Targeted IS-element sequencing uncovers transposition dynamics during selective pressure in enterococci

Insertion sequences (IS) are simple transposons implicated in the genome evolution of diverse pathogenic bacterial species. Enterococci have emerged as important human intestinal pathogens with newly adapted virulence potential and antibiotic resistance. These genetic features arose in tandem with large-scale genome evolution mediated by mobile elements. Pathoadaptation in enterococci is thought to be mediated in part by the IS element IS256 through gene inactivation and recombination events. However, the regulation of IS256 and the mechanisms controlling its activation are not well understood. Here, we adapt an IS256-specfic deep sequencing method to describe how chronic lytic phage infection drives widespread diversification of IS256 in E. faecalis and how antibiotic exposure is associated with IS256 diversification in E. faecium during a clinical human infection. We show through comparative genomics that IS256 is primarily found in hospital-adapted enterococcal isolates. Analyses of IS256 transposase gene levels reveal that IS256 mobility is regulated at the transcriptional level by multiple mechanisms in E. faecalis, indicating tight control of IS256 activation in the absence of selective pressure. Our findings reveal that stressors such as phages and antibiotic exposure drives rapid genome-scale transposition in the enterococci. IS256 diversification can therefore explain how selective pressures mediate evolution of the enterococcal genome, ultimately leading to the emergence of dominant nosocomial lineages that threaten human health.

Insertion sequences (IS) are simple transposons implicated in the genome evolution of diverse pathogenic bacterial species. Enterococci have emerged as important human intestinal pathogens with newly adapted virulence potential and antibiotic resistance. These genetic features arose in tandem with large-scale genome evolution mediated by mobile elements. Pathoadaptation in enterococci is thought to be mediated in part by the IS element IS256 through gene inactivation and recombination events. However, the regulation of IS256 and the mechanisms controlling its activation are not well understood. Here, we adapt an IS256-specfic deep sequencing method to describe how chronic lytic phage infection drives widespread diversification of IS256 in E. faecalis and how antibiotic exposure is associated with IS256 diversification in E. faecium during a clinical human infection. We show through comparative genomics that IS256 is primarily found in hospital-adapted enterococcal isolates. Analyses of IS256 transposase gene levels reveal that IS256 mobility is regulated at the transcriptional level by multiple mechanisms in E. faecalis, indicating tight control of IS256 activation in the absence of selective pressure. Our findings reveal that stressors such as phages and antibiotic exposure drives rapid genome-scale transposition in the enterococci. IS256 diversification can therefore explain how selective pressures mediate evolution of the enterococcal genome, ultimately leading to the emergence of dominant nosocomial lineages that threaten human health.

Author summary
Insertion sequence (IS) elements are simple transposons that are ubiquitous in bacteria. In enterococci, which includes medically relevant species such as Enterococcus faecalis and Enterococcus faecium, the IS element IS256 is widespread and has been implicated in pathogenesis and antibiotic resistance. Despite the importance of IS256 to the biology of enterococci, we know little about how this element is regulated and diversifies

Introduction
Enterococci, including the human commensals Enterococcus faecalis and Enterococcus faecium, are opportunistic pathogens of public health concern due to their acquisition of antibiotic resistance and virulence traits [1]. Enterococcal infections are a leading cause of infective endocarditis [2,3] and 30% of nosocomial infections are vancomycin resistant [4,5]. Antibiotic resistance in enterococci is often encoded on mobile genetic elements, including plasmids and composite transposons [6]. Plasmids, especially those containing Inc18, Rep_3, and RepA-N replicon-containing plasmids, frequently encode resistance genes to diverse antibiotics, including chloramphenicol, erythromycin, and gentamicin [7,8]. Composite transposons are typically organized with a cargo gene (such as an antibiotic resistance gene) flanked by insertion sequence (IS) elements. Examples in the enterococci include Tn4001 or Tn4031, which carries gentamicin resistance genes [9] and Tn1546 and Tn1547, which carry the vancomycin resistance vanA and vanB operons, respectively [10,11].
To combat clinically relevant enterococcal strains, new therapeutics are urgently needed. Recently, the concept of using bacterial viruses (bacteriophages or phages) to treat multi-drug resistant (MDR) bacteria is being revisited [12]. Phage therapy case studies show that these viruses can be efficacious against refractory MDR bacteria in human patients [13][14][15][16][17]. However, bacterial resistance to phage infection develops quickly [18][19][20][21] and the lessons learned from the past century of antibiotic use suggest that bacteria will gain resistance to phage infection after widespread phage treatment [22]. Acquired phage resistance, although beneficial to bacteria in the face of phage pressure, can come with fitness costs that dampen antibiotic resistance and virulence [12,[23][24][25]. Therefore, it is imperative to fully understand the mechanisms that promote phage resistance in bacteria, and the phenotypic outcomes of phage resistance. With this knowledge it will be possible to use phage resistance as a tool that can be leveraged against bacteria to enhance current antibacterial therapeutics.
Enterococci have remarkably diverse genomes containing numerous mobile elements that contribute to their adaptation and evolution [26][27][28]. These include plasmids, prophages, and transposable IS elements. Pathogenic enterococci frequently contain multiple IS elements, and IS elements appear to be a feature of newly-adapted nosocomial strains [29]. In the widely studied nosocomial type-strain E. faecalis V583, there are 38 IS elements consisting of 11 different types [26]. IS256 is the most abundant IS element in the E. faecalis V583 genome, with six chromosomal copies and four plasmid copies. IS256 is also common in other Gram-positive bacteria, including staphylococci, where it has been widely studied [30][31][32]. In the staphylococci, insertion of IS256 into different genes causes clinically-relevant phenotypes, such as small colony variation [33], biofilm formation [31], and antibiotic resistance [34]. The transcription factor σ-B negatively regulates IS256 transposition in the staphylococci through the production of a 3' antisense RNA [31,35]. IS256 is a crucial component for enterococcal genome adaptation. In E. faecalis V583, the virulence factor cytolysin is attenuated by IS256 and related IS905 insertions [36]. Additionally, copies of IS256 in the E. faecalis V583 pheromone-responsive conjugative plasmids pTEF1 and pTEF2 recombine with chromosomal IS256 copies to mobilize broad regions of the E. faecalis genome [37]. A related IS element, IS16, has been shown to be highly abundant in hospital-adapted E. faecium isolates, suggesting that IS16 may aid in the success of E. faecium as a nosocomial pathogen [38]. Considering IS elements are likely involved in the pathoadaptation of the enterococci, little is known about how these elements are regulated and what events lead to their activation.
In this work, we investigated both the regulation and dynamics of IS256 transposition in E. faecalis and E. faecium. We found that IS256 is present in multiple, genetically disparate isolates of both species, and is enriched in hospital adapted lineages. At steady state, IS256 produces a wealth of low abundance insertions throughout the chromosome and is regulated at the transcriptional level to tightly control activation. We discovered that phage infection and antibiotic use in different biological settings are associated with increased diversity of IS256 insertions. In the case of phage infection, cell populations with diversified IS256 insertions maintained phage genomes for an extended period of time and a subpopulation of these cells chronically shed phage particles throughout growth. Phage shedding provided a competitive advantage during co-culture with phage-susceptible enterococci, but this advantage was lost when co-cultured with phage-resistant enterococci. This suggests that phage genome carriage is a strong selective pressure that drives IS256 diversification and can be used to occupy an environmental niche when competing with phage-susceptible bacterial strains. Lastly, we investigated how IS256 diversifies enterococcal genomes in a chronically infected human, and found that IS256 insertion abundances in vivo coincided with specific antibiotic use. Together, this work sheds light on how IS elements are regulated and diversify enterococcal genomes during phage predation and clinical infection, and provides evidence for phage carriage as an important selective pressure that promotes IS256 diversification. Furthermore, this work suggests that therapeutic use of both phages and antibiotics could cause rapid and widespread enterococcal genome evolution for which the physiological consequences are unknown.

