Acquisition of Spacers from Foreign Prokaryotic Genomes by CRISPR-Cas Systems in Natural Environments

Abstract

Clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated (Cas) systems of bacteria and archaea provide immunities against mobile genetic elements, like viruses. In addition, protospacer analyses revealed a very specific acquisition of CRISPR spacers derived from genomes of related species or from closely interacting episymbiont genomes as recently shown for subsurface archaea. However, the origin of most of the spacers that can be found in CRISPR-Cas systems from natural environments has not been deciphered. Here, by analyzing CRISPR-Cas systems of metagenome-assembled genomes (MAGs) from two subsurface environments spanning more than 1 Tb of sequencing data, we show that a substantial proportion of CRISPR spacers are acquired from DNA of other prokaryotes inhabiting the same environment. As such, we found that the number of respective spacers can be up to three times higher than the number of self-targeting spacers. Statistical analyses demonstrated that the acquisition of CRISPR spacers from other prokaryotic genomes is partly explained by the relative abundance of the MAG containing the protospacer, as well as by other factors, such as the total number of CRISPR arrays present in a MAG with the respective spacers. Further, we found that spacer acquisition from foreign prokaryotic DNA occurs in almost all types of CRISPR-Cas systems, but shows preferences for subtypes of CRISPR-Cas systems that differ across the investigated ecosystems. Taken together, our results shed new light on the diversity of CRISPR spacers in natural microbial communities and provide an explanation for some of the many unmatched spacers in public databases.

Keywords: CRISPR spacer; metagenomics; population genomics; protospacer.

Fig. 1.

Phylogenetic distribution of spacer-to-protospacer matches (80% sequence similarity) within foreign prokaryotic genomes based on trees of MAGs (GTDB-Tk, Chaumeil et al. 2022) in GB (a) and WB (b). For results based on 90% sequence similarity for spacer-to-protospacer matches, please see Fig. S2. (c) Distribution of spacer clusters that match at least one MAG cluster in the respective taxonomic rank. Self-targeting spacers refer to spacer clusters matching the same MAG cluster. InterMAG cluster refers to all spacer clusters that match prokaryotic MAG clusters in the respective dataset including those MAG clusters that fall into the same species. Interspecies to interphylum refer to spacer clusters that match MAG clusters of different taxonomic ranks, taking into account the respective hierarchy. For instance, Interclass refers to all spacer clusters that match across classes but also phyla and domain as these are included in that data and are thus indicated by different colors. Please note that the bars are the cumulative number of spacer clusters matching across taxonomic ranks and include all higher taxonomic ranks (also denoted by the numbers next to the bars).

Fig. 2.

Schematic representation of newly suggested CRISPR-Cas interferences with foreign prokaryotic nucleic acids. Bacteria can acquire foreign prokaryotic DNA via various uptake mechanisms. While (the uptake of) environmental DNA can promote HGT and/or provide a source of nutrients and energy, it is intriguing to consider that it may also enable the acquisition of spacers from foreign prokaryotic DNA. Transcription and processing of the CRISPR array generate small crRNAs and allow for the assembly of the effector complexes. Following the recognition of foreign prokaryotic DNA, Cas nucleases allow for the degradation of the target sequence, e.g. to potentially prevent genome spoilage (i.e. the impairment of cellular fitness caused by genomic rearrangements, deletions and insertions). Image created in BioRender ( Sures 2025; https://BioRender.com/jzxgcl3).

Fig. 3.

Scatter plots showing the correlation of the number of spacers matching a MAG cluster and the normalized relative abundance of the same MAG cluster in all samples of GB (a) and WB (b), the correlation of the number of prokaryote-targeting spacers in a MAG cluster and the number of CRISPR arrays in a MAG cluster (c, d), the correlation of the number of prokaryote-targeting spacers in a MAG cluster and the number of total spacers in a MAG cluster (e, f) and the correlation of the total number of spacers in a MAG cluster and the number of CRISPR arrays in a MAG cluster (g, h). Spearman's rank correlation coefficients R and corresponding P-values were calculated for each correlation. Gray dashed lines show the linear regression fitted to all values of a panel. Extreme values smaller than the 0.5th and greater than the 95.5th percentile are highlighted in gray.