Anja van Brabant Smith

Advances in CRISPR-Cas9 genome engineering: lessons learned from RNA interference

The discovery that the machinery of the Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-Cas9 bacterial immune system can be re-purposed to easily create deletions, insertions and replacements in the mammalian genome has revolutionized the field of genome engineering and re-invigorated the field of gene therapy. Many parallels have been drawn between the newly discovered CRISPR-Cas9 system and the RNA interference (RNAi) pathway in terms of their utility for understanding and interrogating gene function in mammalian cells. Given this similarity, the CRISPR-Cas9 field stands to benefit immensely from lessons learned during the development of RNAi technology. We examine how the history of RNAi can inform todays challenges in CRISPR-Cas9 genome engineering such as efficiency, specificity, high-throughput screening and delivery for in vivo and therapeutic applications.

Systematic evaluation of antibody-mediated siRNA delivery using an industrial platform of THIOMAB–siRNA conjugates

Delivery of siRNA is a key hurdle to realizing the therapeutic promise of RNAi. By targeting internalizing cell surface antigens, antibody–siRNA complexes provide a possible solution. However, initial reports of antibody–siRNA complexes relied on non-specific charged interactions and have not been broadly applicable. To assess and improve this delivery method, we built on an industrial platform of therapeutic antibodies called THIOMABs, engineered to enable precise covalent coupling of siRNAs. We report that such coupling generates monomeric antibody–siRNA conjugates (ARCs) that retain antibody and siRNA activities. To broadly assess this technology, we generated a battery of THIOMABs against seven targets that use multiple internalization routes, enabling systematic manipulation of multiple parameters that impact delivery. We identify ARCs that induce targeted silencing in vitro and extend tests to target prostate carcinoma cells following systemic administration in mouse models. However, optimal silencing was restricted to specific conditions and only observed using a subset of ARCs. Trafficking studies point to ARC entrapment in endocytic compartments as a limiting factor, independent of the route of antigen internalization. Our broad characterization of multiple parameters using therapeutic-grade conjugate technology provides a thorough assessment of this delivery technology, highlighting both examples of success as well as remaining challenges.

The ORFeome Collaboration: a genome-scale human ORF-clone resource

To the Editor: Here we describe the ORFeome Collaboration (OC) open reading frame (ORF) clone collection, created by the OC (http://www.orfeomecollaboration.org/), an international collaboration of academic and commercial groups committed to providing genome-scale clone resources for human genes via worldwide commercial and academic clone distributors. Proteins are the predominant functional modules determining the fate of cells, tissues and organisms. An encyclopedic understanding of cellular physiology requires protein expression for proteinprotein interaction screening, cellular functional screening, validation of knockout and knockdown phenotypes, and numerous other approaches. Performing such studies on individual proteins or at the proteome scale requires a comprehensive collection of human protein expression clones. Our collection comprises ORF clones (Supplementary Note) and covers 17,154 RefSeq and Ensembl genes, nearly 73% of human RefSeq genes (http://www.ncbi.nlm.nih.gov/refseq/rsg/) and 79% of the highly curated Consensus Coding DNA Sequence Project (CCDS) human genes (http://www.ncbi.nlm.nih.gov/CCDS/ CcdsBrowse.cgi) (Fig. 1a and Supplementary Data). The collection includes clones of transcript variants for 6,304 (37%) of those genes. All major functional categories of human genes are substantially represented (Fig. 1b). All clones are provided in the Gateway vector format (Life Technologies), permitting high-throughput, precise and directional transfer of ORFs to a large variety of vectors for protein expression in biological systems such as Escherichia coli, yeast and mammals or using cell-free protein expression1 (Supplementary Note). OC clones were generated primarily by PCR amplification of the ORF from full-length, sequence-verified human cDNA clones of the Mammalian Gene Collection2 or the German cDNA Consortium3; ORFs were also prepared by directed RT-PCR cloning4 or DNA synthesis2. All 5′ and 3′ untranslated regions were excluded, permitting direct expression of ORFs as fusions to aminoor carboxy-terminal polypeptides, or as native protein, after transfer to a Gatewayexpression vector1. The clones are designed to maintain the correct reading frame for both aminoand carboxy-fusion proteins. Among all genes represented in the OC collection, 64% of clones are without stop codons, 5% have stop codons, and 31% are present in both versions. Each OC clone was isolated from a single colony and is fully sequenced. Individual clone sequences have been deposited in the GenBank, EMBL and DDBJ databases. The OC website provides a searchable database with annotation of all OC clones, their respective genes, and clone confidence levels based on CCDS and RefSeq annotations (Supplementary Note) along with links to the UCSC and RIKEN browsers (http://genome.ucsc.edu/cgi-bin/hgGateway and http://fantom.gsc.riken.jp/zenbu/gLyphs/#config), which provide graphical representations of the gene structures and transcripts. OC clones are distributed via a good faith agreement, giving unrestricted clone access to all scientists worldwide. The OC website lists OC clone distributors. The value of the OC resource has been demonstrated in numerous studies covering a broad range of applications. These include large-scale binary protein-protein interaction mapping5, production of recombinant human proteins6, mapping of co-complex associations, fluorescent protein tagging for human protein localization in mammalian cells and microscopy-based functional screening of proteins, development of disease-specific protein interaction Figure 1 | RefSeq and Ensembl genes and functional gene categories represented in the OC. (a) Numbers of protein-coding genes represented in the OC collection from RefSeq (blue) and Ensembl (green) gene catalogs. The table summarizes these numbers, together with OC coverage for RefSeq-only and Ensembl-only genes. (b) Numbers of human RefSeq genes represented in the OC collection versus in the human genome, compared in nine functional categories; percentages of genes in the OC are presented above the bars. The methods used to calculate the gene numbers in each category are explained in the Supplementary Note and contrasted to the standard Gene Ontology categories. An expanded list of the top Gene Ontology categories is also provided in the Supplementary Note. The data underlying the graphs are provided as Supplementary Data. a

