Systematic interrogation of CRISPR antimicrobials in Klebsiella pneumoniae reveals nuclease-, guide- and strain-dependent features influencing antimicrobial activity.

CRISPR-Cas systems can be utilized as programmable-spectrum antimicrobials to combat bacterial infections. However, how CRISPR nucleases perform as antimicrobials across target sites and strains remains poorly explored. Here, we address this knowledge gap by systematically interrogating the use of CRISPR antimicrobials using multidrug-resistant and hypervirulent strains of Klebsiella pneumoniae as models. Comparing different Cas nucleases, DNA-targeting nucleases outperformed RNA-targeting nucleases based on the tested targets. Focusing on AsCas12a that exhibited robust targeting across different strains, we found that the elucidated modes of escape varied widely, restraining opportunities to enhance killing. We also encountered individual guide RNAs yielding different extents of clearance across strains, which were linked to an interplay between improper gRNA folding and strain-specific DNA repair and survival. To explore features that could improve targeting across strains, we performed a genome-wide screen in different K. pneumoniae strains that yielded guide design rules and trained an algorithm for predicting guide efficiency. Finally, we showed that Cas12a antimicrobials can be exploited to eliminate K. pneumoniae when encoded in phagemids delivered by T7-like phages. Altogether, our results highlight the importance of evaluating antimicrobial activity of CRISPR antimicrobials across relevant strains and define critical parameters for efficient CRISPR-based targeting.

Microbial defenses against mobile genetic elements and viruses: Who defends whom from what?

Prokaryotes have numerous mobile genetic elements (MGEs) that mediate horizontal gene transfer (HGT) between cells. These elements can be costly, even deadly, and cells use numerous defense systems to filter, control, or inactivate them. Recent studies have shown that prophages, conjugative elements, their parasites (phage satellites and mobilizable elements), and other poorly described MGEs encode defense systems homologous to those of bacteria. These constitute a significant fraction of the repertoire of cellular defense genes. As components of MGEs, these defense systems have presumably evolved to provide them, not the cell, adaptive functions. While the interests of the host and MGEs are aligned when they face a common threat such as an infection by a virulent phage, defensive functions carried by MGEs might also play more selfish roles to fend off other antagonistic MGEs or to ensure their maintenance in the cell. MGEs are eventually lost from the surviving host genomes by mutational processes and their defense systems can be co-opted when they provide an advantage to the cell. The abundance of defense systems in MGEs thus sheds new light on the role, effect, and fate of the so-called "cellular defense systems," whereby they are not only merely microbial defensive weapons in a 2-partner arms race, but also tools of intragenomic conflict between multiple genetic elements with divergent interests that shape cell fate and gene flow at the population level.

Mutations in Cas9 Enhance the Rate of Acquisition of Viral Spacer Sequences during the CRISPR-Cas Immune Response.

CRISPR loci and their associated (Cas) proteins encode a prokaryotic immune system that protects against viruses and plasmids. Upon infection, a low fraction of cells acquire short DNA sequences from the invader. These sequences (spacers) are integrated in between the repeats of the CRISPR locus and immunize the host against the matching invader. Spacers specify the targets of the CRISPR immune response through transcription into short RNA guides that direct Cas nucleases to the invading DNA molecules. Here we performed random mutagenesis of the RNA-guided Cas9 nuclease to look for variants that provide enhanced immunity against viral infection. We identified a mutation, I473F, that increases the rate of spacer acquisition by more than two orders of magnitude. Our results highlight the role of Cas9 during CRISPR immunization and provide a useful tool to study this rare process and develop it as a biotechnological application.

Cas9 off-target binding to the promoter of bacterial genes leads to silencing and toxicity.

