CRISPR-Cas effector specificity and cleavage site determine phage escape outcomes

Abstract

CRISPR-mediated interference relies on complementarity between a guiding CRISPR RNA (crRNA) and target nucleic acids to provide defense against bacteriophage. Phages escape CRISPR-based immunity mainly through mutations in the protospacer adjacent motif (PAM) and seed regions. However, previous specificity studies of Cas effectors, including the class 2 endonuclease Cas12a, have revealed a high degree of tolerance of single mismatches. The effect of this mismatch tolerance has not been extensively studied in the context of phage defense. Here, we tested defense against lambda phage provided by Cas12a-crRNAs containing preexisting mismatches against the genomic targets in phage DNA. We find that most preexisting crRNA mismatches lead to phage escape, regardless of whether the mismatches ablate Cas12a cleavage in vitro. We used high-throughput sequencing to examine the target regions of phage genomes following CRISPR challenge. Mismatches at all locations in the target accelerated emergence of mutant phage, including mismatches that greatly slowed cleavage in vitro. Unexpectedly, our results reveal that a preexisting mismatch in the PAM-distal region results in selection of mutations in the PAM-distal region of the target. In vitro cleavage and phage competition assays show that dual PAM-distal mismatches are significantly more deleterious than combinations of seed and PAM-distal mismatches, resulting in this selection. However, similar experiments with Cas9 did not result in emergence of PAM-distal mismatches, suggesting that cut-site location and subsequent DNA repair may influence the location of escape mutations within target regions. Expression of multiple mismatched crRNAs prevented new mutations from arising in multiple targeted locations, allowing Cas12a mismatch tolerance to provide stronger and longer-term protection. These results demonstrate that Cas effector mismatch tolerance, existing target mismatches, and cleavage site strongly influence phage evolution.

Fig 1. Effects of mismatched crRNAs on Cas12a-mediated phage defense.

(A) Schematic of crRNA expression and processing by FnCas12a and crRNA phage target locations. crRNA mismatches were introduced by mutating individual nucleotides in the spacer sequence. After expression of the pre-crRNA, Cas12a processes it into a guiding crRNA that partially matches the lambda phage genome targets upstream of gene J and in the coding region of gene L. See S1a for target and crRNA spacer sequences. (B) Measurement of phage protection provided by crRNAs with and without target mismatches. Spot assays were performed with bacteria expressing FnCas12a and a crRNA construct that either perfectly matches the lambda phage genome (perfect) or has a crRNA mismatch (MM) at a position in the spacer (position x, sequences shown in S1A Fig). A non-targeting crRNA construct (NT) was used as a negative control. Lambda phage was spotted on cells with 10-fold decreasing concentration at each spot going from left to right. Expression of FnCas12a and pre-crRNAs were controlled by a stronger inducible PBAD promoter or a weaker constitutive promoter. (C) Observed rate constants for in vitro cleavage by Cas12a armed with crRNAs containing target mismatches. Plasmids bearing target sequences for gene J or L were used to measure Cas12a cleavage. Mismatch positions or perfect crRNAs (P) are indicated on the horizontal axis. Data from 3 replicates are plotted. * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, ns = no significant difference compared to the perfect crRNA based on unpaired two-tailed t test. See S1B and S1C Fig and S1 Data for gels and quantification. See S1 Fig for crRNA and target sequences, representative gels, and fit data. (D) Growth curves for E. coli expressing mismatched crRNAs following phage infection. Bacteria containing the P_BAD FnCas12a expression plasmid and various crRNA expression plasmids were inoculated in liquid culture and induced immediately. Lambda phage was added 1.5 h after inoculation and OD₆₀₀ measurements were taken every hour. Bacteria expressed no cRNA, a crRNA with no mismatches to the target (perfect) or a crRNA with a mismatch at the indicated position (position x). A no phage condition was performed as a negative control. The average of 2 replicates is plotted, with error bars representing standard deviation. See S2 Data for quantified data.

