Artificial riboswitches for gene expression and replication control of DNA and RNA viruses

Abstract

Aptazymes are small, ligand-dependent self-cleaving ribozymes that function independently of transcription factors and can be customized for induction by various small molecules. Here, we introduce these artificial riboswitches for regulation of DNA and RNA viruses. We hypothesize that they represent universally applicable tools for studying viral gene functions and for applications as a safety switch for oncolytic and live vaccine viruses. Our study shows that the insertion of artificial aptazymes into the adenoviral immediate early gene E1A enables small-molecule-triggered, dose-dependent inhibition of gene expression. Aptazyme-mediated shutdown of E1A expression translates into inhibition of adenoviral genome replication, infectious particle production, and cytotoxicity/oncolysis. These results provide proof of concept for the aptazyme approach for effective control of biological outcomes in eukaryotic systems, specifically in virus infections. Importantly, we also demonstrate aptazyme-dependent regulation of measles virus fusion protein expression, translating into potent reduction of progeny infectivity and virus spread. This not only establishes functionality of aptazymes in fully cytoplasmic genetic systems, but also implicates general feasibility of this strategy for application in viruses with either DNA or RNA genomes. Our study implies that gene regulation by artificial riboswitches may be an appealing alternative to Tet- and other protein-dependent gene regulation systems, based on their small size, RNA-intrinsic mode of action, and flexibility of the inducing molecule. Future applications range from gene analysis in basic research to medicine, for example as a safety switch for new generations of efficiency-enhanced oncolytic viruses.

Fig. 1.

Schematic outline of the strategy for aptazyme-mediated shutdown of viral genes and functions. (A) Key features of adenovirus and measles virus infection cycle. Target genes for aptazyme control and their influence on viral infection are indicated. (B) Shutdown of viral gene expression is achieved using a cis-acting synthetic RNA switch, specifically an OFF-switch aptazyme. This ligand-dependent self-cleaving ribozyme is inserted into the UTRs of the viral transcription unit. During infection, the viral genes are transcribed in the absence (expression ON), but not in the presence (expression OFF) of the small-molecule ligand (blue dots). This allows replication control of adenoviruses through regulation of the E1A gene and control of progeny infectivity of measles viruses by regulation of the F gene.

Fig. 2.

Aptazyme-mediated control of adenoviral E1A gene expression and DNA replication. (A) Schematic outline of oncolytic adenoviruses containing a regulatable E1A transcription unit. Insertion sites for the P1-F5 aptazyme (Az) or the parental ribozyme (Rz) were in the 5′- and/or 3′-UTR (nomenclature: 5′/3′/5′3′) of the E1A gene (orange box) driven by the adenoviral E1A promoter (blue arrow). Brown box, polyadenylation signal. ctrl, no aptazyme insertion. LITR/RITR, left/right inverted terminal repeat; Ψ, packaging signal; Δ24E1A, E1A gene with a 24-bp deletion for tumor selectivity; ΔE3, E3 region deleted (not required for virus replication); other viral genes and elements are not shown (dashed lines). (B and C) SK-MEL-28 cells were infected with the oncolytic adenovirus variants using 10 TCID₅₀/cell (B) or 1 TCID₅₀/cell (C). Cells were cultured in the absence (−) or presence (+) of 3 mM theophylline and harvested at indicated time points postinfection (p.i.). (B) For analysis of E1A expression, total lysates (20 μg) were assayed by Western blotting for E1A protein expression; human β-actin was detected as loading control. Multiple bands for E1A represent multiple splice variants. Time points of cell harvest were chosen according to replication kinetics of adenoviruses in these cells. Results for the Rz viruses are shown at 48 h p.i. to demonstrate stringent inhibition of gene expression by the ribozyme. (C) Viral genome copy numbers were determined by qPCR and are presented relative to cellular DNA content as determined for each sample individually. Symbols show mean values, error bars SD of three samples (n = 3). Dashed line, background signal of mock-infected cells.

Fig. 3.

