Part:BBa_K5195012

The 2024 Stanford iGEM team designed a fusion protein consisting of mNeonGreen ([1]BBa_K5195004), a P2A linker ([2]BBa_K1442039), a truncated version of the DUX4 protein coding for only the first 217 amino acids (DUX4-DBD) ([3]BBa_K5195002), and a Krüppel associated box (KRAB) transcriptional repressor domain ([4]BBa_K5195001). This design was created in order be able to quantify how much DUX4-DBD protein would be produced in cells upon transfection. Because the P2A linker causes ribosomal “skipping” during translation, this effectively separates the mNeonGreen and DUX4-DBD-KRAB proteins. In designing the coding sequence in this way, we enabled quantification of DUX4-DBD without directly attaching another large protein to it.

This coding sequence was used in a plasmid containing a doxycycline-inducible Tet-On Tight TRE promoter and the coding sequence for the full length DUX4 protein. Primers were designed to amplify the entire original plasmid, Addgene Plasmid #99281 (https://www.addgene.org/99281/) except for the end of the DUX4-fl coding sequence to exclude the transcriptional activation domain. Then, an mNeonGreen-P2A fragment, replicated by PCR from an existing plasmid pK381, provided to us by one of our mentors, Dr. Phillip Kyriakakis, from the Stanford Coleman lab, was gel-purified and then Gibson Assembled with the DUX4-DBD fragment to create our imNG-P2A-DUX4-DBD plasmid. This plasmid was then linearized and a KRAB coding sequence from a gene block was Gibson Assembled to it to create the final plasmid, which was confirmed to be correct by sequencing. The resulting plasmid is shown below with the mNG-P2A-DUX4-DBD-KRAB coding sequence highlighted, along with the linearized fragment gel band (outlined by the red box) at 9285 bp:

Usage and Biology

Located on the human chromosome 4q35, the DUX4 gene a transcription factor encoding the protein DUX4 and is found within the chromosome’s D4Z4 repeat array[1]. The DUX4 full length protein (DUX4-fl) is composed of two double homeodomains, which are responsible for sequence-specific binding to DNA, along with a transcriptional activation domain (TAD) that activates harmful downstream genes[2].

In patients with facioscapulohumeral muscular dystrophy (FSHD), there is a reactivation of the DUX4 transcription factor, which is normally silenced following early embryonic development.

By over-expressing DUX4-DBD, a shortened isoform of DUX4-fl that lacks the TAD, the 2024 Stanford iGEM team hopes to competitively inhibit DUX4-fl binding to the DUX4 binding sites, replacing DUX4-fl binding with DUX4-DBD. Because it lacks the TAD, DUX4-DBD should not activate the downstream genes in the way that DUX4-fl does.

Attaching a KRAB transcriptional repressor domain serves to further prevent DUX4 from activating downstream target genes. The KRAB, or Krüppel-associated box domain, is a powerful transcriptional repression domain found in certain zinc finger proteins. Once the Dux4-DBD binds to the Dux4-fl target sites, the KRAB domain will act to further repress transcription by recruiting a co-repressor complex to those specific sites. This complex includes the KAP1 protein, which then in turn recruits chromatin-modifying enzymes such as histone deacetylases and histone methyltransferases, leading to a more compact and closed chromatin structure and making the DNA less accessible to transcriptional machinery. This way, even if there were to be a time where Dux4-DBD unbinds and Dux4-fl has another chance to bind, it wouldn’t be able to access the DNA due to its repressed state. Essentially, by replacing pathogenic Dux4-fl’s transactivation domain with a repression domain, we aim to create a more potent therapeutic which can hopefully require a smaller dose in order to achieve the same effect.

The below graphic demonstrates our approach of overexpressing DUX4-DBD-KRAB to competitively inhibit DUX4-fl. As an excess of DUX4-DBD-KRAB is introduced, it will be more likely to bind to DUX4 binding sites, thereby stopping DUX4-fl from binding:

Experimental Approach

For testing this device, we first conducted a series of transfections HEK293T cells, seeded at a density of 100,000 cells/well in 24-well plates. Three different plasmids were used - one containing the mNeonGreen-P2A-DUX4-DBD sequence, one containing DUX4-FL, and a DUX4 reporter containing DUX4 binding sites upstream of mScarlet. These plasmids were cloned with PCR and assembled via Gibson Assembly, then transformed into DH5a or DH10b bacteria, miniprepped, and sequenced to verify that the resulting DNA sequences were the same as the intended ones. Varying ratios of DUX4-DBD to DUX4 full length (DUX4-fl) were transfected alongside a DUX4 reporter that produces mScarlet (red) fluorescent protein upon the activation of the DUX4 binding sites located upstream of the fluorescent protein in the plasmid. 48 hours, the plates were imaged using a fluorescent microscope and a fluorescence plate reader assay was performed.

Proof of expression and function of DUX4-DBD

This transfection tested the potential for DUX4-DBD to act as a competitive inhibitor, because with more DUX4-DBD added, we expect the mScarlet fluorescent signal produced by the reporter to decrease because DUX4-DBD binds to the same region of DNA as DUX4-fl, but lacks the transcriptional activation domain responsible for recruiting transcriptional machinery to activate downstream gene expression. Our transfection results are shown below:

Here, the red fluorescent protein, our DUX4 reporter readout, decreases with increasing ratios of DUX4-DBD to DUX4-fl. The increasing amount of DUX4-DBD is implied from the increasing mNeonGreen (GFP) signal.

