Determining the Specificity of Cascade Binding, Interference, and Primed Adaptation In Vivo in the Escherichia coli Type I-E CRISPR-Cas System

Abstract

In clustered regularly interspaced short palindromic repeat (CRISPR)-Cas (CRISPR-associated) immunity systems, short CRISPR RNAs (crRNAs) are bound by Cas proteins, and these complexes target invading nucleic acid molecules for degradation in a process known as interference. In type I CRISPR-Cas systems, the Cas protein complex that binds DNA is known as Cascade. Association of Cascade with target DNA can also lead to acquisition of new immunity elements in a process known as primed adaptation. Here, we assess the specificity determinants for Cascade-DNA interaction, interference, and primed adaptation in vivo, for the type I-E system of Escherichia coli Remarkably, as few as 5 bp of crRNA-DNA are sufficient for association of Cascade with a DNA target. Consequently, a single crRNA promotes Cascade association with numerous off-target sites, and the endogenous E. coli crRNAs direct Cascade binding to >100 chromosomal sites. In contrast to the low specificity of Cascade-DNA interactions, >18 bp are required for both interference and primed adaptation. Hence, Cascade binding to suboptimal, off-target sites is inert. Our data support a model in which the initial Cascade association with DNA targets requires only limited sequence complementarity at the crRNA 5' end whereas recruitment and/or activation of the Cas3 nuclease, a prerequisite for interference and primed adaptation, requires extensive base pairing.IMPORTANCE Many bacterial and archaeal species encode CRISPR-Cas immunity systems that protect against invasion by foreign DNA. In the Escherichia coli CRISPR-Cas system, a protein complex, Cascade, binds 61-nucleotide (nt) CRISPR RNAs (crRNAs). The Cascade complex is directed to invading DNA molecules through base pairing between the crRNA and target DNA. This leads to recruitment of the Cas3 nuclease, which destroys the invading DNA molecule and promotes acquisition of new immunity elements. We made the first in vivo measurements of Cascade binding to DNA targets. Thus, we show that Cascade binding to DNA is highly promiscuous; endogenous E. coli crRNAs can direct Cascade binding to >100 chromosomal locations. In contrast, we show that targeted degradation and acquisition of new immunity elements require highly specific association of Cascade with DNA, limiting CRISPR-Cas function to the appropriate targets.

Keywords: CRISPR; Cas3; Cascade; PAM; priming; protospacer.

FIG 1

Extensive off-target Cascade binding in E. coli. (A) Binding profile of Cse1 across the E. coli genome, as determined by ChIP-seq, for Cse1-FLAG₃ cells (AMD543) carrying a plasmid expressing crRNA targeting the lacZ promoter region (pCB380). The graph indicates the relative levels of sequence read coverage (see Materials and Methods for details) across the genome in a Cse1 ChIP-enriched sample. Bar, 1 Mbp. The location of the on-target binding site at the lacZ promoter is shown by the red arrow. (B) Binding profile of Cas5 across the E. coli genome, as determined by ChIP-seq, for FLAG₃-Cas5 cells (AMD554) with a plasmid expressing crRNA targeting the araB promoter region (pCB381). The location of the on-target binding site at the araC promoter is shown by the red arrow. (C) Enriched sequence motif associated with off-target Cascade binding sites in cells targeting the lacZ promoter, as determined by MEME. The likely PAM sequence is indicated. The number of identified motifs and the MEME E value are shown. (D) Enriched sequence motif associated with off-target Cascade binding sites in cells targeting the araB promoter, as determined by MEME. The likely PAM sequence is indicated.

FIG 2

Endogenous crRNAs drive Cascade association with over 100 chromosomal sites. (A) Binding profile of Cse1 across the E. coli genome, as determined by ChIP-seq, for Cse1-FLAG₃ cells (AMD543) expressing only endogenous crRNAs. The location of the binding site within yggX is shown by the red arrow. (B) Enriched sequence motifs associated with Cascade binding sites in cells expressing only endogenous crRNAs. The four motifs are associated with four of the CRISPR-I spacers, as indicated. The likely PAM sequence is also indicated. The number of identified motifs and the MEME E value are shown. (C) Comparison of Cascade binding events in Cse1-FLAG₃ cells with both CRISPR arrays intact (AMD543) or with CRISPR-II deleted (LC060). Sequence read coverage is shown for CRISPR-1⁺ CRISPR-2⁺ cells (AMD543) and CRISPR-I⁺ ΔCRISPR-II cells (LC060), for all ChIP-seq peaks identified for either strain. (D) Binding profile of Cse1 across the E. coli genome, as determined by ChIP-seq, for Cse1-FLAG₃ cells expressing only endogenous crRNAs but with CRISPR-I deleted (LC077).

