Selective Inactivation of Functional RNAs by Ribozyme-Catalyzed

Dec 2, 2016 - Authors would like to thank the members of the Burke lab for helpful discussion. ... Wang, Yen-Hsiang; Wei, Kathy Y.; Smolke, Christina ...
0 downloads 0 Views 2MB Size
Research Article pubs.acs.org/synthbio

Selective Inactivation of Functional RNAs by Ribozyme-Catalyzed Covalent Modification Raghav R. Poudyal,†,# Malak Benslimane,‡ Melissa P. Lokugamage,§ Mackenzie K. Callaway,§,∇ Seth Staller,† and Donald H. Burke*,†,∥,⊥,§ †

Department of Biochemistry, ‡Genetics Area Program, §Department of Biological Engineering, ∥Department of Molecular Microbiology and Immunology, and ⊥Bond Life Sciences Center, University of Missouri Columbia, Coumbia, Missouri, 65211, United States S Supporting Information *

ABSTRACT: The diverse functions of RNA provide numerous opportunities for programming biological circuits. We describe a new strategy that uses ribozyme K28min to covalently tag a specific nucleobase within an RNA or DNA target strand to regulate and selectively inactivate those nucleic acids. K28min variants with appropriately reprogrammed internal guide sequences efficiently tagged multiple sites from an mRNA and from aptamer and ribozyme targets. Upon covalent modification by the corresponding K28min variant, an ATP-binding aptamer lost all affinity for ATP, and the fluorogenic Mango aptamer lost its ability to activate fluorescence of its dye ligand. Modifying a hammerhead ribozyme near the catalytic core led to loss of almost all of its substrate-cleaving activity, but modifying the same hammerhead ribozyme within a tertiary stabilizing element that reduces magnesium dependence only impaired substrate cleavage at low magnesium concentration. Thus, ribozyme-mediated covalent modification can be used both to selectively inactivate and to fine-tune the activities of targeted functional RNAs, analogous to the effects of post-translational modifications of proteins. Ribozyme-catalyzed covalent modification could therefore be developed to regulate nucleic acids components of synthetic and natural circuits.

R

channeling in bacteria leading to increase yields of metabolites,8 and riboswitches have been engineered and assembled with fluorogenic RNA aptamers to build RNA-based biosensors.9 To regulate the performance of these RNA components, many synthetic circuits have relied on mechanisms of strand displacement, small molecule-induced allostery and RNA cleavage. In contrast, biological proteins are commonly regulated through post-translational covalent modifications, such as phosphorylation, acetylation, ADP-ribosylation, and ubiquitination. In principle, covalent modification could be similarly employed to regulate RNA molecules post-transcriptionally. We present a novel means of regulating functional RNAs by using ribozyme-mediated covalent modification. Specifically, we describe a modular ribozyme that can be engineered to covalently modify target nucleic acids and turn

NA molecules perform critical regulatory roles in all organisms, sometimes exploiting precise three-dimensional folds, and sometimes exploiting intra- or intermolecular base pairing. Because RNAs can be easily engineered for expression in cells, they provide attractive opportunities for building artificial biological parts. In particular, RNA enzymes (ribozymes) and small molecule binding RNA aptamers have been developed as sensors, regulators, and actuators. 1 Combinatorial selections2−4 have greatly expanded their diversity, utility5 and engineering opportunities far beyond the examples found in nature. Ribozymes, aptamers, and other RNA modules are increasingly utilized as effectors and for circuit readout. For example, genetic circuits that use ribozymes and aptamers have been used to fine-tune gene expression and metabolite flux,6 and fluorogenic RNA aptamers have enabled live cell imaging and real-time visualization of metabolite accumulation and transport.7 Scaffolding RNAs have been used to localize enzymes near each other to promote substrate © XXXX American Chemical Society

Received: August 9, 2016

A

DOI: 10.1021/acssynbio.6b00222 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Research Article

ACS Synthetic Biology

Figure 1. Design rationale for engineering cis-acting ribozyme K28(1−77)C into trans-acting ribozyme K28min. (A) Secondary structure of ribozyme K28(1−77)C. Yellow circle denotes the major modification site. (B) Reaction catalyzed by ribozyme K28(1−77)C and K28min. (C) Generic design for ribozyme K28min in which the internal guide sequence (IGS) (green) can be changed to program substrate specificity (blue). DNA, RNA, and 2′F-ribose acceptor substrates are all utilized equivalently. The target nucleotide for modification is highlighted by a yellow circle.