Genetic tools derived from the Cas9 RNA-guided nuclease are providing essential capabilities to study and engineer bacteria. While the importance of off-target effects was noted early in Cas9's application to mammalian cells, off-target cleavage by Cas9 in bacterial genomes is easily avoided due to their smaller size. Despite this, several studies have reported experimental setups in which Cas9 expression was toxic, even when using the catalytic dead variant of Cas9 (dCas9). Specifically, dCas9 was shown to be toxic when in complex with guide RNAs sharing specific PAM (protospacer adjacent motif)-proximal sequence motifs. Here, we demonstrate that this toxicity is caused by off-target binding of Cas9 to the promoter of essential genes, with silencing of off-target genes occurring with as little as 4 nt of identity in the PAM-proximal sequence. Screens performed in various strains of Escherichia coli and other enterobacteria show that the nature of toxic guide RNAs changes together with the evolution of sequences at off-target positions. These results highlight the potential for Cas9 to bind to hundreds of off-target positions in bacterial genomes, leading to undesired effects. This phenomenon must be considered in the design and interpretation of CRISPR-Cas experiments in bacteria.

Computational design of novel Cas9 PAM-interacting domains using evolution-based modelling and structural quality assessment.

We present here an approach to protein design that combines (i) scarce functional information such as experimental data (ii) evolutionary information learned from a natural sequence variants and (iii) physics-grounded modeling. Using a Restricted Boltzmann Machine (RBM), we learn a sequence model of a protein family. We use semi-supervision to leverage available functional information during the RBM training. We then propose a strategy to explore the protein representation space that can be informed by external models such as an empirical force-field method (FoldX). Our approach is applied to a domain of the Cas9 protein responsible for recognition of a short DNA motif. We experimentally assess the functionality of 71 variants generated to explore a range of RBM and FoldX energies. Sequences with as many as 50 differences (20% of the protein domain) to the wild-type retained functionality. Overall, 21/71 sequences designed with our method were functional. Interestingly, 6/71 sequences showed an improved activity in comparison with the original wild-type protein sequence. These results demonstrate the interest in further exploring the synergies between machine-learning of protein sequence representations and physics grounded modeling strategies informed by structural information.

Improving sequence-based modeling of protein families using secondary-structure quality assessment.

MOTIVATION: Modeling of protein family sequence distribution from homologous sequence data recently received considerable attention, in particular for structure and function predictions, as well as for protein design. In particular, direct coupling analysis, a method to infer effective pairwise interactions between residues, was shown to capture important structural constraints and to successfully generate functional protein sequences. Building on this and other graphical models, we introduce a new framework to assess the quality of the secondary structures of the generated sequences with respect to reference structures for the family. RESULTS: We introduce two scoring functions characterizing the likeliness of the secondary structure of a protein sequence to match a reference structure, called Dot Product and Pattern Matching. We test these scores on published experimental protein mutagenesis and design dataset, and show improvement in the detection of nonfunctional sequences. We also show that use of these scores help rejecting nonfunctional sequences generated by graphical models (Restricted Boltzmann Machines) learned from homologous sequence alignments. AVAILABILITY AND IMPLEMENTATION: Data and code available at https://github.com/CyrilMa/ssqa. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Generating functional protein variants with variational autoencoders.

The vast expansion of protein sequence databases provides an opportunity for new protein design approaches which seek to learn the sequence-function relationship directly from natural sequence variation. Deep generative models trained on protein sequence data have been shown to learn biologically meaningful representations helpful for a variety of downstream tasks, but their potential for direct use in the design of novel proteins remains largely unexplored. Here we show that variational autoencoders trained on a dataset of almost 70000 luciferase-like oxidoreductases can be used to generate novel, functional variants of the luxA bacterial luciferase. We propose separate VAE models to work with aligned sequence input (MSA VAE) and raw sequence input (AR-VAE), and offer evidence that while both are able to reproduce patterns of amino acid usage characteristic of the family, the MSA VAE is better able to capture long-distance dependencies reflecting the influence of 3D structure. To confirm the practical utility of the models, we used them to generate variants of luxA whose luminescence activity was validated experimentally. We further showed that conditional variants of both models could be used to increase the solubility of luxA without disrupting function. Altogether 6/12 of the variants generated using the unconditional AR-VAE and 9/11 generated using the unconditional MSA VAE retained measurable luminescence, together with all 23 of the less distant variants generated by conditional versions of the models; the most distant functional variant contained 35 differences relative to the nearest training set sequence. These results demonstrate the feasibility of using deep generative models to explore the space of possible protein sequences and generate useful variants, providing a method complementary to rational design and directed evolution approaches.

