CRISPRs extending their reach: prokaryotic RNAi protein Cas9 recruited for gene regulation

A Cas protein from the CRISPR defence system against foreign DNA, also functions in endogenous gene regulation. Sampson et al (2013) have revealed that in pathogenic Francisella bacteria, the Cas9 protein guided by small RNAs represses the mRNA of a lipoprotein. This novel mechanism of post‐transcriptional regulation enables the infecting bacteria to evade the TLR2‐based innate immune response of its host. Thus, reminiscent of eukaryotic RNAi where some proteins facilitate both genome defence and gene regulation, a central prokaryotic RNAi protein not only destroys invading DNA but also controls mRNA expression.

Both bacteria and eukaryotes use small RNAs to regulate their own genes and guide the destruction of unwanted foreign nucleic acids. In eukaryotes, many proteins function in both these processes, for example, argonaute proteins associate with either microRNAs or siRNAs to regulate endogenous messengers or degrade the RNA of attacking viruses, respectively (Czech and Hannon, 2011). In bacteria, however, the small RNA‐based pathways of gene regulation or genome defence are very different (Figure 1). Gene expression control is typically achieved at the post‐transcriptional level by a plethora of small noncoding RNAs (sRNAs), which base pair with the 5′ end of target mRNAs to repress or activate protein synthesis. The RNA chaperone Hfq, which facilitates the underlying short RNA interactions, is often required for these regulations (Vogel and Luisi, 2011). Although sRNAs and Hfq control many physiological processes, they seem to have few, if any, roles in protective recognition of intrusive nucleic acids (on the contrary, Hfq is even a host factor for the replication of an RNA phage).

Such defence is the task of so‐called CRISPR/Cas loci that confer adaptive, sequence‐based immunity against bacterial plasmids and phages. These loci produce small processed CRISPR RNAs (crRNAs) that assemble with CRISPR‐associated (Cas) proteins into complexes to target genome invaders, a process also known as prokaryotic RNAi (Sorek et al, 2013). In contrast to eukaryotic RNAi, which targets RNA, all three bacterial CRISPR/Cas systems cleave DNA, meaning that if they were to regulate endogenous genes they would inevitably destroy the bacterial chromosome. However, writing in Nature, Sampson et al (2013) now report a post‐transcriptional control mechanism in Francisella novicida whereby a chromosomal virulence gene is regulated by a Cas protein and CRISPR‐associated small RNAs.

Francisella species are highly infective, airborne pathogens of humans and animals, which cause the fatal disease tularaemia or severe systemic infections. Survival of F. novicida inside host cells crucially depends on its ability to evade recognition by toll‐like receptor 2 (TLR2), which senses bacterial lipoproteins and triggers a proinflammatory response. Strikingly, one of the F. novicida genes required for the evasion of TLR2 sensing encodes the CRISPR‐associated endonuclease Cas9 (Sampson et al, 2013).

Cas9 is a marker protein of so‐called type II CRISPR systems, which are generally associated with vertebrate pathogens and commensals (Sorek et al, 2013). Type II systems exhibit a unique crRNA biogenesis pathway involving a trans‐encoded small RNA (tracrRNA) and RNase III as a host factor (Deltcheva et al, 2011; Zhang et al, 2013). Moreover, whereas Type I and III systems use multiple‐protein complexes for CRISPR interference, the Cas9 protein of Type II suffices for crRNA‐mediated target destruction, and this property is now being exploited for genome editing and control in diverse heterologous systems including human cells (Horvath and Barrangou, 2013).

In their study, Sampson et al, 2013 discover that Cas9, but not other proteins of the Francisella CRISPR/Cas locus, represses the synthesis of a bacterial lipoprotein encoded elsewhere in the F. novicida chromosome. Unexpectedly, this involves destabilization of the lipoprotein mRNA, suggesting that in this case Cas9 acts at the RNA rather than the DNA level. Recognition of the target mRNA is achieved by a new ‘scaRNA’ (small CRISPR/Cas associated RNA), which may function as an adapter and base pair with both the tracrRNA and the target's ribosome‐binding site. Genetic analyses and RNA co‐immunoprecipitation experiments indicate that all three components, that is, Cas9, tracRNA, and scaRNA, are essential for target regulation. In Cas9 itself, a new conserved arginine‐rich motif (ARM) distinct from the previously characterized HNH and RuvC‐like nuclease motifs for DNA targeting is essential for the protein's activity on RNA. A mechanism is proposed whereby the scaRNA guides recognition of the lipoprotein mRNA, and the simultaneous recruitment of Cas9 via tracrRNA may bring in an as yet unknown RNase to degrade this target.

Additional experiments provide evidence that the joint activities of Cas9, tracrRNA, and scaRNA are necessary for Francisella virulence. Strains mutated in these genes cause macrophages secrete high levels of proinflammatory cytokine IL‐6, but this effect is lost upon additional genetic inactivation of the target lipoprotein. The Cas9‐mediated evasion of the TLR2 signalling seems to be most important during the early phase of infection since Cas9, tracrRNA, and scaRNA are highly induced 1 h into Francisella infection of phagosomes. As final proof of the importance of Cas9‐mediated lipoprotein repression, lack of this new system for gene regulation attenuates bacterial in mice, to the extent that Francisella strains lacking Cas9 or its companion small RNAs can be utilized as live vaccines.

The current study presents evidence that bacteria may use their CRISPR loci to regulate endogenous genes, a possibility that was raised by other recent studies (e.g., Zegans et al, 2009). Nonetheless, several questions remain to be addressed. First, the lipoprotein mRNA in question shows some unusual behaviour: when general transcription was inhibited with rifampicin, this transcript continued to accumulate (Sampson et al, 2013), as if its biogenesis differed from other mRNAs. Does the Cas9 pathway act on a specific subset of unusual mRNAs? Deep sequencing of Cas9‐associated mRNAs similar to the approach that has provided an atlas of potential Hfq targets (Chao et al, 2012) will help address this question, in both Francisella and other pathogenic bacteria with a Cas9 gene. Second, how exactly is the target recognized? A direct scaRNA–mRNA interaction remains to be proven, and the base pairing predicted seems weaker than in other regulatory RNA interactions. Third, what is the nuclease that causes destabilization of the target mRNA? Sampson et al (2013) tested a few obvious RNases of Francisella but there is no smoking gun. However, the essential RNase E was not tested, perhaps due to the lack of a suitable mutant. If RNase E is involved, it is tempting to speculate that Cas9 may also activate genes, by protecting mRNAs from RNase E cleavage as previously shown in Hfq‐dependent regulation (Papenfort et al, 2013). Fourth, how is scaRNA made to begin with? What is its size and how stable and abundant is it? The authors have detected it by qRT–PCR but other methods such as a northern blot will give more confidence regarding its nature and expression.

These open questions notwithstanding, the work by Sampson et al (2013) is a major advance in our understanding of the ever growing scope of gene regulation by small RNAs in bacteria. Foremost, however, the present study is a prime example of how bacteria have found a new function for a multifaceted RNA‐binding protein, transforming their Cas9 sword for genome defence into a ploughshare for endogenous gene regulation.