Thirty-five years of research into ribozymes and nucleic acid catalysis: where do we stand today?

Recent advances in enzymatic carbon-carbon bond formation

Enzymatic carbon-carbon (C-C) bond formation reactions have become an effective and invaluable tool for designing new biological and medicinal molecules, often with asymmetric features. This review provides a systematic overview of key C-C bond formation reactions and enzymes, with the focus of reaction mechanisms and recent advances. These reactions include the aldol reaction, Henry reaction, Knoevenagel condensation, Michael addition, Friedel-Crafts alkylation and acylation, Mannich reaction, Morita-Baylis-Hillman (MBH) reaction, Diels-Alder reaction, acyloin condensations via Thiamine Diphosphate (ThDP)-dependent enzymes, oxidative and reductive C-C bond formation, C-C bond formation through C1 resource utilization, radical enzymes for C-C bond formation, and other C-C bond formation reactions.

Software Infrastructure for Next-Generation QM/MM-ΔMLP Force Fields

We present software infrastructure for the design and testing of new quantum mechanical/molecular mechanical and machine-learning potential (QM/MM-ΔMLP) force fields for a wide range of applications. The software integrates Amber's molecular dynamics simulation capabilities with fast, approximate quantum models in the xtb package and machine-learning potential corrections in DeePMD-kit. The xtb package implements the recently developed density-functional tight-binding QM models with multipolar electrostatics and density-dependent dispersion (GFN2-xTB), and the interface with Amber enables their use in periodic boundary QM/MM simulations with linear-scaling QM/MM particle-mesh Ewald electrostatics. The accuracy of the semiempirical models is enhanced by including machine-learning correction potentials (ΔMLPs) enabled through an interface with the DeePMD-kit software. The goal of this paper is to present and validate the implementation of this software infrastructure in molecular dynamics and free energy simulations. The utility of the new infrastructure is demonstrated in proof-of-concept example applications. The software elements presented here are open source and freely available. Their interface provides a powerful enabling technology for the design of new QM/MM-ΔMLP models for studying a wide range of problems, including biomolecular reactivity and protein-ligand binding.

Snapshots of the first-step self-splicing of Tetrahymena ribozyme revealed by cryo-EM

Tetrahymena ribozyme is a group I intron, whose self-splicing is the result of two sequential ester-transfer reactions. To understand how it facilitates catalysis in the first self-splicing reaction, we used cryogenic electron microscopy (cryo-EM) to resolve the structures of L-16 Tetrahymena ribozyme complexed with a 11-nucleotide 5'-splice site analog substrate. Four conformations were achieved to 4.14, 3.18, 3.09 and 2.98 Å resolutions, respectively, corresponding to different splicing intermediates during the first enzymatic reaction. Comparison of these structures reveals structural alterations, including large conformational changes in IGS/IGSext (P1-P1ext duplex) and J5/4, as well as subtle local rearrangements in the G-binding site. These structural changes are required for the enzymatic activity of the Tetrahymena ribozyme. Our study demonstrates the ability of cryo-EM to capture dynamic RNA structural changes, ushering in a new era in the analysis of RNA structure-function by cryo-EM.

Scaling Catalytic Contributions of Small Self-Cleaving Ribozymes

Nucleolytic ribozymes utilize general acid-base catalysis to perform phosphodiester cleavage. In most ribozyme classes, a conserved active site guanosine is positioned to act as general base, thereby activating the 2'-OH group to attack the scissile phosphate (γ-catalysis). Here, we present an atomic mutagenesis study for the pistol ribozyme class. Strikingly, "general base knockout" by replacement of the guanine N1 atom by carbon results in only 2.7-fold decreased rate. Therefore, the common view that γ-catalysis critically depends on the N1 moiety becomes challenged. For pistol ribozymes we found that γ-catalysis is subordinate in overall catalysis, made up by two other catalytic factors (α and δ). Our approach allows scaling of the different catalytic contributions (α, β, γ, δ) with unprecedented precision and paves the way for a thorough mechanistic understanding of nucleolytic ribozymes with active site guanines.

Scaling Catalytic Contributions of Small Self-Cleaving Ribozymes

Nucleolytic ribozymes utilize general acid-base catalysis to perform phosphodiester cleavage. In most ribozyme classes, a conserved active site guanosine is positioned to act as general base, thereby activating the 2'-OH group to attack the scissile phosphate (γ-catalysis). Here, we present an atomic mutagenesis study for the pistol ribozyme class. Strikingly, "general base knockout" by replacement of the guanine N1 atom by carbon results in only 2.7-fold decreased rate. Therefore, the common view that γ-catalysis critically depends on the N1 moiety becomes challenged. For pistol ribozymes we found that γ-catalysis is subordinate in overall catalysis, made up by two other catalytic factors (α and δ). Our approach allows scaling of the different catalytic contributions (α, β, γ, δ) with unprecedented precision and paves the way for a thorough mechanistic understanding of nucleolytic ribozymes with active site guanines.

The Medicinal Chemistry of Artificial Nucleic Acids and Therapeutic Oligonucleotides

Nucleic acids play a central role in human biology, making them suitable and attractive tools for therapeutic applications. While conventional drugs generally target proteins and induce transient therapeutic effects, nucleic acid medicines can achieve long-lasting or curative effects by targeting the genetic bases of diseases. However, native oligonucleotides are characterized by low in vivo stability due to nuclease sensitivity and unfavourable physicochemical properties due to their polyanionic nature, which are obstacles to their therapeutic use. A myriad of synthetic oligonucleotides have been prepared in the last few decades and it has been shown that proper chemical modifications to either the nucleobase, the ribofuranose unit or the phosphate backbone can protect the nucleic acids from degradation, enable efficient cellular uptake and target localization ensuring the efficiency of the oligonucleotide-based therapy. In this review, we present a summary of structure and properties of artificial nucleic acids containing nucleobase, sugar or backbone modifications, and provide an overview of the structure and mechanism of action of approved oligonucleotide drugs including gene silencing agents, aptamers and mRNA vaccines.

Overview of Methods for Large-Scale RNA Synthesis

In recent years, it has become clear that RNA molecules are involved in almost all vital cellular processes and pathogenesis of human disorders. The functional diversity of RNA comes from its structural richness. Although composed of only four nucleotides, RNA molecules present a plethora of secondary and tertiary structures critical for intra and intermolecular contacts with other RNAs and ligands (proteins, small metabolites, etc.). In order to fully understand RNA function it is necessary to define its spatial structure. Crystallography, nuclear magnetic resonance and cryogenic electron microscopy have demonstrated considerable success in determining the structures of biologically important RNA molecules. However, these powerful methods require large amounts of sample. Despite their limitations, chemical synthesis and in vitro transcription are usually employed to obtain milligram quantities of RNA for structural studies, delivering simple and effective methods for large-scale production of homogenous samples. The aim of this paper is to provide an overview of methods for large-scale RNA synthesis with emphasis on chemical synthesis and in vitro transcription. We also present our own results of testing the efficiency of these approaches in order to adapt the material acquisition strategy depending on the desired RNA construct.

Control of RNA with quinone methide reversible acylating reagents

Caging RNA by polyacylation (cloaking) has been developed recently as a simple and rapid method to control the function of RNAs. Previous approaches for chemical reversal of acylation (uncloaking) made use of azide reduction followed by amine cyclization, requiring ∼2-4 h for the completion of cyclization. In new studies aimed at improving reversal rates and yields, we have designed novel acylating reagents that utilize quinone methide (QM) elimination for reversal. The QM de-acylation reactions were tested with two bioorthogonally cleavable motifs, azide and vinyl ether, and their acylation and reversal efficiencies were assessed with NMR and mass spectrometry on model small-molecule substrates as well as on RNAs. Successful reversal both with phosphines and strained alkenes was documented. Among the compounds tested, the azido-QM compound A-3 displayed excellent de-acylation efficiency, with t1/2 for de-acylation of less than an hour using a phosphine trigger. To test its function in RNA caging, A-3 was successfully applied to control EGFP mRNA translation in vitro and in HeLa cells. We expect that this molecular caging strategy can serve as a valuable tool for biological investigation and control of RNAs both in vitro and in cells.

