Team:Michigan/Description
Project Description
Our project uses a reporter plasmid to serve as the readout for a competitive binding assay between Cas9 variants. These variants can be modified versions of the Cas9 protein from the same species, or can be Cas9 proteins from different species. One Cas9 variant is active, while the other variant contains mutations that render it a deactivated Cas9 (dCas9). Both Cas9 variants have guide RNA (gRNA) sequences on the plasmid that target them to the same region of the reporter plasmid, which expresses cyan fluorescent protein (CFP). When the active cas9 binds the plasmid, it cleaves the DNA. The double stranded break results in degradation of the plasmid (Cui et al, 2016). If the dCas9 binds the plasmid, it protects the DNA. That binding permits expression of CFP and prevents binding by active Cas9. Examining fluorescence levels allows for measurement of the average amount of plasmid per cell as a proxy of competitive binding to of the enzymes to the DNA. By comparing how frequently the plasmid is degraded or protected, the process reveals how effectively the Cas9 and dCas9 variants bind target DNA in a competitive assay.
Why?
The most accurate way to determine the editing frequency of Cas9 variants is to test them individually in the cell type or organism of interest. However, that can take a significant amount of time and resources and could result in batch effects if the experimental conditions differ for cells edited by the different variants. Our assay is not designed to examine editing, but allows for a comparison of DNA binding by Cas9 variants within E.coli. We created this competitive binding assay so that fellow iGem teams or other researchers could easily compare the ability of two Cas9 variants to bind to target DNA within the same cell.
The measurements within experiments of fluorescence from cells provides a quantitative method within cells. Fluorescence anistropy (https://www.nature.com/protocolexchange/protocols/557) or an EMSA (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2757439/) could be used to examine binding. However, neither could do so within cells and in a competitive manner. By utilizing a competitive assay, we believe subtle differences between variants could be revealed within cells. We envision use of the method by researchers interested in the kinetics of binding or the ability of engineered proteins to bind to target DNA within cells.
Controls:
If only an active Cas9 variant is expressed, then it should be able to bind and cleave the target sequence in the reporter plasmid, leading to degradation of the plasmid and reduced CFP. Examining the cells to ensure the level of fluorescence is not above the background of cells containing only the reporter plasmid.
If only a dCas9 variant is expressed, then no copies of the reporter plasmid should be cleaved. That would result in the average fluorescence per cell being the same as for cells containing only the reporter plasmid.
If an active and deactive version of the same Cas9 variant are expressed, then they should show the same binding efficiency for the reporter plasmid. The level of fluorescence in these cells would serve as the baseline level of fluorescence when dCas9 and the active Cas9 bind equally well.
How it works?
The deactivated Cas9 contains two point mutations, one in each of the RuvC1 and HNH nuclease domains. In the case of SpCas9, these mutations are D10A and H840A. These mutations prevent the dCas9 from cleaving the target DNA. However, it can still form a complex with the gRNA and bind the target DNA. When bound to the target DNA, the dCas9 renders the sequence inaccessible, so the active Cas9 cannot bind to the sequence. Both the dCas9 and Cas9 variant bind to the same DNA sequence upstream of the coding sequence of CFP.
Consider comparing two variants, Cas9 #1 and Cas9 #2. For simplification, Cas9 #1 will be shortened to C1, dCas9 #1 to dC1, Cas9 #2 to C2, and dCas9 #2 to dC2.
In order to exaggerate the differences in kinetics and binding, we selected Cas9 enzymes from two different species of bacteria. This means that they will be structurally different and require different PAM binding sites. Differences in PAM are particularly powerful in effecting the kinetics of binding since PAM searching is required to find the sequence and PAM binding provides the energy required to separate and compare the DNA.
First, control experiments would be performed by expressing C1 and dC1 in the same cells, or by expressing C2 and dC2 in the same cells. Since the same Cas9 variant is used in its active and deactive form, the binding affinities of each form should be the same. That establishes a baseline fluorescence level when binding affinities of the Cas9 and dCas9 are the same.
Next, express C1 and dC2 together in cells. If the fluorescence in cells expressing these proteins is roughly the same as in the control experiments, then C1 and dC2 likely have very similar binding affinities. If fluorescence in cells expressing these proteins is greater than in the control experiments, then less of the reporter plasmid is being degraded. That suggests that dC2 is binding to the target DNA with greater affinity than C1. On the other hand, lower fluorescence levels than in control experiments suggest that more of the reporter plasmid is being degraded. That indicates that C1 is binding with greater affinity than dC2.
At the same time, express dC1 and C2 together in cells. The results in these cells should be the opposite of those in the cells expressing C1 and dC2. For example, if cells expressing C1 and dC2 were more fluorescent than the control cells, than the cells expressing dC1 and C2 should be less fluorescent, meaning more of the reporter plasmid is degraded. That corroborates the idea that C1 (and dC1) have a lower binding efficiency than C2 (and dC2).
Project Design
We started our design process by setting our team’s design goals for the project:
- The Testing Model’s Cas9s had to bind to the same target sequence via their gRNA in order to build a competitive system.
- The Testing Model needed two different Cas9s that would show different affinities and efficiencies, allowing one Cas9 to outcompete the other.
- The Testing Model needed to be measured in an easy and effective way to allow other iGem teams to replicate and use the model for further modification of CRISPR Cas9.