IS256 is common within enterococcal lineages associated with hospital adaptation and pathogenesis
IS256 is an important factor driving the pathoadaptation of E. faecalis [36], yet it is unknown how widely IS256 is distributed among E. faecalis strains. To answer this, we searched all available E. faecalis genomes from NCBI RefSeq (2065 genomes) for IS256 transposases with 100% amino acid similarity to the IS256 transposase copies found in E. faecalis V583. A total of 232 genomes contained one or more IS256 sequences (Fig 1A). The sporadic distribution of IS256 within the E. faecalis phylogenetic tree suggests that IS256 is a recent addition to many E. faecalis genomes and has arisen within multiple discrete lineages of this species. To further characterize the distribution of IS256 across E. faecalis strains, we performed in silico multilocus sequence typing (MLST) on all E. faecalis genomes and compared the proportions of genomes with and without IS256 in each sequence type (ST) (Fig 1B). IS256 occurs within a variety of ST clades and is most abundant in ST6, ST103, ST778, and ST388 genomes, all of which are associated with hospital-adapted, opportunistic pathogenic lineages (S1A Table)

PLOS PATHOGENS
IS256 diversification in the enterococci IS256 ( Fig 1C and S1C Table). IS256-containing genomes are enriched for multiple virulence factors compared to genomes lacking IS256, demonstrating that IS256 preferentially occurs in hospital-acquired and virulent E. faecalis strains. We next expanded this analysis to include all E. faecium genomes from RefSeq (2306 genomes) and similar to E. faecalis, E. faecium IS256 is primarily found in the nosocomial lineages ST17, ST664, ST736, and ST18 (Fig 1D and 1E and S1B Table) [43,44]. Although IS256 is significantly enriched in virulent E. faecium genomes, this was less than E. faecalis (Fig 1F and S1D Table). Finally, IS256 is more frequently found in E. faecium genomes compared to E. faecalis genomes, confirming prior research (Fig 1G) [45]. Together these data show that IS256 is widely distributed within enterococci and is preferentially found in nosocomial and virulent isolates.

IS-Seq identifies widespread movement of IS256 in E. faecalis that is transcriptionally and translationally controlled
IS256 sequences were identified as hot-spots in E. faecalis V583 that facilitated the integration of the endogenous plasmids pTEF1 and pTEF2, leading to the mobilization of the pathogenicity island and other chromosomal regions [37]. Instances of IS256 insertions leading to the inactivation of E. faecalis genes involved in diverse phenotypes have been described [19,[46][47][48]. However, we lack a complete understanding of IS256 mobility across the E. faecalis genome during selective pressure that would potentially drive IS256 activation and genome diversification. To assess genome-wide IS256 insertions in E. faecalis we adapted a next-generation sequencing (NGS) enrichment technique for use with IS256. This technique, IS-Seq, uses IS256 amplicon enrichment during NGS library construction followed by read mapping of IS256-chromosomal junctions to identify specific IS256 insertion locations (S1 Fig) [ . IS-Seq of wild-type (WT) E. faecalis V583 under steady state conditions identified the known IS256 locations in the chromosome and in the three pTEF plasmids (Figs 2A and S2). One of the IS256 copies on pTEF1 lacks a canonical left inverted repeat (IR), preventing binning of the IS-Seq reads originating from this locus. Numerous low abundance IS256 insertions are observed throughout the E. faecalis genome, indicating promiscuous movement of the element within subpopulations of cells (Fig 2A). The three biological replicate cultures tested in Fig 2A show that each E. faecalis population analyzed has a unique repertoire of IS256 insertions. We hypothesize that these low abundance steady-state insertion events provide a mechanism for rapid genome diversification following exposure to selective pressure.
To understand how IS256 is regulated in E. faecalis under steady state conditions, we investigated possible mechanisms of repression of the IS element. First, we assessed the transcription of the IS256 transposase (tnp) gene. During growth in rich media, IS256 gene transcription modestly increases during exponential growth, and reaches 1/10 th of the levels of the reference 16S rRNA gene (Fig 2B). Considering there are 10 copies of IS256 and 4 copies of 16S in E. faecalis V583, each IS256 copy achieves 1/25 th the transcriptional activity of a single 16S gene if all copies are expressed similarly. Thus, the relatively weak IS256 promoter may limit over-activation of the element. Next, we investigated whether IS256 is further transcriptionally or translationally regulated. Other IS elements have been reported to be controlled by antisense small RNAs (asRNA) [52]. These asRNAs repress transposase translation through different mechanisms, including binding to sense transcripts at or before the start codon and ribosome binding site to prevent translation of the sense transcript or targeting the tnp transcript for RNase III cleavage [53]. We reanalyzed a publicly available stranded RNA-Seq dataset of E. faecalis V583 [54] by aligning reads to an IS256 element with delineation of the sense and antisense alignments ( Fig 2C). This dataset was built from short sequencing read fragments originating from the start of mRNAs. The transcriptional start site (TSS) of IS256 has not been experimentally confirmed in any bacterial species and is predicted to reside in the 5' terminus of the element [55]. We demonstrate that IS256 has two TSSs indicated by high sense read mapping coverage in distinct peaks. Additionally, each sense peak is overlayed by a higher antisense peak, demonstrating that antisense inhibition is present at both TSSs. If translation initiates at the site 2 TSS this would produce a truncated IS256 Tnp, here after referred to as IS256 Tnp site 2. Upon further investigation of these TSSs and their asRNAs, we found that the site 1 asRNA is likely under the control of a canonical σ-70 promoter. To identify if this asRNA controls IS256 Tnp translation, we constructed an IS256 Tnp-GFP translational reporter. This construct consists of an IS256 Tnp coding sequence fused in frame at its C- .01, determined with t-test from three biological replicates). E) RT-qPCR of IS256 tnp RNA levels of WT IS256-GFP and IS256-GFP ΔasRNA (* p<0.05, determined with t-test from three biological replicates). F) qPCR of IS256 circular intermediates in E. faecalis OG1RF carrying either an empty vector, wild type (WT) IS256, catalytically inactive IS256 (ΔDDE), IS256 ΔasRNA, or IS256 asRNA site 2 only (Δsite 1) (**** p<0.0001, determined with one-way ANOVA. Comparisons were made to WT IS256 samples and corrected for multiple comparisons using the Dunnett method). https://doi.org/10.1371/journal.ppat.1011424.g002

PLOS PATHOGENS
IS256 diversification in the enterococci terminus to a green fluorescence protein (GFP) gene. GFP fluorescence serves as a proxy for translation of the IS256 Tnp. In addition, we built a version of the IS256 Tnp-GFP fusion lacking the predicted -10 element of the site 1 asRNA. We transformed these constructs into E. faecalis OG1RF, a strain that lacks native IS256 elements and IS256-derived asRNAs in its genome, and measured GFP fluorescence in these cells (Fig 2D). In cells lacking the site 1 asRNA, there is significantly higher GFP fluorescence compared to cells with an intact site 1 asRNA promoter, demonstrating that these asRNAs lower the abundance of the IS256 Tnp. This phenotype was fully complemented. We also measured the RNA levels of the IS256 tnp gene from the same constructs and found that there were significantly more tnp transcripts in the IS256-GFP ΔasRNA cells (Fig 2E). These results show that in E. faecalis IS256 is regulated by multiple mechanisms at the mRNA level.
To extend our findings, we compared the abundance of IS256 circular intermediates between different IS256 constructs in E. faecalis OG1RF (Fig 2F). IS256 elements circularize during transposition, which can be used to measure active transposition [30,31]. Using a multi-copy plasmid, we compared the rate of IS256 circularization in E. faecalis OG1RF carrying the following constructs: WT IS256, an inactive IS256 gene lacking the coding sequences for the two catalytic aspartic acid residues of the Tnp's DDE motif (ΔDDE), an IS256 lacking the promoter for the site 1 asRNA (IS256 ΔasRNA), and an IS256 lacking the 5' UTR and coding sequences upstream of the site 2 promoter (IS256 Δsite 1). The IS256 site 2 construct retains the 5' IR. Diagrams of these constructs are shown in S3 Fig. Surprisingly, we found that all constructs produced significantly less IS256 circles compared to WT IS256. This suggests that the site 2 Tnp product and IS256 ΔasRNA are inactive for transposition, even though loss of the asRNA increases IS256 Tnp transcription. To confirm these findings, we complemented the asRNA and found that this restored the phenotype and increased IS256 circular intermediates above WT IS256 levels. Additionally, we confirmed that IS256 circles require an active IS256 element as there were significantly less IS256 circles produced by the inactive IS256 element.

IS256 selection during phage predation
Our previous work suggested that phage predation increases IS256 transposition in E. faecalis V583 and is used as a mechanism of phage resistance [18]. To determine if phage selective pressure results in the diversification of IS256 elements in E. faecalis, we challenged E. faecalis V583 with the phage phi19 [18]. Following phi19 exposure, we isolated single colonies that were resistant to phi19. Whole genome sequencing (WGS) of these phage-resistant isolates predicted new IS256 insertions in phi19 resistant isolates (referred to as 19RS strains) (Fig 3A). Among the new IS256 insertions in 19RS strains, we found an insertion in the 3' end of epaX, a glycosyltransferase gene involved in the biosynthesis of the enterococcal polysaccharide antigen that is required for infection by phi19 [18]. To verify this insertion and its orientation, we performed PCR using genomic DNA (gDNA) from four 19RS strains and WT E. faecalis V583 (Fig 3B). We found that all 19RS strains harbored an IS256 insertion in epaX in both the forward and reverse complement orientations. This IS256 insertion likely prevents functional EpaX from being expressed blocking phage infection. Recently, Lossouarn et al. identified IS256 insertions inactivating epaX, confirming that an IS256 insertion in epaX is sufficient to prevent phage infection [48]. These results demonstrate that 19RS strains contain a novel IS256 insertion not found in WT E. faecalis V583. These results also suggest that 19RS strains, even though they were isolated as individual colonies, are heterogeneous populations with varied abundances of IS256 insertions.
To test if IS256 provides an advantage to E. faecalis against phage infection, we used phi19 to infect E. faecalis OG1RF carrying either an empty vector (EV), WT IS256, or the inactive version of IS256 (ΔDDE) (Fig 3C). We found that E. faecalis cells containing a functional IS256 gene were more resistant to phage infection compared to cells containing the catalytically dead Tnp or the empty vector, indicating that IS256 carriage increases mutation frequency. To understand if IS256 activation was initiated by phi19 infection, we quantified the number of IS256 circular intermediates. During the course of phi19 infection, IS256 circular intermediates significantly increased relative to a mock-infected cells ( Fig 3D).
WGS analysis indicates that phage predation increases IS256 across the E. faecalis genome. However, this method is low resolution. To quantify the exact locations and abundances of IS256 insertions following phage pressure, we performed IS-Seq on 19RS strains and compared these to WT E. faecalis V583 (Fig 4A). We expected that each strain would have substantial IS256 insertional variability, given that 19RS strains contained an IS256 insertion in multiple orientations ( Fig 3B). We performed IS-Seq on gDNA collected from three biological replicates for both WT E. faecalis V583 and 19RS strains. The three replicates were used to calculate the average peak intensity at each insertion site. IS-Seq revealed a genome-wide  Table). Pairwise comparisons determined that 19RS strains are significantly enriched in the number of novel IS256 insertions compared WT E. faecalis V583. These comparisons were determined using the DESeq2 [56], which accounts for over-dispersion and provides a rigorous multiple testing statistic. We used raw sequencing read counts of each insertion for this analysis ( Fig  4B). The results of this analysis are detailed in S2 and S3 Tables. New IS256 insertions in the pTEF plasmids were also found in the 19RS strains (S4A- S4C Fig). Additionally, we measured IS256 circular intermediates in our IS-Seq dataset by aligning IS256 reads (which originate