FIG 3

CRISPR-I spacer 8 is responsible for the majority of Cascade binding in cells expressing only endogenous crRNAs. (A) Comparison of Cse1-FLAG₃ binding events in cells with both CRISPR arrays intact (AMD543) and in cells with CRISPR-I deleted (LC077) that express CRISPR-I spacer 8 from a plasmid (pLC008). Sequence read coverage is shown for all ChIP-seq peaks identified for either strain. ChIP-seq peaks associated with the CRISPR-I spacer 8 motif (first motif listed in Fig. 2B) are shown in orange. (B) Predicted base-pairing interaction between CRISPR-I spacer 8 and a protospacer within yggX. The PAM is underlined. (C) ChIP-qPCR measurement of Cse1 binding at wild-type (i and iii; AMD566) and mutant (ii and iv; LC099) protospacers in yggX for cells expressing wild-type (i and ii; pLC008) or mutant (iii and iv; pLC010) CRISPR-I spacer 8 from a plasmid. The mutations in spacer 8 restored base-pairing potential with the mutant protospacer, as indicated. Values represent averages of results from three independent replicate experiments. Error bars show 1 standard deviation from the mean.

FIG 4

Off-target Cascade binding events are not associated with interference or primed adaptation. (A) Relative levels of efficiency of transformation of a cas3-expressing plasmid (pAMD191) into LC103 cells expressing spacer 8 from a native CRISPR-I array and containing (i) empty pBAD24 (“No Protospacer”), (ii) a plasmid with a protospacer that base pairs perfectly with spacer 8 (pLC022), (iii) a plasmid with a protospacer that has only partial base pairing with CRISPR-I spacer 8 (pLC021; the protospacer sequence matches the off-target Cascade binding site in yggX), or (iv) a plasmid with a protospacer that base pairs perfectly with CRISPR-I spacer 2 (pLC057). Note that crRNAs were expressed from the chromosome, since both CRISPR arrays are intact in these strains. Transformation efficiency was calculated relative to that of empty pBAD33, as described in Materials and Methods. Values represent averages of results from three independent replicate experiments. Error bars show 1 standard deviations from the means. The calculated transformation efficiency for protospacers ii and iv was 0, but the limit of detection in this assay was 3e⁻⁵. (B) PCR amplification of the start of the CRISPR-II array to detect primed adaptation in cells expressing CRISPR-I spacer 8 (AMD536) and cas3 (pAMD191) and with (i) a protospacer that base pairs perfectly with CRISPR-I spacer 8 (pLC022), (ii) the protospacer from yggX that has only partial base pairing with CRISPR-I spacer 8 (pLC021), or (iii) empty vector (pBAD24). L, molecular weight ladder, with marker sizes indicated. The expected PCR product sizes are indicated.

FIG 5

Assessment of Cascade binding, interference, and primed adaptation for a panel of protospacer variants. (A) Relative Cse1 association (in strain LC099, which expresses CRISPR-I spacer 8 crRNA from the native CRISPR-I array) for each of 13 protospacer variants. The protospacer variants are (i) an optimal protospacer that has a perfect match to CRISPR-I spacer 8 and an AAG PAM (pLC023); (ii) CCG PAM (pLC027); (iii) ATT PAM (pLC029); (iv) mismatches at positions 1 to 3 (pLC031; the wild type is CTG); (v) mismatches at positions 1 and 3 (pLC033); (vi) mismatches at positions 1 and 3 (pLC035); (vii) mismatches at positions 2 and 3 (pLC032); (viii) mismatches at positions 2 and 3 (pLC034); (ix) mismatches across positions 25 to 32 (pLC024); (x) mismatches across positions 19 to 32 (pLC025); (xi) mismatches across positions 1 to 6 (pLC026); (xii) mismatches across positions 1 to 6 and positions 25 to 32 (pLC028); and (xiii) mismatches across positions 7 to 24 (pLC030). Values represent averages of results from five independent replicate experiments. Error bars show 1 standard deviation from the mean. (B) Relative efficiency of interference (see Materials and Methods for details) for each of the indicated protospacer variants (pLC023 to pLC035). Values represent averages of results from three independent replicate experiments. Error bars show 1 standard deviation from the mean. (C) Levels of primed adaptation in AMD688 cells expressing CRISPR-I spacer 8 from a native CRISPR-I array and cas3 (pAMD191) and with each of the 13 indicated protospacer variants (pLC023 to pLC035). Adaptation was measured by monitoring the conversion of a yfp reporter construct into the actively expressed state (see Materials and Methods for details). Values shown represent the percentages of the population that converted to YFP⁺. Background levels of adaptation were detected for variants ii, x, xi, xii, and xiii.