Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.


Dicing defence in bacteria

A newly discovered variation in the process by which bacteria resist invasion by foreign nucleic acids provides an intriguing parallel between the defence mechanisms of the different domains of life. See Article p.602

Viruses and mobile genetic elements such as plasmids pose a threat to the genomic integrity of organisms from all domains of life, whether they be eukaryotes (organisms with discrete cell nuclei, for example plants and animals), bacteria or archaea. Some of the most exciting findings made over the past 15 years have been the RNA-based defence systems of eukaryotes (RNA interference, or RNAi) and of bacteria and archaea (CRISPR systems). In both cases, RNA transcripts are processed by endonuclease enzymes into short guide RNAs that then target invaders that have complementary nucleic-acid sequences. The Dicer endonuclease is essential for RNAi processing1. On page 602 of this issue, Deltcheva et al.2 describe a processing pathway for CRISPR RNA that has intriguing similarities to the RNAi pathway, including dependence on the bacterial equivalent of Dicer, RNase III. A trans-acting RNA, tracrRNA, targets the cleavage process that ultimately produces mature guide RNAs.

The sequencing of bacterial and archaeal genomes led to recognition of clustered, regularly interspaced short palindromic repeats (CRISPR). These palindromic repeats are separated by unique spacers that are frequently homologous to portions of bacteriophage or plasmid genomes (Fig. 1). The spacers function as a bacterial memory of past invasions, and endow the cell with the ability to restrict a second invasion by the same phage or plasmid.

Figure 1: Flexibility in CRISPR processing systems.

In CRISPR arrays of DNA, conserved repeats (squares) alternate with variable spacers (diamonds). During the acquisition/memory step, new spacers are incorporated as a result of invasion of the cell by a phage or plasmid. During the processing steps required to confer immunity, RNA transcripts of the CRISPR array are cleaved to give mature guide RNAs, which then target for destruction invading genomes that match the spacer. The related proteins CasE, Csy4 and Cas6 carry out the initial processing step in many organisms. Deltcheva et al.2 describe a different pathway for guide-RNA maturation that operates in Streptococcus pyogenes and certain other bacteria (boxed area). This pathway involves a trans-acting RNA (tracrRNA) that hybridizes with the spacers, leading to cleavage by RNase III. Csn1 aids in the process, and an undefined step yields mature guide RNAs.

Spacers arise through the capture of invading sequences, with short regions of the invader's genome being incorporated between repeats to create the CRISPR array (acquisition step, Fig. 1; see refs 3, 4 for reviews). The CRISPR region is transcribed as a single, long RNA. As with RNAi, the effective RNA species for the CRISPR defence system is not the initial transcript, but a short, processed piece. For CRISPR, the processed RNA contains the sequence-specific spacer and some of the flanking repeats. This guide RNA is then used to direct the destruction of invading nucleic acids (DNA in some cases, RNA in others)5,6.

How does processing occur? The CRISPR arrays are associated with genes that encode a set of CRISPR-associated proteins (Cas). One of the difficulties in untangling the CRISPR pathway is the variability between species — and even strains — in the multiple families of associated proteins, implying multiple variants in the machinery for the acquisition step and subsequent immunity7. Some organisms have 20 proteins at a CRISPR locus, whereas others have only four. Purified Cas proteins from the bacteria Escherichia coli and Pseudomonas aeruginosa, and from the archaeon Pyrococcus furiosus, can all cleave the precursor RNA within the spacers to generate guide RNAs (processing steps, Fig. 1); therefore, specific Cas proteins are necessary and sufficient for RNA cleavage in these organisms8,9,10.

Streptococcus pyogenes, the organism studied by Deltcheva et al.2, has a CRISPR system with only four Cas-protein genes and none of the protein families that have been implicated in processing in other systems. Instead, the authors find that S. pyogenes uses the bacterial RNase III to process CRISPR RNA; targeting of cleavage is determined by an antisense RNA, tracrRNA, with complementarity to the CRISPR spacer (boxed area in Fig. 1).