PLOS PATHOGENS
IS256 diversification in the enterococci from the left terminus of IS256) to the right terminus of an IS256 element. We found that 19RS strains have a significantly greater number of IS256 circular intermediates compared to WT E. faecalis V583 (Fig 4C).
We focused on two genomic regions that experienced a high level of IS256 insertions in the 19RS strains. The first was the epa locus. In addition to numerous novel insertions in epaX, we found insertions within other epa genes including epaR, epaS, epaAA and epaAB. (S5A Fig). IS256 insertions in these epa genes would alter the teichoic acid decoration of the core rhamnose-containing Epa, likely altering the surface chemistry of E. faecalis resulting in phage resistance [18]. We also identified multiple insertions in the vex/vnc operon (S5B Fig), which is suggested to have both virulence and antibiotic resistance functions [57,58].
E. faecalis is a native member of the intestinal microbiota and can outgrow and cause disease during intestinal dysbiosis. Phages have been proposed as a treatment option for difficult to control enterococcal intestinal blooms [59]. To determine if phage predation supported IS256 diversification in the intestine, we colonized mice with E. faecalis V583 and orally challenged these animals with phi19 (S6 Fig). We performed IS-Seq on fecal gDNA from these animals and found that phi19 exposure in the intestine also increased new IS256 insertions, albeit to a lesser extent than in vitro (Figs 4D-4F and S4D-S4F, and S3 Table). We found that similar to cells exposed to phi19 in vitro, IS256 insertions within the same location in epaX and the vex/vnc operons are observed in the mouse intestine following phage administration (S5C and S5D Fig). This suggests that phage selection plays a role in IS256 diversification in the context of the intestine.

Enterococci with expanded IS256 genomes chronically shed phi19 phages
Phage genomes can be maintained in their host bacterium in a variety of configurations, including chromosomally integrated or as cytosolic episomes [60]. In phage-bacteria communities, sensitive hosts, resistant hosts, and hosts containing phage genomes coexist at a consistent ratio [60]. Sensitive hosts are infected with phage particles and perish increasing the number of phage particles, become resistant, or become lysogenized. Over time, a subset of lysogenized and/or resistant bacteria are desensitized to the phage infection and the cycle repeats. This phenomenon is termed carrier state [60]. Phage genomes can also be in an intermediate state between lytic replication and lysogenic conversion referred to as pseudolysogeny [60]. Such phages enter the lytic cycle spontaneously or following cell stress, and reports indicate that lysogenic phages can be continuously shed from cells with little to no reduction in viability of the overall bacterial population [61,62]. During the analysis of bacterial colonies originating from E. faecalis 19RS strains, we discovered that the phi19 genome is maintained in these phage-resistant bacteria and the bacterial population can grow at a stable rate. This carriage of the phi19 genome may explain its ability to influence IS256 transposition. We found that 19RS strains continually release infectious phage particles during different stages of growth ( Fig 5A). These particles were confirmed to be phi19 using PCR (Fig 5B). We discovered that~10% of 19RS colonies produced infectious phages (Fig 5C). This suggests that a minority of each population maintains phage resistance yet produces infectious phage progeny. Loss of phage resistance in this population would lead to phage infection and population collapse. Colonies that produced zones of clearing were isolated and serially passaged. Cells that initially produced zones of clearing continued to produce zones for at least three passages (S7 Fig), demonstrating that phi19 is stably carried in this subpopulation. Next, we reanalyzed reads from our WGS experiments and found that we could recover the phi19 genome with high coverage from 19RS strains (Fig 5D). WT E. faecalis V583 did not have any detectable phi19 genomic DNA. To determine if the phage genome was integrated into the E. faecalis chromosome or one of its three endogenous plasmids, we performed long-read Oxford Nanopore sequencing on the 19RS strains. This technique provides reads up to 20kb and can be used to find large genomic insertions and rearrangements. We first searched for reads with phage genomic DNA content and then identified if any of these reads contained host genomic DNA. Every 19RS strain had reads containing phage and host hybrid genomic reads. This ranged from 0.2-10% of total phi19 reads ( Fig 5E). While this may be interpreted as evidence of phi19 integration, Nanopore reads have a low rate of chimeric read formation. Chimeric reads are reads containing two sequences combined as a result of a technical artifact during library preparation. Chimeric read formation in Nanopore libraries is reported as 1.7% [63] of total reads, which is in agreement with the proportion of phi19 reads containing both phage and host DNA. Thus, we believe that the phage genome is not integrated into the host genome in 19RS cells but is otherwise maintained as an independent genome that could be considered temperate. Collectively, these results show that 19RS strains release phages from a subpopulation of cells, which may impose stress on E. faecalis influencing IS256 mobility.

IS256 diversification in the enterococci
Considering that 19RS strains have IS256-rich genomes and release infectious phages, we hypothesized that phage carriage may increase the competitiveness of 19RS strains when cocultured with phage sensitive strains. We performed competition assays where WT E. faecalis or 19RS strains were competed against E. faecalis OG1RF or an OG1RF ΔepaOX mutant strain, without the addition of exogenous phages to the cultures (Fig 6A and 6B). The E. faecalis OG1RF epaOX gene is homologous to E. faecalis V583 epaX [64] and E. faecalis OG1RF ΔepaOX is resistant to phi19 infection. 19RS strains that were shedding phi19 outcompeted E. faecalis OG1RF. However, they were unable to outcompete E. faecalis OG1RF ΔepaOX. These results demonstrate that phi19 carriage provides a selective advantage to E. faecalis cells when competing with related bacteria that are phage susceptible but if competitors are phage resistant, phi19 carriage comes with a significant fitness cost. To better understand the fitness cost of phage carriage, we compared the growth rate of E. faecalis 19RS strains, V583, and OG1RF in monoculture. E. faecalis 19RS strains have a growth defect compared to V583 and OG1RF when measuring optical density ( Fig 6C) and doubling time (Fig 6D).

