Critical roles for ‘housekeeping’ nucleases in type III CRISPR-Cas immunity

  1. Lucy Chou-Zheng
  2. Asma Hatoum-Aslan

  • Curated by eLife

    eLife logo

    eLife assessment

    CRISPR-Cas systems are essential components of an adaptive immune system that protects bacteria and archaea from infection by foreign genetic elements like phages and plasmids. The work presented here demonstrates that some CRISPR systems (i.e., type III-A) rely on host nucleases (i.e., RNase R and PNPase) for faithful processing of CRISPR RNAs. Collectively, this work expands the fundamental understanding of how nucleases involved in RNA metabolism contribute to the adaptive immune response in bacteria.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

CRISPR-Cas systems are a family of adaptive immune systems that use small CRISPR RNAs (crRNAs) and CRISPR-associated (Cas) nucleases to protect prokaryotes from invading plasmids and viruses (i.e., phages). Type III systems launch a multilayered immune response that relies upon both Cas and non-Cas cellular nucleases, and although the functions of Cas components have been well described, the identities and roles of non-Cas participants remain poorly understood. Previously, we showed that the type III-A CRISPR-Cas system in Staphylococcus epidermidis employs two degradosome-associated nucleases, PNPase and RNase J2, to promote crRNA maturation and eliminate invading nucleic acids (Chou-Zheng and Hatoum-Aslan, 2019). Here, we identify RNase R as a third ‘housekeeping’ nuclease critical for immunity. We show that RNase R works in concert with PNPase to complete crRNA maturation and identify specific interactions with Csm5, a member of the type III effector complex, which facilitate nuclease recruitment/stimulation. Furthermore, we demonstrate that RNase R and PNPase are required to maintain robust anti-plasmid immunity, particularly when targeted transcripts are sparse. Altogether, our findings expand the known repertoire of accessory nucleases required for type III immunity and highlight the remarkable capacity of these systems to interface with diverse cellular pathways to ensure successful defense.

Article activity feed

  1. Version published to 10.7554/elife.81897 on eLife
  2. eLife

    Author Response

    Reviewer #1 (Public Review):

    In this manuscript, Chou-Zheng and Hatoum-Aslan follow up on their previous studies that have characterized the collaborations between the type III-A CRISPR-Cas10 Csm complex and various cellular housekeeping nucleases. The authors have previously demonstrated that the Csm complex associates with several nucleases that are implicated in RNA degradation via pulldown and mass spectrometry analysis. They also previously showed that some of these enzymes, including PNPase, are important for CRISPR RNA (crRNA) maturation and for robust anti-phage defense. They now show that a second housekeeping enzyme, RNase R, is required for crRNA maturation. PNPase and RNase R act in concert to produce the mature crRNA. The authors also analyze the interactions between Csm5 and both housekeeping proteins. Finally, they demonstrate that PNPase and RNase R are important for robust anti-plasmid activity when using crRNAs that are complementary to low-abundance transcripts.

    This is a well-written paper with clear figures and well-described experiments and results. The experiments in Figures 1 and 2 demonstrating the importance of RNase R for crRNA maturation are excellent. The biochemistry experiments in Figure 2 are especially convincing, in which the authors were able to reconstitute the concerted activities of RNase R and PNPase for crRNA biogenesis. The experiments in Figure 5 implicating PNPase and RNase R in robust anti-plasmid activity when targeting low-abundance transcripts are also clear and convincing, and the result is intriguing. Overall, these experiments provide a new example in a growing list of co-opted host proteins that are important for crRNA biogenesis and CRISPR-mediated defense.

    Thank you for your thoughtful review of our manuscript and comments overall!

    I do have some concerns about experiments in Figures 3 and 4 analyzing interactions between PNPase or RNase R and the Csm5 subunit of the Csm complex, and I believe that some of the authors' conclusions are not fully supported by the evidence presented in these experiments. These concerns, along with a question about their model, are detailed below.

