Avalilação PREreview de Viral SSB-bound ssDNA activates the bacterial anti-phage defense system DARNA

The authors present a very intriguing anti-phage defense system named DARNA (ssDNA-Activated RNase), which provides abortive defense against diverse phages via cleaving cellular tRNAs at the anti-codon loop. The authors follow up on a previously published screen which identified the defense profile of this system. The mechanism of defense mediated by DARNA is uncovered through forward genetics, RNA sequencing, biochemical, and structural studies. Curiously, DARNA is activated by various structurally distinct phage single-stranded DNA-binding proteins (SSBs) in combination with ssDNA in vitro. The data supports the claim that in vitro SSBs are necessary, but not sufficient for DARNA activation. Since they have differing biophysical properties, structural insights into the mechanism of SSB activation would clarify how divergent proteins could lead to the same phenotype. Additionally, further data would help to understand how an inter-hexameric interface-disrupted mutant (F6A+N33A+E347A) still confers defense. Lastly, a key component of the in vitro reconstitution is a short synthetic ssDNA oligonucleotide named “34 nt.” While this acts as a potent activator of DARNA in vitro, it is not representative of its likely endogenous ssDNA activating sequence during phage infection. While in vitro and structural data demonstrate that DARNA binds ssDNA, there is limited evidence showing that this occurs during infection. All in all, this preprint contains compelling data, demonstrating a new mode of molecular pattern sensing during phage infection which confers broad defense.

Major Points

• In Supplementary Figure S4, the authors demonstrate by mass photometry that structure-guided mutations (F6A+N33A+E347A) disrupt the DARNA dodecamer. However, it appears that the mutant can still provide moderate protection against phages T4 and T5, even though in vitro its RNase activity is gone. Therefore, the phage spot assay data contradicts the suggested model that the dodecamer is required for activity and thus phage defense. It is possible that while the dodecamer is the functional unit in vitro, DARNA might behave differently during infection. It would be interesting to know if this mutant induces toxicity when co-expressed with the different phage SSBs. This would greatly improve the understanding of the system and show that while a dodecamer can bind a synthetic ssDNA oligonucleotide in vitro, a dimer is still sufficient to provide defense and be activated by SSB in vivo.

• Through in vitro reactions, the authors show that a synthetic ssDNA oligonucleotide 34 bases in length is sufficient to activate DARNA-mediated tRNA cleavage. However, it is not shown how the authors arrived at this specific oligonucleotide sequence especially since there are no identical sequences present in the genome of the bacteria they studied. The DARNA system functions in vivo and therefore must have a ssDNA-binder within the cell during infection, which could be of bacterial or phage origin. Additional experiments, such as immunoprecipitation of DARNA from phage-infected cells followed by NGS of co-precipitated DNA, could potentially yield information on the endogenous ssDNA binding sequence. This information could not only provide a potentially more efficient activator but also show whether DARNA is activated by the same sequence during T3, T4, T5, and T7 phage infections.

• In Figure 1C, the authors show that co-expression with phage SSBs alone induces DARNA-mediated toxicity. While the phage SSBs alone are not toxic to the bacteria, it is not clear if DARNA alone is toxic. This is especially important since later figures show that DARNA can cleave tRNA in the absence of phage SSB activators. This panel would benefit from a control with DARNA expression alone to support the claim that phage SSB activation is responsible for the toxicity. It is known that many bacterial immune systems, when overexpressed, have a growth effect. Therefore, this experiment would be strengthened by showing that DARNA has no background toxicity and the observed toxicity is solely through co-expression with SSBs.

Minor Points

Figure 1:

• The main takeaway from Figure 1 is that the DARNA defense system is activated by structurally diverse phage SSBs. However, how these phage SSBs differ is not shown. Moving the structures and alignments from Figure S1B to Figure 1 would significantly aid in understanding the main claim.

• Through phage plaque, bacterial growth, and colony formation assays, the authors show that SSBs from phages T3, T4, T5, and T7 can all activate DARNA-mediated defense or toxicity. While the data is very convincing, not all mentioned phage SSBs are represented in each assay. The preprint as a whole would benefit from including all SSBs in each experiment, and either including the data in the main figures or moving the data to the supplement. It will be interesting to know if each SSB functions similarly in each experiment.