Aptazyme-mediated control of adenoviral replication, particle production, and oncolysis. For analysis of aptazyme-regulated production of infectious viral progenies (A), viral cytotoxicity (B), or viral genome replication (C and D), SK-MEL-28 (A–C, melanoma), Colo 829 (D, melanoma), or NCH89 (D, glioblastoma) cells were infected using the indicated oncolytic adenovirus variants (see Fig. 2A). Infected cells were cultured in the absence (gray bar) or presence of 3 mM (black bar) theophylline added either after infection or 6 d p.i. (dashed bar). Cells were harvested at indicated time points p.i. Significance for theophylline-dependent regulation for individual oncolytic adenovirus variants is indicated with *P < 0.05, ***P < 0.001. Columns show mean values, error bars SD of three samples (n = 3). (A) Cells were infected at 1 TCID₅₀/cell. Total infectious progeny particles produced were determined by titration in HEK293 cells. Dashed line, input virus as measured for infected cells harvested 1 h p.i. (B) Cells were infected at 50 TCID₅₀/cell, allowing determination of viral cytotoxicity by staining of surviving cells 11 d p.i. with crystal violet. Crystal violet intensity was quantified and normalized to mock-infected cells (mock, set as 100% for theophylline-treated and untreated groups, dashed line). (C and D) Viral genome copy numbers were determined at indicated time points p.i. (C, 0.01 TCID₅₀/cell; D, 1 TCID₅₀/cell) by qPCR and are presented relative to cellular DNA content as determined for each sample individually. Dashed line, background signal of mock-infected cells. Infections were performed at low titers (C) to enable several virus replication cycles without eradication of cell monolayers. Time points of cell harvest in D differed between cell types because of differences in adenovirus replication kinetics. Adenovirus replication is faster in Colo 829 cells than in NCH89 or SK-MEL-28 cells as determined in a pilot time-course experiment.

Fig. 4.

Aptazyme-mediated control of measles virus infectivity and spread. (A) Schematic outline of F expression plasmids (pCG-F, pCG-F ctrl/5′3′/inAz5′3′) and measles virus (MV) genomes (MV-EGFP-F, MV-EGFP-F ctrl/5′3′/inAz5′3′) containing a regulatable F transcription unit. Insertion sites for the P1-F5 aptazyme (Az) or the inactive mutant thereof (inAz) were in the 5′- and 3′-UTRs of the F gene (yellow box). Blue arrow, CMV promoter; brown box, polyadenylation signal; ctrl, modified pCG-F plasmid with introduced restriction sites for Az/Rz insertion (red bar); red dot in inAz 5′3′, inactivating point mutation. For all MV variants, EGFP is inserted upstream of the N ORF in an additional transcription unit. N, P, M, H, and L represent MV genes. (B–E) Vero (B–D) and SK-MEL-28 (B–E) cells were infected with indicated MV variants at 0.03 ciu/cell or 0.3 ciu/cell (B, SK-MEL-28) and cultured in the absence or presence of 1.5 mM (B and E) or 3 mM (C) or various (D) theophylline concentrations. Cells were harvested at indicated time points. Columns/symbols show mean values, error bars SD of three samples (n = 3). (B) For analysis of MV F protein expression, total lysates (40 μg) were assayed by Western blotting for MV F expression; MV N and β-actin were detected as internal controls. Time points of cell harvest differed between cell types because of a more rapid cell lysis by measles virus infection in Vero cells. (C) Produced infectious progeny particles in cells and supernatant were determined by titration on Vero cells. (D) Concentration-dependent production of infectious progeny particles. Cells were harvested 48 h p.i. and subjected to titration of total infectious progenies on Vero cells. (E) For assaying long-term control of MV infection, infections at indicated low multiplicities of infection (MOI) were performed allowing several rounds of virus replication and titers were determined at the indicated time points. (D and E) Significance for theophylline-dependent regulation for individual MV variants is indicated with *P < 0.05, **P < 0.01, ***P < 0.001.