gene expression in response to dynamic morphogen gradients, provided the spatial average of the morphogens falls within the region of bistability at the single cell level. When we add sender devices, the resulting patterning circuit produces theoretically predicted self-organised gene expression domains in response to a single gradient. We develop computational models of our synthetic circuits parameterised to timecourse fluorescence data, providing both a theoretical and experimental framework for engineering morphogen-induced spatial patterning in cell populations.

Turing Patterning in Stratified Domains.

Reaction-diffusion processes across layered media arise in several scientific domains such as pattern-forming E. coli on agar substrates, epidermal-mesenchymal coupling in development, and symmetry-breaking in cell polarization. We develop a modeling framework for bilayer reaction-diffusion systems and relate it to a range of existing models. We derive conditions for diffusion-driven instability of a spatially homogeneous equilibrium analogous to the classical conditions for a Turing instability in the simplest nontrivial setting where one domain has a standard reaction-diffusion system, and the other permits only diffusion. Due to the transverse coupling between these two regions, standard techniques for computing eigenfunctions of the Laplacian cannot be applied, and so we propose an alternative method to compute the dispersion relation directly. We compare instability conditions with full numerical simulations to demonstrate impacts of the geometry and coupling parameters on patterning, and explore various experimentally relevant asymptotic regimes. In the regime where the first domain is suitably thin, we recover a simple modulation of the standard Turing conditions, and find that often the broad impact of the diffusion-only domain is to reduce the ability of the system to form patterns. We also demonstrate complex impacts of this coupling on pattern formation. For instance, we exhibit non-monotonicity of pattern-forming instabilities with respect to geometric and coupling parameters, and highlight an instability from a nontrivial interaction between kinetics in one domain and diffusion in the other. These results are valuable for informing design choices in applications such as synthetic engineering of Turing patterns, but also for understanding the role of stratified media in modulating pattern-forming processes in developmental biology and beyond.

Stochastic pulsing of gene expression enables the generation of spatial patterns in Bacillus subtilis biofilms.

Stochastic pulsing of gene expression can generate phenotypic diversity in a genetically identical population of cells, but it is unclear whether it has a role in the development of multicellular systems. Here, we show how stochastic pulsing of gene expression enables spatial patterns to form in a model multicellular system, Bacillus subtilis bacterial biofilms. We use quantitative microscopy and time-lapse imaging to observe pulses in the activity of the general stress response sigma factor σB in individual cells during biofilm development. Both σB and sporulation activity increase in a gradient, peaking at the top of the biofilm, even though σB represses sporulation. As predicted by a simple mathematical model, increasing σB expression shifts the peak of sporulation to the middle of the biofilm. Our results demonstrate how stochastic pulsing of gene expression can play a key role in pattern formation during biofilm development.

DNA-based communication in populations of synthetic protocells.

Developing molecular communication platforms based on orthogonal communication channels is a crucial step towards engineering artificial multicellular systems. Here, we present a general and scalable platform entitled 'biomolecular implementation of protocellular communication' (BIO-PC) to engineer distributed multichannel molecular communication between populations of non-lipid semipermeable microcapsules. Our method leverages the modularity and scalability of enzyme-free DNA strand-displacement circuits to develop protocellular consortia that can sense, process and respond to DNA-based messages. We engineer a rich variety of biochemical communication devices capable of cascaded amplification, bidirectional communication and distributed computational operations. Encapsulating DNA strand-displacement circuits further allows their use in concentrated serum where non-compartmentalized DNA circuits cannot operate. BIO-PC enables reliable execution of distributed DNA-based molecular programs in biologically relevant environments and opens new directions in DNA computing and minimal cell technology.

Computing with biological switches and clocks.

The complex dynamics of biological systems is primarily driven by molecular interactions that underpin the regulatory networks of cells. These networks typically contain positive and negative feedback loops, which are responsible for switch-like and oscillatory dynamics, respectively. Many computing systems rely on switches and clocks as computational modules. While the combination of such modules in biological systems leads to a variety of dynamical behaviours, it is also driving development of new computing algorithms. Here we present a historical perspective on computation by biological systems, with a focus on switches and clocks, and discuss parallels between biology and computing. We also outline our vision for the future of biological computing.

Scaling up genetic circuit design for cellular computing: advances and prospects.