Atypical organizations and epistatic interactions of CRISPRs and cas clusters in genomes and their mobile genetic elements.

Prokaryotes use CRISPR-Cas systems for adaptive immunity, but the reasons for the frequent existence of multiple CRISPRs and cas clusters remain poorly understood. Here, we analysed the joint distribution of CRISPR and cas genes in a large set of fully sequenced bacterial genomes and their mobile genetic elements. Our analysis suggests few negative and many positive epistatic interactions between Cas subtypes. The latter often result in complex genetic organizations, where a locus has a single adaptation module and diverse interference mechanisms that might provide more effective immunity. We typed CRISPRs that could not be unambiguously associated with a cas cluster and found that such complex loci tend to have unique type I repeats in multiple CRISPRs. Many chromosomal CRISPRs lack a neighboring Cas system and they often have repeats compatible with the Cas systems encoded in trans. Phages and 25 000 prophages were almost devoid of CRISPR-Cas systems, whereas 3% of plasmids had CRISPR-Cas systems or isolated CRISPRs. The latter were often compatible with the chromosomal cas clusters, suggesting that plasmids can co-opt the latter. These results highlight the importance of interactions between CRISPRs and cas present in multiple copies and in distinct genomic locations in the function and evolution of bacterial immunity.

On-target activity predictions enable improved CRISPR-dCas9 screens in bacteria.

The ability to block gene expression in bacteria with the catalytically inactive mutant of Cas9, known as dCas9, is quickly becoming a standard methodology to probe gene function, perform high-throughput screens, and engineer cells for desired purposes. Yet, we still lack a good understanding of the design rules that determine on-target activity for dCas9. Taking advantage of high-throughput screening data, we fit a model to predict the ability of dCas9 to block the RNA polymerase based on the target sequence, and validate its performance on independently generated datasets. We further design a novel genome wide guide RNA library for E. coli MG1655, EcoWG1, using our model to choose guides with high activity while avoiding guides which might be toxic or have off-target effects. A screen performed using the EcoWG1 library during growth in rich medium improved upon previously published screens, demonstrating that very good performances can be attained using only a small number of well designed guides. Being able to design effective, smaller libraries will help make CRISPRi screens even easier to perform and more cost-effective. Our model and materials are available to the community through crispr.pasteur.fr and Addgene.

Gene silencing with CRISPRi in bacteria and optimization of dCas9 expression levels.

The catalytic null mutant of the Cas9 endonuclease from the bacterial CRISPR immune system, known as dCas9, can be guided by a small RNA to bind DNA sequences of interest and block gene transcription in a strategy known as CRISPRi. This powerful gene silencing method has already been used in a large number of species and in high throughput screens. Here we provide detailed design rules, methods and novel vectors to perform CRISPRi experiments in S. aureus and in E. coli, using the well characterized dCas9 protein from S. pyogenes. In particular, we describe a vector based on plasmid pC194 which is broadly used in Firmicutes, as well as a vector based on the very broad host-range rolling circle plasmid pLZ12, reported to replicate in both Firmicutes and Proteobacteria. A potential caveat of adapting dCas9 tools to various bacterial species is that dCas9 was shown to be toxic when expressed too strongly. We describe a method to optimize the expression level of dCas9 in order to avoid toxicity while ensuring strong on-target repression activity. We demonstrate this method by optimizing a pLZ12 based vector originally developed for S. aureus so that it can work in E. coli. This article should provide all the resources required to perform CRISPRi experiments in a broad range of bacterial species.

Cas9 specifies functional viral targets during CRISPR-Cas adaptation.