Identification of over 200-fold more hairpin ribozymes than previously known in diverse circular RNAs

Self-cleaving ribozymes are catalytic RNAs that cut themselves at a specific inter-nucleotide linkage. They serve as a model of RNA catalysis, and as an important tool in biotechnology. For most of the nine known structural classes of self-cleaving ribozymes, at least hundreds of examples are known, and some are present in multiple domains of life. By contrast, only four unique examples of the hairpin ribozyme class are known, despite its discovery in 1986. We bioinformatically predicted 941 unique hairpin ribozymes of a different permuted form from the four previously known hairpin ribozymes, and experimentally confirmed several diverse predictions. These results profoundly expand the number of natural hairpin ribozymes, enabling biochemical analysis based on natural sequences, and suggest that a distinct permuted form is more biologically relevant. Moreover, all novel hairpins were discovered in metatranscriptomes. They apparently reside in RNA molecules that vary both in size—from 381 to 5170 nucleotides—and in protein content. The RNA molecules likely replicate as circular single-stranded RNAs, and potentially provide a dramatic increase in diversity of such RNAs. Moreover, these organisms have eluded previous attempts to isolate RNA viruses from metatranscriptomes—suggesting a significant untapped universe of viruses or other organisms hidden within metatranscriptome sequences.

Hovlinc is a recently evolved class of ribozyme found in human lncRNA

Although naturally occurring catalytic RNA molecules-ribozymes-have attracted a great deal of research interest, very few have been identified in humans. Here, we developed a genome-wide approach to discovering self-cleaving ribozymes and identified a naturally occurring ribozyme in humans. The secondary structure and biochemical properties of this ribozyme indicate that it belongs to an unidentified class of small, self-cleaving ribozymes. The sequence of the ribozyme exhibits a clear evolutionary path, from its appearance between ~130 and ~65 million years ago (Ma), to acquiring self-cleavage activity very recently, ~13-10 Ma, in the common ancestors of humans, chimpanzees and gorillas. The ribozyme appears to be functional in vivo and is embedded within a long noncoding RNA belonging to a class of very long intergenic noncoding RNAs. The presence of a catalytic RNA enzyme in lncRNA creates the possibility that these transcripts could function by carrying catalytic RNA domains.

RNA therapeutics in ophthalmology - translation to clinical trials

The use of RNA interference technology has proven to inhibit the expression of many target genes involved in the underlying pathogenesis of several diseases affecting various systems. First established in in vitro and later in animal studies, small interfering RNA (siRNA) and antisense oligonucleotide (ASO) therapeutics are now entering clinical trials with the potential of clinical translation to patients. Gene-silencing therapies have demonstrated promising responses in ocular disorders, predominantly due to the structure of the eye being a closed and compartmentalised organ. However, although the efficacy of such treatments has been observed in both preclinical studies and clinical trials, there are issues pertaining to the use of these drugs which require more extensive research with regards to the delivery and stability of siRNAs and ASOs. This would improve their use for long-term treatment regimens and alleviate the difficulties experienced by patients with ocular diseases. This review provides a detailed insight into the recent developments and clinical trials that have been conducted for several gene-silencing therapies, including ISTH0036, SYL040012, SYL1001, PF-04523655, Sirna-027, QR-110, QR-1123, QR-421a and IONIS-FB-L_RX in glaucoma, dry eye disease, age-related macular degeneration, diabetic macular oedema and various inherited retinal diseases. Our aim is to explore the potential of these drugs whilst evaluating their associated advantages and disadvantages, and to discuss the future translation of RNA therapeutics in ophthalmology.

Small molecule recognition of disease-relevant RNA structures

Targeting RNAs with small molecules represents a new frontier in drug discovery and development. The rich structural diversity of folded RNAs offers a nearly unlimited reservoir of targets for small molecules to bind, similar to small molecule occupancy of protein binding pockets, thus creating the potential to modulate human biology. Although the bacterial ribosome has historically been the most well exploited RNA target, advances in RNA sequencing technologies and a growing understanding of RNA structure have led to an explosion of interest in the direct targeting of human pathological RNAs. This review highlights recent advances in this area, with a focus on the design of small molecule probes that selectively engage structures within disease-causing RNAs, with micromolar to nanomolar affinity. Additionally, we explore emerging RNA-target strategies, such as bleomycin A5 conjugates and ribonuclease targeting chimeras (RIBOTACs), that allow for the targeted degradation of RNAs with impressive potency and selectivity. The compounds discussed in this review have proven efficacious in human cell lines, patient-derived cells, and pre-clinical animal models, with one compound currently undergoing a Phase II clinical trial and another that recently garnerd FDA-approval, indicating a bright future for targeted small molecule therapeutics that affect RNA function.

