Project Design

Overview

Through model and expert driven feedback, we designed systems that allow for the stabilization of promoters and on-the-fly control of protein expression at desired levels.

Engineering Principles

While designing our systems, we focused on the engineering principles governing both the functionality of the iFFL network motif and that of our individual parts, as members of a functional whole.

It is integral to maintain the balance between different parameters determining the behaviour of the iFFL in order to achieve the stabilization of a promoter as desired. We made sure that the systems are orthogonal and do not interface with the native genetic circuitry of the chassis to the degree that would cause the systems not to function as expected [1].

Having a pivotal role in the design process of functional parts and projects, we made sure to include debugging in our day-to-day problem solving efforts. Both in experiments, part design and modelling, constantly cataloguing and revising our progress, along with feedback from analyzing our results, helped us pinpoint our mistakes and find ways to tackle them. Our debugging efforts were focused on deciphering what went wrong with the constructs that did not function as expected.

In order for the parts to be easily integrated and successfully used by other iGEM teams, their behaviour has to be predictable and standardized. Working toward this, we conducted a series of experiments to characterize and tune our systems, while keeping in mind the various future applications and uses they might have. Along those lines, we, also, made sure that our measurements were expressed in relative units, calibrated with a fluorescein curve or calibration beads.

Stabilized Promoters

Stabilized promoters are regulatory genetic elements that achieve gene expression independent of gene/plasmid copy number. They can be designed with the implementation of a type I incoherent feedforward Loop (iFFL) network motif that renders the Gene of Interest (GOI) expression independent of the plasmid’s copy number. Copy number positively affects GOI expression, while its repressor, being under the same positive influence, compensates for this change. If the Hill coefficient that characterizes the repression is 1, and thus the repression is perfectly non-cooperative, it results in the disassociation of the GOI’s expression from the copy number.

Each stabilized promoter is characterized by its strength, the gene of interest expression level and error, the relative change in the gene of interest expression as the copy number increases from the lowest to the highest copy number measured [2]. As the expression level of the repressor increases, the stabilization error and the promoter’s strength decrease, leading to weaker but more stable expression across different copy numbers.

Reaching the desired expression level can be achieved by two different approaches. Either by re-tuning the expression level of the repressor, and thus the stabilization level, or by on-the-fly regulation by induction from a riboswitch regulatory element.

System Design

TAL Effector Stabilized Promoters

The first system we used was a set of TAL Effector stabilized promoters, described by Shapiro et. al. [2], in order to investigate their function and potential uses. TAL Effectors bind to specific promoter sequences and prevent the recruitment of RNA polymerase in order to create a type I incoherent Feedforward Loop (iFFL) network motif that renders the expression of the downstream sequences independent of the plasmid’s copy number.

We used TALEsp1 (TALE stabilized promoter 1) and TALEsp2 (TALE stabilized promoter 2) in order to achieve the desired expression level and Error. TALE proteins bind as monomers in specific regions inside the promoter thus inhibiting transcription initiation. The repression level is determined by the operator’s site in the promoter. Specifically, TALEsp1 has stronger expression and higher Error level when compared to TALEsp2, since TALEsp2 binding site includes a nucleotide from the -10 promoter region. To characterize TALE stabilized promoters we measured their expression level in different copy numbers. We, also, conducted an Interlaboratory Collaboration in order to investigate if the TALEsp1 and TALEsp2 reduce interlaboratory variations.

Inducibility

Wanting to expand on the TAL Effector stabilized promoters system, we designed a tool that would allow, on top of the stabilization of a promoter, its induction and activation to the desired expression level.

Ιnitially, we considered the implementation of the LacI repressor that represses transcription by binding to a specific DNA sequence. This system is well characterized and is widely used for transcription regulation.

Since we wanted complete control over the expression of the stabilized promoter we intended to introduce a LacOsym binding site near the promoter region, besides where the TAL Effector recognized and interacted with the promoter sequence. During our time designing the system we took note of possible steric effects between the TAL Effector and the LacI repressor while bound to the Promoter Sequence.

In order to conceptualize and better understand the implications of possible steric effects on the stabilization of the promoter, a model of ODEs describing the primary processes of gene expression and the effects of such interactions was set up and simulated on SimBiology.

In order to conceptualize and better understand the implications of possible steric effects on the stabilization of the promoter, a model of ODEs describing the primary processes of gene expression and the effects of such interactions was set up and simulated on SimBiology.The results pointed towards a loss of the non-cooperative nature of the TAL Effector repression, depending on the IPTG induction level. At that point, during an enlightening conversation with Thomas H. Segall-Shapiro , we noted our concern about the induction via the LacI repressor and the issues it might cause. He confirmed our concerns and encouraged us to use another type of regulator.