The above graph displays the plate reader assay results from 4 replicates of this transfection. It supports the visual results from the fluorescence microscope, showing that with increasing DUX4-DBD to DUX4-fl ratios, the DUX4 reporter expression level is driven down very close to its baseline level.

To obtain more precise, single-cell level data from our transfection experiments, we used flow cytometry to view individual cells and their fluorescence, which gives us a continuum of signals. The flow data supports our idea that as DUX4-DBD:DUX4-fl ratio increases, the corresponding reporter activity in those cells also decreases.

In the above figure, C12, C6, C1, and B8 represent different ratios of DUX4-DBD to DUX4-fl from highest to lowest (C12 is 25:1 DBD:FL, C6 is 10:1, C1 is 1:1, and B8 is 0:1).

These experiments were completed in HEK293T cells, which, though are easy to work with, were not the most representative of the environment our construct is ultimately aimed to be used in. Therefore, we tried to repeat our experiments in C2C12 cells, which is a mouse myoblast cell line. However, we noticed that our transfection efficiency was decreased, and though this was improved after optimization of transfection conditions, it was more difficult to visually see the trend of the DUX4 reporter decreasing with increased DUX4-DBD. The C2C12 transfection visual results are shown below:

Our fluorescence plate reader assay for the C2C12 transfection demonstrated the same trend of increasing DUX4-DBD decreasing DUX4-fl binding:

Proof of function of DUX4-DBD-KRAB

In order to test the specific changes to DUX4-fl binding with the added KRAB domain, the previous HEK293T cell transfections were repeated with the same conditions but using DUX4-DBD-KRAB in addition to repeating the same condiitions with DUX4-DBD. The following graph shows the results from these transfections: DBD-KRAB consistently knocks down DUX4 fluorescent reporter activity at greater levels than DBD alone (note the logarithmic scale).

The following graph shows this same transfection but instead analyzing fold knockdown of the DUX4 mScarlet reporter, where greater knockdown corresponds to a decrease DUX4-fl binding to the reporter. mNG-P2A-DUX4-DBD-KRAB appears to cause a greater reporter knockdown than mNG-P2A-DUX4-DBD. However, this trend must still be more thoroughly tested with more repeats of the experiment and values collected at other ratios of DUX4-DBD to DUX4-fl.

We further characterized the DUX4-DBD-KRAB construct in C2C12 cells by repeating the transfection conditions. The following fluorescence images show that with DUX-DBD-KRAB, there appears to be a greater knockdown of reporter signal compared with only DUX4-DBD:

Results from our fluorescence plate reader assay comparing the two constructs are also shown in the graph below:

Applications

These experiments validate both the expression and function of the DUX4-DBD-KRAB construct. The DUX4 reporter used in these experiments provided a readout of DUX4-fl binding and its recruitment of RNA polymerase to initiate transcription of mScarlet, simulating the way that the transcriptional activating domain in DUX4-fl recruits transcriptional machinery to activate harmful downstream genes. Because the green fluorescence is correlated with the amount of DUX4-DBD-KRAB present, and as green fluorescence increases, red fluorescence decreases, this mNG-P2A-DUX4-DBD-KRAB coding sequence demonstrates promising steps towards overexpressing DUX4-DBD-KRAB to competitively inhibit DUX4-fl as a novel therapeutic approach for treating FSHD.

This construct, with its component parts all originating from fully human proteins, modulates the underlying cause of FSHD without the need to delete or inactivate the DUX4 gene itself, or introduce bacterial proteins. This approach of coupling competitive inhibition with transcriptional repression to control pathogenic gene expression has broader potential to extend to other diseases caused by gain-of-function mutations or misexpression of toxic proteins. The modularity of fusing engineered proteins with functional domains opens possibilities for novel gene therapies, expanding the potential to create treatments for more complex diseases that currently have no cure.

References

[1] U.S. National Library of Medicine. (n.d.). Dux4 double homeobox 4 [homo sapiens (human)] - gene - NCBI. National Center for Biotechnology Information. https://www.ncbi.nlm.nih.gov/gene/100288687

[2] Bosnakovski, D., Gearhart, M. D., Toso, E. A., Ener, E. T., Choi, S. H., & Kyba, M. (2018). Low level DUX4 expression disrupts myogenesis through deregulation of myogenic gene expression. Scientific Reports, 8(1). https://doi.org/10.1038/s41598-018-35150-8

[3] Mitsuhashi, H., Ishimaru, S., Homma, S., Yu, B., Honma, Y., Beermann, M. L., & Miller, J. B. (2018). Functional domains of the FSHD-associated DUX4 protein. Biology Open. https://doi.org/10.1242/bio.033977

[4]Himeda, C. L., Jones, T. I., & Jones, P. L. (2016). CRISPR/dCas9-mediated Transcriptional Inhibition Ameliorates the Epigenetic Dysregulation at D4Z4 and Represses DUX4-fl in FSH Muscular Dystrophy. Molecular therapy : the journal of the American Society of Gene Therapy, 24(3), 527–535. https://doi.org/10.1038/mt.2015.200