New Deoxyribozymes for the Native Ligation of RNA

Deoxyribozymes (DNAzymes) are small, synthetic, single-stranded DNAs capable of catalyzing chemical reactions, including RNA ligation. Herein, we report a novel class of RNA ligase deoxyribozymes that utilize 5'-adenylated RNA (5'-AppRNA) as the donor substrate, mimicking the activated intermediates of protein-catalyzed RNA ligation. Four new DNAzymes were identified by in vitro selection from an N₄₀ random DNA library and were shown to catalyze the intermolecular linear RNA-RNA ligation via the formation of a native 3'-5'-phosphodiester linkage. The catalytic activity is distinct from previously described RNA-ligating deoxyribozymes. Kinetic analyses revealed the optimal incubation conditions for high ligation yields and demonstrated a broad RNA substrate scope. Together with the smooth synthetic accessibility of 5'-adenylated RNAs, the new DNA enzymes are promising tools for the protein-free synthesis of long RNAs, for example containing precious modified nucleotides or fluorescent labels for biochemical and biophysical investigations.

Splitting aptamers and nucleic acid enzymes for the development of advanced biosensors

In analogy to split-protein systems, which rely on the appropriate fragmentation of protein domains, split aptamers made of two or more short nucleic acid strands have emerged as novel tools in biosensor set-ups. The concept relies on dissecting an aptamer into a series of two or more independent fragments, able to assemble in the presence of a specific target. The stability of the assembled structure can further be enhanced by functionalities that upon folding would lead to covalent end-joining of the fragments. To date, only a few aptamers have been split successfully, and application of split aptamers in biosensing approaches remains as promising as it is challenging. Further improving the stability of split aptamer target complexes and with that the sensitivity as well as efficient working modes are important tasks. Here we review functional nucleic acid assemblies that are derived from aptamers and ribozymes/DNAzymes. We focus on the thrombin, the adenosine/ATP and the cocaine split aptamers as the three most studied DNA split systems and on split DNAzyme assemblies. Furthermore, we extend the subject into split light up RNA aptamers used as mimics of the green fluorescent protein (GFP), and split ribozymes.

Catalytic Nucleic Acids: Biochemistry, Chemical Biology, Biosensors, and Nanotechnology

Since the initial discovery of ribozymes in the early 1980s, catalytic nucleic acids have been used in different areas. Compared with protein enzymes, catalytic nucleic acids are programmable in structure, easy to modify, and more stable especially for DNA. We take a historic view to summarize a few main interdisciplinary areas of research on nucleic acid enzymes that may have broader impacts. Early efforts on ribozymes in the 1980s have broken the notion that all enzymes are proteins, supplying new evidence for the RNA world hypothesis. In 1994, the first catalytic DNA (DNAzyme) was reported. Since 2000, the biosensor applications of DNAzymes have emerged and DNAzymes are particularly useful for detecting metal ions, a challenging task for enzymes and antibodies. Combined with nanotechnology, DNAzymes are key building elements for switches allowing dynamic control of materials assembly. The search for new DNAzymes and ribozymes is facilitated by developments in DNA sequencing and computational algorithms, further broadening our fundamental understanding of their biochemistry.

