Evolutionary Sweeps of Subviral Parasites and Their Phage Host Bring Unique Parasite Variants and Disappearance of a Phage CRISPR-Cas System

Abstract

Vibrio cholerae is a significant threat to global public health in part due to its propensity for large-scale evolutionary sweeps where lineages emerge and are replaced. These sweeps may originate from the Bay of Bengal, where bacteriophage predation and the evolution of antiphage counterdefenses is a recurring theme. The bacteriophage ICP1 is a key predator of epidemic V. cholerae and is notable for acquiring a CRISPR-Cas system to combat PLE, a defensive subviral parasite encoded by its V. cholerae host. Here, we describe the discovery of four previously unknown PLE variants through a retrospective analysis of >3,000 publicly available sequences as well as one additional variant (PLE10) from recent surveillance of cholera patients in Bangladesh. In recent sampling we also observed a lineage sweep of PLE-negative V. cholerae occurring within the patient population in under a year. This shift coincided with a loss of ICP1's CRISPR-Cas system in favor of a previously prevalent PLE-targeting endonuclease called Odn. Interestingly, PLE10 was resistant to ICP1-encoded Odn, yet it was not found in any recent V. cholerae strains. We also identified isolates from within individual patient samples that revealed both mixed PLE(+)/PLE(-) V. cholerae populations and ICP1 strains possessing CRISPR-Cas or Odn with evidence of in situ recombination. These findings reinforce our understanding of the successive nature of V. cholerae evolution and suggest that ongoing surveillance of V. cholerae, ICP1, and PLE in Bangladesh is important for tracking genetic developments relevant to pandemic cholera that can occur over relatively short timescales. IMPORTANCE With 1 to 4 million estimated cases annually, cholera is a disease of serious global concern in regions where access to safe drinking water is limited by inadequate infrastructure, inequity, or natural disaster. The Global Task Force on Cholera Control (GTFCC.org) considers outbreak surveillance to be a primary pillar in the strategy to reduce mortality from cholera worldwide. Therefore, developing a better understanding of temporal evolutionary changes in the causative agent of cholera, Vibrio cholerae, could help in those efforts. The significance of our research is in tracking the genomic shifts that distinguish V. cholerae outbreaks, with specific attention paid to current and historical trends in the arms race between V. cholerae and a cooccurring viral (bacteriophage) predator. Here, we discover additional diversity of a specific phage defense system in epidemic V. cholerae and document the loss of a phage-encoded CRISPR-Cas system, underscoring the dynamic nature of microbial populations across cholera outbreaks.

Keywords: CRISPR-Cas; Vibrio cholerae; bacteriophages; cholera; evolution.

FIG 1

Surveillance of V. cholerae PLE and ICP1 anti-PLE loci in Bangladesh. (A) Total counts of patient stool samples collected and analyzed by month and year of collection (239 total samples) (Data Set S1). (B) Percentage of stool samples in a given month that screened PCR positive for ICP1-encoded anti-PLE counterdefense loci CRISPR-Cas (n = 36), odn (n = 32), or both (n = 3). (C) Number of whole-genome-sequenced ICP1 isolates (44 total) from stool samples over the surveillance period by the type of anti-PLE counterdefense locus CRISPR-Cas or odn. (D) Number of whole-genome-sequenced V. cholerae isolates (148 total) from stool samples over the surveillance period by PLE variant type. PLE(−) indicates no PLE detected.

FIG 2

Genetic, temporal, and spatial variability of PLEs in >3,000 V. cholerae genomes collected between 1916 and 2019. (A) Timeline of PLE-positive strains colored by PLE variant type from 3,363 V. cholerae whole-genome sequences. The lower histogram provides the total number of strains per year, while the upper histogram represents a stacked percentage of those that are positive for each PLE. Because of the small number of sequenced genomes from isolates prior to 1960, those genomes were collapsed into a single column. Metadata for all strains analyzed can be found in Data Set S2. (B) Overview of the geographical occurrences of PLE(+) isolates by country of isolation. Each individual isolate is indicated with a colored dot corresponding to the PLE type, as in panel A, and grouped with all others for a given country. The relatively large numbers of isolates from Bangladesh are represented in the inset. (C) Global alignment of all 10 PLEs using progressiveMauve. The order was determined by phylogenetic distance between PLEs (Fig. S1). The mauve-colored regions are conserved across all PLE sequences, while the light gray regions are unique to that specific variant. All other regions vary in conservation between two or more PLEs and were assigned random colors from the variable palette. Annotated genes are shown to scale as black arrows, and several genes with known functions are labeled (§, lidI).