    1. The authors used the structure of S. thermophilus Csm5 to guide their design of truncations to probe potential intrinsically disordered regions (IDR1 and IDR2) that may be sites of interaction with PNPase or RNase R. Since the authors submitted their manuscript, an AlphaFold predicted structure of the S. epidermidis Csm5 has been released on the AlphaFold Protein Structure Database. In this model, the IDR2 region is predicted by AlphaFold to be a beta strand at the center of a beta sheet, rather than a disordered region. If the prediction is accurate, deletion of this strand could cause Csm5 to misfold, making it difficult to interpret what causes loss of interaction with PNPase (i.e. deletion of a specific interaction surface versus misfolding of the overall tertiary structure). In light of this, the discussion surrounding these experiments should be altered to include more caveats about the truncations, and conclusions based on this experiment should be softened.

    While this manuscript was under review, several cryo-EM structures of the Cas10-Csm complex from S. epidermidis were solved and reported (Smith et al, 2022, Structure). In the unbound complex (PDB ID 7V02), IDR2 of Csm5 does indeed overlap with a short beta strand, but it is flanked by loops/unstructured regions. In addition, of the 46 residues that we deleted in the Csm546 mutant, 20 residues are unresolved in the experimentally-determined structure, supporting the notion that this region is generally flexible. Also, it is unlikely that this and the other Csm5 deletion mutants are misfolded because they all retain the ability to associate with the complex (Fig. 4B), and we were able to readily purify the mutant with the largest deletion (Csm546) without any issues (Fig. 5). To address this concern, we added panel D in Figure 4-figure supplement 1, which highlights the IDR regions in Csm5 from the recently-published S. epidermidis Cas10-Csm complex structure and integrated the observations mentioned above in the narrative (lines 241-247 in the marked-up revised manuscript). We also softened the conclusions based on these experiments (lines 276-278 in the marked-up revised manuscript): “Taken together, these results suggest that the IDR2 region of Csm5 likely plays a role in the recruitment and stimulation of PNPase, while the binding site for RNase R may reside elsewhere in Csm5”.

    1. The native gels testing interactions between Csm5 and RNase R show a slight change in mobility of RNase R upon the addition of Csm5. Although I agree with the authors' interpretation that this shift could be due to transient interactions between Csm5 and RNase R, it is also possible that the mobility of RNase R is affected simply based on the addition of a large excess of a second protein, even without a specific interaction between the two proteins. As a result, the evidence for direct interaction with Csm5 is limited. Discussion of how RNaseR is recruited by the Csm complex could contain more possible explanations. For example, it is possible that the interaction between RNase R and the Csm complex is mediated by another protein (e.g. PNPase could bridge interaction between the two) or that such an interaction could be stabilized by intermediate crRNA or target RNA binding by the Csm complex.

    Thank you for this comment. To help rule out the possibility that excess Csm5 could cause a shift of any protein nonspecifically, we included a control in the original manuscript in which the same native gel assay was performed with BSA and Csm5, and found that Csm5 fails to cause an upward shift in BSA (Figure 3-figure supplement 1). In addition, to bolster the claim of a direct interaction between Csm5 and RNase R, we performed an additional pulldown assay (Figure 3-figure supplement 2). Details are described under the essential revisions point number 3 above. Regarding the other possibilities mentioned, it is unlikely that PNPase is bridging the interaction with RNase R because when we delete PNPase from cells, we still get some maturation (Fig. 1E and Chou-Zheng and Hatoum-Aslan, eLife, 2019). Also, in the reconstituted system, RNase R can still perform some level of maturation on its own (Fig. 2D). These observations argue against the need for bridging interactions with PNPase. Furthermore, maturation occurs in the absence of target RNA, ruling out the possibility that target RNA bridging is necessary for RNase R-mediated crRNA maturation. However, we agree with the reviewer that it is possible that other components of the Cas10-Csm complex may help to recruit and stabilize the interaction with RNase R in vivo, and this possibility was already mentioned in the narrative in the original submission, although we did not explicitally state the intermediate crRNA as one such component (lines 213-215 and again in lines 413-416 in the marked-up revised manuscript). We have replaced “subunits” with “components” in line 415 to be more inclusive of this possibility. Since this is all still speculative, we opt not to elaborate further on this point in the current manuscript. Needless to say, we are actively pursuing other more quantitative assays to measure the interactions between Csm5 and PNPase/RNase R and hope to have such data available in a follow-up manuscript.

    1. On lines 367-391, the authors propose a model for how PNPase and RNase R may contribute to defense against foreign DNA through their recruitment by the Csm complex to the target transcript. However, their experiments do not test whether PNPase and RNase R must interact with the Csm complex to support anti-plasmid activity. Indeed, it may make more sense for free RNase R to be involved in defense, similar to how free activated Csm6 degrades transcripts non-specifically, rather than only cleaving transcripts in close proximity to the Csm complex. The authors could expand their discussion to mention the possibility that free RNase R or PNPase are acting in anti-plasmid defense.