Molecular simulations of the pistol ribozyme: unifying the interpretation of experimental data and establishing functional links with the hammerhead ribozyme

The pistol ribozyme (Psr) is among the most recently discovered RNA enzymes and has been the subject of experiments aimed at elucidating the mechanism. Recent biochemical studies have revealed exciting clues about catalytic interactions in the active site not apparent from available crystallographic data. The present work unifies the interpretation of the existing body of structural and functional data on Psr by providing a dynamical model for the catalytically active state in solution from molecular simulation. Our results suggest that a catalytic Mg2+ ion makes inner-sphere contact with G33:N7 and outer-sphere coordination to the pro-RP of the scissile phosphate, promoting electrostatic stabilization of the dianionic transition state and neutralization of the developing charge of the leaving group through a metal-coordinated water molecule that is made more acidic by a hydrogen bond donated from the 2'OH of P32. This model is consistent with experimental activity-pH and mutagenesis data, including sensitivity to G33(7cG) and phosphorothioate substitution/metal ion rescue. The model suggests several experimentally testable predictions, including the response of cleavage activity to mutations at G42 and P32 positions in the ribozyme, and thio substitutions of the substrate in the presence of different divalent metal ions. Further, the model identifies striking similarities of Psr to the hammerhead ribozyme (HHr), including similar global fold, organization of secondary structure around an active site three-way junction, catalytic metal ion binding mode, and guanine general base. However, the specific binding mode and role of the Mg2+ ion, as well as a conserved 2'-OH in the active site, are interrelated but subtly different between the ribozymes.

The origin of the central dogma through conflicting multilevel selection

The central dogma of molecular biology rests on two kinds of asymmetry between genomes and enzymes: informatic asymmetry, where information flows from genomes to enzymes but not from enzymes to genomes; and catalytic asymmetry, where enzymes provide chemical catalysis but genomes do not. How did these asymmetries originate? Here, we show that these asymmetries can spontaneously arise from conflict between selection at the molecular level and selection at the cellular level. We developed a model consisting of a population of protocells, each containing a population of replicating catalytic molecules. The molecules are assumed to face a trade-off between serving as catalysts and serving as templates. This trade-off causes conflicting multilevel selection: serving as catalysts is favoured by selection between protocells, whereas serving as templates is favoured by selection between molecules within protocells. This conflict induces informatic and catalytic symmetry breaking, whereby the molecules differentiate into genomes and enzymes, establishing the central dogma. We show mathematically that the symmetry breaking is caused by a positive feedback between Fisher’s reproductive values and the relative impact of selection at different levels. This feedback induces a division of labour between genomes and enzymes, provided variation at the molecular level is sufficiently large relative to variation at the cellular level, a condition that is expected to hinder the evolution of altruism. Taken together, our results suggest that the central dogma is a logical consequence of conflicting multilevel selection.

From biomarkers to therapeutic targets-the promises and perils of long non-coding RNAs in cancer

Biomarker-driven personalized cancer therapy is a field of growing interest, and several molecular tests have been developed to detect biomarkers that predict, e.g., response of cancers to particular therapies. Identification of these molecules and understanding their molecular mechanisms is important for cancer prognosis and the development of therapeutics for late stage diseases. In the past, significant efforts have been placed on the discovery of protein or DNA-based biomarkers while only recently the class of long non-coding RNA (lncRNA) has emerged as a new category of biomarker. The mammalian genome is pervasively transcribed yielding a vast amount of non-protein-coding RNAs including lncRNAs. Hence, these transcripts represent a rich source of information that has the potential to significantly contribute to precision medicine in the future. Importantly, many lncRNAs are differentially expressed in carcinomas and they are emerging as potent regulators of tumor progression and metastasis. Here, we will highlight prime examples of lncRNAs that serve as marker for cancer progression or therapy response and which might represent promising therapeutic targets. Furthermore, we will introduce lncRNA targeting tools and strategies, and we will discuss potential pitfalls in translating these into clinical trials.

Therapeutic Potential of Ribozymes

Recent advances in RNA engineering during the last two decades have supported the development of RNA-based therapeutics targeting a variety of human diseases. The broad scope of these emerging drugs clearly demonstrates the versatility of RNA. Ribozymes have been seen as promising candidates in this area. However, efficient intracellular application of ribozymes remains challenging, and other strategies appear to have outperformed ribozymes as molecular drugs. Nevertheless, trans-cleaving ribozymes have been applied for specific cleavage of target mRNAs in order to inhibit undesired gene expression. Furthermore, ribozymes have been engineered to allow site-directed RNA sequence alterations, enabling the correction of genetic misinformation at the RNA level. This chapter provides an overview of ribozyme-based strategies, highlighting the promises and pitfalls for potential therapeutic applications.