Clustered regularly interspaced short palindromic repeat (CRISPR) loci and their associated (Cas) proteins provide adaptive immunity against viral infection in prokaryotes. Upon infection, short phage sequences known as spacers integrate between CRISPR repeats and are transcribed into small RNA molecules that guide the Cas9 nuclease to the viral targets (protospacers). Streptococcus pyogenes Cas9 cleavage of the viral genome requires the presence of a 5'-NGG-3' protospacer adjacent motif (PAM) sequence immediately downstream of the viral target. It is not known whether and how viral sequences flanked by the correct PAM are chosen as new spacers. Here we show that Cas9 selects functional spacers by recognizing their PAM during spacer acquisition. The replacement of cas9 with alleles that lack the PAM recognition motif or recognize an NGGNG PAM eliminated or changed PAM specificity during spacer acquisition, respectively. Cas9 associates with other proteins of the acquisition machinery (Cas1, Cas2 and Csn2), presumably to provide PAM-specificity to this process. These results establish a new function for Cas9 in the genesis of prokaryotic immunological memory.

Genome-wide CRISPR-dCas9 screens in E. coli identify essential genes and phage host factors.

High-throughput genetic screens are powerful methods to identify genes linked to a given phenotype. The catalytic null mutant of the Cas9 RNA-guided nuclease (dCas9) can be conveniently used to silence genes of interest in a method also known as CRISPRi. Here, we report a genome-wide CRISPR-dCas9 screen using a starting pool of ~ 92,000 sgRNAs which target random positions in the chromosome of E. coli. To benchmark our method, we first investigate its utility to predict gene essentiality in the genome of E. coli during growth in rich medium. We could identify 79% of the genes previously reported as essential and demonstrate the non-essentiality of some genes annotated as essential. In addition, we took advantage of the intermediate repression levels obtained when targeting the template strand of genes to show that cells are very sensitive to the expression level of a limited set of essential genes. Our data can be visualized on CRISPRbrowser, a custom web interface available at crispr.pasteur.fr. We then apply the screen to discover E. coli genes required by phages λ, T4 and 186 to kill their host, highlighting the involvement of diverse host pathways in the infection process of the three tested phages. We also identify colanic acid capsule synthesis as a shared resistance mechanism to all three phages. Finally, using a plasmid packaging system and a transduction assay, we identify genes required for the formation of functional λ capsids, thus covering the entire phage cycle. This study demonstrates the usefulness and convenience of pooled genome-wide CRISPR-dCas9 screens in bacteria and paves the way for their broader use as a powerful tool in bacterial genomics.

Conditional tolerance of temperate phages via transcription-dependent CRISPR-Cas targeting.

A fundamental feature of immune systems is the ability to distinguish pathogenic from self and commensal elements, and to attack the former but tolerate the latter. Prokaryotic CRISPR-Cas immune systems defend against phage infection by using Cas nucleases and small RNA guides that specify one or more target sites for cleavage of the viral genome. Temperate phages include viruses that can integrate into the bacterial chromosome, and they can carry genes that provide a fitness advantage to the lysogenic host. However, CRISPR-Cas targeting that relies strictly on DNA sequence recognition provides indiscriminate immunity both to lytic and lysogenic infection by temperate phages-compromising the genetic stability of these potentially beneficial elements altogether. Here we show that the Staphylococcus epidermidis CRISPR-Cas system can prevent lytic infection but tolerate lysogenization by temperate phages. Conditional tolerance is achieved through transcription-dependent DNA targeting, and ensures that targeting is resumed upon induction of the prophage lytic cycle. Our results provide evidence for the functional divergence of CRISPR-Cas systems and highlight the importance of targeting mechanism diversity. In addition, they extend the concept of 'tolerance to non-self' to the prokaryotic branch of adaptive immunity.

Tuning dCas9's ability to block transcription enables robust, noiseless knockdown of bacterial genes.

Over the past few years, tools that make use of the Cas9 nuclease have led to many breakthroughs, including in the control of gene expression. The catalytically dead variant of Cas9 known as dCas9 can be guided by small RNAs to block transcription of target genes, in a strategy also known as CRISPRi. Here, we reveal that the level of complementarity between the guide RNA and the target controls the rate at which RNA polymerase "kicks out" dCas9 from the target and completes transcription. We use this mechanism to precisely and robustly reduce gene expression by defined relative amounts. Alternatively, tuning repression by changing dCas9 concentration is noisy and promoter-strength dependent. We demonstrate broad applicability of this method to the study of genetic regulation and cellular physiology. First, we characterize feedback strength of a model auto-repressor. Second, we study the impact of amount variations of cell-wall synthesizing enzymes on cell morphology. Finally, we multiplex the system to obtain any combination of fractional repression of two genes.