Product formation was monitored by denaturing electrophoresis on an organomercurial (APM) polyacrylamide gel, which retains thiophosphorylated nucleic acids at the APM interface while allowing molecules without an exposed sulfur to pass through.13,14 Even without optimization, 7 of the 11 engineered K28min ribozymes efficiently modified their model DNA substrates. Yields for most of these were indistinguishable from yields with the native d3.1 DNA substrate (Figure S1B). As observed previously,12 2′OH moieties are not required in the acceptor strand. These data indicate that ribozyme K28min can be engineered to target many other substrates with a reasonably high unoptimized initial hit rate. Targeting RNA Aptamers for Selective Inactivation by Ribozyme K28min. RNA components such as aptamers can regulate circuit behavior or report on circuit operation through many mechanisms. We reasoned that aptamers with strategically placed 5′-GGA-3′ triplets could be targeted for selective inactivation by K28min variants by using appropriately engineered IGS sequences. To this end, we first engineered K28min to target G8 within a well-studied ATP aptamer (Figure 2A,B) that has arisen in several independent selections.15−17 The G8 nucleotide is essential for aptamer function. In NMR structures, G8 acts as a hydrogen bond donor to the ATP ligand and as part of a “GNRA-like” structural motif in which the bound ATP ligand serves as the “A” in the fourth position of this structural element18,19 (Figure 2A,C). Aptamer RNA transcripts were incubated with ribozyme K28minATPapt and donor GTPγS. Ribozyme-mediated modification was evaluated on denaturing APM gels followed by

off or attenuate their activities. Ribozyme K28(1−77)C (Figure 1A)10,11 catalyzes both thiophosphoryl and nucleotidyl transfer to generate a large adduct on the second G of a required GGA triplet (Figure 1B). The acceptor and catalytic functions can be separated into two strands, making it a trans-acting ribozyme K28min (Figure 1C).12 We show here that ribozyme K28min can be engineered to modify diverse functional RNA molecules, such as an ATP-binding aptamer, the fluorogenic Mango aptamer and a hammerhead ribozyme. Importantly, the respective functions of these RNA components are attenuated or shut off upon modification. This is the first demonstration of ribozyme-mediated regulation of functional RNA through formation of stable adducts, and it sets the stage for building new ribozymes that similarly regulate other RNA devices.



RESULTS Facile Reprogramming of Target Specificity. In the two-stranded, trans-acting ribozyme K28min, the catalytic core binds to the acceptor strand through stems I and II, which act as an internal guide sequence (IGS).12 We reasoned that the IGS could be readily reprogrammed to target other RNA substrates (Figure 1C), and that the resulting stable adducts would interfere with the functions of those RNAs. As an initial test of this design concept, we altered the IGS to target each of the 11 GGA sites within enhanced green fluorescent protein (eGFP) mRNA. These ribozymes were incubated with GTPγS and with synthetic DNA substrates (19−22 nt) corresponding to each target site (Supporting Information Figure S1A). B

DOI: 10.1021/acssynbio.6b00222 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Research Article

ACS Synthetic Biology

Figure 2. Regulation of ATP-binding aptamer by ribozyme K28minATPapt. (A) Secondary structure of the ATP aptamer. Nucleotides involved in ligand recognition are shown in blue circles. The G8 residue targeted by K28min is denoted by the orange wedge. (B) Secondary structure of ribozyme/substrate complex for ATP aptamer. (C) The target residue is in position to hydrogen bond with the bound AMP ligand (PDB: 1RAW, image rendered in PyMol), in addition to stabilizing overall structure. (D) Affinity elution profiles for unmodified (green) and modified (red) ATP aptamer. Orange arrow indicates the start for elution with 1 mM ATP.

