Adapting to change : on the mechanism of type I-E CRISPR-Cas defence

Host-pathogen interactions are among the most prevalent and evolutionary important interactions known today. The predation of prokaryotes by their viruses is happening on an especially large scale and had a major influence on the evolutionary history of prokaryotes. Since most viruses are lytic at some point in their life-cycle, there is a high selection pressure for prokaryotes to develop defense mechanisms. As described in Chapter 1, the CRISPR-Cas system is a relatively recently discovered defense system and is also the first adaptive defense system discovered in prokaryotes. CRISPR-Cas systems are widespread, occurring in the majority of archaea and also a considerable fraction of bacteria. This diversity is also reflected in the diversity of different types of CRISPR-Cas systems, currently being divided into 6 major types with a large number of subtypes. The type I-E system of Escherichia coli is a well-studied model system and of high relevance, since it is a major subtype of type I systems which make up around 50 % of all discovered CRISPR-Cas systems. CRISPR-Cas systems basically comprise the CRISPR array, made up of repeats and foreign derived spacers, and a set of cas genes. Immunity is commonly divided into three functional stages, adaptation, expression and interference. Adaptation is the acquisition of new spacers from the foreign nucleic acid and its incorporation into the CRISPR array. During expression, the CRISPR array is transcribed, processed and assembled with Cas proteins into CRISPR RNA (crRNA) guided ribonucleoprotein complexes (crRNP). Interference is the detection, binding and destruction of foreign nucleic acids by the crRNP and in type I systems the Cas3 nuclease. The type I-E system contains another function, called primed adaptation. Primed adaptation is a more rapid and efficient version of regular (naïve) adaptation. In addition to the adaptation machinery, primed adaptation also requires the interference machinery. Chapter 2 describes and compares a fundamental feature of most, if not all, CRISPR-Cas systems and also many other small RNA based systems. The mode of action of small RNAs relies on protein-assisted base pairing of the guide RNA with target mRNA or DNA to interfere with their transcription, translation or replication. Several unrelated classes of small non-coding RNAs have been identified including eukaryotic RNA silencing associated small RNAs, prokaryotic small regulatory RNAs and prokaryotic CRISPR (clustered regularly interspaced short palindromic repeats) RNAs. All three groups identify their target sequence by base pairing after finding it in a pool of millions of other nucleotide sequences in the cell. In this complicated target search process, a region of 6 to 12 nucleotides of the small RNA termed the ‘seed’ plays a critical role. The seed is often a structurally pre-ordered region that increases accessibility and lowers the energy barrier of RNA-DNA duplex formation. Furthermore, the length of the seed is optimally chosen to allow rapid probing and also rejection of potential target sites. The seed is a perfect example of parallel evolution, showing that nature comes up with the same strategy independently multiple times. Chapter 3 provides a description and protocol of the Electrophoretic Mobility Shift Assay (EMSA) and its use for studying crRNPs. EMSA is a straightforward and inexpensive method for the determination and quantification of protein–nucleic acid interactions. It relies on the different mobility of free and protein-bound nucleic acid in a gel matrix during electrophoresis. Nucleic acid affinities of crRNPs can be quantified by calculating the dissociation constant (Kd ). Protocols for two types of EMSA assays are described using the Cascade ribonucleoprotein complex from Escherichia coli as an example. One protocol uses plasmid DNA as substrate, while the other uses short linear oligonucleotides. Plasmids can be easily visualized with traditional DNA staining, while oligos have to be radioactively labelled using the 32Phosphate isotope. The EMSA method and these protocols are applied throughout the other chapters of this thesis. Chapter 4 focusses on the processes of interference and primed adaptation, specifically on their tolerance of mutations. Invaders can escape Type I-E CRISPR-Cas immunity in E. coli by making point mutations in the protospacer (especially in the seed) or its adjacent motif (PAM), but hosts quickly restore immunity by integrating new spacers in a positive feedback process termed priming. Here, we provide a systematic analysis of the constraints