Precision engineering of biological function with large-scale measurements and machine learning.

As synthetic biology expands and accelerates into real-world applications, methods for quantitatively and precisely engineering biological function become increasingly relevant. This is particularly true for applications that require programmed sensing to dynamically regulate gene expression in response to stimuli. However, few methods have been described that can engineer biological sensing with any level of quantitative precision. Here, we present two complementary methods for precision engineering of genetic sensors: in silico selection and machine-learning-enabled forward engineering. Both methods use a large-scale genotype-phenotype dataset to identify DNA sequences that encode sensors with quantitatively specified dose response. First, we show that in silico selection can be used to engineer sensors with a wide range of dose-response curves. To demonstrate in silico selection for precise, multi-objective engineering, we simultaneously tune a genetic sensor's sensitivity (EC50) and saturating output to meet quantitative specifications. In addition, we engineer sensors with inverted dose-response and specified EC50. Second, we demonstrate a machine-learning-enabled approach to predictively engineer genetic sensors with mutation combinations that are not present in the large-scale dataset. We show that the interpretable machine learning results can be combined with a biophysical model to engineer sensors with improved inverted dose-response curves.

Best Practices for DNA Template Preparation Toward Improved Reproducibility in Cell-Free Protein Production.

Performance variability is a common challenge in cell-free protein production and hinders a wider adoption of these systems for both research and biomanufacturing. While the inherent stochasticity and complexity of biology likely contributes to variability, other systematic factors may also play a role, including the source and preparation of the cell extract, the composition of the supplemental reaction buffer, the facility at which experiments are conducted, and the human operator (Cole et al. ACS Synth Biol 8:2080-2091, 2019). Variability in protein production could also arise from differences in the DNA template-specifically the amount of functional DNA added to a cell-free reaction and the quality of the DNA preparation in terms of contaminants and strand breakage. Here, we present protocols and suggest best practices optimized for DNA template preparation and quantitation for cell-free systems toward reducing variability in cell-free protein production.

The genotype-phenotype landscape of an allosteric protein.

Allostery is a fundamental biophysical mechanism that underlies cellular sensing, signaling, and metabolism. Yet a quantitative understanding of allosteric genotype-phenotype relationships remains elusive. Here, we report the large-scale measurement of the genotype-phenotype landscape for an allosteric protein: the lac repressor from Escherichia coli, LacI. Using a method that combines long-read and short-read DNA sequencing, we quantitatively measure the dose-response curves for nearly 105 variants of the LacI genetic sensor. The resulting data provide a quantitative map of the effect of amino acid substitutions on LacI allostery and reveal systematic sequence-structure-function relationships. We find that in many cases, allosteric phenotypes can be quantitatively predicted with additive or neural-network models, but unpredictable changes also occur. For example, we were surprised to discover a new band-stop phenotype that challenges conventional models of allostery and that emerges from combinations of nearly silent amino acid substitutions.

Sparse estimation of mutual information landscapes quantifies information transmission through cellular biochemical reaction networks.

Measuring information transmission from stimulus to response is useful for evaluating the signaling fidelity of biochemical reaction networks (BRNs) in cells. Quantification of information transmission can reveal the optimal input stimuli environment for a BRN and the rate at which the signaling fidelity decreases for non-optimal input probability distributions. Here we present sparse estimation of mutual information landscapes (SEMIL), a method to quantify information transmission through cellular BRNs using commonly available data for single-cell gene expression output, across a design space of possible input distributions. We validate SEMIL and use it to analyze several engineered cellular sensing systems to demonstrate the impact of reaction pathways and rate constants on mutual information landscapes.

Measurements drive progress in directed evolution for precise engineering of biological systems.

Precise engineering of biological systems requires quantitative, high-throughput measurements, exemplified by progress in directed evolution. New approaches allow high-throughput measurements of phenotypes and their corresponding genotypes. When integrated into directed evolution, these quantitative approaches enable the precise engineering of biological function. At the same time, the increasingly routine availability of large, high-quality data sets supports the integration of machine learning with directed evolution. Together, these advances herald striking capabilities for engineering biology.

Evolving Bacterial Fitness with an Expanded Genetic Code.

Since the fixation of the genetic code, evolution has largely been confined to 20 proteinogenic amino acids. The development of orthogonal translation systems that allow for the codon-specific incorporation of noncanonical amino acids may provide a means to expand the code, but these translation systems cannot be simply superimposed on cells that have spent billions of years optimizing their genomes with the canonical code. We have therefore carried out directed evolution experiments with an orthogonal translation system that inserts 3-nitro-L-tyrosine across from amber codons, creating a 21 amino acid genetic code in which the amber stop codon ambiguously encodes either 3-nitro-L-tyrosine or stop. The 21 amino acid code is enforced through the inclusion of an addicted, essential gene, a beta-lactamase dependent upon 3-nitro-L-tyrosine incorporation. After 2000 generations of directed evolution, the fitness deficit of the original strain was largely repaired through mutations that limited the toxicity of the noncanonical. While the evolved lineages had not resolved the ambiguous coding of the amber codon, the improvements in fitness allowed new amber codons to populate protein coding sequences.

Addicting diverse bacteria to a noncanonical amino acid.

Engineered orthogonal translation systems have greatly enabled the expansion of the genetic code using noncanonical amino acids (NCAAs). However, the impact of NCAAs on organismal evolution remains unclear, in part because it is difficult to force the adoption of new genetic codes in organisms. By reengineering TEM-1 ß-lactamase to be dependent on a NCAA, we maintained bacterial NCAA dependence for hundreds of generations without escape.

Structure-based non-canonical amino acid design to covalently crosslink an antibody-antigen complex.

Engineering antibodies to utilize non-canonical amino acids (NCAA) should greatly expand the utility of an already important biological reagent. In particular, introducing crosslinking reagents into antibody complementarity determining regions (CDRs) should provide a means to covalently crosslink residues at the antibody-antigen interface. Unfortunately, finding the optimum position for crosslinking two proteins is often a matter of iterative guessing, even when the interface is known in atomic detail. Computer-aided antibody design can potentially greatly restrict the number of variants that must be explored in order to identify successful crosslinking sites. We have therefore used Rosetta to guide the introduction of an oxidizable crosslinking NCAA, l-3,4-dihydroxyphenylalanine (l-DOPA), into the CDRs of the anti-protective antigen scFv antibody M18, and have measured crosslinking to its cognate antigen, domain 4 of the anthrax protective antigen. Computed crosslinking distance, solvent accessibility, and interface energetics were three factors considered that could impact the efficiency of l-DOPA-mediated crosslinking. In the end, 10 variants were synthesized, and crosslinking efficiencies were generally 10% or higher, with the best variant crosslinking to 52% of the available antigen. The results suggest that computational analysis can be used in a pipeline for engineering crosslinking antibodies. The rules learned from l-DOPA crosslinking of antibodies may also be generalizable to the formation of other crosslinked interfaces and complexes.