IS256 mobilization occurs in enterococcal genomes during human infection
IS elements contribute to genome diversification and adaptive plasticity that is important for bacterial evolution [65]. Environmental pressure is a strong driver of IS element mobility [66][67][68], and the work described here expands this knowledge to phage predation. Antibiotics are a strong selective pressure that can promote bacterial evolution, and IS256 insertions have been tied to pathoadaptation in enterococci during human infection [36]. Furthermore, antibiotics can increase IS256 transposition [69]. To assess if environmental cues such as antibiotics may also be driving IS element diversification, we investigated whether enterococcal populations within a single individual infected with E. faecium experiences IS256 mobility during therapeutic intervention. Stools samples were collected at various timepoints over 109 days, during which the individual was treated with regimens containing daptomycin, vancomycin, and the vancomycin analog oritavancin for recurrent E. faecium bacteremia. IS-Seq was performed on these samples across this time series (Fig 7A). Nearly all of the insertions found using IS-Seq occurred in contigs which were generated from E. faecium (S8 Fig). We found that these E. faecium populations had a larger proportion of IS256 insertions following oritavancin therapy compared to daptomycin and/or vancomycin therapy. Comparing differentially abundant insertions during oritavancin treatment revealed multiple enriched insertions in the coding sequences of a putative Rib transcriptional regulator, and the vex/vncRS operon also identified as a IS256 insertion hotspot in E. faecalis 19RS strains (Fig 7B and 7C). These two operons are associated with antibiotic resistance and virulence in E. faecium and IS256 insertions likely render these genes nonfunctional. These insertions increase in sequencing depth during the course of treatment and peak during oritavancin administration. The vex/vncRS operon in E. faecalis V583 contains an IS element belonging to the ISL3 family [26], indicating that the genome evolution of fecal E. faecium that cause blood stream infections may follow a similar trajectory to E. faecalis V583 [26]. We also found that IS256 regularly interrupts an aminoglycoside resistance operon (S9 Fig). Oscillating insertions both in coding and non-coding regions in this operon suggest that this region may be an IS256 insertion hotspot. Prophage abundance also increased during oritavancin treatment (S10 Fig). Overabundant prophages had nucleotide sequence homology to the prophage genomes from E. faecium strains, but lacked homology to enterococcal prophages known to carry platelet binding protein genes [70] or to enterococcal prophages associated with increased colonization ability [71]. These results suggest that antibiotic treatment can also drive the diversification of multiple classes of mobile elements in the human host. Lastly, we found that oritavancin-treated samples had an increase in IS256 circle formation indicating that in addition to novel insertion events, active IS256 transposition is occurring in fecal enterococci (Fig 7D). Overall, these findings suggest that IS256 transposition is frequent in the intestine during E. faecium infection, and this transposition likely impacts the antibiotic resistance and virulence profiles of the infecting bacterial population.

PLOS PATHOGENS
IS256 diversification in the enterococci Discussion IS elements directly shape bacterial physiological responses involved in antibiotic resistance and virulence, and are key drivers of genome evolution [72][73][74]. Considering the importance of IS element biology to bacterial fitness, the selective pressures that guide IS element regulation and mobility remain poorly defined. Using E. faecalis, we extend our knowledge of IS elements by detailing the population biology, regulation, and transposition dynamics of IS256. IS256 is a widely distributed IS element in Gram-positive pathogens. Through the use of bacterial genetics, comparative genomics, and IS-Seq we have revealed that IS256 insertion events are strongly dictated by phage predation and the mammalian host environment. Together these results suggest that IS256 mobility creates a genetically flexible genome that enables enterococci to rapidly adapt to diverse environmental conditions. IS256 is common in E. faecalis genomes and is found more frequently in hospital adapted lineages (Fig 1). Consistent with this idea, IS256 element abundance is a defining feature of hospital adapted E. faecium isolates [43,44]. Movement of IS elements between cells requires other mobile elements, such as phages and plasmids, as IS elements lack any transforming or

