Detection and characterization of spacer integration intermediates in type I-E CRISPR-Cas system

Abstract

The adaptation against foreign nucleic acids by the CRISPR-Cas system (Clustered Regularly Interspaced Short Palindromic Repeats and CRISPR-associated proteins) depends on the insertion of foreign nucleic acid-derived sequences into the CRISPR array as novel spacers by still unknown mechanism. We identified and characterized in Escherichia coli intermediate states of spacer integration and mapped the integration site at the chromosomal CRISPR array in vivo. The results show that the insertion of new spacers occurs by site-specific nicking at both strands of the leader proximal repeat in a staggered way and is accompanied by joining of the resulting 5'-ends of the repeat strands with the 3'-ends of the incoming spacer. This concerted cleavage-ligation reaction depends on the metal-binding center of Cas1 protein and requires the presence of Cas2. By acquisition assays using plasmid-located CRISPR array with mutated repeat sequences, we demonstrate that the primary sequence of the first repeat is crucial for cleavage of the CRISPR array and the ligation of new spacer DNA.

Figure 1.

Southern analysis of CRISPR locus after induction of cas1–cas2 expression. (A and B) The experimental procedure to test for intermediate states in the course of spacer acquisition. Escherichia coli BL21-AI cells were transformed with pCR001 harboring cas1–cas2 operon as indicated in (A). CRISPR array located on the genomic DNA (gDNA) is colored in red. Genomic DNA was isolated either from cycle 1 or cycle 2 cells with or without induction of plasmid-encoded cas1–cas2 expression, respectively (see text for details). DNA samples were digested with restriction enzyme HindIII, separated on a 0.7% agarose gel (20 μg DNA in each lane) and blotted onto nylon membrane. (C) The fragments with the CRISPR locus were visualized by hybridization of radiolabeled probes complementary to the upstream region of the leader DNA (probe 9, Supplementary Table S4) or to the second spacer of the CRISPR locus (probe 10, Supplementary Table S4). In addition to the native CRISPR locus with a length of 1953 nt, shorter bands were obtained (indicated by arrow heads) when cas1–cas2 expression was induced (lanes 4 and 5).

Figure 2.

Southern analyses with genomic DNA after induction of spacer integration. Genomic DNAs were prepared from BL21-AI strains harboring either pCR001 (expressing wild-type Cas1 and Cas2 proteins) or pCR002 (expressing D221A Cas1 and wild-type Cas2), grown for 24 h without (−) or with IPTG/arabinose-induction (+). 27.5 μg (in C and D) or 30 μg (in F) of Dra- or Ban-digested DNA were separated on 10% denaturing polyacrylamide gel and blotted onto nylon membrane. The fragments were visualized by hybridization with radiolabeled oligonucleotides complementary to the non-template or template strand of the leader DNA (probes 1 and 2), the second spacer sequence (probes 3 and 4), or the first spacer (probes 5 and 6). Autoradiograms of the Southern blots obtained with DraI-digested (C, F) or BanI-digested genomic DNA (D, F) are shown. Cas1 and Cas2-dependent cleavage products and their lengths are indicated by arrows. (A and B) The schemes depict the expected lengths of the DNA fragments in the case of a putative Cas1 and Cas2-directed staggered cut at the first repeat sequence. (E) Modified model for the intermediates is shown, which considers a concerted cut and ligation of a new spacer DNA (S0, green). R: repeat; S: spacer.

Figure 3.

Verification of the gaps. Southern blots were performed as described in Figure 2. Radiolabeled probes complementary to the repeat sequence of the non-template (probe 7) or template strand (probe 8) were hybridized against Dra- or BanI-digested genomic DNA. Schemes of intermediates with one new spacer (S0, green) (A) or two new spacer (S0, green and S02, magenta) (B). (C) Autoradiograms of Southern blots with Dra- (lanes 1–3 and 8–10) and BanI-digested DNA (lanes 5–7 and 12–14) using probe 7 or 8, as indicated.

Figure 4.

Effects of mutations within the first repeat sequence on spacer acquisition. (A) A staggered cut at the first repeat sequence can occur either in a sequence- or structure-specific manner. Putative cruciform structure of the wild-type or mutated repeat-sequences are shown. The arrows indicate the positions of the nicks as deduced from Southern analysis and sequencing of expanded CRISPR arrays. (B) 1:50 or 1:500 dilutions of E. coli BL21-AI cultures transformed with pCR003WT (lanes 1–4) or pCR003RM1, grown for 18 h after induction of cas1–cas2 expression (cycle 1), were analyzed by PCR using primers 10 and 15 (amplification of the genomic array) or primers 16 and 17 (amplification of the synthetic array on the plasmids). The PCR products were separated on 1.2% agarose gels. A second round of spacer acquisition was started through transfer of aliquots of the culture into fresh IPTG/arabinose-containing medium and incubation for additional 18 h (cycle 2). The PCR products of the parental CRISPR locus (parental) and those with new spacer DNA (extended) are indicated. (C) The same as in (B) but using cells transformed with pCR003RM2 (lanes 6–9).

Figure 5.

Southern analyses of plasmid-located CRISPR arrays after induction of cas1–cas2 expression. (A) The scheme depicts the expected lengths of DNA fragments in the case of a Cas1 and Cas2-directed cleavage and ligation of new spacer DNA (S0). The plasmid DNA was linearized either with EcoRI (120 bp upstream of the leader-repeat junction) or with KpnI (126 bp downstream of the first repeat). One microgram of each linearized plasmids were separated on 10% denaturing polyacrylamide gels. (B) Southern analysis with radiolabeled oligonucleotides (probes 11 and 12, Supplementary Table S4) against the leader DNA of EcoRI linearized plasmids is shown. Lanes 1, 5, 7, 11: length marker; lanes 2, 8: wild-type plasmids isolated from cells without induction of cas1–cas2 expression; lanes 3, 9: wild-type plasmids isolated from cells with induction of cas1–cas2 expression; lanes 4, 10: pCR003RM1 plasmids isolated from cells with induction of cas1–cas2 expression; lanes 6, 12: pCR003RM2 plasmids isolated from cells with induction of cas1–cas2 expression (C) The same as in (B) but with radiolabeled oligonucleotides against the spacer (probes 13 and 14, Supplementary Table S4) of KpnI linearized plasmids.