In our biological experiments, we tested the use of PhAge deRived growth SlowErs (PARSE) for the goal of making a genetically-tunable growth modulation system. During the typical bacteriophage infection process, a multitude of strategies are employed in order to enact replicative, transcriptional and translational downregulation of the host cells, to allow for metabolic flux toward phage growth and development. As such, bacteriophages require the use of many proteins that bind and inactivate vital bacterial proteins, such as helicase loaders and DNA polymerases to halt replication as well as elements of the RNA polymerase holoenzyme to hinder transcriptional output. A side effect of these processes is the lowering of the overall growth rate of the bacteria. Thus, we reasoned that precise genetic control of these “growth-slowers” outside of the bacteriophage context could constitute a reliable method with which to tune the growth rate of bacteria.
The intended outcome of our project is a growth modulation system that enables regulation of bacterial growth. Bacteriophages utilise growth-slowing (GS) genes to direct the host bacteria's metabolic flux towards generation of more bacteriophages instead of replication. A selection of these GS genes were utilised in our constructs, and their individual descriptions can be found here.
Our Constructs
The backbone we utilised was pET28-A with BsaI sites inserted via PCR amplification. Initially, we used pJUMP28, which encodes sfGFP by default. However, when we attempted to insert vsfGFP into the same plasmid we were unable to select the correct ligation product. We attempted to swap the selection process of pJUMP29 (lacZ gene) into pJUMP28, however after encountering some difficulty we decided to use pET28-1A as our backbone instead.
Figure 1a shows a construct of GS gene (gp2) inserted into pET28, so that its expression is IPTG inducible. Additionally, a pET28-vsfGFP structure, as illustrated in Figure 1b, was also created, utilising the same promotion system. This facilitated our ability to assess the growth rate at various levels of IPTG expression. As a result, our biomodelling team was able to employ this information to show it is possible to calibrate a promoter's strength in comparison to an inducible promoter. This in turn can be used for growth rate predictions based on a construct with a known promoter strength and one of our characterised growth inhibitors.
Furthermore, we conducted tests on vsfGFP constructs under the regulation of the Anderson promoters (as shown in Figure 2), a group of promoters whose strengths are well-established relative to one another. This allowed us to illustrate a method by which users can calibrate their own inducible promoter and employ it to anticipate the growth rate when combining a growth-inhibiting gene with their inducible promoter under specified inducer molecule concentrations
Applications
Controlling bacterial co-cultures
iGEM Imperial College London's Ecolibrium (2016) [2] employed quorum sensing machinery to control the expression of gp2, one of the bacterial GSs we have characterised, to enable the maintenance of stable bacterial co-cultures. By characterising a range of phage-derived GS proteins, PARSE would build on their work and enable a better prediction of the impacts of expressing these proteins on bacterial co-cultures, and furthermore could allow for precise calculations of conditions required for desired ratios of species within co-cultures.
Bacterial co-cultures are phenomenally useful in many bioindustrial contexts, particularly where the product of one bacteria feeds into the metabolism of another. Thus, an increased fidelity in manipulating co cultures may enhance the rate at which certain products of interest are produced. Most novel though, it could relieve issues where co-existing bacterial populations within a reactor vessel grow at different rates, which often leads to wash out of the slower growing bacteria. Dr Karunakaran stated that the use of controllable GSs in this way would be very useful for her research into developing bioplastics.
Easig the maintenance of bioreactors
A growth modulating system controlled by quorum sensing machinery would enable a steady-state optical density (OD) to be maintained. Dr Mukherjee and Dr Karunakaran, academics at the University of Sheffield, both expressed that this kind of genetic system would contribute to overcoming current issues with batch bioreactors. These are often favoured in industry due to ease of regular sterilisation to prevent a build up of culture within the outflow pipes, which can cause damage to the bioreactor over time. The major disadvantage of batch cultures, however, is that it does not enable as many engineering approaches to modulating the bacterial cell growth. Our system creates a biological solution to this problem by preventing bacterial aggregation using a quorum responsive GS system, in addition to growth modulation without needing to flush bacteria out.