Fig 2. crRNA mismatches cause emergence of diverse lambda phage mutations.

(A) Schematic of workflow for determining the genetic diversity of phage exposed to interference by Cas12a. Phage samples were collected from liquid cultures at various time points and the target region was PCR amplified. Mutations were observed using MiSeq high-throughput sequencing of these amplicons. (B) Line graph tracking the fraction of phage with mutated target sequences over time. Samples were taken from liquid cultures at time points after phage infection. The “0 h” samples were taken directly after addition of phage to the culture tubes. The average of 2 replicates are plotted with error bars representing standard deviation. See S3 Data for quantified data. (C) Heat maps showing the location of enriched phage mutations in target regions at the 8 h time point for gene J or L targets. Z-scores for abundance of single-nucleotide variants, including nucleotide identity changes or deletions, were determined for each sample relative to the non-targeted control phage population. Experiments were performed using λ_vir phage (WT) and λ_vir phage with the red operon deleted (Δred). Enriched sequences indicate high Z-scores. Z-scores range from 0 (white) to 10.3 (darkest red). Single-nucleotide deletions are shown at adjacent position to the 3′ side. Positions with crRNA mismatches are labeled with solid black boxes in the heat map. A thin outline indicates that the majority of sequences contain single point mutations at these positions while a thick outline indicates that the majority of sequences contain multiple point mutations at these positions. See S3 Data for quantification of variant abundance.

Fig 3. Two PAM-distal mismatches are more deleterious than seed mismatches.

(A) Spot assays performed using E. coli expressing FnCas12a and a crRNA that perfectly matches the lambda phage gene J target (perfect) or has mismatches at the indicated positions. Three types of second mismatches were added and the type of the mismatch is indicated in parenthesis next to the position number. See S7A Fig for crRNA spacer sequences. (B) Observed rate constants for in vitro cleavage by Cas12a armed with crRNAs containing 2 target mismatches. Cleavage was measured for plasmid DNA containing a gene J target. The types of mismatches for the second mismatch are indicated. * P ≤ 0.05, **** P ≤ 0.0001, ns P > 0.05 compared to the perfectly matching crRNA based on unpaired two-tailed t test. See S7B and S7C Fig and S1 Data for gels, and quantified and fit data. (C) Phage spot assays for target mutant phages isolated upon challenge with Cas12a programmed with a position 15 mismatched crRNA targeting gene L. Spot assays were performed with E. coli expressing a non-targeting crRNA (NT), a crRNA that perfectly matched wild-type phage (Perfect), or the crRNA with a mismatch at position 15 (MM15). Phage mutations were in the seed (A2T) or PAM-distal (G17T) region. (D) Observed rate constants for in vitro Cas12a cleavage of plasmids bearing wild-type (WT), seed mutant (A2T), or PAM-distal mutant (G17T) gene L target sequences. Cas12a cleavage was measured for both the perfectly matched crRNA (P) or the MM15 crRNA (15). Significance was tested pairwise for all crRNA/target combination by unpaired two-tailed t test. * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001, ns P > 0.05. Pairwise comparisons for which P value are not indicated had a P < 0.0001. See S8 Fig and S1 Data for crRNA and target sequences, gels, and quantified and fit data. (E) Schematic of competition assay. Two mutant phages, A2T and G17T, were mixed at approximately equal titers. E. coli expressing Cas12a and the position 15 mismatched crRNA were infected with a dilution series of the mutant phage mix. Lysates were harvested and the proportion of each mutant phage was determined by high-throughput sequencing. (F) Ratio of seed mutant (A2T) to PAM-distal mutant (G17T) following lysis of cultures infected with a dilution series of the mixed phage. E. coli expressed Cas12a and either a non-targeting (NT, red) or position 15 mismatched (MM15, blue) crRNA. “Mix” indicates the initial mixture of phage. The average of 3 replicates is shown, with error bars indicating standard deviation. See S6 Data for raw and quantified data.

Fig 4. Cas9 challenge does not cause emergence of PAM-distal mutants.