PLOS PATHOGENS
IS256 diversification in the enterococci transducing ability [27]. The pattern of intermittent IS256 abundance in the enterococci is likely due to periodic exposure to IS256-containing mobile elements. In other words, genetically-related strains may interact with IS256-containing mobile elements at different rates, resulting in the dissemination of IS256 to only a subset of the population. The enrichment of IS256 among hospital adapted strains may reflect the need for increased genomic plasticity in these strains as they adapt to an altered host environment [44,75]. Pathogens can utilize IS elements to form pseudogenes and to condense genome size through IS-mediated recombination, which can inhibit immune detection through loss of virulence factors and reduces the genome's metabolic burden [74,75].
Regulation of IS256 has not been investigated in enterococci and is key to understanding this element's biology. Here we show that IS256 is basally active in E. faecalis V583 and is regulated by multiple mechanisms at the transcriptional level (Fig 2B-2E). This is important because overactive IS256 could be catastrophic to E. faecalis viability and layers of regulation likely temper the elements activity until appropriate conditions require IS256 activation. Transposase protein levels are correlated to transposition frequency in E. coli [76,77]. Additionally, we show that IS256 has two different TSSs as part of this regulation and that both sites are regulated by asRNAs. The protein product occurring from the site 2 TSS is predicted to be truncated. Other IS elements encode two or more protein coding frames and these different protein products can contribute to IS element regulation through inhibition [73,78]. To better understand this, we built a truncated version of the IS256 Tnp and showed that it is inactive (Fig 2F). Previous work has shown that similarly truncated IS256 variants are able to bind to the element's IRs, but structural studies indicate that this truncation likely loses key dimerization domains essential for transposition [79,80]. We hypothesize that IS256 Tnp site 2 binds to the IRs to prevent full-length IS256 Tnp from binding and activating transposition.
asRNAs are known to regulate transposase protein levels in the E. coli, but to our knowledge, these have not yet been investigated in the enterococci. Here, we show that the asRNA produced from site 1 in IS256 negatively regulates the RNA levels of the IS256 Tnp. asRNAs can repress transcription through formation of transcriptional repressors or by targeting the mRNA for degradation by RNAse III [53,81]. Additionally, loss of this asRNA led to a dramatic decrease in IS256 circular intermediates (Fig 2F). We believe that this asRNA plays a crucial role in controlling IS256 transposition.
Using IS-Seq as a readout of basal IS256 diversity, we show that IS256 insertions function as a mechanism of genome-wide mutation (Fig 2A). This is the highest resolution IS-Seq map created, to our knowledge, and demonstrates that this technique can identify low-abundance insertions with high confidence [50,51]. Baseline IS256 movement likely adds mutational diversity to an enterococcal population which allows for adaption to changing environmental conditions through selection.
Previous work from our group determined that E. faecalis infection by the phage phi19 was associated with elevated of IS256 insertions [18] (Fig 3A). This was found using WGS that was based on less than 10 reads per insertion. We extended this observation by performing IS-Seq to accurately identify these novel insertion sites both in vitro and in vivo. We found that E. faecalis challenged with phi19 (19RS strains) had a major increase in new IS256 insertions ( Fig  4A and 4B). 19RS strains were chronically infected by phi19 and cells within the population carried the phi19 genome. Chronic phage infection may provide a continuous selective pressure that aids in IS256 mobility. This correlated with 19RS strains harboring elevated IS256 circular intermediates (Fig 4C). These circles are formed during IS256 copy-paste transposition and have been studied extensively [30].
We also observed that in a mouse model of E. faecalis V583 intestinal colonization, exposure to phi19 resulted in a similar diversification of IS256 (Fig 4D-4F). IS256 movement was less robust in the mouse intestine compared to the levels seen in vitro. This may be for several reasons, including a lower frequency of insertion events per location in vivo and a lower number of samples sequenced. An additional explanation is that Epa is crucial for bacterial colonization and epa mutations may lead to these mutants being outcompeted in the intestine [18,82].
Although we found evidence for phage-mediated IS256 activation in E. faecalis 19RS strains, we also determined that selection is important for the location of IS256 insertions ( Fig  4G and 4H). Cells that were not exposed to phi19 had lower IS256 read density in epaX. EpaX is a glycosyltransferase involved in teichoic acid decoration of the core rhamnose-containing Epa exopolysaccharide backbone. Mutations in epaX result in poor phage adsorption [18]. We found no other insertions in genes known to be involved in phage infection indicating that epaX and other epa genes are likely hotspots for IS256 integration following phage pressure. In light of these discoveries, we hypothesize that initial phage infection may select for epa-IS256 insertions and that prolonged phage exposure though chronic infection leads to increased and non-specific IS256 transposition as a stress response. epa mutations have been shown to sensitize E. faecalis to cell wall targeting antibiotics, blood killing, and reduces virulence in animal models [18,83,84]. This suggests that even though phage carrier state rapidly evolves the E. faecalis population, it could also reduce the populations virulence potential. We suspect that phage carriage and release will continually pressure intermittently resistant cells and force fully resistant cells to maintain high levels of resistance. Carrier state is also associated with genetic adaption in the Bacteroides, a prevalent member of the gut microbiota. Bacteroides coexist with crAss-phages in a constant cycle of infection, resistance, and re-sensitization. To counter crAss-phage, Bacteroides will reversibly invert the genetic sequences of capsule genes to prevent phage adsorption and attachment [ , an in-depth analysis of IS-mediated mutations during human infection had never been performed. Previously, such dynamics were studied by laboriously sequencing individual isolates. While this approach can successfully identify new IS element insertion sites, it overlooks IS element mobility within the entirety of the bacterial population. Here, we show that IS-Seq can be used to study bacterial adaptation at the population level and can reveal multiple insertions within such a population (Fig 7). Following oritavancin therapy an increase in IS256 insertions occurred in the vex/ vncRS operon, which has been implicated in both virulence [58] and vancomycin resistance [57]. Additionally, there is an increase in insertions in a putative transcriptional regulator that may regulate a Rib/Alpha adhesin [89,90]. Loss of these potential virulence factors may lead to modulation of virulence and avoidance of immune activation or detection. We predict that these insertions arise from an increase in IS256 mobility given the increase in IS256 circular intermediates (Fig 7D), followed by selection for beneficial insertions. This is likely similar to phage selection for insertions in the epa operon.
Additionally, we found highly abundant IS256 insertions in an aminoglycoside acetyltransferase. Vancomycin is reported to act synergistically with gentamicin against enterococci [91], likely due to increased gentamicin transport through the vancomycin-permeabilized cell wall [92]. In recent years, enterococci with aminoglycoside resistance have emerged [93] with acquisition of resistance genes carried on an IS256 composite transposon [94]. Our results suggest that genetic modulation of aminoglycoside resistance fluctuates in the host during infection, with different highly abundant insertions prominent at different timepoints. This data suggests that aminoglycoside resistance is repeatedly modulated and the intestinal enterococcal population regularly samples different versions of this aminoglycoside resistance operon. Future studies to identify if glycopeptide treatment potentiates aminoglycoside sensitivity through IS256 insertions is warranted. We additionally show that oritavancin treatment is associated with IS256 diversification in this patient. Oritavancin is an analog of vancomycin, but is also reported to disrupt cell membrane integrity and possess additional mechanisms for inhibiting cell wall synthesis [95]. This may explain the difference in IS256 expandability, as this antibiotic may impose strong selective pressure resulting in a broader range of mutations. Likewise, oritavancin treatment was associated with an increase in IS256 circles. In Staphylococcus aureus, vancomycin treatment increases IS256 transposition [35,69], reflecting a similar glycopeptideinduced IS256 activation. Enterococcal IS256 diversification is associated with both antibiotic exposure and phage predation. Both phages and antibiotics have been shown to mobilize conjugative transposons in Vibrio cholera. This suggests that the signaling mechanisms underlying enterococcal IS256 mobilization following externally applied stressors may be similar.
In summary, we show how the widespread and clinically-relevant IS element IS256 creates genome diversification during phage predation and human infection using sensitive NGS IS-Seq. This work provides a high-resolution picture of IS256 movement during both steadystate and stress inducing conditions. Understanding the outcomes of these IS256 insertion events will be critical to the study of the evolution of enterococcal pathogenesis in response to both phage and antibiotic therapies. We also show that IS256 is regulated at the transcriptional level in E. faecalis and have discovered that IS256 mobility is biased for highly mobilizable insertion events in nosocomial and virulent enterococcal isolates, further demonstrating that IS elements are domesticated and specialized in their choice of bacterial host.

Ethics statement
Animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Colorado (protocol number 00253).

Bacteria and bacteriophages
E. faecalis strains were grown in Brain Heart Infusion (BHI) broth with aeration at 37˚C. A list of bacterial and bacteriophage strains can be found in S4

PLOS PATHOGENS
IS256 diversification in the enterococci were identified by calling all predicted ORFs in all downloaded RefSeq genomes using prodigal v2.6.3 [102] and running diamond blastp v0.9.21 [103] against this database with known E. faecalis and E. faecium virulence factors from the Virulence Factor Database [104]. All further analyses and statistics were performed in R.