Automatic elements of recombinant protein production
Additionally, there would be potential for the arrest of growth that our characterised GS enact to allow for the shunting of metabolic growth towards more efficient protein production. This could be as simple as encoding both a GS and a gene of interest (GOI) under the same promoter, such that upon inducer addition growth rate is hindered whilst protein production is triggered. Moreover, this could be further modified by placing both the GOI and GS under a quorum sensitive promoter: in this way, bacterial cell density controls the shunting of metabolic flux towards protein production, rather than the user, which may remove some labour intensity from protein production workflow.
Figure 3 shows a potential way the GSs and quorum sensing could interact to allow control of the growth. The IPTG inducible promoter pLac is used as an example but this could be swapped for any inducible promoter. If the bacteria already have quorum sensing encoded into their genome, this system could be utilised in parallel to the plasmid of Figure 3, as this would allow the bacteria to regulate their own growth metabolic flux, or it would allow addition of IPTG to regulate it manually. This offers an opportunity for an automated control of growth system, which can also be adjusted manually.
Dr. Karunakaran corroborated this suggestion, and remarked that this would have even more utility if our GS genes could be used in a variety of industrially relevant organisms such as E. coli, Vibrio natriegens, and cyanobacteria such as Synechocystis and some Streptomyces species. Making PARSE more context-independent would enable this; research into the mechanistic detail of GS proteins, and their impact on important cellular processes is vital for the identification of target systems and proteins that are conserved across more species. It is worth noting that our system would likely need a eukaryotic variation if porting to higher organisms, since the proteomic diversity between eukaryotes and prokaryotes would make finding a single shared target protein for growth modulation virtually impossible.
Expanding the potential of directed evolution
iGEM Sheffield's rEvolver (2022) proposed a novel system with which in vivo directed evolution could be performed. The system was composed of three plasmids: a mutagenic plasmid encoding MutaT7, which allowed for transcription-coupled, targeted diversification of a GOI; a selection plasmid in which the output of a growth modulation gene was controlled by the fitness output of the GOI; and a third plasmid that encoded CRISPR knockdown to dampen the cells' innate DNA repair machinery. A significant challenge within this field is linking the fitness of an arbitrary evolving gene to the overall fitness of the cell - a challenge whose solution was proposed by the inclusion of the growth modulation selection plasmid: as the GOI gets fitter (according to externally placed parameters), the cells grow faster and thus outcompete less evolved neighbours. Mechanistically, one of the ways this could work is if the output of a GS was modulated by a transcriptional regulator, which itself would be regulated by a biosensor for the fitness of the evolving gene - in other words, growth slowing is relieved as the fitness of the GOI increases.
In characterising the GS gene expression under inducible promoters (and by extension, the ability to model their action under the Anderson promoters) and their effects on growth rate, PARSE represents a major leap forward toward successfully building rEvolver’s selection plasmid and therefore toward making in vivo directed evolution more accessible. Additionally, the modelling of the action of a range of GSs could allow for some control over the extent of evolution that occurs, whereby less effective GSs may be preferred if the evolutionary landscape to be explored should be relatively small. Some further work to assess the validity of this application further would be to repeat our experiments, but include the presence of a genetic NOT gate (such as pT181, as rEvolver originally proposed) which would respond to heightened GOI fitness by downregulating the associated GS gene. The data could then be incorporated into the PARSE calibration procedure, such that the extra step in the genetic pathway would be included in modelling considerations and allow for the precise control of the degree of evolution.
References
[1] F. Arhin et al., “Antimicrobial drug discovery through bacteriophage genomics,” Nature biotechnology, vol. 22, no. 2, pp. 185–191, 2004, doi: 10.1038/nbt932.
[2] iGEM Imperial College London (2016) Ecolibrium. Available at: https://2016.igem.org/Team:Imperial_College (Accessed: 15 August 2023).
[3] iGEM Sheffield (2022) rEvolver. Available at: https://2022.igem.wiki/sheffield/index.html (Accessed: 1 April 2023).