Synthetic biology aims to engineer and redesign biological systems for useful real-world applications in biomanufacturing, biosensing and biotherapy following a typical design-build-test cycle. Inspired from computer science and electronics, synthetic gene circuits have been designed to exhibit control over the flow of information in biological systems. Two types are Boolean logic inspired TRUE or FALSE digital logic and graded analog computation. Key principles for gene circuit engineering include modularity, orthogonality, predictability and reliability. Initial circuits in the field were small and hampered by a lack of modular and orthogonal components, however in recent years the library of available parts has increased vastly. New tools for high throughput DNA assembly and characterization have been developed enabling rapid prototyping, systematic in situ characterization, as well as automated design and assembly of circuits. Recently implemented computing paradigms in circuit memory and distributed computing using cell consortia will also be discussed. Finally, we will examine existing challenges in building predictable large-scale circuits including modularity, context dependency and metabolic burden as well as tools and methods used to resolve them. These new trends and techniques have the potential to accelerate design of larger gene circuits and result in an increase in our basic understanding of circuit and host behaviour.

Synthesizing and tuning stochastic chemical reaction networks with specified behaviours.

Methods from stochastic dynamical systems theory have been instrumental in understanding the behaviours of chemical reaction networks (CRNs) arising in natural systems. However, considerably less attention has been given to the inverse problem of synthesizing CRNs with a specified behaviour, which is important for the forward engineering of biological systems. Here, we present a method for generating discrete-state stochastic CRNs from functional specifications, which combines synthesis of reactions using satisfiability modulo theories and parameter optimization using Markov chain Monte Carlo. First, we identify candidate CRNs that have the possibility to produce correct computations for a given finite set of inputs. We then optimize the parameters of each CRN, using a combination of stochastic search techniques applied to the chemical master equation, to improve the probability of correct behaviour and rule out spurious solutions. In addition, we use techniques from continuous-time Markov chain theory to analyse the expected termination time for each CRN. We illustrate our approach by synthesizing CRNs for probabilistically computing majority, maximum and division, producing both known and previously unknown networks, including a novel CRN for probabilistically computing the maximum of two species. In future, synthesis techniques such as these could be used to automate the design of engineered biological circuits and chemical systems.

A Mechanistic Model for Predicting Cell Surface Presentation of Competing Peptides by MHC Class I Molecules.

Major histocompatibility complex-I (MHC-I) molecules play a central role in the immune response to viruses and cancers. They present peptides on the surface of affected cells, for recognition by cytotoxic T cells. Determining which peptides are presented, and in what proportion, has profound implications for developing effective, medical treatments. However, our ability to predict peptide presentation levels is currently limited. Existing prediction algorithms focus primarily on the binding affinity of peptides to MHC-I, and do not predict the relative abundance of individual peptides on the surface of antigen-presenting cells in situ which is a critical parameter for determining the strength and specificity of the ensuing immune response. Here, we develop and experimentally verify a mechanistic model for predicting cell-surface presentation of competing peptides. Our approach explicitly models key steps in the processing of intracellular peptides, incorporating both peptide binding affinity and intracellular peptide abundance. We use the resulting model to predict how the peptide repertoire is modified by interferon-γ, an immune modulator well known to enhance expression of antigen processing and presentation proteins.

Model reduction enables Turing instability analysis of large reaction-diffusion models.

Synthesizing a genetic network which generates stable Turing patterns is one of the great challenges of synthetic biology, but a significant obstacle is the disconnect between the mathematical theory and the biological reality. Current mathematical understanding of patterning is typically restricted to systems of two or three chemical species, for which equations are tractable. However, when models seek to combine descriptions of intercellular signal diffusion and intracellular biochemistry, plausible genetic networks can consist of dozens of interacting species. In this paper, we suggest a method for reducing large biochemical systems that relies on removing the non-diffusible species, leaving only the diffusibles in the model. Such model reduction enables analysis to be conducted on a smaller number of differential equations. We provide conditions to guarantee that the full system forms patterns if the reduced system does, and vice versa. We confirm our technique with three examples: the Brusselator, an example proposed by Turing, and a biochemically plausible patterning system consisting of 17 species. These examples show that our method significantly simplifies the study of pattern formation in large systems where several species can be considered immobile.

Predicting DNA hybridization kinetics from sequence.