ribozyme (HHRz) variants have been designed to cleave themselves (cis) or other RNAs (trans) for numerous applications. Minimal HHRz’s require elevated concentrations of Mg2+ for optimal activity, while extended forms include tertiary interactions that stabilize them to be active in the lowMg2+ intracellular environment.25,26 Ribozyme RzB-4 was previously engineered to cleave the 5′ leader of HIV-1 transcripts near the primer-binding sequence at physiological magnesium concentrations and temperature.27 This HHRz is a variant of ribozyme RzB, whose crystal structure was recently reported,28,29 and it contains two 5′-GGA-3′ sites (Figure 4A). Site 1 lies within a hexaloop at the apex of stem II and provides part of the stabilizing tertiary interaction. Site 2 is close to the active site between stems II and III29 and includes nucleotide G12, which has been implicated as the general base in the ribozyme catalytic mechanism.28−32 RzB-4 transcripts were incubated with appropriate variants of ribozyme K28min and GTPγS donor to prepare HHRz strands modified at Site 1 or Site 2. Both modified and unmodified HHRz transcripts were gel extracted to further characterize their RNA cleavage properties. Formation of a stable adduct at site 2 greatly impaired substrate cleavage under all conditions, but especially at low magnesium concentration (Figure 4B). In contrast, when RzB-4 was modified at site I, the ribozyme was inactive at 0.5 mM MgCl2, but its activity was restored almost completely by raising magnesium concentration to 20 mM (Figure 4B,C). These data indicate that covalent modification of one ribozyme by another can be used not only for complete on−off switching, but also for fine-tuning the activity or metal ion dependence of the targeted RNA. Engineering Constraints. Finally, to guide future engineering of ribozyme K28min variants for specific applications,

extraction of both the unmodified and K28minATPapt-modified aptamer transcripts. These were then each passed through an ATP−agarose affinity resin. Unmodified aptamer was able to bind to the resin and was subsequently eluted by free ATP. In contrast, K28minATPapt-modified ATP aptamer was unable to bind to the ATP resin and came off during the load fractions (Figure 2D), establishing that the stable adduct blocked the aptamer’s ability to bind ATP. We next engineered K28min to modify aptamers that form fluorescent complexes with dyes. The recently described Baby Spinach20,22 and Mango21 aptamers activate fluorescence from otherwise nonfluorescent dyes by restricting their motion and stabilizing the planar form that favors fluorescence.23,24 Both aptamers bind their ligands through guanosine quadruplex structures. The Spinach aptamer quadruplex has one 5′-GGA-3′ site, and the Mango aptamer has three sites (Figure 3A). Among these sites, site 2 within the Mango aptamer was modified especially well by ribozyme K28minMango2 (Figure 3B). The functional effects of modification at Mango Site 2 were characterized further by gel extracting modified and unmodified aptamers and measuring their fluorescent properties. In the presence of unmodified Mango aptamer RNA, dye fluorescence was activated with the expected emission maximum near 530 nm. In contrast, the emission spectrum for the dye in the presence of Site 2-modified Mango aptamer is indistinguishable from that of buffer with or without dye (Figure 3C). As with the ATP aptamer above, this selective inactivation of the Mango aptamer provides proof of principle for using ribozyme-catalyzed covalent modification to “switch off” RNA aptamers. Site-Specific, Differential Regulation of a Hammerhead Ribozyme by Covalent Modification. Hammerhead C

DOI: 10.1021/acssynbio.6b00222 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Research Article

ACS Synthetic Biology

Figure 3. Regulation of fluorescent Mango aptamer by ribozyme K28minMango. (A) Secondary structure of Mango aptamer core (redrawn from ref 21. Copyright 2014 American Chemical Society). Residues forming the G quadruplex are shown in blue circles; numbers denote the two residues targeted by ribozyme K28min. Secondary structures of Mango-K28minMango complexes are shown in Figure S2. (B) Reactions were performed for the native K28min ribozyme/substrate complex (lanes A) and fluorescent aptamers Baby Spinach (lanes B), and Mango Aptamer (lanes C). Retention of radioactive signal at the APM interface (arrow) indicates thiophosphorylated product formed. Sizes of unreacted materials are indicated to the right of the gel. (C and D) Fluorescence of unmodified and K28min modified Mango aptamers; 200 nM thiazol orange-biotin dye was present in both tubes (C) and in three of the four spectra (D), as indicated.

of cis-acting ribozyme K28(1−77)C, these results point to the feasibility of optimizing ribozyme turnover through the further engineering of Stem II in trans-acting K28min variants.

we sought to define any sequence constraints that might exist beyond those established previously for the secondary structure of the parent ribozyme K28.10 First, single-nucleotide deletions were introduced at each of the 13 unpaired nucleotides of the ribozyme core within self-modifying ribozyme K28(1−77)C. Three of these internal deletions were tolerated especially well (last three nucleotides before stem III), suggesting that slightly smaller variants of K28min could be developed. Little or no product yield was observed for many of the remaining deletions (Figure S3), consistent with conservation patterns upon reselection from mutated libraries10 that indicated critical requirements for most of the unpaired nucleotides. Second, we explored the impact of topological rearrangements that extend the IGS to the 5′ side of the acceptor strand GGA triplet. Ribozyme-target pairs carrying an additional 8bp stem (“stem IV”) were active as long as several unpaired nucleotides were included between the core structure and the new stem IV. Little or no product was observed when stem IV was continuous with either stem III or the GGA target site (Suppl Figure 4). K28min variants with extended IGS sequences could potentially facilitate invasion into, and efficient targeting of, highly structured targets that sequester the GGA target site. Finally, truncations were introduced within stem II of ribozyme K28(1−77)C. Product yields were comparable to that of the original ribozyme for three of nine of these variants (Figure S5). Although these truncations were evaluated in the context



DISCUSSION AND CONCLUSIONS In this study we demonstrate utilization of an RNA enzyme to regulate and selectively inactivate other functional RNAs through formation of stable, bulky adducts. Ribozyme K28min can modify the second guanosine nucleobase within a 5′-GGA-3′ triplet. Specificity is readily controlled by programming complementary sequences into the ribozyme IGS. No intrinsic lower limit was observed in the allowable length of stem II of the IGS, and K28min could potentially be further reduced in size through deletions at certain unpaired nucleotides. We were able to modify 7 out of 11 sites that contained 5′-GGA-3′ derived from the eGFP mRNA sequence. The other four ribozymes may be limited by self-structure. Nevertheless, other reprogrammed ribozymes modified the ATP binding aptamer and the fluorogenic Mango aptamer and completely turned off their activities. Complete “on → off” switching could potentially be useful to control the fate of genetic circuits that require strong control over the expression and activity of RNA-based sensors and reporters. We also showed that trans-cleaving hammerhead ribozyme RzB-4 can be modified at two sites using K28min. When RzB-4 was modified within a loop that is required for tertiary interaction D

DOI: 10.1021/acssynbio.6b00222 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Research Article

ACS Synthetic Biology

Figure 4. Selective inactivation of trans-cleaving RzB-4 ribozyme by K28min. (A) Structure of the RzB-4 ribozyme/substrate complex. Cleavage substrate is shown in blue; black arrow denotes the cleavage site; numbers denote modification sites for ribozyme K28min. Secondary structures of RzB-4/K28minHHRz complexes are shown in Figure S2. Nucleotides in the loop of Stem II (black) interact with unpaired nucleotides in Stem I (green) to form tertiary interactions that are important for cleavage activity at low magnesium concentrations. Nucleotides shown in red are invariable nucleotides. (B) Denaturing PAGE of substrate RNA strand cleavage by modified and unmodified RzB-4 ribozymes at the indicated magnesium concentrations. Reactions were performed for 18 h. (C and D). Comparison of cleavage activity at early time points for unmodified RzB4 (red) and K28minSiteI-modified RzB-4 (blue) in 0.5 mM MgCl2 (C) and in 20 mM MgCl2 (D).

Louis). Radiolabeled nucleotides [γ-32P]ATP and [α-32P]CTP were purchased from PerkinElmer (Waltham, MA). Ribozymes and their substrates were either internally labeled by transcribing with 33 nM [α-32P] CTP or 5′ end-labeled by transcribing with nonradioactive CTP followed by dephosphorylation with alkaline phosphatase and treatment with [γ-32P]ATP and PNK (New England Biolabs). Trilayered APM gels containing 100 μg/mL [(N-acryloylamino)phenyl]mercuric chloride in the middle layer were prepared as described14,39 to capture sulfur-containing polynucleotides generated by thiophosphoryl transfer from GTPγS. Product Formation by K28(1−77)C Ribozyme Mutants. Internally labeled K28(1−77)C variants (12 000 to 24 000 cpm) were heated at 85 °C for 5 min. A 5X reaction buffer (1X = 25 mM Tris pH 8.0, 30 mM MgCl2, 10 μM CuCl2, 200 mM KCl and 15 mM NaCl) was added, and the tubes were incubated at room temperature for 5 min and moved to ice. Reactions were initiated by adding GTPγS to a final concentration of 1 mM. Tubes were then incubated at room temperature for 6 h unless stated otherwise. The reactions were quenched by adding an equal volume of 90% formamide and 50 mM EDTA, and products were separated on 8% or 10% denaturing trilayered APM gels (8 M urea). Autoradiographs were obtained with a FLA-9000 GE phosphorimager (FujiFilm) and analyzed with MultiGauge software (version 3.0). ATP Binding by ATP Aptamer. ATP affinity elution was performed by using a C8-linked ATP-agarose column (Sigma A2767). In brief, the ATP-agarose was resuspended in 1X wash buffer (25 mM Tris pH 7.5, 10 mM MgCl2, 200 mM KCl), then loaded into a 1 cc syringe to give approximately 200 μL of bed volume, with glass wool in the end of the syringe to prevent agarose from flowing out. The column was equilibrated by washing with 1 mL of washing buffer. 5′-Radiolabeled ATP

(site 1), it was essentially inactive at low magnesium concentration, but high magnesium concentrations restored the activity, indicating that only the divalent metal ion sensitivity had been affected by the covalent adduct. In contrast, when RzB-4 was modified close to the active site (site 2), its ability to cleave its RNA substrate was severely impaired at both high and low magnesium concentrations. These data indicate that modification of elements required for tertiary interactions can be utilized for fine-tuning the activity of desired function. Controllable RNA devices are an established paradigm for synthetic genetic circuits. Our study demonstrates that stable covalent adducts can be used to inactivate those devices. This mode of regulation is reminiscent of covalent, post-translational modifications of proteins, such as phosphorylation and acetylation, which are very common in biological regulatory pathways. Ribozymes for acyl33,34 and phosphoryl35−37 transfer have been isolated through in vitro evolution, and deoxyribozymes (DNAzymes) have been described that phosphorylate small peptides.38 Ribozyme K28min demonstrates the general feasibility of regulating RNA devices through ribozymecatalyzed, post-transcriptional RNA modification. In principle, nucleic acids for phosphoryl and acyl transfers and other reactions can be developed to regulate RNA devices or proteins in living and cell-free systems by identifying sequences that optimize multiple-turnover rates under appropriate reaction conditions. Methods. Oligodeoxynucleotides were purchased from Integrated DNA Technologies (Coralville, IA). Ribozymes and aptamers were transcribed in vitro using wild type or Y639F mutant phage T7 RNA polymerase, which were overexpressed as His-tagged fusion proteins in bacteria and purified in house. GTPγS was purchased from Sigma (St. E

DOI: 10.1021/acssynbio.6b00222 ACS Synth. Biol. XXXX, XXX, XXX−XXX

ACS Synthetic Biology aptamer was modified by ribozyme K28minATPapt and extracted from the APM layer by cutting the gel slice and soaking the slice in 300 mM NaOAc, 5 mM DTT. The supernatants were then ethanol precipitated and resuspended in water. Unreacted ATP aptamer was also extracted from the same gel to control for any variations introduced by the experimental method. The modified and unreacted ATP aptamers were resuspended in 50 μL of 1X wash buffer, heated at 85 °C for 2 min and cooled at room temperature for 10 min to allow proper folding, and then loaded onto the column. The column was washed 7 times with 300 μL of 1X wash buffer, followed by elution in fractions of 300 μL of 1X wash buffer containing 1 mM ATP. Radioactive signal for 32P was counted for each of the fractions in the scintillation counter (PerkinElmer Tri-Carb 2800TR), using a scintillation cocktail. Mango Aptamer Fluorescence. K28min-modified and unmodified Mango aptamer were resuspended in 10 mM sodium phosphate pH 7.2, 140 mM KCl, 1 mM MgCl2 and 0.05% Tween-20. The fluorescence emission was scanned from 508 to 650 nm in a 96-well plate using a PerkinElmer Enspire 2300 Multiplate Reader. The concentration of the ThiazoleOrange Biotin Dye (Gift from Peter Unrau, Simon Fraser University) was 200 nM. RzB-4 Activity Assays. Nonradiolabeled RzB-4 ribozyme was modified by K28min as above and extracted from the APM layer. The cleavage reactions were assembled with 400 nM ribozyme and