Lecturers

If one wishes to deploy an organism in the environment or a plant/animal host one most take into account both the challenges and power in the microbial consortia that inevitably preexist in these locations. Here we will discuss dissection and assembly of natural and artificial consortia of microbes and phage and their possible engineering for various applications.

In this lecture we will discuss the design of algorithms for the dissection of regulatory networks, including ARACNe, for the reverse engineering of transcriptional networks, MINDy, for the reverse engineering of signal transduction networks, and PrePPI for the dissection of protein-protein interaction networks using protein structure information. We will then address the use of regulatory networks inferred by these algorithms to elucidate master regulator proteins that mechanistically control the transcriptional state of the cell, using the VIPER algorithm, elucidating genetic and epigenetic variants contributing to the aberrant activity of master regulator proteins, using the DIGGIT algorithm, and using network based methodologies to elucidate drug mechanisms of action, with the VIPER and DeMAND algorithms. Finally, we will discuss how integrative use of multiple algorithms can be used to dissect the mechanism of action of specific proteins of interest, such as KRAS, using the OncoSig algorithm.

In this second lecture we will discuss how network-based methodologies can be used to systematically identify, validate, and pharmacologically target of a new class of therapeutic targets represented by Master Regulator proteins, whose concerted aberrant activity within tightly regulated modules (tumor checkpoints) is responsible for the mechanistic implementation and maintenance of specific tumor cell states. These methodologies have led to development of two NY CLIA certified tests, OncoTarget and OncoTreat, that leverage large-scale drug-perturbation assays to systematically identify drugs and drug combinations whose mechanism of action is specifically effective in abrogating the activity of individual aberrantly active proteins or of entire tumor checkpoints, on an individual patient basis. We will specifically discuss how these methodologies have been used to create a new class of RNA-based, N of 1 clinical trials for patients with aggressive malignancies who failed to respond to multiple lines of therapy. Finally, we will discuss how these approaches can be used to target tumor heterogeneity, as manifested by tumor cells with distinct transcriptional states within the same tumor mass. For instance, we will discuss targeting the mesenchymal and proneural states of glioblastoma at the individual cell level as well as the stem-like progenitor cells in breast cancer patients.

In vitro directed evolution allows biomolecules with new and useful properties to be engineered—mimicking natural evolution on an experimentally accessible time scale by creating large libraries of DNA mutants using PCR and then carrying out a high-throughput assay for variants with improved function. To provide a breakthrough in the complexity of libraries that can be readily searched experimentally for synthetic biology and to allow systems to be directly engineered in the cell, my laboratory is engineering S. cerevisiaeso that both the mutagenesis and selection steps of directed evolution can be carried out entirely in vivo, under conditions of sexual reproduction. We have built a modular chemical complementation assay, which provides a selection for diverse chemistry beyond that natural to the cell using themes and variations on the yeast two-hybrid assay. In addition, we devised a heritable recombination system, for simultaneous mutagenesis and selection in vivounder conditions of sexual reproduction. Finally, we have begun to utilize these mutagenesis and selection technologies to engineer yeast to carry out new functions themselves ranging from being a biosensor, to a therapeutic, to a self-organizing community.

The fluorescent proteins revolutionized our ability to study protein function directly in the cell by enabling individual proteins to be selectively labeled through genetic encoding of a fluorescent tag.  As researchers seek to make increasingly sophisticated dynamic measurements of protein function in the cell to unravel molecular mechanism, we designed a chemical tag to combine the advantages of genetic encoding with a modular organic fluorophore.  With TMP-tag, the protein of interest is tagged with E. coli dihydrofolate reductase, which can subsequently be labeled with a cell permeable trimethoprim-fluorophore conjugate.  Here we demonstrate that TMP-tag is a robust cellular reagent.  We present recent results exploiting the modular nature of the chemical tag to generate TMP-tags for specific applications in single-molecule, super-resolution, and multi-color imaging.  We look forward to innovations at the interface of chemical tag technology and spectroscopy for biological imaging.

