CRISPRTarget: bioinformatic prediction and analysis of crRNA targets

Abstract

The bacterial and archaeal CRISPR/Cas adaptive immune system targets specific protospacer nucleotide sequences in invading organisms. This requires base pairing between processed CRISPR RNA and the target protospacer. For type I and II CRISPR/Cas systems, protospacer adjacent motifs (PAM) are essential for target recognition, and for type III, mismatches in the flanking sequences are important in the antiviral response. In this study, we examine the properties of each class of CRISPR. We use this information to provide a tool (CRISPRTarget) that predicts the most likely targets of CRISPR RNAs (http://bioanalysis.otago.ac.nz/CRISPRTarget). This can be used to discover targets in newly sequenced genomic or metagenomic data. To test its utility, we discover features and targets of well-characterized Streptococcus thermophilus and Sulfolobus solfataricus type II and III CRISPR/Cas systems. Finally, in Pectobacterium species, we identify new CRISPR targets and propose a model of temperate phage exposure and subsequent inhibition by the type I CRISPR/Cas systems.

Keywords: CRISPR; Cas; R-loop; bioinformatics; crRNA; horizontal gene transfer; phage resistance; small RNA targets.

Figure 1. Example annotated CRISPRTarget outputs of representatives of type I, II and III CRISPR/Cas systems. The protospacer is the DNA target complementary to the crRNA spacer. The crRNA is displayed as RNA 5′ to 3′ and the base paired protospacer is 3′ to 5′. (A) The predicted spacer 6 crRNA from the type I-F CRISPR1 (CRISPR1_6) in P. aeruginosa PA14 targets Pseudomonas phage JBD67. The output visualizes the 5′-protospacer-GG-3′ PAM and the crRNA with 8 and 20 nt 5′ and 3′ handles, respectively. (B) The CRISPR1_15 from the type II system from Streptococcus thermophilus DGCC7710 WTphi858phi2972+S13S14 matched to Streptococcus phage 5093. The output shows the predicted length of the 3′ handle, based on Streptococcus pyogenes, and the 5′-WTTCTNN-protospacer-3′ PAM. (C) Spacer 1 from the type III-A system from Staphylococcus epidermidis RP62a targeting plasmid pGO1. The output was adjusted to display the 8 nt 5′ handle with an entire mature crRNA length of 43 nt and no PAMs were scored. Yellow sequences include spacer and protospacer, blue indicates flanking sequences and PAMs are shown in green.

Figure 2. Flowchart of the steps in CRISPRTarget (details are in the Materials and Methods). Input is predictions of the CRISPR arrays, selected databases and initial parameters. This input is processed and the spacers screened using BLASTn for matches against the databases. The flanks of these matches are extended and PAMs and handles analyzed in an interactive manner. Output is as a text/spreadsheet format, or as a graphical display (HTML).

Figure 3. CRISPRTarget input. Several formats are accepted. The BLASTn parameters for the initial screen are defined at this step. They default to values that favor a gapless match, but some mismatches. The output may be refined and reordered (Fig. 4) after it is obtained.

Figure 4. Graphical output of CRISPRTarget. The output of a search for the targets of the Streptomyces thermophilus DGCC7710 CRISPR array. The direction of transcription is known; however, both strands are shown in diagram, as if the direction of transcription was unknown. Two relatively low-scoring matches using these interactive settings are shown (rank 44–45). They have good spacer-protospacer base pairing but lack a WTTCTNN PAM. Match 45 is to a phage to which this strain is sensitive (Φ2972). Yellow indicates spacer/protospacer, blue shows flanking sequences and mismatches between the crRNA and the target DNA protospacer are indicated in red.

Figure 5.Pectobacterium prophages are targeted by CRISPR/Cas. (A) Prophage ϕPCC21_1 is targeted by spacers in P. atrosepticum. (B) P. atrosepticum SCRI1043 (top, 2761697–2811697) compared with ϕPCC21_1 in P. carotovorum subsp carotovorum PCC21 (bottom, phage coordinates: PCC21_018470–019020 from 2092807–2135244. PCC21 is reversed for clarity). (C) Prophage ϕECA29 is targeted by spacers in P. carotovorum subsp carotovorum PCC21. (D) P. carotovorum subsp carotovorum PCC21 (top, PCC21_017190–017500 from 1936500–1976500. PCC21 is reversed) compared with ϕECA29 (HAI9) in P. atrosepticum SCRI1043 (bottom, ECA2598-ECA2637 from 2935264–2966783). (E) Prophage ϕPC1_1 is targeted by a spacer in P. carotovorum subsp carotovorum PCC21. (F) P. carotovorum subsp carotovorum PCC21 (top, PCC21_027150–027460 from 3058299–3095299) compared with ϕPC1_1 in P. carotovorum subsp carotovorum PC1 (bottom, PC1_2622–2666 from 2989228–3022511). (G) Prophage ϕPCC21_1 is targeted by spacers in P. wasabiae. (H) P. wasabiae WPP163 (top, 2291600–2341600) compared with ϕPCC21_1 in P. carotovorum subsp carotovorum PCC21 (bottom, phage coordinates: PCC21_018470–019020 from 2092807–2135244). (I) Prophage ϕPC1_2 is targeted by spacers in P. wasabiae. (J) P. wasabiae WPP163 (top, 1192372–1236372) compared with ϕPC1_2 in P. carotovorum subsp carotovorum PC1 (bottom, phage coordinates: PC1_3152–3199 from 3573374–3608557. PC1 is reversed). Prophages (K) ϕECA29 and (L) ϕPC1_2 are targeted by P. wasabiae spacers. Genome comparisons were generated using Easyfig; genes are cyan arrows, putative prophage regions are purple and spacer target locations indicated with asterisks. Homologous regions by BLASTn are shown in shades of gray.