Since interference with the regulatory effects of TALE repressors was obstructive to their intended function, as predicted by the model, we chose to abstain from implementing transcriptional regulators, and thus moved to translational regulatory elements, like the Riboswitch, instead.

We chose the theophylline riboswitch, since, in contrast to the adenine riboswitch, which we also considered, is completely orthogonal and thus provides greater predictability in the outcome of the induction. In order to allow for more precise control of the expression levels a riboswitch with high dynamic range was required .

We came in contact with Prof. Howard Salis to ask for advice about the riboswitch design. He noted that the post and pre-aptamer regions where important, and different RNA sequences might have varying effects on the behaviour of the riboswitch due to possible changes in the secondary structure of the aptamer. He advised us to use a fusion protein as a marker that would maintain the first 99 ribonucleotides (33 amino-acids) of the reporter used in the study, luciferase, if we wanted to use a different one like sfGFP, and avoid aberrant aptamer secondary structures that would, potentially, deteriorate its functionality.

We implemented two different riboswitches, BBa_K2839006 and Theo-27 [3] in the TALEsp1 system. As a reporter, we chose sfGFP fused with the first 99 nucleotides of luciferase, due to the fact that it uses luciferase as the reporter. The fusion protein can be, also, used with the 12.1 riboswitch [4], because it does not affect the aptamer region. We selected sfGFP over other available marker since its maturation and fluorescence is unaffected by fusion partner misfolding [5].

CRISPRi

While the TALE-stabilized promoters function with great precision [2], in order to stabilize a large number of promoters, (e.g. in a complex metabolic pathway or genetic circuit) one would need either more than one TAL Effector proteins or to redesign the promoter, changing the TALE binding site to effectively stabilize promoters with the desired strength-error trade-off and to the required expression level. This would increase the metabolic burden imposed on the cell by the expression of multiple proteins at high levels. This sets a limit to the usage of stabilized promoters.

To tackle this issue we decided to use CRISPRi to repress and stabilize promoters. It utilizes the ability of the catalytically dead variant of Cas9, dCas9, to bind to specific sequences, in order to prevent transcription initiation by blocking RNAP recruitment [6]. CRISPRi, being a relatively new addition to the arsenal of gene regulation factors, possesses great potential in controlling gene expression due to dCas9’s strong affinity and specificity to the target sequence, determined solely by its complementarity with the 20 first nucleotides of the single guide RNA (sgRNA) sequence [7]. The only limitation is the need for the presence of a protospacer adjacent motif (PAM) for the dCas9 enzyme to recognize [8]. Afterwards, it interrogates the upstream sequence for complementarity with the sgRNA. The aforementioned complementarity determines the binding affinity of the sgRNA-dCas9 repressor complex to the determined sequence [7].

As the output promoter, we used an Anderson promoter, since the Anderson library is well characterized and all promoters include a 5’-NGG-3’ PAM sequence just upstream of the -10 site. This allows for easy design of the sgRNA [8] and ensures tight repression [9][10].

Choosing where each component should be expressed from, we set up a model to simulate the system. Initially, we had 3 options for the design of our system, regarding dCas9 expression site: the plasmid with the sfGFP marker, the genome, or the insertion of a second compatible plasmid. Results showed that the expression of both the sgRNA and the dCas9 from a single plasmid ( Topology A ), in which the promoter to be stabilized resides, broke the system and exhibited behaviour similar to repression with a Hill coefficient of 2. Another option was the genome insertion of dCas9 ( Topology B ), which would offer increased system stability, but, when simulated in our model, expression of dCas9 proved inadequate.

We decided to use a double plasmid system ( Topology C ), co-expressing the sgRNA along with sfGFP and expressing dCas9 on a separate vector. The plasmid containing the dCas9 expression cassette is a low-medium copy number vector (p15A ORI) and dCas9 expression rate is controlled by a pTet promoter, therefore its output depends on the usage of the external inducer, doxycycline.

In this system, the sfGFP promoter’s stabilization was designed to depend only on the expression level of the sgRNA. Therefore, we expressed dCas9 on the highest level that creates saturating condition without inhibiting growth due to excessive metabolic burden [11]. In order to investigate the effect of dCas9 production on the growth rate of the DH5alpha E.coli strain, we conducted a growth assay, expressing dCas9 on different levels, via induction with Doxycycline.

Regarding the sgRNA design, in order to achieve the required repression level, we had 2 options: either constitutively express a sgRNA partially complementary to the target sequence or induce the expression of a fully complementary sgRNA to its target. We decided to induce the expression of sgRNA so that we could create a response function from which we can acquire valuable data on the behaviour of the system [1]. Control of sgRNA expression was achieved through a L-Rhamnose inducible promoter, pRha [12]. Full complementarity of sgRNA to its target sequence, in combination with dCas9 overexpression and lack of saturation with sgRNA, made it possible to control the repression levels by adding L-Rhamnose.

