Recombination between phages and CRISPR-cas loci facilitates horizontal gene transfer in staphylococci

Abstract

CRISPR (clustered regularly interspaced short palindromic repeats) loci and their associated (cas) genes encode an adaptive immune system that protects prokaryotes from viral¹ and plasmid² invaders. Following viral (phage) infection, a small fraction of the prokaryotic cells are able to integrate a small sequence of the invader's genome into the CRISPR array¹. These sequences, known as spacers, are transcribed and processed into small CRISPR RNA guides^3-5 that associate with Cas nucleases to specify a viral target for destruction^6-9. Although CRISPR-cas loci are widely distributed throughout microbial genomes and often display hallmarks of horizontal gene transfer^10-12, the drivers of CRISPR dissemination remain unclear. Here, we show that spacers can recombine with phage target sequences to mediate a form of specialized transduction of CRISPR elements. Phage targets in phage 85, ΦNM1, ΦNM4 and Φ12 can recombine with spacers in either chromosomal or plasmid-borne CRISPR loci in Staphylococcus, leading to either the transfer of CRISPR-adjacent genes or the propagation of acquired immunity to other bacteria in the population, respectively. Our data demonstrate that spacer sequences not only specify the targets of Cas nucleases but also can promote horizontal gene transfer.

Figure 1.. Transfer of CRISPR-Cas elements through spacer-mediated transduction.

(A) Transduction of CRISPR-adjacent loci (blue arrow) after phage recombination with a chromosomal CRISPR-cas locus; “S” CRISPR spacer, “PS” protospacer. (B) Transducing particle production from S. aureus strain 08BA02176 tagged with an erythromycin resistance cassette. Liquid cultures were infected at a MOI (multiplicity of infection) of 50. Wild-type phage (ctl) or phage with the target site of the first spacer of the CRISPR array inserted at 15 kb (P1) or 20 kb (P2) positions on the phage genome were used. Wild-type phage was also used to infect a strain with a chromosomally inserted phage-targeting spacer (C1). Mean + STD of 3 biological replicates are reported. Two-tailed unpaired t-test was used to calculate P value, *p = 0.023, ****p = 0.00058 (C) Transduction of a plasmid containing a CRISPR-cas locus after recombination with the phage genome. J1 and J2, recombination junctions. (D) Growth and phage titers of infected cultures of bacteria containing plasmids with either the type II-A CRISPR-Cas system (CRISPR⁺) or the empty vector control (CRISPR⁻). Liquid cultures were infected at a MOI of 1 with ΦNM4γ4. The growth of cultures was determined by measurement of absorbance at 600 nm (A₆₀₀). Titers, plaque forming units/ml (pfu/ml), and levels of transducing-immune phage particles, colony forming units/ml (cfu/ml), were determined every time point. No transducing-immune particles were detected using a vector control. Mean ± STD of 3 biological replicates are reported. (E) Levels of transduction during CRISPR adaption were determined by mixing cells at a 1 CRISPR⁺ (naïve, without a targeting spacer) to 5 CRISPR⁻ ratio and infecting the mix with ΦNM4γ4 at an MOI of 1. Cultures were collected 20 hours post infection and survivors resulting from the acquisition of new spacers or the transduction of the adapted pCRISPR were measured by enumerating colonies on plates containing different antibiotic combinations. As a control, the experiment was repeated after mixing non-CRISPR resistant cells with naïve CRISPR⁺ staphylococci. Mean ± STD of 3 biological replicates are reported.

Figure 2.. Spacers sequences determine frequency of pCRISPR transduction.

(A) Cultures containing pCRISPR were infected with ΦNM4γ4 at a MOI of 1. Expanded CRISPR arrays were analyzed by next generation sequencing. The reads per million (RPM) values of the acquired spacers were plotted against the ΦNM4γ4 genome (blue line, CRISPR-resistant). Lysates containing ΦNM4γ4 as well as pCRISPR transducing particles were collected at 20 hours post infection and were used to infect cells without CRISPR-cas at an MOI of 1. Cells were plated to collect pCRISPR transductants and their spacer content was analyzed by next-generation sequencing (green line, CRISPR-transduced). The ratio of transduced spacers over resistant spacers was also plotted for each acquired spacer (black line). Select spacers with low-(L), intermediate-(I), and high (H1, H2) ratios, along with phage pac site, are indicated. Mean of 4 replicates was reported. (B) Transduction efficiency, measured as the ratio of transductant colony forming units (cfu) to the total number of plaque forming units (pfu) of phage in the lysates obtained after infection of staphylococci harboring pCas9 carrying the L, I, H1 and H2 spacers, or no targeting spacer as a control (ctl). Mean + STD of 3 biological replicates are reported. Two-tailed unpaired t-test was used to calculate P values; ns, not significant (p = 0.1554), *p (H1) = 0.024, *p (H2) = 0.024 (C). Same as (B) but measuring the transduction of pSpacer plasmids; i.e. not carrying cas9. Mean + STD of 3 biological replicates are reported. One-way ANOVA was used to calculate P values; ns, not significant (p = 0.59). (D) PCR products after amplification of J1 and J2 junctions (Fig. 1C) either from lysates (phage DNA) or infected cells (cell DNA) obtained after infection of cells harboring pCas9 (+Cas9) or pSpacer (-Cas9) plasmids with L, I, H1 or H2 spacers. These results were representative of three independent experiments. (E) Next generation sequencing of phage DNA harvested after infection of cells pCas9(H1). Reads were aligned to the putative J1 junction. Reads per million (RPM) for each nucleotide within this region are shown. Dotted lines indicate a 75-nucleotide sequence that is unique to the junction. Results from a single experiment are shown. (F) Same as (E) but for the J2 junction. Results from a single experiment are shown.

Figure 3.. Spacers that mediate high pCRISPR transduction provide poor immunity to the host.

(A) Cell survival of cultures containing pCas9 with spacers L, I, H1 or H2, monitored as growth after infection with ΦNM4γ4 at a MOI of 1. No spacer (ctl) and no phage experiments were used as controls. Growth was recorded by following the absorbance at 600 nm (A₆₀₀) of the cultures. Mean - STD of 3 biological replicates are reported. (B) Cell survival of cultures lacking CRISPR-Cas immunity, monitored as growth after infection with a cell lysate obtained after infection of staphylococci harboring pCRISPR (or an empty vector as a control) with ΦNM4γ4 at a MOI of 1. This lysate contains both phage as well as plasmid transducing plasmids. Growth was recorded by following the absorbance at 600 nm (A₆₀₀) of the cultures. Mean ± STD of 5 biological replicates are reported. (C) Transduction efficiency of plasmid #2, measured as transductant cfu / total pfu present in lysates of staphylococci harboring two pCas9 plasmids with different spacers, collected 90 minutes after infection with ΦNM4γ4 at a MOI of 50. Limit of detection is 1.5 cfu/ml (dotted line). Mean + STD of 3 biological replicates are reported. Two-tailed unpaired t-test was used to calculate P values, *p = 0.034, **p = 0.0012 (D) Model for spacer-mediated p CRISPR transduction. During the type II-A CRISPR-Cas immune response, efficient spacers fully protect the host and prevent the generation of both phage progeny and transducing particles. In contrast, spacers that mediate poor immunity allow the lysis of some of the infected hosts, enabling the release of pCRISPR transducing particles generated after the recombination of the plasmid and the phage genome at the spacer/protospacer sequences.