The minimal cell is the hydrogen atom of cellular biology. Such a cell, because of its simplicity and absence of redundancy would be a platform for investigating just what biological components are required for life, and how those parts work together to make a living cell. Since the late 1990s, our team at the Venter Institute has been developing a suite of synthetic biology tools that enabled us to build what previously has only been imagined, a minimal cell. Specifically, a bacterial cell with a genome that expresses only the minimum set of genes needed for the cell to divide every two hours that can be grown in pure culture. That minimal cell has about half of the genes that are in the bacterium on which it was based, Mycoplasma mycoides JCVI syn1.0, the so-called synthetic bacteria we reported on in 2010. We used transposon bombardment to identify non-essential genes, and genes needed to maintain rapid growth in M. mycoides. Based on those data, we designed and synthesized a reduced genome in eight overlapping segments. All segments were individually viable when combined with wild type versions of the seven other segments. Combinations of reduced segments that were not viable allowed us to identify synthetic lethal pairs of genes. These occur when two genes each encode an essential function. Those findings required re-design and re-synthesis of some reduced genome segments. Three cycles of design, synthesis, and testing, with retention of quasi‐essential genes, produced synthetic bacterium JCVI‐Syn3.0 (531 kb, 474 genes), which has a genome smaller than that of any autonomously replicating cell found in nature. Synthetic bacterium JCVI-Syn3.0 retains almost all genes involved in synthesis and processing of macromolecules. Surprisingly, it also contained 149 genes with unknown biological functions, suggesting the presence of undiscovered functions essential for life. This minimal cell is a versatile platform for investigating the core functions of life, and for exploring whole‐genome design. Since it was initially reported in 2016, we have identified functions for about 50 of the original 149 genes of unknown function and are in the process of developing a computational model of the cell.

This work was supported by Synthetic Genomics, Inc., DARPA Living Foundries contract HR0011-12-C-0063, and the J. Craig Venter Institute.

In recent years, the JCVI Synthetic Biology team, which is best known for construction of bacterial cells programmed with synthetic genomes, has also taken on projects that required synthesis and booting up of small and large viral genomes, grand scale genome engineering of yeast chromosomes, and design, synthesis, and installation of human artificial chromosomes. Many of the methods we employed for these projects were derived from lessons learned through our project to makes the aforementioned synthetic bacterial and minimal bacterial cell. Other approaches had to be developed just for these eukaryotic genome and chromosome projects. I will present four vignettes about our sometime successful and sometime unsuccessful efforts.

• Rapid synthesis and booting up of Influenza virus genomes to speed and improve vaccine makers capacity to address new pandemics
• Construction of a 155 kb Herpes Simplex Virus genome as a yeast centromeric plasmid and infectious virus production using that plasmid.
• Efforts to build and boot up a chromosome encoding only the genes essential for life for Kluyveromyces marxianus, which is a yeast and the fastest growing known eukaryote.
• Construction in yeast of a fully synthetic human artificial chromosome with a novel synthetic centromere and its installation into mammalian cells.

Synthetic biology has emerged as a novel discipline that tries to engineer biology in a more controllable, predictable and standardize manner. Bringing such characteristics to microbial systems allow us to expand the tools available to engineering metabolism in a more powerful way. This talk will present through some examples, how cutting-edge synthetic biology tools such as combinatorial assembly methods or CRISPR can be used to improve microbial bioproduction.

After several decades of engineering microbes for biotechnological applications, we have realized that overengineer strains often present undesirable features that limit the improvement of current production systems. Moreover, natural systems evolve as ecosystems and not as an individual to allow better performance in the environment. The use of microbial communities in the food and feed industry is clearly effective and it is responsible the most of our fermentable products. Now, thanks to synthetic biology we can engineer microbial communities in a way not possible in single organisms. In this talk, the concepts and tools to create synthetic communities will be discussed and applications of engineered communities will be presented.

Understanding how behaviours at one scale emerge from interactions at another scale is a central theme in mathematical biology. These lectures will focus on this problem in the context of collective cell migration. This first lecture will develop a hybrid agent-based model for cranial neural crest migration in normal development. The model will consist of a partial differential equation to describe the diffusion and kinetics of a signalling growth factor, VEGF, and a discrete model to describe the motion of cells in response to VEGF signals. The model is used to test a number of hypotheses on how this system can yield migration of cells at the population level and model predictions are validated by experimental studies.

Angiogenesis is the process by which the body forms new blood vessels. This can occur as part of wound repair, or in response to a solid tumour requiring more nutrient. This lecture will describe the multitude of ways in which this process can be analysed with a series of models of increasing complexity.