Hybridization is a key molecular process in biology and biotechnology, but so far there is no predictive model for accurately determining hybridization rate constants based on sequence information. Here, we report a weighted neighbour voting (WNV) prediction algorithm, in which the hybridization rate constant of an unknown sequence is predicted based on similarity reactions with known rate constants. To construct this algorithm we first performed 210 fluorescence kinetics experiments to observe the hybridization kinetics of 100 different DNA target and probe pairs (36 nt sub-sequences of the CYCS and VEGF genes) at temperatures ranging from 28 to 55 °C. Automated feature selection and weighting optimization resulted in a final six-feature WNV model, which can predict hybridization rate constants of new sequences to within a factor of 3 with ∼91% accuracy, based on leave-one-out cross-validation. Accurate prediction of hybridization kinetics allows the design of efficient probe sequences for genomics research.

Host genotype and time dependent antigen presentation of viral peptides: predictions from theory.

The rate of progression of HIV infected individuals to AIDS is known to vary with the genotype of the host, and is linked to their allele of human leukocyte antigen (HLA) proteins, which present protein degradation products at the cell surface to circulating T-cells. HLA alleles are associated with Gag-specific T-cell responses that are protective against progression of the disease. While Pol is the most conserved HIV sequence, its association with immune control is not as strong. To gain a more thorough quantitative understanding of the factors that contribute to immunodominance, we have constructed a model of the recognition of HIV infection by the MHC class I pathway. Our model predicts surface presentation of HIV peptides over time, demonstrates the importance of viral protein kinetics, and provides evidence of the importance of Gag peptides in the long-term control of HIV infection. Furthermore, short-term dynamics are also predicted, with simulation of virion-derived peptides suggesting that efficient processing of Gag can lead to a 50% probability of presentation within 3 hours post-infection, as observed experimentally. In conjunction with epitope prediction algorithms, this modelling approach could be used to refine experimental targets for potential T-cell vaccines, both for HIV and other viruses.

The Role of Multiscale Protein Dynamics in Antigen Presentation and T Lymphocyte Recognition.

T lymphocytes are stimulated when they recognize short peptides bound to class I proteins of the major histocompatibility complex (MHC) protein, as peptide-MHC complexes. Due to the diversity in T-cell receptor (TCR) molecules together with both the peptides and MHC proteins they bind to, it has been difficult to design vaccines and treatments based on these interactions. Machine learning has made some progress in trying to predict the immunogenicity of peptide sequences in the context of specific MHC class I alleles but, as such approaches cannot integrate temporal information and lack explanatory power, their scope will always be limited. Here, we advocate a mechanistic description of antigen presentation and TCR activation which is explanatory, predictive, and quantitative, drawing on modeling approaches that collectively span several length and time scales, being capable of furnishing reliable biological descriptions that are difficult for experimentalists to provide. It is a form of multiscale systems biology. We propose the use of chemical rate equations to describe the time evolution of the foreign and host proteins to explain how the original proteins end up being presented on the cell surface as peptide fragments, while we invoke molecular dynamics to describe the key binding processes on the molecular level, including those of peptide-MHC complexes with TCRs which lie at the heart of the immune response. On each level, complementary methods based on machine learning are available, and we discuss the relationship between these divergent approaches. The pursuit of predictive mechanistic modeling approaches requires experimentalists to adapt their work so as to acquire, store, and expose data that can be used to verify and validate such models.

A spatially localized architecture for fast and modular DNA computing.

Cells use spatial constraints to control and accelerate the flow of information in enzyme cascades and signalling networks. Synthetic silicon-based circuitry similarly relies on spatial constraints to process information. Here, we show that spatial organization can be a similarly powerful design principle for overcoming limitations of speed and modularity in engineered molecular circuits. We create logic gates and signal transmission lines by spatially arranging reactive DNA hairpins on a DNA origami. Signal propagation is demonstrated across transmission lines of different lengths and orientations and logic gates are modularly combined into circuits that establish the universality of our approach. Because reactions preferentially occur between neighbours, identical DNA hairpins can be reused across circuits. Co-localization of circuit elements decreases computation time from hours to minutes compared to circuits with diffusible components. Detailed computational models enable predictive circuit design. We anticipate our approach will motivate using spatial constraints for future molecular control circuit designs.