Microbial-derived products as potential new antimicrobials.

Due to the continuing global concerns involving antibiotic resistance, there is a need for scientific forums to assess advancements in the development of antimicrobials and their alternatives that might reduce development and spread of antibiotic resistance among bacterial pathogens. The objectives of the 2nd International Symposium on Alternatives to Antibiotics were to highlight promising research results and novel technologies that can provide alternatives to antibiotics for use in animal health and production, assess challenges associated with their authorization and commercialization for use, and provide actionable strategies to support their development. The session on microbial-derived products was directed at presenting novel technologies that included exploiting CRISPR-Cas nucleases to produce sequence-specific antimicrobials, probiotics development via fecal microbiome transplants among monogastric production animals such as chickens and mining microbial sources such as bacteria or yeast to identify new antimicrobial compounds. Other research has included continuing development of antimicrobial peptides such as newly discovered bacteriocins as alternatives to antibiotics, use of bacteriophages accompanied by development of unique lytic proteins with specific cell-wall binding domains and novel approaches such as microbial-ecology guided discovery of anti-biofilm compounds discovered in marine environments. The symposium was held at the Headquarters of the World Organisation for Animal Health (OIE) in Paris, France during 12-15 December 2016.

Consequences of Cas9 cleavage in the chromosome of Escherichia coli.

The RNA-guided Cas9 nuclease from CRISPR-Cas systems has emerged as a powerful biotechnological tool. The specificity of Cas9 can be reprogrammed to cleave desired sequences in a cell's chromosome simply by changing the sequence of a small guide RNA. Unlike in most eukaryotes, Cas9 cleavage in the chromosome of bacteria has been reported to kill the cell. However, the mechanism of cell death remains to be investigated. Bacteria mainly rely on homologous recombination (HR) with sister chromosomes to repair double strand breaks. Here, we show that the simultaneous cleavage of all copies of the Escherichia coli chromosome at the same position cannot be repaired, leading to cell death. However, inefficient cleavage can be tolerated through continuous repair by the HR pathway. In order to kill cells reliably, HR can be blocked using the Mu phage Gam protein. Finally, the introduction of the non-homologous end joining (NHEJ) pathway from Mycobacterium tuberculosis was not able to rescue the cells from Cas9-mediated killing, but did introduce small deletions at a low frequency. This work provides a better understanding of the consequences of Cas9 cleavage in bacterial chromosomes which will be instrumental in the development of future CRISPR tools.

Impact of Different Target Sequences on Type III CRISPR-Cas Immunity.

UNLABELLED: Clustered regularly interspaced short palindromic repeat (CRISPR) loci encode an adaptive immune system of prokaryotes. Within these loci, sequences intercalated between repeats known as "spacers" specify the targets of CRISPR immunity. The majority of spacers match sequences present in phages and plasmids; however, it is not known whether there are differences in the immunity provided against these diverse invaders. We studied this issue using the Staphylococcus epidermidis CRISPR system, which harbors spacers matching both phages and plasmids. We determined that this CRISPR system provides similar levels of defense against the conjugative plasmid pG0400 and the bacteriophage CNPX. However, whereas antiplasmid immunity was very sensitive to the introduction of mismatches in the target sequence, mutations in the phage target were largely tolerated. Placing the phage and plasmid targets into a vector that can be both conjugated and transduced, we demonstrated that the route of entry of the target has no impact on the effect of the mismatches on immunity. Instead, we established that the specific sequences of each spacer/target determine the susceptibility of the S. epidermidis CRISPR system to mutations. Therefore, spacers that are more resistant to mismatches would provide long-term immunity against phages and plasmids that otherwise would escape CRISPR targeting through the accumulation of mutations in the target sequence. These results uncover an unexpected complexity in the arms race between CRISPR-Cas systems and prokaryotic infectious genetic elements. IMPORTANCE: CRISPR-Cas loci protect bacteria and archaea from both phage infection and plasmid invasion. These loci harbor short sequences of phage and plasmid origin known as "spacers" that specify the targets of CRISPR-Cas immunity. The presence of a spacer sequence matching a phage or plasmid ensures host immunity against infection by these genetic elements. In turn, phages and plasmids constantly mutate their targets to avoid recognition by the spacers of the CRISPR-Cas immune system. In this study, we demonstrated that different spacer sequences vary in their ability to tolerate target mutations that allow phages and plasmids to escape from CRISPR-Cas immunity. These results uncover an unexpected complexity in the arms race between CRISPR-Cas systems and prokaryotic infectious genetic elements.