Andy Hesketh, Marta Vergnano, Chris Wan, & Steve Oliver

Cambridge Systems Biology Centre & Department of Biochemistry, University of Cambridge, Cambridge UK

The understanding of metabolic interactions between microorganisms is being advanced significantly by the use of computer simulations that employ genome-scale metabolic models. However, there is also a need to construct biological models to enable the study of refractory interactions, for instance those between pathogenic bacteria and their plant and animal hosts. In many of these interactions, the “alarmone”, ppGpp, is synthesised in response to nutrient starvation signals and is an important factor in the establishment and maintenance of infections. In order to study this aspect of pathogen-host interactions, we engineeredSaccharomyces cerevisiae to inducibly synthesise the bacterial signalling nucleotides cdiGMP, cdiAMP and ppGpp so as to characterise the effects that these nucleotides exert on eukaryotic cell function. Extensive in vivo interactions between the bacterial signalling molecules and eukaryotic gene function occur, resulting in outcomes ranging from growth inhibition to death.  We found that yeast mitochondrial function and chromatin remodelling and modification processes serve as important buffers against their activity. Expression of the human Mesh1 ppGpp hydrolase enzyme in yeast resulted in clearance of ppGpp from the cells, restoring wild-type growth. This suggests that higher eukaryotes may have evolved systems for eliminating damaging bacterial nucleotides from their cells.

The unity of molecular and cell biology across the Eukaryotes makes yeasts favourable organisms with which to model both infectious and systemic diseases of humans. 

In order to model infectious diseases, we have set up systems in which the infectious agent and the human host are modelled either within the same or separate yeast cells. For instance, many tropical diseases are caused by eukaryotic pathogens, including protozoa and nematode worms, with a very similar biochemistry to humans. This makes it difficult to identify agents that will kill the parasite but not the patient. The search for such drugs is also hampered by the fact that many of these parasites are difficult or impossible to culture in the laboratory. This is where yeast comes in. We have developed a robust, fully automated method to screen for potential anti-parasitic drugs. This system uses separate yeast strains whose growth is dependent on the expression of coding sequences specifying target enzymes from either the parasite or the human host. The yeast system thus permits multiple parasite targets to be screened in parallel and is also able to exclude compounds that do not discriminate between host and parasite enzymes. 

Our experiments on yeast purine nucleotide metabolism generated data that could not be accurately predicted by existing models of yeast metabolism. Investigation of this deficiency led us to increase the representation of iron metabolism in the model and to discover that yeast has potential for modelling the cellular processes involved in the onset and early progression of Parkinson’s disease. I shall discuss other examples of using yeast to model neurodegenerative diseases, including Friedreich’s Ataxia and motor neurone disease. 

This lecture will introduce the principle of genetically-encoded biosensors and their applications in synthetic biology.  The various types of detection elements that can be used (nucleic acids, proteins, and whole-cells) and the advantages and disadvantages of different circuit designs will be discussed.  Examples from the recent literature will be used to illustrate the principles of biosensor design with particular focus on microbial biosensors.

This lecture follows on from part 1 to discuss the key performance characteristics of a genetically-encoded biosensor including signal-to-noise ratio, sensitivity, and dynamic range, as well as the design features that can be used to modulate these characteristics.  Basic models of biosensor circuits  will also be covered to enable students to gain an understanding of how these can be applied to design of a biosensor based on specifications of its function.

DNA origami, in which a long scaffold strand is assembled with a large number of short staple strands into parallel arrays of double helices, has proven a powerful method for custom nanofabrication. Although diverse shapes in 2D are possible, the single-layer rectangle has proven the most popular, as it features fast and robust folding and modular design of staple strands for simple abstraction to a regular pixel surface. Here we introduce a barrel architecture, built as stacked rings of double helices, that retains these appealing features, while extending construction into 3D. We demonstrate hierarchical assembly of a 100 megadalton barrel that is ~90 nm in diameter and ~270 nm in height, and that provides a rhombic-lattice canvas of a thousand pixels each, with a pitch of 9 nm, on its inner and outer surfaces. Complex patterns rendered on these surfaces were resolved using up to twelve rounds of exchange PAINT super-resolution fluorescence microscopy. We envision these structures as versatile nanoscale pegboards for applications requiring complex 3D arrangements of matter. One application I will discuss is the use of DNA barrels to scaffold the manufacture large lipid nanodiscs suitable for cryoEM structure determination of large membrane proteins and their complexes embedded in a native lipid bilayer environment.

DNA origami, in which a long scaffold strand is assembled with a large number of short staple strands into parallel arrays of double helices, has proven a powerful method for custom nanofabrication of shapes up to 100 nm in size. However, the scaffold represents about half the mass of an origami, therefore the origami size is restricted by the length of the scaffold. However, it is impractical and prohibitively expensive to scale the length of the scaffold. Here I will discuss a strategy, that we call crisscross ultracooperative assembly, that combines all-or-nothing scaffold-dependent initiation of folding with scaffold-independent growth, therefore allowing for sizes unbounded by the length of the scaffold. A major application will be digital counting of molecular analytes, where each molecular detection event triggers growth of a single filament resolvable by low-cost microscopy. I also will discuss a method we have developed for scalable isolation of ssDNA from dsDNA, which will be enabling for crisscross assembly and other applications of DNA nanotechnology and DNA biotechnology.

Systems and synthetic biology deal with substantial challenges due to uncertainty that more traditional disciplines such as physics or engineering do not face to the same extent. This has two main reasons: (i) despite the increasing availability of quantitative datasets, our knowledge about essentially all cellular networks is incomplete, regarding quantitative parameters and functional relations between components alike; and (ii) single cells exhibit variable behavior due to random effects, even in populations of genetically identical cells in a homogeneous environment. The first lecture will address how uncertainty in components and their interactions (i.e., biological hypotheses) is reflected in uncertain structures of mathematical models, which is a particular challenge for mechanistic, dynamic modelling. The focus is on computational methods for systematically representing uncertainty in dynamic mathematical models, either for inferring unknown quantities and mechanisms by combining models and data, or for deriving robust designs of synthetic gene circuits. Example applications include the inference of mechanisms of yeast nutrient signaling and the design of synthetic circuits with dynamic properties such as adaptation.

A key step for understanding heterogeneity in cell populations is to disentangle sources of cell-to-cell and intra-cellular variability. Single-cell time-lapse data provides potential means for this, but single-cell analysis with dynamic models is a challenging open problem. Most of the existing inference methods address only single-gene expression or neglect correlations between processes that underlie heterogeneous cell behaviors. The focus of the second lecture will be a simple, flexible, and scalable method for estimating cell-specific and population-average parameters to characterize sources and effects of cell-to-cell variability. The framework relies on non-linear mixed effects models of cellular networks, and its accuracy is demonstrated with a published model and dataset. An application to endocytosis in yeast demonstrates that dynamic models of realistic size can be developed for the analysis of single-cell data. Combined with sensitivity analysis for identifying which biological sub-processes quantitatively and dynamically contribute to cell-to-cell variability, this application shows that shifting the focus from single reactions or parameters to nuanced and time-dependent contributions of sub-processes helps biological interpretation.

Genomics has undergone a resolution revolution, such that the nucleic acid content of individual cells are now sequenced routinely in high-throughput mode. Thus we can define the cellular composition of tissues in an unbiased and comprehensive way using single cell transcriptomics, revealing the molecular fingerprint of cell states and their predicted signalling circuits in tissues across development and disease. I will illustrate this with the human healthy and asthmatic lung, as well as the maternal-fetal interface in the first trimester placenta/decidua

Self-assembling protein containers are promising delivery vehicles for cellular and gene therapy applications, but the ability to predict how mutations alter self-assembly and other particle properties remains a significant challenge. Here, we combine comprehensive codon mutagenesis with high-throughput sequencing to characterize the assembly-competency of all single amino acid variants of several self-assembling proteins, including those from a virus-like particle, a bacterial microcompartment, and a secretion system. The revealed a high-resolution fitness landscape that challenged several conventional protein design assumptions on the composition of linkers, mutability of pore-lining residues, and more. Using the same approach with additional comprehensive mutagenesis strategies and selective pressures on the virus-like particle identified several new variants that permit efficient chemical and post-translational modifications and have altered stability. For example, the wild-type virus-like particle is acid tolerant down to pH 2, but we identified a variant with a single point mutation that confers stability at neutral pH but acid-triggered disassembly. Acid sensitivity is highly desirable in targeted delivery to improve the efficiency of endosomal release. Additional insights will be shared, both from the virus-like particles and from the other two self-assembling model proteins.

Protein production is a multi-billion dollar industry, with applications in pharmaceuticals, industrial enzyme use (eg in laundry detergent), and biomaterials. Bacteria are receiving renewed interest as protein production hosts because of their fast growth and tractability. The Salmonella enterica Type III Secretion System secretes non-native proteins at product titers of up to 400 mg/L in rich media, but is highly sensitive to environmental and growth conditions and therefore not robust. This system is nonetheless an ideal process for protein production applications because it is non-essential for bacterial metabolism and allows for target proteins to cross both bacterial membranes in one step, via characteristic needle-like protein structures. We engineered an optimized, more robust secreting strain of Salmonella for the high-titer production of a variety of biochemically challenging heterologous proteins, such as degradation-prone biopolymer proteins, antibodies, and toxic antimicrobial peptides.  To do so required manipulating regulation, engineering the protein components of the apparatus, and optimizing media composition and growth parameters. Now that the system is producing protein at commercially-relevant titers and quality, we also established a process to demonstrate the new strain is “generally recognized as safe” (GRAS) so it can be applied across industries.