(A) Observed rate constants for cleavage of a target plasmid bearing a wild type (WT), seed mutant (G2T) and PAM-distal mutant (A17T) gene L target sequence. Cas9 cleavage was measured for both the perfectly matched crRNA (P) or the MM15 crRNA (15). Significance was tested pairwise for all crRNA/target combination by unpaired two-tailed t test. * P ≤ 0.05, *** P ≤ 0.001, ns P > 0.05. Pairwise comparisons for which P value are not indicated had a P < 0.0001. See S9 Fig and S1 Data for crRNA and target sequences, gels, and quantified and fit data. (B) Growth curves of E. coli expressing Cas9 and sgRNAs bearing either a non-targeting sequence, the perfectly matching spacer sequence (perfect), or a spacer containing mismatch at the indicated position with respect to the PAM. Phage was added when the cells reached mid log phase at approximately 2 h after inoculation. The average of 3 replicates is plotted for each condition, with error bars representing standard deviation. See S2 Data for quantified data. (C) Heat maps showing the location of enriched phage mutations in target regions at the 8 h time point for gene J or L targets after Cas9-mediated selection. Z-scores for abundance of single-nucleotide variants, including nucleotide identity changes or deletions, were determined for each sample relative to the non-targeted control phage population. Enriched sequences indicate high Z-scores. Z-scores range from 0 (white) to 7.0 (darkest red). See S7 Data for quantification of variant abundance.

Fig 5. Combined mismatches are necessary for complete phage escape.

(A) Schematic for experiment to test the impact of MOI on escape phage diversity. Cultures expressing Cas12a and the position 3 mismatched crRNA targeting gene J were infected with lambda phage at varied MOIs. Mutant phages in lysates were detected by high-throughput sequencing. (B) Heat map showing the position of phage mutations that arose when infecting bacteria expressing seed mismatch crRNA at different MOIs. Phage was harvested from liquid cultures containing bacteria expressing FnCas12a and a crRNA with a C-T mismatch at position 3. Phage was added to the culture at mid-log phase at a range of MOIs starting at 0.15 and serial 2-fold dilutions from 1/2 to 1/32 and an additional sample at an MOI of 1.5 × 10⁻³. Phage was harvested 5 h after infection. High-throughput sequencing was used to determine the percent of phages in each that had a mutation in the target region. The heat map shows the positions in the target that were enriched with mutations. These positions are colored darker red according to their Z-score relative to the control phage population. Z-scores range from 0 (white) to 10.1 (darkest red). The position of the initial crRNA mismatch is indicated in solid black. See S8 Data for quantification of variant abundance. (C) Spot assays using phage isolated from liquid cultures as described in (A) on bacteria expressing a matching crRNA. Phage harvested in (A) was 10-fold serial diluted and spotted on bacteria with a crRNA matching the wild-type lambda phage genome target (matching crRNA) or bacteria without a crRNA guiding Cas12a (no crRNA). Wild-type phage controls were spotted on these same bacterial strains. Phages harvested from the lowest MOI cultures were omitted due to their low titer which prevented visible plaque formation on the CRISPR active E. coli strain. See S11B Fig for full plates. (D) Schematic for experiment shown in panel (E). Wild-type or mutant phage populations were used for spot assays on plates with lawns of E. coli expressing Cas12a and either the perfect or the seed mismatched crRNA to determine whether the combination of the preexisting mismatch and newly acquired target mutations are necessary for complete escape from Cas12a targeting. (E) Spot assays using mutationally diverse phage on bacteria expressing crRNAs with and without mismatches. Phage was isolated from the liquid culture as described in (A) that was initially infected with phage diluted 1:8. Mutated phage and unmutated control phage (WT) were then used for spot assays on bacterial lawns expressing FnCas12a and a matching crRNA (perfect), a crRNA with the original seed mismatch, or no crRNA as negative control.

Fig 6. Multiple mismatched crRNAs provide more protection than individual mismatched crRNAs.

