Spacer acquisition from RNA mediated by a natural reverse transcriptase-Cas1 fusion protein associated with a type III-D CRISPR-Cas system in Vibrio vulnificus

Abstract

The association of reverse transcriptases (RTs) with CRISPR-Cas system has recently attracted interest because the RT activity appears to facilitate the RT-dependent acquisition of spacers from RNA molecules. However, our understanding of this spacer acquisition process remains limited. We characterized the in vivo acquisition of spacers mediated by an RT-Cas1 fusion protein linked to a type III-D system from Vibrio vulnificus strain YJ016, and showed that the adaptation module, consisting of the RT-Cas1 fusion, two different Cas2 proteins (A and B) and one of the two CRISPR arrays, was completely functional in a heterologous host. We found that mutations of the active site of the RT domain significantly decreased the acquisition of new spacers and showed that this RT-Cas1-associated adaptation module was able to incorporate spacers from RNA molecules into the CRISPR array. We demonstrated that the two Cas2 proteins of the adaptation module were required for spacer acquisition. Furthermore, we found that several sequence-specific features were required for the acquisition and integration of spacers derived from any region of the genome, with no bias along the 5'and 3'ends of coding sequences. This study provides new insight into the RT-Cas1 fusion protein-mediated acquisition of spacers from RNA molecules.

Figure 1.

Characteristics of the VvYJ016 type III-D CRISPR operon. (A) Schematic diagram of the type III-D CRISPR–Cas loci in VvYJ016. The operon consists of a canonical five-gene cassette putatively encoding the type III-B Cmr effector complex (indicated by a beige background) followed by the gene encoding Cas6, which is involved in crRNA maturation. The opposite strand carries the adaptation module, which consists of an operon of three genes encoding RT-Cas1 and two Cas2 proteins, located between two CRISPR arrays containing three and nine spacers (CRISPR02). Four ancillary genes (two csx1 and two csx16) and two genes of unknown function (VVA1542 and VVA1551) complete this CRISPR-island. The black arrows indicate the array promoters identified. (B) Sequence alignment of the leader sequence and first direct repeat of the VvYJ016 CRISPR01 and CRISPR02 arrays. The nucleotide sequence of the first repeat is framed by a gray box. The putative promoter of the two sequences is framed by a blue box. The red letters represent the bases differing between the two sequences. (C) Determination of the level of transcription of the leader sequence of CRISPR01 and CRISPR02. β-Galactosidase activity was measured for the empty plasmid (pMP220) and the two complete arrays, in both orientations with respect to the lacZ gene (sense: pCA1s; pCA2s for CRISPR01 and CRISPR02, respectively; antisense: pCA1as; pCA2as for CRISPR01 and CRISPR02, respectively). Errors bars show the standard deviation for three biological replicates.

Figure 2.

Spacer acquisition by the VvYJ016 adaptation operon in the heterologous E. coli system. (A) Schematic diagram of the high-throughput spacer acquisition assay. Overexpression of the adaptation operon in E. coli HMS 174 (DE3) followed by the extraction of plasmid DNA, two rounds of PCR/purification of the expanded CRISPR array, and deep sequencing, analysis and characterization of the spacers identified. (B) Frequency of new spacer detection per million reads for the wildtype RT-Cas1, RT active site mutant (YAAA), Cas1 domain mutants E517A and E597A and the ΔCas2A, ΔCas2B and ΔCas2A-B mutants. The bars indicate the range for three biological replicates.

Figure 3.

Characterization of the spacers acquired by the VvYJ016 adaptation operon. (A) Coverage of spacers aligning with the E. coli HMS174 (DE3) genome and a representative locus. Identical alignments represent recurrent spacers acquired in independent biological samples (n = 11). (B) Strand bias in pools of newly acquired spacers relative to the source transcript. Proportion of newly acquired spacers with the Wild-type RT-Cas1 in the sense or antisense strand of coding genes or in intergenic regions of the E. coli genome (n = 11). (C) Histogram showing normalized counts of E. coli genome or pAGDt-Op439 plasmid spacers, by length. (D) GC content distribution of genome- and plasmid-aligned spacers. The dotted lines represent the GC content of the plasmid (light gray) and the genome (dark gray). For C and D, the bars indicate the range for the three assays in which the largest numbers of spacers were detected (>10,000 newly acquired spacers per experiment).

Figure 4.

Spacer composition and position relative to coding sequences. (A) GC content (above) and nucleotide probabilities (below) at each position along the wild-type RT-Cas1–acquired protospacers. Given the variation of protospacer length, two panels are shown, with the spacer anchored 5′and 3′ at positions 15 and 35, respectively. Spacer (gray background) and flanking (white background) nucleotides are shown. The dark gray background indicates asymmetry at the two ends of the spacers, with a stretch of positions rich in ‘AT’ (see text). The particular bias towards ‘C’ within the spacer observed is indicated by a red line. The ‘GC’ content is shown for spacers aligning with the genome (blue) and plasmid (yellow). (B) Gene body coverage of spacer alignments along the length of transcripts. The relative position corresponds to the percentile of coding sequence length ± 300 bp of the adjacent genomic regions. Dotted lines in A and B represent the mean error of alignment for the three assays in which the largest numbers of spacers were detected (>10,000 newly acquired spacers per experiment).

Figure 5.

Spacer acquisition from RNA in the VVYJ016 type III-D system. (A) Schematic diagram of td intron-containing constructs. We determined whether the spacers originated from RNA, using a self-splicing transcript that produces an RNA sequence junction not encoded by DNA. Newly acquired spacers containing this exon junction may be considered to have been acquired from an RNA target. (B) RNA derived from the newly acquired exon junction-spanning spacer (blue). The splice site is indicated by a blue triangle. Red arrows indicate that the spacer are in the antisense orientation relative to the direction of transcription of the td intron. At the bottom, the highlighted sequence of one of the splice junction-containing spacers located in the CRISPR array is indicated.