    Thank you for this suggestion. The following statement has been added to the discussion (lines 393-395 in the marked-up revised manuscript): “Once recruited by the complex, PNPase and RNase R may degrade nucleic acids in the vicinity nonspecifically, similarly to Csm6.”

    Reviewer #2 (Public Review):

    This work follows up on an earlier publication that showed PNPase and RNase J2 play important roles in CRISPR RNA processing (doi: 10.7554/eLife.45393). Here, the authors show that RNase R also plays a critical role in CRISPR RNA maturation. In addition, they show that RNase R and PNPase are both recruited to the type III CRISPR complex (Cas10-Csm) via direct interactions with the Cmr5 subunit and that deletion of an intrinsically disordered region (IDR2) on Cmr5 selectively inhibits PNPase recruitment but not RNase R. The authors show unquantified stimulation of PNPase nuclease activity by Cmr5. Phage challenge assays are performed to test the impact of PNPase and RNase R deletion mutations on CRISPR-Cas mediated phage defense. Contrary to expectation, over-expression of the CRISPR system in cells that contain a deletion of PNPase and/or RNase R, maintain robust anti-phage immunity. The interpretation of this experiment is that RNase R and PNPase may be dispensable in an over-expression system that produces high (non-natural) concentrations of the Csm complex. They test this idea using a system that expresses the CRISPR-Cas components off of a chromosomally encoded locus (strain RP62a) and challenge these cells using a plasmid conjugation assay. In this iteration, deletion of PNPase has no impact on CRISPR performance, while deletion of RNase R "exhibited a moderate" attenuation of the immune response. In contrast, to either single gene deletion, the PNPase and RNase R double mutant showed a near complete loss of immunity.

    Overall, the paper provides convincing evidence that PNPase and RNase R are involved in crRNA processing, and that they are recruited to the type III complex via Cmr5. The work on RNase R is entirely new and the role of PNPase is expanded. The role of cellular RNases in CRISPR RNA biogenesis is important, though some of the results are subtle and some of the biochemistry would benefit from a more quantitative analysis.

    Thank you for your thorough assessment and comments overall.

  3. eLife

    eLife assessment

    CRISPR-Cas systems are essential components of an adaptive immune system that protects bacteria and archaea from infection by foreign genetic elements like phages and plasmids. The work presented here demonstrates that some CRISPR systems (i.e., type III-A) rely on host nucleases (i.e., RNase R and PNPase) for faithful processing of CRISPR RNAs. Collectively, this work expands the fundamental understanding of how nucleases involved in RNA metabolism contribute to the adaptive immune response in bacteria.

  4. eLife

    Reviewer #1 (Public Review):

    In this manuscript, Chou-Zheng and Hatoum-Aslan follow up on their previous studies that have characterized the collaborations between the type III-A CRISPR-Cas10 Csm complex and various cellular housekeeping nucleases. The authors have previously demonstrated that the Csm complex associates with several nucleases that are implicated in RNA degradation via pulldown and mass spectrometry analysis. They also previously showed that some of these enzymes, including PNPase, are important for CRISPR RNA (crRNA) maturation and for robust anti-phage defense. They now show that a second housekeeping enzyme, RNase R, is required for crRNA maturation. PNPase and RNase R act in concert to produce the mature crRNA. The authors also analyze the interactions between Csm5 and both housekeeping proteins. Finally, they demonstrate that PNPase and RNase R are important for robust anti-plasmid activity when using crRNAs that are complementary to low-abundance transcripts.

    This is a well-written paper with clear figures and well-described experiments and results. The experiments in Figures 1 and 2 demonstrating the importance of RNase R for crRNA maturation are excellent. The biochemistry experiments in Figure 2 are especially convincing, in which the authors were able to reconstitute the concerted activities of RNase R and PNPase for crRNA biogenesis. The experiments in Figure 5 implicating PNPase and RNase R in robust anti-plasmid activity when targeting low-abundance transcripts are also clear and convincing, and the result is intriguing. Overall, these experiments provide a new example in a growing list of co-opted host proteins that are important for crRNA biogenesis and CRISPR-mediated defense.