Systematic analysis of CRISPR-Cas9 mismatch tolerance reveals low levels of off-target activity.

The discovery that the bacterial clustered regularly interspaced short palindromic repeats (CRISPR)-CRISPR-associated protein 9 (Cas9) acquired immune system can be utilized to create double-strand breaks (DSBs) in eukaryotic genomes has resulted in the ability to create genomic changes more easily than with other genome engineering techniques. While there is significant potential for the CRISPR-Cas9 system to advance basic and applied research, several unknowns remain, including the specificity of the RNA-directed DNA cleavage by the small targeting RNA, the CRISPR RNA (crRNA). Here we describe a novel synthetic RNA approach that allows for high-throughput gene editing experiments. This was used with a functional assay for protein disruption to perform high-throughput analysis of crRNA activity and specificity. We performed a comprehensive test of target cleavage using crRNAs that contain one and two nucleotide mismatches to the DNA target in the 20mer targeting region of the crRNA, allowing for the evaluation of hundreds of potential mismatched target sites without the requirement for the off-target sequences and their adjacent PAMs to be present in the genome. Our results demonstrate that while many crRNAs are functional, less than 5% of crRNAs with two mismatches to their target are effective in gene editing; this suggests an overall high level of functionality but low level of off-targeting.

Optimized PCR Conditions and Increased shRNA Fold Representation Improve Reproducibility of Pooled shRNA Screens

RNAi screening using pooled shRNA libraries is a valuable tool for identifying genetic regulators of biological processes. However, for a successful pooled shRNA screen, it is imperative to thoroughly optimize experimental conditions to obtain reproducible data. Here we performed viability screens with a library of ∼10 000 shRNAs at two different fold representations (100- and 500-fold at transduction) and report the reproducibility of shRNA abundance changes between screening replicates determined by microarray and next generation sequencing analyses. We show that the technical reproducibility between PCR replicates from a pooled screen can be drastically improved by ensuring that PCR amplification steps are kept within the exponential phase and by using an amount of genomic DNA input in the reaction that maintains the average template copies per shRNA used during library transduction. Using these optimized PCR conditions, we then show that higher reproducibility of biological replicates is obtained by both microarray and next generation sequencing when screening with higher average shRNA fold representation. shRNAs that change abundance reproducibly in biological replicates (primary hits) are identified from screens performed with both 100- and 500-fold shRNA representation, however a higher percentage of primary hit overlap between screening replicates is obtained from 500-fold shRNA representation screens. While strong hits with larger changes in relative abundance were generally identified in both screens, hits with smaller changes were identified only in the screens performed with the higher shRNA fold representation at transduction.

Versatility of chemically synthesized guide RNAs for CRISPR-Cas9 genome editing

The CRISPR-Cas9 system has become the most popular and efficient method for genome engineering in mammalian cells. The Streptococcus pyogenes Cas9 nuclease can function with two types of guide RNAs: the native dual crRNA and tracrRNA (crRNA:tracrRNA) or a chimeric single guide RNA (sgRNA). Although sgRNAs expressed from a DNA vector are predominant in the literature, guide RNAs can be rapidly generated by chemical synthesis and provide equivalent functionality in gene editing experiments. This review highlights the attributes and advantages of chemically synthesized guide RNAs including the incorporation of chemical modifications to enhance gene editing efficiencies in certain applications. The use of synthetic guide RNAs is also uniquely suited to genome-scale high throughput arrayed screening, particularly when using complex phenotypic assays for functional genomics studies. Finally, the use of synthetic guide RNAs along with DNA-free sources of Cas9 (mRNA or protein) allows for transient CRISPR-Cas9 presence in the cell, thereby resulting in a decreased probability of off-target events.

High-content analysis screening for cell cycle regulators using arrayed synthetic crRNA libraries

The CRISPR-Cas9 system has been utilized for large-scale, loss-of-function screens mainly using lentiviral pooled formats and cell-survival phenotypic assays. Screening in an arrayed format expands the types of phenotypic readouts that can be used to now include high-content, morphology-based assays, and with the recent availability of synthetic crRNA libraries, new studies are emerging. Here, we use a cell cycle reporter cell line to perform an arrayed, synthetic crRNA:tracrRNA screen targeting 169 genes (>600 crRNAs) and used high content analysis (HCA) to identify genes that regulate the cell cycle. Seven parameters were used to classify cells into cell cycle categories and multiple parameters were combined using a new analysis technique to identify hits. Comprehensive hit follow-up experiments included target gene expression analysis, confirmation of DNA insertions/deletions, and validation with orthogonal reagents. Our results show that most hits had three or more independent crRNAs per gene that demonstrated a phenotype with consistent individual parameters, indicating that our screen produced high-confidence hits with low off-target effects and allowed us to identify hits with more subtle phenotypes. The results of our screen demonstrate the power of using arrayed, synthetic crRNAs for functional phenotypic screening using multiparameter HCA assays.

Amide linkages mimic phosphates in RNA interactions with proteins and are well tolerated in the guide strand of short interfering RNAs.

Abstract While the use of RNA interference (RNAi) in molecular biology and functional genomics is a well-established technology, in vivo applications of synthetic short interfering RNAs (siRNAs) require chemical modifications. We recently found that amides as non-ionic replacements for phosphodiesters may be useful modifications for optimization of siRNAs. Herein, we report a comprehensive study of systematic replacement of a single phosphate with an amide linkage throughout the guide strand of siRNAs. The results show that amides are surprisingly well tolerated in the seed and central regions of the guide strand and increase the silencing activity when placed between nucleosides 10 and 12, at the catalytic site of Argonaute. A potential explanation is provided by the first crystal structure of an amide-modified RNA–DNA with Bacillus halodurans RNase H1. The structure reveals how small changes in both RNA and protein conformation allow the amide to establish hydrogen bonding interactions with the protein. Molecular dynamics simulations suggest that these alternative binding modes may compensate for interactions lost due to the absence of a phosphodiester moiety. Our results suggest that an amide can mimic important hydrogen bonding interactions with proteins required for RNAi activity and may be a promising modification for optimization of biological properties of siRNAs.

The Power Decoder Simulator for the Evaluation of Pooled shRNA Screen Performance

RNA interference screening using pooled, short hairpin RNA (shRNA) is a powerful, high-throughput tool for determining the biological relevance of genes for a phenotype. Assessing an shRNA pooled screen’s performance is difficult in practice; one can estimate the performance only by using reproducibility as a proxy for power or by employing a large number of validated positive and negative controls. Here, we develop an open-source software tool, the Power Decoder simulator, for generating shRNA pooled screening experiments in silico that can be used to estimate a screen’s statistical power. Using the negative binomial distribution, it models both the relative abundance of multiple shRNAs within a single screening replicate and the biological noise between replicates for each individual shRNA. We demonstrate that this simulator can successfully model the data from an actual laboratory experiment. We then use it to evaluate the effects of biological replicates and sequencing counts on the performance of a pooled screen, without the necessity of gathering additional data. The Power Decoder simulator is written in R and Python and is available for download under the GNU General Public License v3.0.

A Single Amide Linkage in the Passenger Strand Suppresses Its Activity and Enhances Guide Strand Targeting of siRNAs

Potential in vivo applications of RNA interference (RNAi) require suppression of various off-target activities. Herein, we report that replacement of a single phosphate linkage between the first and second nucleosides of the passenger strand with an amide linkage almost completely abolished its undesired activity and restored the desired activity of guide strands that had been compromised by unfavorable amide modifications. Molecular modeling suggested that the observed effect was most likely due to suppressed loading of the amide-modified strand into Ago2 caused by inability of amide to adopt the conformation required for the backbone twist that docks the first nucleotide of the guide strand in the MID domain of Ago2. Eliminating off-target activity of the passenger strand will be important for improving therapeutic potential of RNAi.