Adapting to new threats: the generation of memory by CRISPR-Cas immune systems.

Clustered, regularly interspaced, short palindromic repeats (CRISPR) loci and their associated genes (cas) confer bacteria and archaea with adaptive immunity against phages and other invading genetic elements. A fundamental requirement of any immune system is the ability to build a memory of past infections in order to deal more efficiently with recurrent infections. The adaptive feature of CRISPR-Cas immune systems relies on their ability to memorize DNA sequences of invading molecules and integrate them in between the repetitive sequences of the CRISPR array in the form of 'spacers'. The transcription of a spacer generates a small antisense RNA that is used by RNA-guided Cas nucleases to cleave the invading nucleic acid in order to protect the cell from infection. The acquisition of new spacers allows the CRISPR-Cas immune system to rapidly adapt against new threats and is therefore termed 'adaptation'. Recent studies have begun to elucidate the genetic requirements for adaptation and have demonstrated that rather than being a stochastic process, the selection of new spacers is influenced by several factors. We review here our current knowledge of the CRISPR adaptation mechanism.

Programmable repression and activation of bacterial gene expression using an engineered CRISPR-Cas system.

The ability to artificially control transcription is essential both to the study of gene function and to the construction of synthetic gene networks with desired properties. Cas9 is an RNA-guided double-stranded DNA nuclease that participates in the CRISPR-Cas immune defense against prokaryotic viruses. We describe the use of a Cas9 nuclease mutant that retains DNA-binding activity and can be engineered as a programmable transcription repressor by preventing the binding of the RNA polymerase (RNAP) to promoter sequences or as a transcription terminator by blocking the running RNAP. In addition, a fusion between the omega subunit of the RNAP and a Cas9 nuclease mutant directed to bind upstream promoter regions can achieve programmable transcription activation. The simple and efficient modulation of gene expression achieved by this technology is a useful asset for the study of gene networks and for the development of synthetic biology and biotechnological applications.

The integron integrase efficiently prevents the melting effect of Escherichia coli single-stranded DNA-binding protein on folded attC sites.

Integrons play a major role in the dissemination of antibiotic resistance genes among bacteria. Rearrangement of gene cassettes occurs by recombination between attI and attC sites, catalyzed by the integron integrase. Integron recombination uses an unconventional mechanism involving a folded single-stranded attC site. This site could be a target for several host factors and more precisely for proteins able to bind single-stranded DNA. One of these, Escherichia coli single-stranded DNA-binding protein (SSB), regulates many DNA processes. We studied the influence of this protein on integron recombination. Our results show the ability of SSB to strongly bind folded attC sites and to destabilize them. This effect was observed only in the absence of the integrase. Indeed, we provided evidence that the integrase is able to counterbalance the observed effect of SSB on attC site folding. We showed that IntI1 possesses an intrinsic property to capture attC sites at the moment of their extrusion, stabilizing them and recombining them efficiently. The stability of DNA secondary structures in the chromosome must be restrained to avoid genetic instability (mutations or deletions) and/or toxicity (replication arrest). SSB, which hampers attC site folding in the absence of the integrase, likely plays an important role in maintaining the integrity and thus the recombinogenic functionality of the integron attC sites. We also tested the RecA host factor and excluded any role of this protein in integron recombination.