DNA extraction
Bacterial genomic DNA was extracted from overnight cultures grown in 5 mL BHI broth using the Zymo BIOmics DNA Miniprep Kit (Zymo, Irvine, CA).

Whole genome sequencing, IS256 identification from whole genome sequencing, and sequencing read alignment to the phi19 genome
Whole genome sequencing of E. faecalis V583 and 19RS strains was performed using Illumina HiSeq 2500 with 150 cycle paired-end sequencing. These genomes were previously described [18], and the reads have been deposited at the European Nucleotide Archive (ENA) as project accession number PRJEB30526. Novel IS256 insertions were found using CLC Workbench (Qiagen, Hilden, Germany) by first mapping reads to an E. faecalis V583 reference genome where all IS256 elements were removed manually. Unmapped reads were collected and mapped against a database of known IS256 elements with mapping parameters that allowed for partial matches (0.4 length fraction and 0.9 fraction similarity). Reads that mapped against IS256 were binned, the section of the read containing the IS256 sequence was removed, and the remaining read was realigned against the E. faecalis V583 reference genome (GCA_000007785.1) to identify IS256 insertion sites.
Reads aligning to the phi19 genome were found by first aligning whole genome sequencing reads to the E. faecalis V583 reference genome using Bowtie2 v2.3.5.1 [105], collecting unmapped reads, and aligning these reads to the phi19 genome.

Oxford Nanopore Technologies (ONT) sequencing
High molecular weight genomic DNA was isolated from E. faecalis V583 and 19RS strains after overnight growth as described above. ONT sequencing was performed by SeqCenter (Pittsburgh, PA) using PCR-free V14 chemistry ligation library preparation and sequenced on a GridION (ONT, Oxford, UK). 113,000 ± 22,400 reads were generated with an average read length of 3.7 ± 0.8 kb. phi19 sequences in reads were found using blastn [106] with a minimum alignment length of 1 kb. Reads passing this threshold were compared to the E. faecalis V583 genome with blastn and positive hits were filtered with a minimum alignment length of 1 kb. The reads have been deposited at the ENA as project accession number PRJEB30526.

IS-Seq analysis
For library preparation, 200 ng of genomic DNA was used as input in the Illumina DNA Prep Kit following the manufacturer's instructions (Illumina, San Diego, CA). For the in vitro and in vivo experiments (Fig 4), genomic DNA was isolated from 5x10 9 E. faecalis cells. Following library construction, IS256 enrichment was performed as follows. First, 10 ng of library DNA was amplified with the p7 primer and the IS-Seq Step1 primer (S5 Table) for 13 cycles using Q5 2X Master Mix (New England Biolabs, Ipswich, MA). These reaction products were diluted 1:100 and 10 μL was added to a PCR reaction with the p7 primer and the IS-Seq Step2 primer for 9 cycles using Q5 2X Master Mix. The products of this second reaction were sequenced on an Illumina NovaSEQ 6000 by the University of Colorado Anschutz Medical Center Genomics Core with paired-end 150 cycle chemistry. 23.3 ± 3.8 M reads, 17.9 ± 2.5 M reads, and 16.3 ± 2.6 M reads were obtained from sequencing the human, mouse, and in vitro IS-Seq libraries, respectively. IS-seq reads have been deposited at the ENA under project accession number PRJEB55280.
Paired end reads obtained from IS-Seq were partitioned into read1 and read2 files, and reads from read1 files were used in the downstream analysis. First, cutadapt v1.18 was used to bin reads with a full IS256 5' terminus. The IS256 5' terminus was then trimmed from these reads using cutadapt with the following flags: '-O 6 -g "^CGTAAAAGGACTGTTATATGG CCTTTTTACTTTTACACAATTATACGGACTTTATC" ' [107]. These trimmed reads were aligned to the E. faecalis V583 reference genome using Bowtie2 and the location of each insertion was determined using samtools v1.13 [108] and bedtools genomecov v2.30.0 [109].
For data generated in Fig 7, contigs were assembled from Illumina sequencing reads from each sample using megahit v. 1.2.7 [110]. These Illumina reads were generated from libraries prepared using the Nextera prep kit (Illumina) and sequenced on an Illumina NextSeq 500 with 150 cycle paired-end sequencing. Duplicate contigs were removed using gclust v1.0 [111] and these representative contigs were used as a reference for aligning IS-Seq reads as described above.
All further downstream analysis was performed in R. Insertion site abundances were normalized using DESeq2 estimateSizeFactors [56] before plotting to account for abnormalities during library preparation and sequencing. Significantly enriched unique insertion sites were found using DESeq2 by providing raw counts to the program and then computing general linearized models for each insertion site with at least five reads in three replicates. Models were fit using either a negative binomial fit, or when appropriate, a local fit. To account for fold change variance during each pairwise comparison, log fold change values were shrunk using the ApeGLM model [112] in DESeq2's lfcshrink function. Each pairwise comparison was performed using a separate call to the DESeq function in order to improve sensitivity, as recommended by the developers [56]. Both significantly enriched and insignificant insertions with at least 100 reads were reported in S2 and S3 Tables and Figs 4B and 4D and 7B.

Stranded RNA-Seq analysis
The E. faecalis V583 RNA-Seq data used for stranded IS256 gene regulation analysis was obtained from SRA experiments SRR7229491, SRR7229489, and SRR7229487 [54]. The reads were aligned to an E. faecalis V583 IS256 element sequence using Bowtie2 and forward and reverse aligning reads were found by parsing the SAM file. Only reads that perfectly aligned to the IS256 reference were retained. The depth of read alignment at each position in IS256 was calculated using samtools depth and the data was visualized in R.

E. faecalis phi19 infection in the mouse intestine
Conventional 6-week-old C57BL/6 mice were divided into two groups (Control: 2 female/ 2 male, Experimental: 2 female/ 2 male). All mice were treated with an antibiotic cocktail (streptomycin [1 mg/ml], gentamicin [1 mg/ml], erythromycin [200 μg/ml]) by first dosing via oral gavage with 100 μl of the cocktail and replacing their drinking water with the same antibiotic cocktail for 1 week. On day 7, antibiotic water was replaced with antibiotic free-drinking water. After 24 hours mice were colonized with E. faecalis V583 suspended in phosphate buffered saline (PBS) by oral gavage (10 9 CFU in 100 μl). After 24 hours mice were orally gavaged daily as follows: 100 μl of 1M sodium bicarbonate (all mice) and 100 μl PBS per control mouse or 10 9 PFU phi19 in 100 μl per experimental mouse. Feces were collected daily and homogenized in 1ml PBS. 10μl of the fecal slurry was serially diluted and plated on BHI agar containing 100 μg/ml gentamicin or 100 μg/ml gentamicin and 10 8 pfu/ml phi19. From the experimental samples, 5 μl of slurry was mixed with 45μl chloroform, spun at 16,363 RCF for 1 min and the supernatant was enumerated for phage particles by agar overlay plaque assay as described previously [19]. E. faecalis colonies were isolated on BHI agar containing 100 μg/ml gentamicin. These cells were grown to stationary phase in 5 mL of BHI broth, after which genomic DNA was isolated and IS-Seq was performed on these samples.