RNase III is a conserved endonuclease that cleaves double-stranded RNA and is involved in the maturation of ribosomal RNA. It is also the bacterial equivalent of the Dicer endonuclease. The new work2 demonstrates the flexibility in CRISPR systems and, for the first time, links RNase III to the bacterial CRISPR defence system. Finally, the use of a core RNase rather than a Cas-specific protein provides a clear example of crosstalk between the CRISPR system and bacterial metabolism.

The antisense RNA that plays an important part here was identified when Deltcheva and colleagues' deep sequencing of S. pyogenes revealed abundant short RNAs all containing a region of complementarity to the CRISPR repeat. These antisense RNAs were encoded close to the CRISPR cluster but on the opposite strand. Deleting the gene for the antisense RNA prevented processing of the CRISPR RNA. In vitro, this antisense RNA can anneal to CRISPR RNA, making a double-stranded RNA that is a substrate for RNase III. In vivo, cleavage requires RNase III and Csn1, one of the Cas proteins specific to this family of CRISPR clusters; Csn1 may facilitate pairing of tracrRNA and CRISPR RNA, and also a second processing step. Finally, Deltcheva et al. find that all of these elements (tracrRNA, Csn1 and RNase III) are necessary for immunity — that is, for S. pyogenes to destroy an invading plasmid containing sequences complementary to the CRISPR spacer.

When the family of CRISPRs containing Csn1 was first defined, Haft et al.7 noted that “A characteristic feature of this subtype is a single copy of the repeat (sometimes direct, sometimes inverted), which appears upstream of the first gene in the locus”. Now we know that Csn1 is essential, as is the nearby upstream repeat, to this unique pathway for guide RNA maturation. The many bacterial species containing this family of proteins all express tracrRNA2, and so are highly likely to use this RNA, and RNase III, to mature their CRISPR RNAs.

We do not yet know whether this CRISPR variant is an evolutionary remnant, predating the acquisition of dedicated Cas proteins, or a more recent variation. Certainly the small number of Cas proteins necessary for the S. pyogenes system to work, coupled with the enticing similarity of the tracrRNA-dependent processing pathway to the eukaryotic Dicer pathway, favours this as an ancestral minimal system. Curiously, all of the species that harbour the genes in this pathway are associated with vertebrates, either as pathogens or commensals7. Perhaps this is a system that has evolved to deal not only with phage invaders, but also with invasion by eukaryotic RNAs (possibly microRNAs).

The recognition of CRISPR was a direct outcome of the sequencing of multiple microbial genomes. Now, deep sequencing2 of the RNA complement of S. pyogenes has demonstrated the existence of a novel pathway for an antisense RNA in CRISPR maturation. We can expect many more roles for regulatory RNAs to emerge as this approach continues to be applied to eukaryotes, bacteria and archaea.


  1. 1

    Jaskiewicz, L. & Filipowicz, W. Curr. Top. Microbiol. Immunol. 320, 77–97 (2008).

    CAS  PubMed  Google Scholar 

  2. 2

    Deltcheva, E. et al. Nature 471, 602–607 (2011).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Karginov, F. V. & Hannon, G. J. Mol. Cell 37, 7–19 (2010).

    CAS  Article  Google Scholar 

  4. 4

    Marraffini, L. A. & Sontheimer, E. J. Nature Rev. Genet. 11, 181–190 (2010).

    CAS  Article  Google Scholar 

  5. 5

    Marraffini, L. A. & Sontheimer, E. J. Science 322, 1843–1845 (2008).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Hale, C. R. et al. Cell 139, 945–956 (2009).

    CAS  Article  Google Scholar 

  7. 7

    Haft, D. H., Selengut, J., Mongodin, E. F. & Nelson, K. E. PLoS Comput. Biol. 1, e60 (2005).

    ADS  Article  Google Scholar 

  8. 8

    Brouns, S. J. J. et al. Science 321, 960–964 (2008).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Carte, J., Wang, R., Li, H., Terns, R. M. & Terns, M. P. Genes Dev. 22, 3489–3496 (2008).

    CAS  Article  Google Scholar 

  10. 10

    Haurwitz, R. E. et al. Science 329, 1355–1358 (2010).

    ADS  CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Susan Gottesman.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Gottesman, S. Dicing defence in bacteria. Nature 471, 588–589 (2011).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing