Control of Lineage-Specific Gene Expression by Functionalized gRNA Barcodes

Abstract

Lineage tracking delivers essential quantitative insight into dynamic, probabilistic cellular processes, such as somatic tumor evolution and differentiation. Methods for high diversity lineage quantitation rely on sequencing a population of DNA barcodes. However, manipulation of specific individual lineages is not possible with this approach. To address this challenge, we developed a functionalized lineage tracing tool, Control of Lineages by Barcode Enabled Recombinant Transcription (COLBERT), that enables high diversity lineage tracing and lineage-specific manipulation of gene expression. This modular platform utilizes expressed barcode gRNAs to both track cell lineages and direct lineage-specific gene expression.

Keywords: barcode; clonal dynamics; heterogeneity; lineage tracing; population variation.

Figure 1.

Lineage-specific activation of gene expression. (a) Generation and lineage specific gene activation of independent barcoded gRNA populations. Three unique barcodes were randomly generated following the GNSNWNSNWNSNWNSNWNSN template and assembled into lentiviral gRNA expression cassettes. Cell lines: HEK293T, Caco2, and MDA-MB-231 were independently transduced with the three different barcode gRNAs and selected for stable integration. The barcoded populations were then cotransfected with each of one of the Recall plasmids, R-A/B/C_GFP and the dCas9-VPR plasmid. GFP expression was assessed 48 h post transfection via flow cytometry. (b) View of the lineage specific expression components. The lineage-specific Recall Plasmid contains a 3× barcode of interest_PAM array and adjacent downstream miniCMV promoter_sfGFP. In the presence of the matching barcode gRNA/dCas9-VPR complex, binding of the barcode arrays by the transcriptional activator dCas9-VPR will drive expression of sfGFP. In the case of mismatching barcode gRNA/dCas9-VPR complex, binding of the barcode arrays will not occur and expression of sfGFP will not be driven. (c) staggered histograms comparing high GFP expression for instances of matching barcode gRNA/Recall plasmid and nominal expression for instances of mismatch. GFP expression was measured via flow cytometry. (d) Error load graphs showing percent positive population activation at a given error rate. (e) Time-lapse fluorescent imaging of Caco2 Bg-A and Bg-B populations transfected with Recall-A_GFP and Recall-A_BAX_GFP along with the red apoptotic marker, Annexin-V red. The matching Bg-A population transfected with R-A_BAX_GFP displays efficient gene expression with GFP expression starting at 12 h followed BAX induced apoptosis present at 24 h. The mismatched Bg-B population does not show either GFP expression or notable induction of apoptosis. (f) Quantification of the number of Annexin-V Red positive cells show that, relative to Bg-A/R-A_GFP, there is a significant fold increase apoptosis in the cell populations transfected with a matching Recall plasmid containing either BAX or the hyper active mutant BAX D71A. Additionally, neither population with a mismatching Recall plasmid containing BAX or the hyperactive mutant displayed an increase in apoptosis over the control.

Figure 2.

Isolation of a single lineage of interest within a high diversity population. (a) High diversity barcoded-gRNA HEK293T cell population was generated with a GNSNWNSNWNSNWNSNWNSN template. The HEK293T Bg-A population was spiked in with the high diversity Bg-Library population to obtain a 1% and 0.1% Bg-A mixed population. Bg-A cells were then isolated from the mixed population via cotransfection of the R-A_GFP plasmid and dCas9-VPR plasmid and cell sorted based off of GFP expression. (b) To initialize sorting gates and determine capture rate, the Bg-Library, Bg-A 1% and Bg-A 0.1% populations were transfected with the Recall components and flow analyzed. Sorting gates were set based off of the Bg-Library to have a 0% error-load, resulting in the capture of 0.51% and 0.062% of cells in the Bg-A 1% and Bg-A 0.1%, respectively. (c) Barcode sequencing of the Bg-library and Bg-A populations as well as the Bg-A 1% and 0.1% isolated populations. Barcode sequencing of the isolated Bg-A 1% and 0.1% population show 92.8% and 99.1% of the reads being Bg-A respectively, nearly identical to the 95.8% Bg-A reads from the sampled Bg-A population. (d) Example experimental workflow utilizing COLBERT. A population of cells are tagged with a library of expressed barcoded gRNAs. The barcoded population can subsequently go through a phenotypic selection of interest. The resulting selected population is then sampled for relative barcode abundance. The barcode analysis would inform lineages of interest and allow for lineage-specific Recall plasmids to be assembled and cotransfected with dCas9-VPR into the resultant population for lineage-specific gene expression. In this instance, GFP is expressed, allowing the lineage of interest to be cell sorted and isolated from the mixed population.