In order to characterize the CRISPRi stabilized promoter we induced the expression of the sgRNA driven by the L-Rhamnose inducible promoter at the desired lowest copy number (psc101). From the response function we can determine the cooperativity of the repression and -choose the expression level of the sgRNA that corresponds to the desired strength of the stabilized promoter.

After characterizing the response function of the input and output promoters, we would have settled on a desired Strength-Error level and replace the rhamnose inducible promoter with a constitutive one, thereby maintaining the expression stable, independent of the plasmids copy number, for all copy numbers without need for addition of more rhamnose to compensate for the increase in the number of RhaS binding sites. Unfortunately, we observed minimal induction when measuring the pRha mediated sgRNA expression so we hypothesized that this was due to the change of a conserved region after the promoter’s +1 site. Therefore, we chose BBa….. for the expression of the sgRNA cassette. We opted for a strong constitutive promoter in order to achieve a low Error value as this may prove valuable when implementing the CRISPRi stabilization system in order to stabilize complex biosynthetic pathways.

Stabilized heterologous AND gate

As we wanted not only to characterize stabilized promoters, but also to implement them into synthetic circuits that exhibit dysfunctionality over different copy numbers, we chose a previously well-characterized heterologous AND gate, that, when introduced into E. coli cells at different copy number backbones, exhibits contradicting behaviour, with fluorescence output seemingly being inversely proportional to the copy number. The circuit consists of the σ54-dependent hrpL promoter which is a part of hrp regulon naturally found at the plant pathogen Pseudomonas syringae.The system is active producing the output signal only if both the co-activating input proteins hrpR and hrpS are expressed forming the HrpR/Hrps complex which binds and induces the output promoter. HrpS is expressed from the TetR repressor promoter, PTet, inducible by anhydrotetracycline and doxycycline while HrpR, from the LuxR activatable promoter, PLux, induced by AHL. Despite being orthogonal to E. coli cells, its exhibits contradictory behaviour stemming from irregularities in the TetR induction level on different copy numbers. While we would expect expression from a low copy number plasmid to yield lower fluorescence relative to that of a medium copy number plasmid, this AND gate, behaves in the opposite way. Due to higher TetR levels expressed from the medium copy number plasmid, the same amount of inducer yields lower expression of HrpS co-activator. This leads to a lower activator complex concentration, since it is determined by the concentration of the subunit with the lowest expression level.

In order to allow for its stabilization across different copy numbers without the need for re-tuning, we attempted to stabilize the expression of TetR at a level that would allow for the same behaviour at both copy numbers investigated.

Unfortunately, the And Gate didn’t function at all.

Attenuator

In search for a more elegant approach to the stabilized promoters, we turned to RNA-based regulation, that does not require the introduction of transgenic proteins that might cause burden-related issues due to overexpression and has a shorter response time. We designed a sequence containing many individually acting units, as part of the same transcript. Firstly, two small RNA (sRNA) separated by a hammerhead ribozyme are transcribed upstream of an attenuator sequence that is itself flanked by hammerhead ribozymes. The sRNAs interact with the attenuator sequence, halting transcription, by forming a terminator loop. Downstream of the attenuator is the RBS along with the coding sequence. This way, the sRNAs are constitutively expressed and, by causing premature transcription termination, repress the transcription of the sequences downstream of the attenuator. Through modeling, we decided to use two copies of the sRNA instead of one since they provide tighter repression and thus lower stabilization error. (graphs) In addition to the gene/copy number independant expression, this design achieves promoter-independent expression. That is to say, a relatively stable expression over a wide range of promoters with different output levels. This is achieved by the inclusion of the sRNA, acting as the repressor, in the same transcript as the coding sequence. Thus, all events occurring before the transcription of the repressor, including the frequency at which the RNA Polymerase (RNAP) is recruited (ή: the rate of RNAP recruitment), are part of the primary input of the iFFL and rendered irrelevant to the expression level achieved.

This design is expected to greatly reduce variations arising from transcription related events, thus diminishing promoter-caused internal noise.

Due to time constraints, we did not manage to submit and characterize this part. It’s model-predicted behavior can be found here.

Experimental design

Colonies from transformation of ligated products (Golden Gate assembly, 3A assembly) were sampled and colony PCR was performed with VR and VF2 primers to determine the size of the insert and thus its validity. For more information on the methods and cloning strategy we used, visit the protocols or parts page respectively.

In order to study and test the predicted function of our systems we designed experiments measuring fluorescence intensity of a reporter gene. The reporter we used was the superfolder Green Fluorescent Protein (sfGFP). We measured single-cell fluorescence in a CyFlow(R) Cube 8 Flow cytometer at FL1 (ex. 488). The majority of the single cell fluorescence was measured during mid log phase of growth, as described in the protocol section of our Notebook page.