RNA extraction and qRT-PCR
RNA was isolated from exponentially growing (OD = 0.3) E. faecalis using a modified protocol for the RNeasy Mini-Prep Kit (Qiagen). First, cell pellets were collected by centrifugation of 5 mL of cell culture at 8,228 RCF for 5 minutes. Cell pellets were resuspended in 1 mL RNAlater and centrifuged again at 8,228 RCF for 10 minutes. The cell pellets were stored at -80˚C until sample processing. To isolate RNA, the pellets were thawed and resuspended in 100 μL TE buffer with 15 mg/mL lysozyme and incubated at room temperature for 30 minutes. Following this, 700 μL Buffer RLT with 0.01% beta-mercaptoethanol was added and the sample was transferred to a Lysing Matrix B tube (MPBio, Irvine, CA). The tubes were bead-beaten in a Mini-Beadbeater-16 (Biospec Products, Bartlesville, OK) on the fastest setting at 30-second intervals with 1-minute rests on ice between cycles. Following bead beating, the supernatant was removed after centrifugation at 16,363 RCF for 30 seconds and processed following the manufacturer's instructions. RNA was eluted from the column in RNase/DNase free H 2 O and any residual DNA contamination was removed with a 1 hour off-column DNase treatment (Qiagen). The RNA was purified following the RNeasy Mini-Prep Kit following manufacturer's instructions. cDNA was synthesized using Qscript Master Mix (Quantabio, Beverley, MA) and 1 μg RNA. For qPCR of gDNA targets, 1 ng of gDNA template was used per reaction. qPCR was run with PowerUp SYBR Green Master Mix (Thermo Fisher, Waltham, MA) on an Applied Biosystems QuantStudio 7 Flex qRT-PCR system. qRT-PCR primers are listed in S5 Table. Total 16S rRNA gene transcripts were used for qRT-PCR normalization and clpX was used for DNA qPCR normalization associated with IS circle quantification.

DNA manipulation and cloning
All cloning primers and DNA constructs used in this study are listed in S5 Table. All plasmid constructs used the pLZ12 shuttle vector backbone [113]. For cloning, all inserts and vectors were amplified using 2X Q5 DNA polymerase Master Mix and assembled using 2X Gibson Assembly Master Mix (New England Biolabs).

Phage carriage identification and serial passaging
To enumerate extracellular viral particles, we collected supernatant from cultures after centrifugation for 1 minute at 10,000 RCF, treated with 1/10 volume chloroform, centrifuged for 1 min at 21,000 RCF and removed the aqueous phase. The resulting supernatant was serially diluted in SM+ buffer (50 mM Tris-HCl, 100 mM NaCl, 8 mM MgSO 4 , 5 mM CaCl 2 [pH 7.4]), and 10 μL was mixed with 130 μL of a 1:10 dilution of E. faecalis V583 cells grown overnight. This mixture was combined with 5 mL of 0.35% Todd Hewitt agar (THA) supplemented with 10mM MgSO 4 , and poured onto THA+10mM MgSO 4 plates.
To identify 19RS E. faecalis cells actively shedding phages, glycerol stocks of phage-resistant cultures were streaked on BHI agar and incubated at 37˚C. After overnight growth, 182 μL of a 1:10 dilution in SM+ of an overnight wild type E. faecalis V583 culture was mixed with 7 mL of 0.35% THA and 10mM MgSO 4 and solidified on THA+10mM MgSO 4 plates for 30 minutes. Individual colonies were picked from the overnight plate patched onto a plate containing E. faecalis V583 embedded in 0.35% THA, with care not to break the surface of the agar, and incubated at 37˚C overnight.
To serially passage phage-positive colonies, colonies producing zones of clearing were patched onto both THA+10mM MgSO 4 plates and THA+10mM MgSO 4 with an E. faecalis V583 top agar layer as described above. The next day, the colony on the THA+10mM MgSO 4 plate was patched onto both types of plates and this was continued for two subsequent passages.

Growth curves
E. faecalis cultures were grown overnight in BHI broth and diluted to an OD 600 of 1 in PBS. 50 μL of each strain was added to 50 mL BHI and incubated at 37˚C. OD 600 readings were taken every 30 minutes on a Tecan Infinite M Plex (Tecan, Männedorf, Switzerland).

Enterococcal isolation from stool samples and gDNA extraction
Enterococcal stool populations were obtained over a span of 109 days from a patient being treated for recurrent E. faecium bacteremia. Stool samples were diluted and plated onto bile esculin azide-Enterococcosel agar (Becton Dickinson, Franklin Lakes, NJ), and 100-1,000 colonies were collected and pooled into BHI containing 16.7% glycerol and stored at -80˚C. One mL aliquots of the frozen population stocks were thawed and washed with BHI. Populations were allowed to briefly expand at 37˚C at 170 rpm for 5.5 hours. The cultures were pelleted and genomic DNA was extracted into DNase free water using the DNeasy Blood and Tissue Kit (Qiagen). gDNA concentrations were determined with the Qubit 1X dsDNA High-Sensitivity Kit and fluorometer (Invitrogen, Thermo Fisher).

Phage analysis from clinical enterococcal blood isolates
Viral genomes were found in assembled contigs using VIBRANT v1.2.1 [114] and reads from clinical samples were mapped against these contigs using Bowtie2. Multiple comparison analysis was performed using DESeq2 as described above.

Statistics and visualization
Graphs were made in Graphpad Prism or in R using dplyr and ggplot2. t-tests and one-way ANOVAs were calculated in Graphpad Prism. Chi-squared tests and insertion site differential abundance and proportional tests (prop.test) were calculated in R. Significantly abundant insertions from the IS-Seq experiments were found with DESeq as described above.

PLOS PATHOGENS
IS256 diversification in the enterococci