• In Figure 1D, the authors co-express T7 SSB WT or variants with DARNA and show effects on bacterial growth over time. While the data is convincing, it is not clear that the background of this experiment is DARNA-expressing cells. This panel would benefit from formatting like panel B, where DARNA expression is noted.

• Figure 1 demonstrates DARNA-mediated defense and activation by phage SSBs. However, the dependence of these phenotypes on DARNA RNase activity is unclear. DARNA may illicit immunity through its RNase activity, or through interactions independent of it. Therefore, this figure would be strengthened through the inclusion of the predicted RNase-dead mutant H122A. This would support the claim that defense and toxicity results from DARNA activity.

Figure 2:

• The authors show RNA-sequencing data in which normalized read ends are mapped onto the MG1655 genome. The data strikingly shows increased read-ends mapping to the location of tRNAs. However, the authors show in Figure S2 that T7 phage tRNAs are unaffected. This data does not support the claim that DARNA acts as a promiscuous RNase. The paper would benefit from some elaboration on why phage tRNAs are mostly unaffected, such as providing information on any differences in the anti-codon loop sequences.

Figure 3:

• The authors provide mass photometry data of the oligomerization states of purified DARNA complex, in which they mention was checked by SDS-PAGE. The SDS-PAGE could be included in a supplemental figure, to validate the purity of the complex.

• Figure 3 panel B shows that a dodecamer of DARNA shifts in mass when presented with the “64 nt” species of ssDNA. The oligonucleotide is not consistent with what’s used in vitro for the tRNA cleavage assays, in Figure 3D, and for the cryoEM models. The proposed mechanism includes ssDNA binding and subsequent RNase activity. However, for these to be linked, it would be beneficial to have the same oligonucleotide used in each experiment. Additional data showing that DARNA can bind the “34 nt” species by mass photometry, or that the “64 nt” species that DARNA binds can also activate tRNA cleavage in vitro, would support the proposed mechanism.

• In Figure 3D, the authors assert that the DARNA dodecameric complex significantly shifts in the presence of the “34 nt” ssDNA species and T4 SSB. However, the figure would benefit from an additional mass photometry run showing that the purified T4 SSB alone can also bind the ssDNA oligonucleotide, as it is known to not only bind ssDNA but also form filaments that cover approximately 7 nt segments. This would support the claim that there is a coupled-binding mechanism. Additionally, T7 SSB is used in the in vitro reconstitution, but is not shown in the mass photometry data to bind to ssDNA or to the ssDNA:DARNA complex.

Figure 4:

• In Figure 3D, the authors present compelling data that the addition of ssDNA and T4 SSB results in a shift in the DARNA dodecamer to a mass which is greater than what is shown for DARNA and ssDNA alone. This suggests that SSB is also a part of this complex. However, there is no structural interrogation of this potential complex, which is likely the functional unit. It may be interesting, as a future direction, to see if incubation with the T4 SSB with ssDNA and DARNA prior to cryoEM results in new density somewhere near the ssDNA.

Figure 5:

• In Figure 5B, the authors show that while increasing amounts of host SSB inhibits background tRNA cleavage by DARNA, adding phage SSBs increases the amount of RNase activity. The authors note that a ssDNA species is being used in the assay, but it is not mentioned which one. This figure would benefit from the mention of which ssDNA species is used.

Figure S4:

• The authors demonstrate that disrupting the dodecameric complex impacts the ability of DARNA to defend against phages T4 and T5 in panel C. In panel D, the authors then show that these mutations also impact the RNase activity of DARNA in vitro with T7 SSB. While this data clearly shows effects of this triple mutation in vivo and in vitro, it would be more convincing if the impact on phage T7 is shown in panel C.

Figure S6:

• The authors show EMSAs for the phage SSBs used in the in vitro studies. The data convincingly demonstrates the ssDNA binding; however, a control is missing that shows the EMSA for the purified E. coli SSB.

I reviewed this preprint as part of the Transparent and Accountable Peer Review training program. This report was written with the support of my research mentor.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.