FIG 3

PLE10 exhibits excision, replication, and packaging in response to ICP1 infection and evades the anti-PLE nuclease Odn. (A) Model showing the steps of the PLE-ICP1 interaction in a V. cholerae host cell. (1) The integrated PLE (in dark blue) excises and circularizes upon ICP1 infection. (2) PLE replicates and (3) is packaged into modified phage particles. (4) PLE transducing-particles are released through cell lysis, which occurs on an accelerated timeline. (B) Agarose gel of PCR products to detect circularized PLE in uninfected V. cholerae and following infection by ICP1²⁰⁰⁶ lacking CRISPR-Cas (the phage isolate previously used to probe PLE circularization [19]). The lane on the far left is the ladder, and that on the far right is the no template control (ntc). (C) Quantification of change in PLE1 (blue) and PLE10 (purple) copy number 30 min after infection with ICP1²⁰⁰⁶ lacking CRISPR-Cas or the ΔhelA mutant (the phage isolate previously used to probe PLE replication [18, 20, 29]). IPTG-inducible plasmid constructs (P_tac) were induced prior to phage infection. EV is the empty vector control. The dashed line indicates no change in copy number compared to the sample taken prior to phage infection (see Materials and Methods). (D) Lysis curves of V. cholerae harboring PLE1 or PLE10 following infection by ICP1²⁰⁰⁶ lacking CRISPR-Cas (the phage isolate previously used to probe lysis kinetics [18, 22]) versus the PLE(−) control. (E) PLE-transducing particles generated during infection with ICP1²⁰⁰⁶ lacking CRISPR-Cas (the phage isolate previously used to probe PLE transduction [18]). (F) Tenfold dilutions of the phage isolate carrying the anti-PLE locus indicated or mutant derivative spotted on V. cholerae with the PLE indicated (bacterial lawns are in gray, zones of killing are in black). Contemporary ICP1 isolates were used in this assay: the CRISPR(+) isolate with spacers against PLE1 and PLE10, ICP1²⁰¹⁷, was recovered from the same patient sample as the original PLE10(+) V. cholerae, and an Odn(+) isolate, ICP1²⁰¹⁹, were used. See Table S1 for a complete description of phage isolates. For experiments in panels B to F, PLEs were transduced into the same genetic background (V. cholerae E7946); see Table S1 for strain details.

FIG 4

PLE(+)/PLE(−) V. cholerae and ICP1 anti-PLE loci heterogeneity within patients. (A) The fraction of total colonies with and without PLE for V. cholerae isolates within individual stool samples as determined by PCR. The number of colonies screened from each patient sample (given an arbitrary number by date of isolation with the prefix D are from icddr,b in Dhaka) are given above each bar. (B) The fraction of ICP1 plaques with the anti-PLE locus indicated recovered from one stool sample from Mathbaria (M76) as determined by PCR. (C) Alignment of six ICP1 isolates from stool sample M76 (panel B) using progressiveMauve. The mauve-colored regions are conserved across all ICP1 sequences, while the light gray regions are unique to that specific isolate. All other regions vary in conservation between two or more isolates and were assigned random colors from the variable palette. Annotated genes are shown as black arrows, and Cas and odn genes are shown in gray. The gray shading highlights regions with odn with associated genes (in phages 160 and 167) and CRISPR-Cas with associated genes (in phages 159, 170, and 166) at ≥95% sequence identity. Isolate 164 demonstrates recombination between the CRISPR-Cas region and the odn-associated region. See Table S1 for a complete description of phage isolates.

FIG 5

Phylogeny of V. cholerae isolates during the surveillance period in Bangladesh reveals the replacement of a previously prevalent PLE(+) lineage with a PLE(−) lineage in 2018 to 2019. A cladogram of 148 V. cholerae isolates based on 532 variable nucleotide genomic positions (Data Set S4). Columns next to the tree contain metadata for (left to right) the PLE variant present, the SXT ICE present, city of isolation, year isolated, and serotype (based on whole-genome sequencing and analysis of wbeT). Column coloring and order is shown in the legend from top to bottom. Strains lacking MGEs (PLE or SXT ICE) are indicated in light gray.