Riboswitch behavior was characterized by induction at different inducer levels independently and under the stabilization of the TALEsp1 repressor BBa_K2839000.

All measurements were performed in biological replicates (n=3), unless otherwise stated. For the characterization of the CRISPRi stabilized promoters we set up fluorescence intensity measurements of the output promoter under the control of an inducible input promoter at different levels of induction. Characterization of the input promoters by measuring the reporter expression at different concentrations of the inducer and corresponding the inducer concentration used for the output promoter characterization with the input promoter expression, results in the response function of the NOT Gate [1].

From the response function we can gather information about the experimentally-derived cooperativity of our system, a strong predictor of its proper function. By creating a response function we obtained insight into the behavior of the system at different repressor expression levels. This allows for the replacement of the inducible input promoter with a constitutive one at a desired expression level, thus tuning the stabilized promoter and allowing for the choice of a desired strength level and subsequent strength-error stade-off. To determine the optimal expression level of dCas9 from the psB3K3 vector, we performed a growth rate analysis of different dCas9 expression levels. From the results we determined a inducer concentration, corresponding to the highest dCas9 expression level that does not greatly hinder the growth rate of the bacteria. Subsequent measurements that included the induction of dCas9 were performed at that concentration of Doxycycline

[1] Brophy, J. A. N., & Voigt, C. A. (2014). Principles of genetic circuit design. Nature Methods. https://doi.org/10.1038/nmeth.2926 [2] Segall-Shapiro, T. H., Sontag, E. D., & Voigt, C. A. (2018). Engineered promoters enable constant gene expression at any copy number in bacteria. Nature Biotechnology. https://doi.org/10.1038/nbt.4111 [3] Espah Borujeni, A., Mishler, D. M., Wang, J., Huso, W., & Salis, H. M. (2016). Automated physics-based design of synthetic riboswitches from diverse RNA aptamers. Nucleic Acids Research, 44(1), 1–13. https://doi.org/10.1093/nar/gkv1289 [4] Lynch, S. A., & Gallivan, J. P. (2009). A flow cytometry-based screen for synthetic riboswitches. Nucleic Acids Research, 37(1), 184–192. https://doi.org/10.1093/nar/gkn924 [5] Pédelacq, J. D., Cabantous, S., Tran, T., Terwilliger, T. C., & Waldo, G. S. (2006). Engineering and characterization of a superfolder green fluorescent protein. Nature Biotechnology. https://doi.org/10.1038/nbt1172 [6] Qi, L. S., Larson, M. H., Gilbert, L. A., Doudna, J. A., Weissman, J. S., Arkin, A. P., & Lim, W. A. (2013). Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell, 152(5), 1173–1183. https://doi.org/10.1016/j.cell.2013.02.022 [7] Sternberg, S. H., Redding, S., Jinek, M., Greene, E. C., & Doudna, J. A. (2014). DNA interrogation by the CRISPR RNA-guided endonuclease Cas9. Nature, 507(7490), 62–67. https://doi.org/10.1038/nature13011 [8] Larson, M. H., Gilbert, L. A., Wang, X., Lim, W. A., Weissman, J. S., & Qi, L. S. (2013). CRISPR interference (CRISPRi) for sequence-specific control of gene expression. Nature Protocols, 8(11), 2180–2196. https://doi.org/10.1038/nprot.2013.132 [9] Bikard, D., Jiang, W., Samai, P., Hochschild, A., Zhang, F., & Marraffini, L. A. (2013). Programmable repression and activation of bacterial gene expression using an engineered CRISPR-Cas system. Nucleic Acids Research, 41(15), 7429–7437. https://doi.org/10.1093/nar/gkt520 [10] Vigouroux, A., Oldewurtel, E., Cui, L., Bikard, D., & van Teeffelen, S. (2018). Tuning dCas9’s ability to block transcription enables robust, noiseless knockdown of bacterial genes. Molecular Systems Biology, 14(3), e7899. https://doi.org/10.15252/MSB.20177899 [11] Nielsen, A. A., & Voigt, C. A. (2014). Multi-input CRISPR/Cas genetic circuits that interface host regulatory networks. Molecular Systems Biology, 10(11), 763–763. https://doi.org/10.15252/msb.20145735 [12] Kelly, C. L., Liu, Z., Yoshihara, A., Jenkinson, S. F., Wormald, M. R., Otero, J., … Heap, J. T. (2016). Synthetic Chemical Inducers and Genetic Decoupling Enable Orthogonal Control of the rhaBAD Promoter. ACS Synthetic Biology, 5(10), 1136–1145. https://doi.org/10.1021/acssynbio.6b00030