5Heterologous gene expression can be a significant burden to cells, consuming resources and 6 causing decreased growth and stability. We describe here an in vivo monitor that tracks E. 7 coli capacity changes in real-time and can be used to assay the burden synthetic constructs and 8 their parts impose. By measuring capacity, construct designs with reduced burden can be 9 identified and shown to predictably outperform less efficient designs, despite having equivalent 10 expression outputs. 11Robust expression of heterologous genes is necessary for many applications in biotechnology and is 12 central to synthetic biology where predictable fine-tuning of expression is typically desired 1-3 . 13However, for engineered bacteria all heterologous expression represents an unnatural load, 14 consuming cellular resources usually allocated to replication, repair and native gene expression 15 ( Figure 1A). Gene expression burden is a well-known phenomenon characterised by decreased 16 growth rates that can predispose synthetic constructs to evolutionary instability and can 17 unexpectedly alter their behaviour 4-10 . Burden presents a major barrier to predictable and stable 18 engineering of cells, yet it is largely an unquantified phenomenon, inferred in most cases by tracking 19 growth rate decline 5, 6, 11 . Recent research has begun to explore burden, demonstrating how its 20 impact varies between different E. coli strains 6, 11 and showing how expression load can be measured 21 in vitro using cell-free extracts 12 . However, an improved way of quantifying how heterologous gene 22 expression imposes burden in vivo has yet to be described, despite the arrival of new models of 23 bacterial growth that outline the importance of expression resources for the cell [13][14][15] . 24To advance in vivo quantification of burden we developed a fluorescence-based method to measure 25 in real-time the gene expression capacity of bacterial genomes. We built integration vectors to insert 26 a 'capacity monitor', a synthetic constitutive green fluorescent protein (GFP) expression cassette, 27 into defined genomic loci of commonly-used E. coli strains ( Figure S1); reasoning that because this 28 cassette lacks regulation, changes in GFP expression due to global expression changes will reflect 29 changes in resource availability 16 . To demonstrate how the capacity monitor improves on using 30 growth rates to assess burden, we measured GFP expression rates from the genome of DH10B E. coli 31 hosting an operon-expressing plasmid induced at different time-points during exponential growth 32 ( Figure 1B). Capacity (determined as GFP production rate per cell) decreases significantly compared 33 to uninduced cells within 30 minutes of construct induction, and this rapid change contrasts with the 34 smaller, slower decreases in growth observed when culture optical density is measured. The fact 35 that capacity changes precede growth rate changes supports the view that decreased expression 36 resources causes growth rate decline and underlin...

Cells use feedback regulation to ensure robust growth despite fluctuating demands for resources and differing environmental conditions. However, the expression of foreign proteins from engineered constructs is an unnatural burden that cells are not adapted for. Here we combined RNA-seq with an in vivo assay to identify the major transcriptional changes that occur in Escherichia coli when inducible synthetic constructs are expressed. We observed that native promoters related to the heat-shock response activated expression rapidly in response to synthetic expression, regardless of the construct. Using these promoters, we built a dCas9-based feedback-regulation system that automatically adjusts the expression of a synthetic construct in response to burden. Cells equipped with this general-use controller maintained their capacity for native gene expression to ensure robust growth and thus outperformed unregulated cells in terms of protein yield in batch production. This engineered feedback is to our knowledge the first example of a universal, burden-based biomolecular control system and is modular, tunable and portable.

No abstract

8Characterisation of genetic control elements is essential for the predictable engineering 9 of synthetic biology systems. The current standard for in vivo characterisation of control 10 elements is through the use of fluorescent reporter proteins such as green fluorescent 11 protein (GFP). Gene expression, however, involves not only protein production but also 12 the production of messenger RNA (mRNA). Here we present the use of the Spinach 13 aptamer sequence, an RNA mimic of GFP, as a tool to characterise mRNA expression in 14 Escherichia coli. We show how the aptamer can be incorporated into gene expression 15 cassettes and how co-expressing it with a red fluorescent protein (mRFP1) allows, for 16 the first time, simultaneous measurement of mRNA and protein levels from engineered 17 constructs. Using flow cytometry, we apply this tool here to evaluate ribosome binding 18 site sequences and promoters and use it to highlight the differences in the temporal 19 behaviour of transcription and translation. 20 21

Cells use feedback regulation to ensure robust growth despite fluctuating demands on resources and different environmental conditions. Yet the expression of foreign proteins from engineered constructs is an unnatural burden on resources that cells are not adapted for. Here we combined multiplex RNAseq with an in vivo assay to reveal the major transcriptional changes in two E. coli strains when a set of inducible synthetic constructs are expressed. We identified that native promoters related to the heat-shock response activate expression rapidly in response to synthetic expression, regardless of the construct. Using these promoters, we built a CRISPR/dCas9-based feedback regulation system that automatically adjusts synthetic construct expression in response to burden. Cells equipped with this general-use controller maintain capacity for native gene expression to ensure robust growth and as such outperform unregulated cells at protein yields in batch production. This engineered feedback is the first example of a universal, burden-based biomolecular control system and is modular, tuneable and portable.