of both direct interference and subsequent priming in E. coli. We have defined a high-resolution genetic map of direct interference by Cascade and Cas3, which includes five positions of the protospacer at 6 nt intervals that readily tolerate mutations. Importantly, we show that priming is an extremely robust process capable of utilizing degenerate target regions with up to at least eleven mutations throughout the PAM and protospacer region. Priming is influenced by the number of mismatches, their position and is nucleotide dependent. Our findings imply that even out-dated spacers containing many mismatches can induce a rapid primed CRISPR response against diversified or related invaders, giving microbes an advantage in the co- evolutionary arms race with their invaders. In Chapter 5 we elucidate the mechanism of priming. Specifically, we determine how new spacers are produced and selected for integration into the CRISPR array during priming. We show that priming is directly dependent on interference. Rapid priming occurs when the rate of interference is high, delayed priming occurs when the rate of interference is low. Using in vitro assays and next generation sequencing, we show that Cas3 couples CRISPR interference to adaptation by producing DNA breakdown products that fuel the spacer integration process in a two-step, PAM-associated manner. The helicase-nuclease Cas3 pre-processes target DNA into fragments of about 30–100 nt enriched for thymine-stretches in their 3’ ends. By reconstituting the spacer integration process in vitro, we show that the Cas1-2 complex further processes these fragments and integrates them sequence- specifically into CRISPR repeats by coupling of a 3’ cytosine of the fragment. Our results highlight that the selection of PAM-compliant spacers during priming is enhanced by the combined sequence specificities of Cas3 and the Cas1-2 complex, leading to an increased propensity of integrating functional CTT-containing spacers. In Chapter 6 we look deeper into a nucleotide specific effect on priming that was discovered in Chapter 4. Immunity is based on the complementarity of host encoded spacer sequences with protospacers on the foreign genetic element. The efficiency of both direct interference and primed acquisition depends on the degree of complementarity between spacer and protospacer. Previous studies focused on the amount and positions of mutations, not the identity of the substituted nucleotide. In Chapter 4, we describe a nucleotide bias, showing a positive effect on priming of C substitutions and a negative effect on priming of G substitutions in the basepairing strand of the protospacer. Here we show that these substitutions rather directly influence the efficiency of interference and therefore indirectly influence the efficiency of interference dependent priming. We show that G substitutions have a profoundly negative effect on interference, while C substitutions are readily tolerated when in the same positions. Furthermore, we show that this effect is based on strongly decreased binding of the effector complex Cascade to G mutants, while C mutants only minimally affect binding. In Chapter 5 we showed a connection between the rate of interference and the time of occurrence of priming. Here, we also quantify the extent of priming and show that priming is very prevalent in a population that shows intermediate levels of interference, while high or low levels of interference lead to a lower prevalence of priming. Chapter 7 describes an attempt to make use of our knowledge about the Cascade complex and develop it into a genome editing tool. The development of genome editing tools has made major leaps in the last decade. Recently, RNA guided endonucleases (RGENs) such as Cas9 or Cpf1 have revolutionized genome editing. These RGENs are the hallmark proteins of class II CRISPR-Cas systems. Here, we have explored the possibility to develop a new genome editing tool that makes use of the Cascade complex from E. coli. This RNA guided protein complex is fused to a FokI nuclease domain to sequence specifically cleave DNA. We validate the tool in vitro using purified protein and two sets of guide RNAs, showing specific cleavage activity. The tool requires two target sites of 32 nt each at a distance of 30-40 nt and inward facing three nucleotide flexible PAM sequences. Cleavage occurs in the middle between the two binding sites and primarily creates 4 nt overhangs. Furthermore, we show that an additional RFP can be fused to FokI-Cascade, allowing visualization of the complex in target cells. Unfortunately, we were not able to successfully apply the tool in vivo in eukaryotic cells.