(A) Schematic of experiment in which 2 crRNAs bearing mismatches at position 3 are expressed from a CRISPR plasmid. The CRISPRs either have 2 identical spacers targeting gene J (J+J) or gene L (L+L), or have 2 different spacers targeting gene J and L (J+L). (B) Number of plaques formed on lawns of bacteria expressing multiple mismatched crRNAs. Plaque assays were performed using bacteria containing a plasmid expressing FnCas12a along with different crRNA expression plasmids. Plasmid expressed either the perfect crRNA (P) or the position 3 mismatched crRNA (3). For the perfect crRNAs, plasmids expressed either 1 (gene J or gene L alone) or 2 (J+L) spacers. For the position 3 mismatched crRNAs, each crRNA expression plasmid contains 2 spacers: both targeting gene J both targeting gene L and 1 spacer targeting each of gene J and gene L (J+L). (-) indicates a negative control in which no crRNA was expressed. Plaque forming units (pfu) was calculated using the number of plaques on each plate and the volume of phage lysate added. The average of multiple replicates (n = 2 to 4) is plotted with error bars representing standard deviation. Unpaired, two-tailed t tests were used to determine the statistical significance of each of the single spacer constructs compared to the double spacer construct, *** P ≤ 0.001, **** P ≤ 0.0001. See S2 Data for quantified data. (C) Growth curves using the same bacterial strains described in (A). Phage was added when the cells reached mid log phase at approximately 2 h after inoculation. The average of 3 replicates is plotted for each condition, with error bars representing standard deviation. See S2 Data for quantified data. (D) Spot assays used to measure the titer of phage over time in phage infection cultures. Spot assays were performed using 10-fold serial dilutions of phage harvested from cultures in (B) that infected bacterial strains with 2 mismatched spacers at different time points on lawns of CRISPR-inactive E. coli. (E) Sequences of both CRISPR targets in single phage plaques for phage harvested from E. coli cultures expressing a double spacer construct. The 2 crRNAs contained mismatches at positions highlighted in black. Target regions for the gene J and gene L target were sequenced for 6 individual plaques using Sanger sequencing. Target sequences are aligned to the WT sequence (top row) and mutations are underlined. See S12B Fig for chromatograms.

Fig 7. Generation of double-mutant phage is driven by insufficiently deleterious mutations.

(A) Schematic of the process for generating and purifying single-mutant phage populations. Wild-type phage was used to challenge E. coli expressing a crRNA with a mismatch to the target in the phage genome in liquid culture. The resulting phage were isolated and used for a plaque assay on lawns of bacteria expressing the same mismatched crRNA. Single plaques were isolated and again used to challenge bacteria expressing a mismatched crRNA in liquid culture, further purifying and propagating single mutants. Finally, single-mutant phages were used to challenge bacteria expressing a perfectly matching crRNA in liquid culture to determine whether second mutations would appear. (B) Growth curves of bacteria expressing a perfectly matching crRNA challenged with wild-type phage and phage with various single target mutations. Locations of the single mutations in the target are labeled (PAM mutant, seed mutant, and PAM-distal mutant). Position and type of mutations are indicated in parenthesis. Three biological replicates are shown separately for each experimental condition. See S2 Data for quantified data. (C) Diagram of initial and selected mutations that appeared when a single-mutant phage was used to challenge bacteria expressing a perfectly matching crRNA. Initial mutants are single mutants that were generated and purified as described in (A). Sequences below arrows show phage mutants that appeared in different biological replicates (rep 1, 2, or 3) after initial mutant phage lysates were used to infect bacteria expressing a perfectly matching crRNA in liquid culture. Positions with ambiguous base calls are indicated with 2 bases (X/Y) at that position. Target sequences were interpreted from Sanger sequencing chromatograms (see S13 Fig). (D) Spot assays challenging bacteria expressing a perfectly matching crRNA with various single- and double-mutant phage lysates. WT phage or phages with the indicated target mutations were spotted on bacteria expressing a non-targeting crRNA (left column) and a perfectly matching crRNA (right column).