    I do have some concerns about experiments in Figures 3 and 4 analyzing interactions between PNPase or RNase R and the Csm5 subunit of the Csm complex, and I believe that some of the authors' conclusions are not fully supported by the evidence presented in these experiments. These concerns, along with a question about their model, are detailed below.

    1. The authors used the structure of S. thermophilus Csm5 to guide their design of truncations to probe potential intrinsically disordered regions (IDR1 and IDR2) that may be sites of interaction with PNPase or RNase R. Since the authors submitted their manuscript, an AlphaFold predicted structure of the S. epidermidis Csm5 has been released on the AlphaFold Protein Structure Database. In this model, the IDR2 region is predicted by AlphaFold to be a beta strand at the center of a beta sheet, rather than a disordered region. If the prediction is accurate, deletion of this strand could cause Csm5 to misfold, making it difficult to interpret what causes loss of interaction with PNPase (i.e. deletion of a specific interaction surface versus misfolding of the overall tertiary structure). In light of this, the discussion surrounding these experiments should be altered to include more caveats about the truncations, and conclusions based on this experiment should be softened.

    2. The native gels testing interactions between Csm5 and RNase R show a slight change in mobility of RNase R upon the addition of Csm5. Although I agree with the authors' interpretation that this shift could be due to transient interactions between Csm5 and RNase R, it is also possible that the mobility of RNase R is affected simply based on the addition of a large excess of a second protein, even without a specific interaction between the two proteins. As a result, the evidence for direct interaction with Csm5 is limited. Discussion of how RNaseR is recruited by the Csm complex could contain more possible explanations. For example, it is possible that the interaction between RNase R and the Csm complex is mediated by another protein (e.g. PNPase could bridge interaction between the two) or that such an interaction could be stabilized by intermediate crRNA or target RNA binding by the Csm complex.

    3. On lines 367-391, the authors propose a model for how PNPase and RNase R may contribute to defense against foreign DNA through their recruitment by the Csm complex to the target transcript. However, their experiments do not test whether PNPase and RNase R must interact with the Csm complex to support anti-plasmid activity. Indeed, it may make more sense for free RNase R to be involved in defense, similar to how free activated Csm6 degrades transcripts non-specifically, rather than only cleaving transcripts in close proximity to the Csm complex. The authors could expand their discussion to mention the possibility that free RNase R or PNPase are acting in anti-plasmid defense.

  5. eLife

    Reviewer #2 (Public Review):

    This work follows up on an earlier publication that showed PNPase and RNase J2 play important roles in CRISPR RNA processing (doi: 10.7554/eLife.45393). Here, the authors show that RNase R also plays a critical role in CRISPR RNA maturation. In addition, they show that RNase R and PNPase are both recruited to the type III CRISPR complex (Cas10-Csm) via direct interactions with the Cmr5 subunit and that deletion of an intrinsically disordered region (IDR2) on Cmr5 selectively inhibits PNPase recruitment but not RNase R. The authors show unquantified stimulation of PNPase nuclease activity by Cmr5. Phage challenge assays are performed to test the impact of PNPase and RNase R deletion mutations on CRISPR-Cas mediated phage defense. Contrary to expectation, over-expression of the CRISPR system in cells that contain a deletion of PNPase and/or RNase R, maintain robust anti-phage immunity. The interpretation of this experiment is that RNase R and PNPase may be dispensable in an over-expression system that produces high (non-natural) concentrations of the Csm complex. They test this idea using a system that expresses the CRISPR-Cas components off of a chromosomally encoded locus (strain RP62a) and challenge these cells using a plasmid conjugation assay. In this iteration, deletion of PNPase has no impact on CRISPR performance, while deletion of RNase R "exhibited a moderate" attenuation of the immune response. In contrast, to either single gene deletion, the PNPase and RNase R double mutant showed a near complete loss of immunity.

    Overall, the paper provides convincing evidence that PNPase and RNase R are involved in crRNA processing, and that they are recruited to the type III complex via Cmr5. The work on RNase R is entirely new and the role of PNPase is expanded. The role of cellular RNases in CRISPR RNA biogenesis is important, though some of the results are subtle and some of the biochemistry would benefit from a more quantitative analysis.

  6. Version published to 10.1101/2022.07.18.500432v1 on bioRxiv