RNA-Cleaving DNAzymes: Old Catalysts with New Tricks for Intracellular and In Vivo Applications

DNAzymes are catalytically active DNA molecules that are normally isolated through in vitro selection methods, among which RNA-cleaving DNAzymes that catalyze the cleavage of a single RNA linkage embedded within a DNA strand are the most studied group of this DNA enzyme family. Recent advances in DNA nanotechnology and engineering have generated many RNA-cleaving DNAzymes with unique recognition and catalytic properties. Over the past decade, numerous RNA-cleaving, DNAzymes-based functional probes have been introduced into many research areas, such as in vitro diagnostics, intracellular imaging, and in vivo therapeutics. This review focus on the fundamental insight into RNA-Cleaving DNAzymes and technical tricks for their intracellular and in vivo applications, highlighting the recent progress in the clinical trial of RNA-Cleaving DNAzymes with selected examples. The challenges and opportunities for the future translation of RNA-cleaving DNAzymes for biomedicine are also discussed.

Structural and Biochemical Properties of Novel Self-Cleaving Ribozymes

Fourteen well-defined ribozyme classes have been identified to date, among which nine are site-specific self-cleaving ribozymes. Very recently, small self-cleaving ribozymes have attracted renewed interest in their structure, biochemistry, and biological function since the discovery, during the last three years, of four novel ribozymes, termed twister, twister sister, pistol, and hatchet. In this review, we mainly address the structure, biochemistry, and catalytic mechanism of the novel ribozymes. They are characterized by distinct active site architectures and divergent, but similar, biochemical properties. The cleavage activities of the ribozymes are highly dependent upon divalent cations, pH, and base-specific mutations, which can cause changes in the nucleotide arrangement and/or electrostatic potential around the cleavage site. It is most likely that a guanine and adenine in close proximity of the cleavage site are involved in general acid-base catalysis. In addition, metal ions appear to play a structural rather than catalytic role although some of their crystal structures have shown a direct metal ion coordination to a non-bridging phosphate oxygen at the cleavage site. Collectively, the structural and biochemical data of the four newest ribozymes could contribute to advance our mechanistic understanding of how self-cleaving ribozymes accomplish their efficient site-specific RNA cleavages.

Two genetic codes: Repetitive syntax for active non-coding RNAs; non-repetitive syntax for the DNA archives

Current knowledge of the RNA world indicates 2 different genetic codes being present throughout the living world. In contrast to non-coding RNAs that are built of repetitive nucleotide syntax, the sequences that serve as templates for proteins share-as main characteristics-a non-repetitive syntax. Whereas non-coding RNAs build groups that serve as regulatory tools in nearly all genetic processes, the coding sections represent the evolutionarily successful function of the genetic information storage medium. This indicates that the differences in their syntax structure are coherent with the differences of the functions they represent. Interestingly, these 2 genetic codes resemble the function of all natural languages, i.e., the repetitive non-coding sequences serve as appropriate tool for organization, coordination and regulation of group behavior, and the non-repetitive coding sequences are for conservation of instrumental constructions, plans, blueprints for complex protein-body architecture. This differentiation may help to better understand RNA group behavioral motifs.

Synthetic Riboswitches: From Plug and Pray toward Plug and Play

In synthetic biology, metabolic engineering, and gene therapy, there is a strong demand for orthogonal or externally controlled regulation of gene expression. Here, RNA-based regulatory devices represent a promising emerging alternative to proteins, allowing a fast and direct control of gene expression, as no synthesis of regulatory proteins is required. Besides programmable ribozyme elements controlling mRNA stability, regulatory RNA structures in untranslated regions are highly interesting for engineering approaches. Riboswitches are especially well suited, as they show a modular composition of sensor and response elements, allowing a free combination of different modules in a plug-and-play-like mode. The sensor or aptamer domain specifically interacts with a trigger molecule as a ligand, modulating the activity of the adjacent response domain that controls the expression of the genes located downstream, in most cases at the level of transcription or translation. In this review, we discuss the recent advances and strategies for designing such synthetic riboswitches based on natural or artificial components and readout systems, from trial-and-error approaches to rational design strategies. As the past several years have shown dramatic development in this fascinating field of research, we can give only a limited overview of the basic riboswitch design principles that is far from complete, and we apologize for not being able to consider every successful and interesting approach described in the literature.

