Thousands of new phages have recently been discovered thanks to viral metagenomics. These phages are extremely diverse and their genome sequences often do not resemble any known phages. To appreciate their ecological impact, it is important to determine their bacterial hosts. CRISPR spacers can be used to predict hosts of unknown phages, as spacers represent biological records of past phage–bacteria interactions. However, no guidelines have been established to standardize host prediction based on CRISPR spacers. Additionally, there are no tools that use spacers to perform host predictions on large viral datasets. Here, we developed a set of tools that includes all the necessary steps for predicting the hosts of uncharacterized phages. We created a database of >11 million spacers and a program to execute host predictions on large viral datasets. Our host prediction approach uses biological criteria inspired by how CRISPR–Cas naturally work as adaptive immune systems, which make the results easy to interpret. We evaluated the performance using 9484 phages with known hosts and obtained a recall of 49% and a precision of 69%. We also found that this host prediction method yielded higher performance for phages that infect gut-associated bacteria, suggesting it is well suited for gut-virome characterization.

CRISPR-Cas systems in prokaryotic cells provide an adaptive immunity against invading nucleic acids. For example, phage infection leads to addition of new immunity (spacer acquisition) and DNA cleavage (interference) in the bacterial model species Streptococcus thermophilus, which primarily relies on Cas9-containing CRISPR-Cas systems. Phages can counteract this defense system through mutations in the targeted protospacers or by encoding anti-CRISPR proteins (ACRs) that block Cas9 interference activity. Here, we show that S. thermophilus can block ACR-containing phages when the CRISPR immunity specifically targets the acr gene. This in turn selects for phage mutants carrying a deletion within the acr gene. Remarkably, a truncated acrIIA allele, found in a wild-type virulent streptococcal phage, does not block the interference activity of Cas9 but still prevents the acquisition of new immunities, thereby providing an example of an ACR specifically inhibiting spacer acquisition.

It contained an error in Fig. 4A, in which a 6-amino acid insertion (positions 101-106) was incorrectly shown in protein AcrIIA6 123, compared to AcrIIA6 D1811. The correct figure now shows that this 6-amino acid sequence is also present in AcrIIA6 D1811.It contained an error in Fig. 4C, in which a 6-amino acid insertion (positions 101-106) was incorrectly highlighted using green colour. This highlighting has been removed from the figure.It contained an error in Fig. 4E, in which a 6-amino acid insertion (positions 101-106) was incorrectly highlighted using a dotted ellipse. The dotted ellipse has been removed from the figure.It contained errors in the legend of Fig. 4, which incorrectly read 'Red and green represent amino acid substitutions and amino acid insertions, respectively' and 'The dotted ellipse highlights the position of the amino acid insertion in AcrIIA6 123 '. The correct version replaces the first sentence with 'Red patches represent amino acid substitutions', and removes the second sentence.It contained an error in a sentence of the 'Results and discussion' section, which incorrectly read 'This 3D model of AcrIIA6 123 highlights a 6-residue insertion (K101-I106) in the loop connecting the β-sheet to the C-terminal α-helix, as well as some amino acid substitutions distributed over the entire structure'. The correct version replaces this sentence with 'This 3D model of AcrIIA6 123 highlights amino acid substitutions distributed over the entire structure'.It contained an error in a sentence of the 'Results and discussion' section, which incorrectly read 'Interestingly, the monomeric AcrIIA6 123 presents three notable features by (1) offering a much smaller binding surface than that of the dimeric AcrIIA6, (2) harboring the amino acid insertion in an RNA-interacting loop, and (3) containing an amino acid substitution (I23 in AcrIIA6 123 instead of N81 in AcrIIA6) that likely disrupts the St1Cas9-RNA-binding interface'. The correct version replaces this sentence with 'Interestingly, the monomeric AcrIIA6 123 presents two notable features by (1) offering a much smaller binding surface than that of the dimeric AcrIIA6, and (2) containing an amino acid substitution (I23 in AcrIIA6 123 instead of N81 in AcrIIA6) that likely disrupts the St1Cas9-RNA-binding interface'.The errors have been corrected in both the PDF and HTML versions of the Article.