Engineering biological systems using automated biofoundries

A High-throughput workflow for CRISPR/Cas9 mediated combinatorial promoter replacements and phenotype characterization in yeast

Due to the rapidly increasing sequence information on gene variants generated by evolution and our improved abilities to engineer novel biological activities, microbial cells can be evolved for the production of a growing spectrum of compounds. For high productivity, efficient carbon channeling towards the end product is a key element. In large scale production systems the genetic modifications that ensure optimal performance cannot be dependent on plasmid-based regulators, but need to be engineered stably into the host genome. Here we describe a CRISPR/Cas9 mediated high-throughput workflow for combinatorial and multiplexed replacement of native promoters with synthetic promoters and the following high-throughput phenotype characterization in the yeast Saccharomyces cerevisiae. The workflow is demonstrated with three central metabolic genes, ZWF1, PGI1 and TKL1 encoding a glucose-6-phosphate dehydrogenase, phosphoglucose isomerase and transketolase, respectively. The synthetic promoter donor DNA libraries were generated by PCR and transformed to yeast cells. A 50% efficiency was achieved for simultaneous replacement at three individual loci using short 60-bp flanking homology sequences in the donor promoters. Phenotypic strain characterization was validated and demonstrated using liquid handling automation and 150 µl cultivation volume in 96-well plate format. The established workflow offers a robust platform for automated engineering and improvement of yeast strains.

Advancing Metabolic Engineering of Saccharomyces cerevisiae Using the CRISPR/Cas System

Thanks to its ease of use, modularity, and scalability, the clustered regularly interspaced short palindromic repeats (CRISPR) system has been increasingly used in the design and engineering of Saccharomyces cerevisiae, one of the most popular hosts for industrial biotechnology. This review summarizes the recent development of this disruptive technology for metabolic engineering applications, including CRISPR-mediated gene knock-out and knock-in as well as transcriptional activation and interference. More importantly, multi-functional CRISPR systems that combine both gain- and loss-of-function modulations for combinatorial metabolic engineering are highlighted.

MACBETH: Multiplex automated Corynebacterium glutamicum base editing method

CRISPR/Cas9 or Cpf1-introduced double strand break dramatically decreases bacterial cell survival rate, which hampers multiplex genome editing in bacteria. In addition, the requirement of a foreign DNA template for each target locus is labor demanding and may encounter more GMO related regulatory hurdle in industrial applications. Herein, we developed a multiplex automated Corynebacterium glutamicum base editing method (MACBETH) using CRISPR/Cas9 and activation-induced cytidine deaminase (AID), without foreign DNA templates, achieving single-, double-, and triple-locus editing with efficiencies up to 100%, 87.2% and 23.3%, respectively. In addition, MACBETH was applied to generate a combinatorial gene inactivation library for improving glutamate production, and pyk&ldhA double inactivation strain was found to improve glutamate production by 3-fold. Finally, MACBETH was automated with an integrated robotic system, which would enable us to generate thousands of rationally engineered strains per month for metabolic engineering of C. glutamicum. As a proof of concept demonstration, the automation platform was used to construct an arrayed genome-scale gene inactivation library of 94 transcription factors with 100% success rate. Therefore, MACBETH would be a powerful tool for multiplex and automated bacterial genome editing in future studies and industrial applications.

Automated Planning Enables Complex Protocols on Liquid-Handling Robots

Robotic automation in synthetic biology is especially relevant for liquid handling to facilitate complex experiments. However, research tasks that are not highly standardized are still rarely automated in practice. Two main reasons for this are the substantial investments required to translate molecular biological protocols into robot programs, and the fact that the resulting programs are often too specific to be easily reused and shared. Recent developments of standardized protocols and dedicated programming languages for liquid-handling operations addressed some aspects of ease-of-use and portability of protocols. However, either they focus on simplicity, at the expense of enabling complex protocols, or they entail detailed programming, with corresponding skills and efforts required from the users. To reconcile these trade-offs, we developed Roboliq, a software system that uses artificial intelligence (AI) methods to integrate (i) generic formal, yet intuitive, protocol descriptions, (ii) complete, but usually hidden, programming capabilities, and (iii) user-system interactions to automatically generate executable, optimized robot programs. Roboliq also enables high-level specifications of complex tasks with conditional execution. To demonstrate the system's benefits for experiments that are difficult to perform manually because of their complexity, duration, or time-critical nature, we present three proof-of-principle applications for the reproducible, quantitative characterization of GFP variants.

Simulation Modeling to Compare High-Throughput, Low-Iteration Optimization Strategies for Metabolic Engineering

Increasing the final titer of a multi-gene metabolic pathway can be viewed as a multivariate optimization problem. While numerous multivariate optimization algorithms exist, few are specifically designed to accommodate the constraints posed by genetic engineering workflows. We present a strategy for optimizing expression levels across an arbitrary number of genes that requires few design-build-test iterations. We compare the performance of several optimization algorithms on a series of simulated expression landscapes. We show that optimal experimental design parameters depend on the degree of landscape ruggedness. This work provides a theoretical framework for designing and executing numerical optimization on multi-gene systems.

Advances in analytical tools for high throughput strain engineering

The emergence of inexpensive, base-perfect genome editing is revolutionising biology. Modern industrial biotechnology exploits the advances in genome editing in combination with automation, analytics and data integration to build high-throughput automated strain engineering pipelines also known as biofoundries. Biofoundries replace the slow and inconsistent artisanal processes used to build microbial cell factories with an automated design-build-test cycle, considerably reducing the time needed to deliver commercially viable strains. Testing and hence learning remains relatively shallow, but recent advances in analytical chemistry promise to increase the depth of characterization possible. Analytics combined with models of cellular physiology in automated systems biology pipelines should enable deeper learning and hence a steeper pitch of the learning cycle. This review explores the progress, advances and remaining bottlenecks of analytical tools for high throughput strain engineering.

Improving the success and impact of the metabolic engineering design, build, test, learn cycle by addressing proteins of unknown function

Rational, predictive metabolic engineering of organisms requires an ability to associate biological activity to the corresponding gene(s). Despite extensive advances in the 20 years since the Escherichia coli genome was published, there are still gaps in our knowledge of protein function. The substantial amount of data that has been published, such as: omics-level characterization in a myriad of conditions; genome-scale libraries; and evolution and genome sequencing, provide means of identifying and prioritizing proteins for characterization. This review describes the scale of this knowledge gap, demonstrates the benefit of addressing the knowledge gap, and demonstrates the availability of interesting candidates for characterization.