Programmable Genome Editing Tools and their Regulation for Efficient Genome Engineering

Targeted genome editing has become a powerful genetic tool for studying gene function or for modifying genomes by correcting defective genes or introducing genes. A variety of reagents have been developed in recent years that can generate targeted double-stranded DNA cuts which can be repaired by the error-prone, non-homologous end joining repair system or via the homologous recombination-based double-strand break repair pathway provided a suitable template is available. These genome editing reagents require components for recognizing a specific DNA target site and for DNA-cleavage that generates the double-stranded break. In order to reduce potential toxic effects of genome editing reagents, it might be desirable to control the in vitro or in vivo activity of these reagents by incorporating regulatory switches that can reduce off-target activities and/or allow for these reagents to be turned on or off. This review will outline the various genome editing tools that are currently available and describe the strategies that have so far been employed for regulating these editing reagents. In addition, this review will examine potential regulatory switches/strategies that can be employed in the future in order to provide temporal control for these reagents.

Breeding novel solutions in the brain: a model of Darwinian neurodynamics

Background: The fact that surplus connections and neurons are pruned during development is well established. We complement this selectionist picture by a proof-of-principle model of evolutionary search in the brain, that accounts for new variations in theory space. We present a model for Darwinian evolutionary search for candidate solutions in the brain.

Methods: We combine known components of the brain – recurrent neural networks (acting as attractors), the action selection loop and implicit working memory – to provide the appropriate Darwinian architecture. We employ a population of attractor networks with palimpsest memory. The action selection loop is employed with winners-share-all dynamics to select for candidate solutions that are transiently stored in implicit working memory.

Results: We document two processes: selection of stored solutions and evolutionary search for novel solutions. During the replication of candidate solutions attractor networks occasionally produce recombinant patterns, increasing variation on which selection can act. Combinatorial search acts on multiplying units (activity patterns) with hereditary variation and novel variants appear due to (i) noisy recall of patterns from the attractor networks, (ii) noise during transmission of candidate solutions as messages between networks, and, (iii) spontaneously generated, untrained patterns in spurious attractors.

Conclusions: Attractor dynamics of recurrent neural networks can be used to model Darwinian search. The proposed architecture can be used for fast search among stored solutions (by selection) and for evolutionary search when novel candidate solutions are generated in successive iterations. Since all the suggested components are present in advanced nervous systems, we hypothesize that the brain could implement a truly evolutionary combinatorial search system, capable of generating novel variants.

Breeding novel solutions in the brain: a model of Darwinian neurodynamics

Background: The fact that surplus connections and neurons are pruned during development is well established. We complement this selectionist picture by a proof-of-principle model of evolutionary search in the brain, that accounts for new variations in theory space. We present a model for Darwinian evolutionary search for candidate solutions in the brain. Methods: We combine known components of the brain - recurrent neural networks (acting as attractors), the action selection loop and implicit working memory - to provide the appropriate Darwinian architecture. We employ a population of attractor networks with palimpsest memory. The action selection loop is employed with winners-share-all dynamics to select for candidate solutions that are transiently stored in implicit working memory. Results: We document two processes: selection of stored solutions and evolutionary search for novel solutions. During the replication of candidate solutions attractor networks occasionally produce recombinant patterns, increasing variation on which selection can act. Combinatorial search acts on multiplying units (activity patterns) with hereditary variation and novel variants appear due to (i) noisy recall of patterns from the attractor networks, (ii) noise during transmission of candidate solutions as messages between networks, and, (iii) spontaneously generated, untrained patterns in spurious attractors. Conclusions: Attractor dynamics of recurrent neural networks can be used to model Darwinian search. The proposed architecture can be used for fast search among stored solutions (by selection) and for evolutionary search when novel candidate solutions are generated in successive iterations. Since all the suggested components are present in advanced nervous systems, we hypothesize that the brain could implement a truly evolutionary combinatorial search system, capable of generating novel variants.