- The Testing Model needed an active Cas9 to degrade the reporter plasmid and a dead Cas9 to protect the reporter plasmid.
- The Testing Model needed to be inducible for the control of Cas9 expression and thus measurement.
We then reviewed literature and the iGEM registry in search of parts that would help us accomplish our goals. Our initial design called for three plasmids, two with different Cas9s and a reporter plasmid for measurement. After consulting Dr. Zhang, we were able to use her Streptococcus pyogenes CRISPR Cas9 (SpCas9) in an inducible system. After consulting the iGEM registry, we could not find a Streptococcus pyogenes Cas9 with any functional modifications that would change binding efficiencies or affinities. In our efforts to find a Cas9, we came across Streptococcus aureus Cas9 (SaCas9) that is known to have function differences (Friedland et al. 2015). However, because there was no SaCas9 in stock, we synthesized SaCas9 through IDT in a similar inducible model as the Zhang SpCas9. To test if this design met our goals, we devised the experiments outlined in detail on the Experiments page.
SpCas9 and SaCas9 are both proteins frequently used for genetic engineering purposes, particularly as part of the CRISPR-Cas9 platform. The easiest distinction to be made is that the proteins are from different species of bacteria (the former from Streptococcus pyogenes and the latter from Staphylococcus aureus), corresponding with their abbreviated names. Although they are comparable in terms of binding/cutting efficiency, SpCas9 and SaCas9 display several key differences in functionality and characterization which may lend them suitable for certain applications. For instance, SpCas9 is slightly larger at 1368 amino acids (as opposed to SaCas9 with 1053 amino acids), requiring its coding sequence to also be longer. This bp length difference, while minor, may impact transformation efficiency if the overall plasmid is too large. The most important of these differences in terms of practical purposes however is the fact that the Cas9 varieties bind different PAM sequences, with SpCas9 recognizing *NGG and SaCas9 recognizing *NGGRRT. This effectively lessens the probability of off-target cuts in SaCas9 in comparison to SpCas9, due to the greater number of complementary base-pairs needed to bind the PI domain of Cas9 to its PAM sequence. In relation to experiments performed, this necessitated the creation of a PAM sequence that would be compatible with both Cas9 types, although this was not particularly difficult due to the first 3 bases being shared between both PAMs.
In order for the both Cas9s to bind to the same target sequence, we needed each PAM sequence to be represented downstream of the sequence. The PAM sequence for Streptococcus pyogenes Cas9 is *NGG and the PAM sequence for Streptococcus aureus Cas9 is *NGGRRT thus we decided on a sequence that was necessary for both Cas9s to bind - TGGGAT. The guide RNA needed to bind to the DNA (crRNA) that is transcribed is the same for both species of Cas9, but the tracrRNA (gRNA scaffold for Cas9/gRNA interaction) is specific to the species of Cas9.
In designing the gRNA, it is important to note which sequence needs to be transcribed for use by the Cas9. In order for the gRNA/Cas9 to recognize and bind to the target DNA, the crRNA must be the same sequence just upstream of the PAM sequence, thus it binds to the complementary DNA of the PAM sequence.
Proof of Concept
The application of synthetic biology and engineering principles to the Cas9 enzyme could have the potential to expand its use as a powerful tool with a wide range of applications. A wide range of papers have been published in recent years describing various modifications to the Cas9 enzyme. Many different methods were used in these papers to measure the changes to the enzyme’s functionality. This makes it tough to compare Cas9 modifications between papers. We saw a need for a standardized testing system so that researchers can more easily understand how their modifications compare to others. With this in mind, we set out to create a testing platform that will allow other researchers to rapidly and effectively test the modifications they are making to the enzymes. This goal led to the idea of developing a competition based assay where there is a definitive winner of each target binding site between two competing Cas9 enzymes. The Guard Assassin CRISPR/Cas9 Assay (GACCA) allows for this direct competition through the use of a three plasmid system. The first plasmid serves as the target plasmid with a CFP reporter downstream from a target site that both Cas9 enzymes will be attempting to bind. The second plasmid will contain one of the Cas9 enzymes with normal nuclease activity (i.e. able to cut the DNA normally) while the third plasmid contains a dCas9 system with deactivated nuclease activity (i.e. able to bind but not cleave the DNA). The core principle is that the dCas9 will be able to bind to the target sequence before the regular Cas9 if its binding affinity or searching ability is better. If this is the case, the plasmid should be protected from cleavage and degradation of the target plasmid. Then the roles of “Assassin” and “Guard” would be flipped so that the Cas9 that was active previously would now be deactivated and vice versa.
We had considerable difficulty cloning our Cas9 sequences into the intended vectors and thus we were not able to generate any data on the actual performance of this system given the short time scale on which we were working. Given more time, we would establish controls that demonstrate baseline fluorescence, the ability of nuclease active Cas9 to cleave the target plasmid, and the ability of deactivated Cas9 (dCas9) to bind but not cut the target sequence. Once these controls are established and standardized, we would then move into the competition experiments. Additionally, we developed a simple computation model in MATLAB that show the potential effect of enzyme modification on the amount of each complex that is able to form. We intend on fitting this model to data that we generate in the future and showing other teams how they can do the same for their own experiments as they test their Cas9 designs.