An RNA Motif That Enables Optozyme Control and Light‐Dependent Gene Expression in Bacteria and Mammalian Cells

Abstract The regulation of gene expression by light enables the versatile, spatiotemporal manipulation of biological function in bacterial and mammalian cells. Optoribogenetics extends this principle by molecular RNA devices acting on the RNA level whose functions are controlled by the photoinduced interaction of a light‐oxygen‐voltage photoreceptor with cognate RNA aptamers. Here light‐responsive ribozymes, denoted optozymes, which undergo light‐dependent self‐cleavage and thereby control gene expression are described. This approach transcends existing aptamer‐ribozyme chimera strategies that predominantly rely on aptamers binding to small molecules. The optozyme method thus stands to enable the graded, non‐invasive, and spatiotemporally resolved control of gene expression. Optozymes are found efficient in bacteria and mammalian cells and usher in hitherto inaccessible optoribogenetic modalities with broad applicability in synthetic and systems biology.


Figure S2. a)
Sequence and secondary structure of sequence 4, sequence 53, and sequence 58, dubbed motif 1, motif 2, and motif 3, respectively.b) Frequency analysis of motif 1, motif 2, and motif 3 of the NmPAL selection. [26]c) Fold-change regulation of gene expression of the optozymes in HEK293T cells.d) Interaction analysis of the self-cleavage deficient optozymes OzS4 and OzS5, the corresponding non-binding variants (OzS4N21, OzS5M21), the selfcleavage deficient hammerhead ribozyme (HHR(m)), motif 3 (58), and its non-binding variant 58M21with NmPAL in light or darkness.e) Regulation of gene expression by light of the optozymes in presence (mCherryPAL) or absence (mCherry) of PAL.norm.gMFI: normalized geometric mean fluorescence intensity f) Fold changes calculated from light vs. dark conditions from (e).g) Relative gene expression of EGFP under the control of HHR, HHR(m), OzS4 or OzS5 quantified by qPCR.UT: untransfected.N = 2, in duplicates.****: p ≤ 0.0001; **: p ≤ 0.01, *: p ≤ 0.05; all other differences were found being non-significant.-A).e) Singlet cells were detected using forward scatter height (FSC-H) vs. forward scatter width (FSC-W).f) EGFP positive cells were identified using FSC-H vs. fluorescein isothiocyanate area (FITC-A).Cells transfected with SH1 and incubated in darkness were set to 3.1 % EGFP positive cells and gating was applied to all other tested samples.Gating strategies to identify EBFP positive cells for CRISPR/dCas9 experiments.g) 30.000 events were recorded and visualized in an SSC-A vs. FSC-A dot plot.h) The main cell population was isolated to exclude cell debris.i) The background fluorescence of untransfected cells was set to 1% using a quadrant gate.The gating strategy was applied to all samples and the amount of EBFP activation was calculated from the percentage of EBFP positive cells given in the top-right corner.The FSC threshold was set to 5.000.HeLa cells were gated for single cells with an auto fluorescence of 1%.The activation was normalized using the feature scaling method.

Tables
GCGGCCGCGUACAUCCAGCUGAUGAGUCCCAAAUAGGACAAAAGGUUCAGCUGCGAGCCUCCUGGAUUCCAGG CGGCCGC Table S4.RNA molecules used in the RiboGreen interaction assay.In bold: motif 3 sequence; in bold and gray: mutation yielding inactive aptamer (M21); in bold and red: mutation yielding inactive HHR.

Figure S3 .
Figure S3.Gating strategies to identify eGFP positive cells for HHR experiments: a) 30.000 events were recorded and plotted in an SSC-A vs. FSC-A dot plot.b) The main cell population was isolated to exclude cell debris.c) The fluorescence threshold of the untransfected population was set to 1% using the quadrant gating method in the FITC-A vs. SSC-H plot.This gating scheme was applied to all the measured samples.Finally, the optozyme activity was elucidated by calculating the geometric mean fluorescence intensity (gMFI) of the threshold fluorescence.Gating strategies to identify EGFP positive cells for shRNA experiments: d) Cell debris was excluded using side scatter area (SSC-A) vs. forward scatter area (FSC-A).e) Singlet cells were detected using forward scatter height (FSC-H) vs. forward scatter width (FSC-W).f) EGFP positive cells were identified using FSC-H vs. fluorescein isothiocyanate area (FITC-A).Cells transfected with SH1 and incubated in darkness were set to 3.1 % EGFP positive cells and gating was applied to all other tested samples.Gating strategies to identify EBFP positive cells for CRISPR/dCas9 experiments.g) 30.000 events were recorded and visualized in an SSC-A vs. FSC-A dot plot.h) The main cell population was isolated to exclude cell debris.i) The background fluorescence of untransfected cells was set to 1% using a quadrant gate.The gating strategy was applied to all samples and the amount of EBFP activation was calculated from the percentage of EBFP positive cells given in the top-right corner.The FSC threshold was set to 5.000.HeLa cells were gated for single cells with an auto fluorescence of 1%.The activation was normalized using the feature scaling method.

Figure S4 .
Figure S4.Images from the photomask experiment, where cells were incubated in illuminated conditions.HEK293 cells stably expressing mCherry-PAL were transfected with a EGFP reporter containing HHR, HHRm, OzS4 or OzS4M21.The spatial repression of EGFP, the mCherry signal and a merged variant of the images is shown.White scale bar: 2000 µm.N = 2 in duplicates.The brightness of the mCherry and EGFP signal was digitally increased by 50 units to improve the visibility of the EGFP signal.Overall, the mCherry signal was found to be lower compared to the EGFP signal and is more intense at locations on the plate, where the cells tend to accumulate.Dark patches showing low mCherry signal could be due to lower cell densities in those areas.

Figure S5 .
Figure S5.Images from the photomask experiment, where cells were incubated in darkness.HEK293 cells stably expressing mCherry-PAL were transfected with a EGFP reporter containing HHR, HHRm, OzS4 or OzS4M21.The spatial expression of EGFP, the mCherry signal and a merged variant of the images is shown.White scale bar: 2000 µm.N = 2 in duplicates.The brightness of the mCherry and EGFP signal was digitally increased by 50 units to improve the visibility of the EGFP signal.Overall, the mCherry signal was found to be lower compared to the EGFP signal and is more intense at locations on the plate, where the cells tend to accumulate.Dark patches showing low mCherry signal could be due to lower cell densities in those areas.

Figure S6 .
Figure S6.Sequence and design of sgRNA with embedded motif 3.

Table S1 .
List of constructs of optozymes in bacterial system.In bold: motif 3 sequence; in bold and red:

Table S2 .
Oligonucleotides used for cloning of optozymes into the bacterial vector system

Table S3 .
List of optozyme constructs used for expression in mammalian cells.In bold: motif 3 sequence; in bold and gray: mutation yielding inactive aptamer (M21); in bold and red: mutation yielding inactive HHR.

Table S5 .
List of shRNA constructs.In bold: motif 3 sequence; in orange: Hinge region nucleotides.

Table S6 .
List of oligonucleotides for the construction of sgRNA 58 and sgRNA 58 sequence.In bold: motif 3 sequence.